Nematic polymer liquid-crystal waveplate for high-power lasers at 1054 nm
Eileen M. Korenic, Stephen D. Jacobs, J. Kelly Houghton, Ansgar Schmid, andFranz Kreuzer
A nematic polymer liquid crystal is used to construct wave plates for use at 1054 nm. Three methods ofwave-plate construction are discussed: double substrate with fiber spacers in homogeneous distribution,double substrate with fiber spacers in annular distribution, and single substrate. The polymer liquidcrystal shows high laser-damage resistance, making it particularly useful for high-peak-power laserapplications. Alignment techniques and measurement of birefringence for the highly viscous polymerare described.
Liquid crystalline materials capable of replacing tradi-tional crystalline solids in the production of waveplates are becoming more desirable because of thehigh cost of the solid crystals typically used.' Theseinclude natural and synthetic mica, quartz, sapphire,magnesium fluoride, and potassium dihydrogen phos-phate (KDP crystal). The laborious and often expen-sive process of high-precision optical polishing of theplates fabricated from the crystals makes their usestill less feasible for large-aperture requirements.In addition, the range of available retardance valuesis limited, and the laser-damage thresholds are notalways known.
Low-molecular-weight liquid-crystal (LMLC) mono-mers have overcome the cost disadvantages of typicalsolid crystals. LMLC's are particularly successful inmeeting specific retardance needs and can exhibithigh resistance to pulsed-laser damage in IR and UVregimes.2 Despite the many successes of LMLC's,they introduce some difficulties of their own. Theseinclude the use of thick glass substrates to supportthe liquid crystal (LO) without bowing, and the epoxysealing of these substrates without long-term trans-mitted wave-front distortion caused by the sealant.
Polymer liquid crystals PLC's) potentially meet
F. Kreuzer is with the Consortium fur Elektrochemische Indus-trie GmbH, D-8000 Munchen 70, Germany. The other authorsare with the Laboratory for Laser Energetics, University ofRochester, 250 East River Road, Rochester, New York 14623-1299.Received 1 March 1993.
the same needs as monomeric LMLC's but withouttheir difficulties. PLC's generally have a high isotro-pic transition temperature. At and above this tem-perature, PLC's can be manipulated like LMLC's.At lower temperatures, the polymers become viscousenough to maintain their chosen configuration. Thepolymer also acts as its own substrate sealant, elimi-nating the need for epoxies.4
In this paper we examine the properties of anematic PLC (NPLC) aligned by using a buffed Nylon6/6 layer.5 Properties studied include viscosity (tem-perature dependence), birefringence (dispersion andtemperature dependence), and laser-damage resis-tance. We describe three techniques to constructNPLC wave plates and evaluate the uniformity andtransmitted wave-front quality of several 50-mm-diameter devices.
2. Properties of Nematic Polymer Liquid Crystal
A. Structure and Physical Properties
The liquid-crystal polymer used for this study,LC360N, is a side-chain polymer with a polysiloxanebackbone (see Fig. 1).6 The polymer's glass transi-tion temperature and clearing temperature are 23 0Cand 88 0C, respectively. Its viscosity at room tem-perature (22 0C) is greater than 3 x 108 cP (3 x 105Pa s), making it sticky to the touch. Because itsdielectric anisotropy is nearly zero, it does not affordalignment by application of an electric field.
B. Viscosity
For a PLC to be useful in optical devices such as waveplates, it must have a high viscosity at the tempera-ture at which the device is to be used to ensure device
Fig. 1. Structure of the NPLC LC360N.bone is shown on the left.
The polysiloxane back-
stability. The PLC must also have a low enoughviscosity at some convenient temperature to facilitateboth alignment and device construction. LC36ONmeets both of these requirements as determined by aviscosity-temperature measurement.
Viscosity was measured by a commercial digitalviscometer equipped with a helipath accessory.7The viscometer rotated a spindle and measured theforce required to maintain a chosen rotational speedagainst a viscous drag of the sample. This force wasdirectly related to the viscosity. The sample con-tainer had an inner diameter of 1.8 cm and the samplevolume was 12 mL. For LC360N, at low tempera-tures (T < 80 C) a T-bar-type spindle (1.09-cm cross-bar length) was used with helical path action. Astand raised and lowered the sample at a rate of 2.22cm/min through a distance of 2.3 cm around therotating spindle to ensure that the spindle continu-ously encountered fresh polymer, i.e., that not previ-ously grooved. At high temperatures (T > 80 'C)the usual disk-type spindle (1.46 cm diameter) with-out the helipath accessory was used. As a result,there was a wider range of measured values for thehigher viscosities (lower temperatures) because of theaction of raising and lowering the spindle. Rota-tional speed was varied from 0.5 to 100 rpm.
