polarization selective hologram-based photonic delay lines

1 December 1998

Ž .Optics Communications 157 1998 225–237

Full length article

Polarization selective hologram-based photonic delay lines

Nicholas Madamopoulos, Nabeel A. Riza )

( )Center for Research and Education in Optics and Lasers CREOL and the Department of Electrical and Computer Engineering, UniÕersityof Central Florida, P.O. Box 162700, 4000 Central Florida BlÕd., Orlando, FL 32826, USA

Received 26 May 1998; revised 20 August 1998; accepted 14 September 1998

Abstract

The use of polarization selective holograms as optical signal routing elements for the implementation of photonic delayŽ . Ž .lines PDLs is proposed. A single bit PDL using ferroelectric liquid crystal FLC devices as active polarization switches

Ž .and polymer dispersed liquid crystal PDLC devices as polarization selective holograms for optical path routing isexperimentally demonstrated and characterized. Different within-channel leakage noise filters for improved PDL perfor-

Ž .mance are discussed and experimentally demonstrated. Record high optical signal-to-leakage noise ratios )45 dB areobtained for both PDL settings using a combination of the proposed noise filters. An alternative reflective PDL architectureis also proposed. This reflective architecture requires half the physical length for each path compared with the transmissivedesign to obtain the same time delays. Other polarization dependent optical router designs based on birefringent-mode

Ž .nematic liquid crystal NLC devices are also proposed. q 1998 Elsevier Science B.V. All rights reserved.

PACS: 42.79.S; 84.40.Zc; 42.70.Df; 42.79.KrŽ .Keywords: Phased array antennas; Polarization-selective holograms; Polymer dispersed liquid crystal PDLC ; Ferroelectric liquid crystal

Ž .FLC ; Photonic beamformer; Delay lines; Microwave radar

1. Introduction

Ž .In the recent years, several photonic delay lines PDLsbased on various optoelectronic technologies have been

w xproposed 1 . PDL applications include phased array anten-w x w x w x w xnas 1–4 , laser radars 5 , astronomy 6 , ultrasound 7

w xand two-photon memories 8 . During the last few years,we have concentrated our work on the implementation ofN-bit PDL systems using electrically controlled polariza-

Ž .tion switching devices such as nematic liquid crystal NLCw x Ž . w x9 and ferroelectric liquid crystal FLC devices 10 .These PDLs use passive polarization components such ascube polarization beam-splitters and beam-combiners foroptical signal routing.

We are currently exploring new photonic technologiesfor the implementation of high performance opticalswitches, routers and combiners, that will form a highperformance switching fabric to be used in different PDL

) Corresponding author. E-mail: [email protected]

architectures. A new promising technology for opticalŽ .signal routing is polarization-selective holograms PSHs .

PSHs are optical elements that have a different phasefunction depending on the state of polarization of theincident light. Fixed PSHs have been recorded in organic

w xdyes for applications in holographic interferometry 11 , indichromated gelatins for substrate mode holographic inter-

w xconnects 12 , and in photorefractive crystals for free spacew xoptical interconnects 13 . Birefringent computer generated

holograms have also been demonstrated for the implemen-tation of 2=2 optical switches for free-space multistage

w xinterconnection networks 14 . Other polarization depen-dent beam routing elements can be implemented usingNLC technology such as birefringent-mode NLC devices

w xpreviously used for active optical beam focusing 15 , andw xcombining 16 .

This paper describes the basic workings of a PSH-basedswitched PDL. A PSH device can be fabricated using anyof the technologies mentioned in the previous paragraph. Aproof of concept experiment for the characterization ofsuch a PDL is also presented. The PSHs used in our

0030-4018r