Because a different speed was optimum at eachtemperature, the viscosity reading was used when thedigital reading was 10% of the maximum rangeavailable for the given spindle-speed combination(results are shown in Fig. 2) The data were fittedempirically to give the relationship
1n(-i/ri) = 1.03 x 10 7 (1/T) 2 - 4.69
x 104 (1/T) + 51.33, (1)
where is the viscosity in centipoise, i is theviscosity in centipoise at the isotropic temperature,and T is the absolute temperature (K). The tempera-ture was maintained by using a circulating-waterjacket8; the water bath temperature supplying thejacket was controllable to ±0.01 C accuracy. Ourresults showed no sharp change in viscosity at thenematic-isotropic transition, as has been reported byothers for lyotropic polymers9 and for thermotropicLMLC's.10 At the nematic-isotropic transition, theapparent viscosity was 6.78 + 0.25 x 103 cP(6.78 ± 0.25 Pa s). At room temperature (22 CC), theempirical fit projects an apparent viscosity of 3 x 103(3 x 105 Pa s), which is comparable with taffy.
Fig. 2. Temperature dependence of NPLC viscosity. Viscosity isdetermined as a function of temperature by a rotating-spindle-typeviscometer. Measurements for temperatures below 80 0C aremade by a T-bar spindle with helical path action. Measurementsfor temperatures at and above 80 0C are made by a disk-typespindle and no helical action.
These data confirm that flow and alignment can beaccomplished at elevated temperatures, with the align-ment frozen in at a glass transition at or above roomtemperature.
C. Birefringence Measurement: Method and Results
The birefringence of the NPLC must be accuratelyknown for one to construct a wave plate with thedesired retardance. Measurements of the ordinaryand extraordinary refractive indices were performedby using a commercial Abbe refractometer.1 Therefractometer was calibrated12 with a fused-silicastandard at each wavelength-temperature combina-tion used. Because of the high viscosity of LC36ONat room temperature, placing it directly on the refrac-tometer stage is inconvenient and polymer alignmentis difficult. For this side-chain polysiloxane, the sidechains are responsible for the birefringence. With-out an alignment layer on the substrate, the sidechains align homeotropically. When a rubbed align-ment layer is used, the side chains align along the rubdirection. We chose the latter configuration becausealignment quality could be checked quickly by view-ing between crossed polarizers. The refractive indi-ces were determined with a cell constructed by sand-wiching the NPLC between S, a high-index glassplate (SF4, nD = 1.76), and S2, a low-index glass plate(microscope slide, nD = 1.5), shown in Fig. 3(a); bothof these had been prepared with a Nylon 6/6 align-ment layer in the following manner.
The glass substrate surfaces intended for contact-ing the NPLC were covered with 88% (reagent grade)formic acid for 30 s, spun at 1900 rpm for 60 s, andthen covered with a 0.2% solution of Nylon 6/6 in
Fig. 3. Schematic diagram of the Abbe refractometer and sample:(a) The refracting prism has the highest refractive index; index-matching fluid has a refractive index between SI and the refractingprism. S1 is a high-index glass substrate. The NPLC is alignedby thermal annealing between substrates prepared with a rubbedpolymer alignment layer. S2 is a glass substrate of arbitraryrefractive index. Source light is directed through the sample ontoa mirror and then to be the observer. The mirror is rotated todetect the critical angle used to calculate the refractive index. (b)The sample cell is oriented on the Abbe prism with the direction ofbuffing orthogonal to the incoming light beam.
88% formic acid for 60 s and spun at 1900 rpm for 2min. The coated substrates were heated on a pro-grammable hot plate'3 at 115 "C (15 "C above theboiling point of formic acid to ensure complete evapo-ration of the solvent) for 1.5 h and allowed to cool toroom temperature. The coated substrate was buffedby a device developed in house (at the Laboratory forLaser Energetics, LLE). It consisted of a spinning(2000-rpm) cylinder, 25 cm long and 10 cm in diam-eter and covered with a polyaramide fiber sheet.' 4
The substrate was held to a platform by vacuum andthen passed beneath the rotating cylinder. Theheight of the roller was adjusted until passing thesubstrate under the stationary roller just caused lightbehind the roller to be extinguished. The directionof buffing was defined to be the direction of the rollermotion (same as the movement of the platform forfirst pass). The substrate was passed forward andbackward under the spinning roller four times.
The NPLC and the SF4 plate were heated to100 "C. The NPLC was dabbed by microspatula
onto the SF4 plate and covered with the unheatedlow-index plate so that the rub directions of thesubstrates were antiparallel. No spacers were used.This cell was placed directly on a hot plate, held at88 0C (Ti) for 24 h, and then cooled from 88 C to20 0C at a rate of 1 0C/h. The thickness of the NPLClayer was determined with a contact micrometers tobe 20 A.m. The same sandwich cell was used for allbirefringence measurements.
S1 was mounted on the very-high-index (nD = 1.92)refractometer prism with a fluid'6 whose refractiveindex was nD = 1.81. The Abbe refractometer andsample arrangement are shown in Fig. 3(a).Although the flatness of substrates S1 and S2 waspoorer than 4X at 633 nm, the NPLC alignmentquality was sufficient for the measurements. Theindex of S2 was not critical. (If the index of S2happened to be higher than the NPLC, S2 would notcouple light into the sandwich cell and the existenceof 2 would simply not register.) The cell wasoriented on the refractometer stage such that thelight from the source would sample both the ordinaryand extraordinary indices. The rub direction wasorthogonal to the incoming light beam, as shown inFig. 3(b). Both n0 and ne measurements were read inthis configuration; five trials of each were averaged.
The ordinary and extraordinary refractive indiceswere measured at 20 C for various wavelengths.The indices, their average values {nave = [(2n0
2 +ne2)/3]1/2l, and the birefringence values are shownwith their standard deviations, UN-1, in Table 1. Thelight sources used were a sodium lamp, a mercurylamp, He-Ne laser, and a Nd:YLF laser. The so-dium and mercury sources were unpolarized. Thepolarized He-Ne and Nd:YLF beams were passedthrough a diffuser to randomize the polarization.The Abbe refractometer makes use of the critical-angle relationship of the sample with the refractom-eter prism. The observer sees a split light field,darker where the critical angle for a given sampleindex is exceeded. The lighter part of the field ofview is illuminated by the light source. To see thesplit field when using an IR source, we held ahandheld IR scope7 at the eyepiece, and measure-ments were then made in the usual manner. Threepoints (X = 546.1, X = 632.8, and X = 1054 nm) wereused to fit the data to the Cauchy equation:
n = A + B/X 2 + C/X4. (2)
The coefficients for n0 and n, are shown in Table 2.The individual curves, seen in Fig. 4(a), were sub-tracted point by point to give the birefringence disper-
Table 2. Coefficients of Cauchy Equation for Dispersion of Ordinaryand Extraordinary Indices of Refraction for NPLC at 20 'C
Index A B C
n.0 1.5071 6.6064 x 103 2.4866 X 108n, 1.6474 1.1344 x 104 9.3661 x 108
sion curve seen in Fig. 4(b). The birefringence disper-sion of crystalline quartz,'8 shown in Fig. 4(b) forcomparison, is 15 times smaller than that ofLC360N.
The refractive indices were determined at 1054 nmfor various temperatures. Temperature was main-tained by using a circulating-water bath8 around therefracting prism. These indices are listed in Table 3.The temperature dependence of the ordinary andextraordinary indices is shown in Fig. 5(a). Thetemperature dependence of the birefringence of ourNPLC (LC360N) and that of several other LC's arelisted in Table 4 and are plotted on Fig. 5(b). Theother LC's shown are another NPLC (methylstilbenePLC19) but at 589.6 nm, a LMLC (K1520,2') at 1054nm, a LLE mixture (60% 18523 and 40% 14627321 byweight) at 1047 nm, and a LMLC (E2002') at 1047 nm.
For LC360N, an empirical second-order polynomial
C
00
4-000C
1.7
1.6
1.5
1-AL
no
540 640 740 840 940 1040Wavelength (nm)
(a)
0.160
c 0.150
a)0)c 0.010
.h 0.009m
NPLC
Quartz
0.008540 640 740 840 940 1040
Wavelength (nm)(b)
Fig. 4. Dispersion of refractive index of the NPLC. (a) Theordinary and extraordinary refractive indices of the NPLC arefitted separately to the Cauchy equation. (b) The ordinary andextraordinary index curves are subtracted point by point to yieldthe dispersion curve. All index dispersion measurements aredetermined at 20 C, Birofringonce dispersion for crystallinequartz1 8 is shown for the same range.
fit of birefringence versus temperature (C) yields ad(An)/dT at 20 "C of (-)0.0003/C. At room tem-perature, the birefringence change per degree iswithin the standard deviation of the birefringencevalue. A wave plate made with this NPLC wouldshow a variation in retardance at room temperatureof only 0.3 nm/"C/Rm of polymer thickness. In azero-order half-wave plate, this represents a retar-dance change per degree of only 0.15 nm. By com-parison, the methylstilbene polymer shows a varia-tion in retardance of 0.03 nm/"C [Lm, but because ofits lower birefringence, it would require a thicker cellto produce the same retardance as our NPLC. TheLMLC's show a wide variation of the temperaturedependence of birefringence. The LMLC K15 exhib-its a very large birefringence change with tempera-ture at 20 C. This is to be expected because itsclearing temperature is approximately 33 C and thed(An)/dT of every LC increases sharply near its ownisotropic temperature. The second example of aLMLC is an in-house (LLE) wave plate mixture (60%18523 and 40% 14627). This mixture has been usedfor LC optical devices at the LLE and has proved to beconvenient for conventional LC cell construction.The mixture does, however, show a higher birefrin-gence change with temperature than does the NPLCLC360N, and for half-wave-plate construction at1054 nm it would exhibit a slightly higher retardancechange with temperature. Finally, a newer LMLC,E200, shows a d(An)/dT at 20 C of (-)0.00009/C,which is an order of magnitude smaller than ourNPLC. A zero-order half-wave plate made of thisLMLC would show only one third the birefringencechange of LC360N, but such a cell would require anexternal epoxy seal. Our NPLC shows promise forproviding temperature-stable retardance in a thin,self-sealing cell.
3. Wave-Plate Construction
Here we describe the fabrication of NPLC wave platesby various substrate and space arrangements. Allresearch with formic acid was performed in a class 10fume hood in a class 1000 clean room. All cell con-struction, excluding degassing, was performed in aclass 10 laminar flow hood in a class 1000 clean room.
For the double-substrate wave plates, our target re-tardance using an 18-jIm-thick path length was 2614nm. This provided a second-order retardance of 506nm, i.e., a second-order half-wave plate at 1054 nm.
A. Double-Substrate, Fiber-Spacer HomogeneousDistribution
The thickness of the NPLC layer was fixed by usingcommercially available glass fiber spacers22 mixedinto the polymer. The fiber-spacer diameters were18 ± 0.3 [Lm, and lengths varied from 20 to 100 ,um.The fiber spacers were added to the polymer for acalculated yield of 150 spacers in the completed cell(0.8 mg of fibers to 3.2 g of polymer). Homogeneousdistribution of the fibers was achieved by heating themixture to 100 "C and stirring mechanically for 15
min. A small amount of the mixture was removedwith a stirring rod and smeared onto a bare micro-scope slide. This smear was checked under a lightmicroscope for homogeneous distribution (i.e., noclumping) of the fibers. Stirring introduced bubblesinto the mixture, so it was degassed at 100 C and < 1Torr until bubbles were no longer visible (approxi-mately 1 h for the given mass of polymer). Thevacuum was applied gradually, 250 Torr every 15min, to prevent excessive foaming of the NPLC.
The wave-plate substrates were two 50-mm-diam-eter, 7-mm-thick borosilicate glass (BK-7) disks pol-ished to a surface flatness of better than X/10 peak to
1.66
0a 1.62
_ 1.5800 1.54
1.50
I I
I I Ie II
0 10 20 30 40 50 60Temperature (C)
(a)
0.18
0.16
0.14
a 0.12UCc 0.10co
- 0.08a)
m 0.06
0.04
0.02
70
I I i I
0 20 40 60 80 100
Temperature (C)(b)
Fig. 5. Temperature dependence of refractive index for theNPLC. (a) ne and no between T = 20'C and T = 70 'C, (b) Anbetween T = 20 'C and T = 70 'C, compared with literature andmeasured data for several other LC compounds. Stars, NPLC(LC360N) at 1054 nm; crosses, methylstilbene NPLC at 589.6nm20 ; circles, K15 at 1054 nm
21 ; squares, 18543/14627 at 1047nm; triangles, E200 at 1047 nm.
valley at 633 nm. Both were coated with a Nylon6/6 alignment layer by using the procedure describedin Subsection 2.C. One substrate was heated to100 'C. The polymer was separately heated to 100 0Cand scooped onto the substrate by using a metalmicrospatula. The other substrate was kept at roomtemperature so that it could be lowered by hand ontothe polymer. The rub directions of the substrateswere antiparallel. The constructed cell was de-gassed at 100 C and < 1 Torr for 24 h to removetrapped air bubbles. The cell was placed on a 5-mm-thick silica square, loaded onto a hot plate, coveredwith a Pyrex glass dish, and held at 88 0C for 24 h; itwas then cooled from 88 C to 20 0C at 1 C/h. Thiswas the same annealing protocol used for the birefrin-gence measurement cell.
The completed cell had 75 fibers distributed overthe full aperture. There was little distortion of theNPLC alignment around these fibers, as shown byviewing at 100 x magnification between crossed polar-izers [see Fig. 6(a)]. Annealing removed all evidenceof fluid flow around the spacers.
B. Double-Substrate, Fiber-Spacer Annulus
An alternative technique was employed to keep thefiber spacers out of the wave plate's clear aperture,where they might otherwise provide sites for laserdamage or contribute to gradients in the transmittedoptical wave front. The fibers were applied in anannulus after the alignment layer had been appliedand buffed. For 50-mm-diameter BK-7 substrates, a32-mm-diameter glass mask was placed over the clearaperture. The fibers were applied by dipping a cot-ton swab into the fibers, tapping them off above thesubstrate, and removing the mask. Then the NPLCwas applied by scooping it at 100 0C onto the clearaperture of the substrate also held at 100 C. As wedescribed in Subsection 3.A, the procedure thencontinued with construction of the double-substratecell. Both parallel and antiparallel construction weretested. Retardance values of the parallel and anti-parallel cells were comparable, but total tilt anglevaried (see Subsection 4.B.2).
The density of fibers in the annular region washigh. Figure 6(b), a photograph taken at 50 x magni-fication between crossed polarizers, shows a densityof 15 fibers/mm2 . This density contributed to theformation of lines of disclination, but this effect wasconfined to the annulus outside of the clear apertureand did not affect transmitted wave-front quality ormeasured retardance.
Form NPLC NPLC LMLC LMLC LMLCTi 88 C ' 100 C 33 C 6 50 C 88.5 CX 1054 nm 589.6 nm 1054 nm 1047 nm 1047 nmAn at 20 0C 0.1452 0.057 0.168 0.042 0.152d(An)/dTat 20 C (1/C) -2.755 x 10-4 -2.970 x 10-5 -4.18 x 10-3 -3.35 x 10-4 -9.492 x 10-5d(tAn)dt per pAm at 20 C (nm/0 C/ m) -0.2755 -0.0297 - -4.18 -0.335 -0.0942d(tAn)/dT for t = 527 nm and 20 C -0.1452 - -2.202 -0.1766 -0.0496
(nm/0 C/pim)Ref. - 19 20, 21 3, 21 21
C. Single Substrate
In addition to the methods described above for sand-wich-type NPLC wave plates, we developed a methodfor producing single-substrate wave plates. Becausea large proportion of the cost of any high-quality LCdevice is attributed to the substrates, a single-substrate wave plate essentially halves the expense ofthe device. A single substrate was prepared with analignment layer as described previously. The singlesubstrates used here were 12 mm thick.
The first procedure tested was spin coating a solu-tion of the NPLC in chloroform solvent (5% wt./wt.)onto the buffed substrate. Solvent alone was drippedby syringe onto the substrate and allowed to sit for 30s. The substrate was then spun at 1800 rpm whilewe dripped solvent (1 mL) onto the substrate for atotal spin of 60 s. Then the 5% solution was drippedby syringe onto the substrate, allowed to sit for 60 s,and spun at room temperature at 1800 rpm, againwith us dripping 1 mL of the solution for the initialpart of the spin, for a total spin of 120 s. The singlesubstrate was annealed under the same protocol asthe double-substrate cells. Viewing between crossedpolarizers showed alignment and visual transmission-extinction contrast; however, retardance measure-ments by laser ratiometer were within the systemnoise levels, suggesting that the NPLC layer was toothin to measure. Microscopic examination showedbeading of the NPLC along the rub direction.Higher concentrations did not spin on uniformly.
A second method was attempted that involveddripping the NPLC-chloroform solution onto thebuffed substrate and allowing the solvent to evapo-rate, leavingbehind the NPLC layer. Low concentra-tions ( < 7%) evaporated so rapidly that the NPLC wasdeposited in uneven patches. Evaporation of higher-concentration solutions left a polymer layer barrier atthe air-solution interface that precluded furtherevaporation.
Consequently, the NPLC had to be applied directlyto the buffed substrate in the following way: Thesubstrate was placed on a silica square and heated to100 °C. The NPLC at 100 °C was scooped along oneedge to form a crescent. A stainless-steel rod ofwound stainless-steel wire,23 designed to give a thick-ness of 6 m, was used to stroke the NPLC across thesubstrate either parallel or perpendicular to the buff
direction for a given trial. To prevent sliding of thesubstrate, we epoxied glass tacks at one corner of thesilica square. The cell was transferred to a plainsilica square, held in an oven at 88 C for 1 h, andthen allowed to cool to room temperature (averagerate of cooling was 20 °C/h). Longer heating (2 h)with slower annealing (3 °C/h) showed no improve-ment in uniformity or alignment. There was nosignificant difference in uniformity or alignment be-tween single-substrate cells made with flat (peak tovalley < X/10) and nonflat (peak to valley 5) sub-strates, nor for stroking the NPLC with the steel rodperpendicular versus parallel to the rub direction.Neither the bulk NPLC nor that on the substrate wasdegassed because previous experiments with degas-sing showed no improvement in appearance.
All trials showed a residual outline of the originalcrescent-shaped NPLC application region. Partialremoval of this nonuniformity was achieved by usinga combination of application techniques. A buffedsubstrate sitting on a silica square with glass tacks atone corner was heated to 100 C. The NPLC, also at100 C, was scooped onto the center of the substrate.A small brush was used to spread the NPLC smoothlyover the surface. Brush strokes were either parallelor perpendicular to the buff direction for a given trial.The warmed rod as above was drawn across thesurface to promote uniform thickness. This cell wasthen held at 110 C for 15 min. The hot cell wasthen spun at 4000 rpm for 30 s to distribute theNPLC evenly over the substrate and to remove anyresidual brush or rod lines. (This process step dic-tated the use of thick single substrates. Their largeheat capacity ensured that the NPLC was warmenough to flow during the entire spin.) This cell washeld in an oven as before at 88 C for 1 h and thenallowed to cool to room temperature. The brush didintroduce some air bubbles, but these disappearedduring the 1 h of heating.
4. Wave-Plate Evaluation
The optical quality of a wave plate depends onundistorted wave-front transmission over the clearaperture and on accuracy and uniformity of re-tardance. Constructed wave plates were first viewedbetween cross polarizers for uniformity of trans-mission and extinction. Table 5 gives transmit-
(0)Fig. 6. Glass fiber spacers (18-pum diameter) in double-substrate wave plates: (a) Photograph at 10Ox magnification between crossedpolarizers shows three fibers in a 0.8-mm2 area in the double-substrate wave plate with homogeneous fiber distribution. There are noapparent disclinations. (b) Photograph at 50x magnification between crossed polarizers shows fibers in a 3-mm2 area in thedouble-substrate wave plate with fiber-spacer annulus. These fibers and the disclinations surrounding them are outside the clear apertureof the wave plate. Other features in both photomicrographs may be attributed to dirt or buffing marks.
ted wave-front distortion with and without the NPLC, substrates ultimately used in the double-substrateretardance, tilt angle at center and over the clear configuration were first assembled with 5-jim Mylar26
apertures (75% of the hard apertures), and thickness spacers and index-matching fluid 6 (nD = 1.516).calculated from total retardance and tilt angle24 for all Three external glass tacks, epoxied across the sides ofthree types of wave plate. the substrates, maintained the cell structure for
interferometry. The transmitted wave-front qualityof this mock cell was used as a standard of comparison
Transmitted wave front was measured by using a for the double-substrate cell subsequently assembled.Fizeau interferometer2 5 at X = 633 nm. To establish The reference for the single substrate was the sub-a transmitted wave-front reference, we assembled the strate itself.
Transmitted wave-front distortion for all threetypes of cell varied from two to four times that of theblank cell. One double-substrate cell, stored verti-cally for 1 month, showed transmitted wave-frontimprovement of 0.1 . Such improvement mightbe desirable but points out the susceptibility of aNPLC like LC36ON to flow under the influence ofgravity when the glass transition temperature is toonear room temperature.
B. Retardance
1. Value and VariationRetardance was measured by using a laser ratiometerwith a Soleil-Babinet compensator27 between crossedGlan-Thompson calcite polarizers. The approxi-mate beam diameter of the laser at the cell locationwas 2 mm. The retardance order was established bythe fiber-spacer diameter. Nine spots were mea-sured and averaged. Single-spot retardance was mea-sured five times at the center as a way to establishreproducibility. Variation was no more than ± 5 nmfor all types of wave plate.
Both types of double-substrate wave plate showedexcellent agreement of expected and average mea-sured retardance values. Because of the near-room-temperature glass transition of the NPLC, any ma-nipulation of the cell by hand contributed to somedegradation of retardation uniformity. Retardanceuniformity was most dependent on fiber-spacer distri-bution. In the homogeneous distribution type, asfew as 75 spacers over the aperture provided excellentgap uniformity, as shown in Fig. 7(a). In the fiberannulus distribution type, the density of fibers wasmore difficult to control. Regions of excessive fiberdeposition led to more severe gap wedges, as shown inFig. 7(b). More uniform and reproducible gaps wereobtained by swabbing fibers directly onto the exposedsubstrate around the mask than by tapping from aswab. Direct swabbing, however, scratched the Ny-lon alignment layer, creating regions of scatteringthat sometimes extended into the clear aperture.
The single-substrate wave plate showed fairly gooduniformity over most of the clear aperture. Varia-tions in thickness, i.e., wedge, were commonly causedby unevenness of stroking with the hand-coater rod,as shown in Fig. 7(c).
For use in a laser system, a retardance deviation ofno more than ± 10% of calculated retardance isdesirable.28 All three types of wave plate showedthis accuracy when averaged over the aperture.However, this specification presumes uniformityacross the clear aperture. Only the double-substratewave plate having fiber spacers in homogeneousdistribution showed sufficient uniformity for use in ahigh-peak-power laser system. Improved unifor-mity of the other types of wave plate may be expectedwith refinements in the application of (a) fibers, in thecase of the annular distribution double-substratewave plate, or (b) the NPLC, in the case of thesingle-substrate wave plate.
2. Tilt AngleTilt angle is the angle that the side chains of theNPLC make with respect to the substrate surface.The tilt angle was measured at the center of the cellfor reproducibility tests and at nine spots for unifor-mity tests by using a phase-retardation method24 inwhich the cell is rotated around an axis perpendicularto the incident beam in a laser ratiometer.
Because the tilt-angle measurement is an averagethrough the cell, parallel construction of the double-substrate cells yielded tilt-angle measurements ofalmost zero, whereas antiparallel construction led totilt-angle measurements of 2-3°. This is consistentwith the induced tilt angle being the same at bothinner surfaces.29 The tilt of the side chains at thefirst surface induces a slight retardance on the inci-dent beam. In parallel construction, this beam thenencounters side chains at the second surface that aretilted in the opposite direction, compensating for theretardance. As the cell is rotated around an axisperpendicular to the incident beam, a minimum inthe transmission indicates a matching of the incident
(c)Fig. 7. Wave-plate retardance contour maps: (a) double-sub-strate homogeneous fibers, (b) double-substrate fiber annulus, (c)single substrate. Values shown represent 5 and 10 nm ofsecond-order retardance and 20 nm of first-order retardancewithin the indicated region for (a), (b), and (c), respectively.Retardance is measured with a Soleil-Babinet compensator.
beam polarization with the side-chain tilt. For theparallel cell this minimum occurs at normal inci-dence, where retardance of one surface just compen-sates for the other. In antiparallel construction nosuch compensation takes place, in which case mini-mum transmission occurs at an off-normal angle ifside-chain tilt exists. Table 5 reports the tilt angles
for a double-substrate wave plate with a homoge-neous distribution of fibers and antiparallel construc-tion and for a double-substrate wave plate with fibersin an annular distribution and parallel construction.The tilt angle of the NPLC on the single substratewas negligible.
3. ThicknessBecause there was a range in diameters of the fiberspacers used for the double-substrate cells, totalretardance was used to calculate the thickness of theNPLC and the fiber thickness was used to establishthe order of the retardance. The retardance used forthickness calculations had to be corrected for thatretardance caused by tilt angle.24 The thicknessesshown in Table 5 are based on these corrected totalretardance values. Even without the correction, mea-sured thickness values of the NPLC layer were wellwithin the range established by the fiber-spacer diam-eter. This result implies that the fibers maintainedthe cell gap without trapping any NPLC between thefiber and the substrate.
5. Laser-Damage Resistance
The NPLC wave plates constructed and evaluatedabove provide numerous advantages over both tradi-tional crystalline solids and LMLC's for typical opti-cal uses. However, the particular NPLC used forthis study provided yet another advantageous prop-erty: high laser-damage resistance.
The laser-damage resistance of the LC36ON wasmeasured at 1054 nm by a technique described ingreater detail elsewhere.2 The NPLC was dissolvedin toluene (at 2% wt./wt. solid content) and sprayedby airbrush onto the small face of a 30-60-90 BK-7prism, the geometry of which prevents backreflectionand subsequent interference effects. Both 1-on-1andN-on-i tests were made. In 1-on-i tests, a singlelaser pulse is incident on a previously nonirradiatedsite. The energy within the pulse is gradually in-creased at each new location until damage is seen.In N-on-1 testing, laser pulses of increasing energyare incident on the same site at 5-s intervals untildamage is seen. The laser used was a Nd:glass,mode-locked, feedback-controlled, Q-switched oscilla-tor whose 1054-nm pulses could be frequency tripledto 351 nm by using KDP crystal. Pulses were 1 0.1 ns in duration in the IR (3 J/pulse). Damage wasdiagnosed as the appearance, within the 3-mm-diameter irradiation spot size, of a bubble that scat-tered light. Under Nomarski microscopic inspec-tion, shown in Fig. 8(a), the NPLC exhibited apparentmultiple-clustered pitting.
The original brown sample of LC36ON was contami-nated with traces of a polymerization catalyst be-lieved to be colloidal metallic platinum, with consider-able potential for causing laser damage. Afterpurification to 1 part in 106 Pt by column chromatog-raphy (purification carried out by Wacker Chemie),the clear, colorless polymer showed a much-improved
laser-damage resistance and the laser-damage sitesshowed far less severe pitting, as shown in Fig. 8(b).
The laser-damage test data are reported in Table 6.The films used in damage testing were 2.5 timesthicker than the NPLC in the second-order waveplates and 13 times thicker than the NPLC layerrequired for a zero-order wave plate. Because laser-damage resistance can be expected to improve forthinner layers, damage thresholds for NPLC waveplates in a working system would likely be evenhigher than those reported here. The laser-damagemeasurements of a Nylon alignment layer,3 buffedunder both hard and light conditions, show resistanceof 5-10 J/cm 2 (Table 6). Damage-resistance levelsof severaljoules per square centimeter are compatiblewith use in many high-peak-power laser applications.
(a)
(b)
Fig. 8. Nomarski photomicrographs (10OX) of the NPLC afterirradiation with a i-ns laser pulse at 1054 nm. (a) Extensivepitting in a 16-pLm layer of unpurified material (39 parts in 106Pt). (b) Reduced incidence of damage in a 45-pLm layer of purifiedmaterial (1.0 parts in 106 Pt).
Table 6. Laser-Damage Resistance of Nematic Polymer LC36ON andNylon 6/6
The components of a NPLC wave plate with a Nylonalignment layer meet this specification.
6. Summary
In summary, we have developed techniques for usinga nematic polymer liquid crystal in the constructionof a wave plate. These techniques include measure-ment of the birefringence of the highly viscous NPLCand alignment by thermal annealing against a buffedNylon 6/6 aligning layer. We were able to constructthree types of wave plate: double substrate withfiber spacers distributed homogeneously, double sub-strate with fiber spacers distributed in an annulus,and single substrate. The viscosity-temperature be-havior of the NPLC permitted alignment at highertemperatures but configuration stability at a lower(room) temperature. The NPLC acted as its ownadhesive to seal the confining substrates together.The high laser-damage resistance of the NPLC indi-cates that this and similar materials will prove usefulfor high-peak-power laser applications.
The wave plates showed promising uniformity andtransmitted wave-front quality, as well as the capabil-ity of meeting specific retardance requirements at acost determined mainly by the glass substrates.Further improvements in device stability could beobtained by using a NPLC with a higher glass-transition temperature (> 40 C to ensure a locked-inconfiguration at room temperature, to permit greaterease of handling, and to eliminate long-term flowunder the influence of gravity). If the NPLC had apositive dielectric anisotropy (Ac), molecular align-ment would be facilitated by permitting annealing inthe presence of an electric field.
The authors thank M. J. Guardalben for providingphotographs of damaged NPLC, K. L. Marshall fordiscussions regarding NPLC alignment and manu-script preparation, and S. Papernov for the discussionof damage-testing methodology. This research wassupported by the U.S. Department of Energy, Officeof Inertial Confinement Fusion under CooperativeAgreement DE-FC03-92SF19460, the University ofRochester, and the New York State Energy Researchand Development Authority.
References and Notes1. S. D. Jacobs, K. A. Cerqua, K. L. Marshall, A. Schmid, M. J.
Guardalben, and K. J. Skerrett, "Liquid-crystal laser optics:design, fabrication, and performance," J. Opt. Soc. Am B 5,1962-1972 (1988).
2. M. Guardalben, A. Bevin, K. Marshall, A. Schmid, and F.Kreuzer, "1053-nm high-field effect in monomeric and poly-meric conjugated systems," Natl. Inst. Stand. Technol. Spec.Pub. 775, 462-469 (1989).
3. A. L. Rigatti, D. M. Dudek, R. G. Carnes, L. D. Lund, K. L.Marshall, S. Papernov, A. W. Schmid, D. J. Smith, and S. D.Jacobs, "Uniformity issues in the manufacture of large-aperture liquid-crystal wave plates," in Conference on Lasersand Electro-Optics, Vol. 11 of 1993 OSA Technical DigestSeries (Optical Society of America, Washington, D.C., 1993),p. 50.
4. S. D. Jacobs, K. L. Marshall, and K. A. Cerqua, "Methods ofmaking composite optical devices employing polymer liquidcrystal," U.S. patent 5,054,888 (8 October 1991).
9. M. C. Muir and R. S. Porter, "Processing rheology of liquidcrystalline polymers: a review," Mol. Cryst. Liq. Cryst. 169,83-95 (1989).
10. G. Marrucci, "Rheology of nematic polymers," in LiquidCrystallinity in Polymers: Principles and Fundamental Prop-erties, A. Ciferri, ed. (VCH, New York, 1991), Chap. 1.
11. Bellingham and Stanley Model 60/HR Abbe refractometer,Bellingham and Stanley Ltd., Longfield Rd., North FarmIndustrial Estate, Tunbridge Wells, Kent TN2 3EY, England.
12. Bellingham and Stanley calibration tables for Abbe refractom-eter Model 60/Hr, Bellingham and Stanley Ltd., Longfield Rd.,North Farm Industrial Estate, Tunbridge Wells, Kent TN23EY, England.
13. Dataplate digital hot plate/stirrer, PMC Industries, Inc., 6335Ferris Square, San Diego, Calif. 92121-3208.
14. TexTech Industries, Main St., North Monmouth, Me. 04265.15. Digimatic indicator IDC Series 543, Mitutoyo Corp., 31-19,
Shiba 5-chome, Minato-ku, Tokyo 108, Japan.16. R. P. Cargille Laboratories, Inc., Cedar Grove, N.J. 07009-
South Vermont St., Palatine, Ill. 60067. The maximumrange of wavelength sensitivity is 1.2 pum.
18. J. M. Bennett and A. T. Glassman, "Polarizer materials," inCRC Handbook of Laser Science and Technology, Section 1.4:Optical Materials, M. J. Weber, ed. (CRC, Boca Raton, Fla.,1986), Vol. IV, Part 2, p. 227.
19. N. A. Vaz, G. W. Smith, G. P. Montgomery, Jr., W. D. Marion,V. Percec, and M. Lee, "Refractive indices of a methylstilbenepolymer liquid crystal," Mol. Cryst. Liq. Cryst. 198, 305-321(1991).
20. K. C. Chu, C. K. Chen, and Y. R. Shen, "Measurement ofrefractive indices and study of isotropic-nematic phase transi-tion by the surface plasmon technique," Mol. Cryst. Liq. Cryst.59, 97-108(1980).
23. K Hand Coater, RK Print-Coat Instruments, Ltd., South ViewLaboratories, Litlington, Royston, Herts SG8 OQZ, England.
24. H. L. Ong, "Cell thickness and surface pretilt angle measure-ments of a planar liquid-crystal cell with obliquely incidentlight," J. Appl. Phys. 71, 140-144 (1992).
27. Soleil-Babinet compensator Model 8-400, Special Optics, 101E. Main St., P.O. Box 163, Little Falls, N.Y. 07424.
28. C. Cotton, Laboratory for Laser Energetics, Rochester, N.Y.14623 (personal communication, 1992).
29. G. Baur, V. Wittwer, and D. W. Berreman, "Determination ofthe tilt angles at surfaces of substrates in liquid crystal cells,"Phys. Lett. A 56, 142-144 (1976).
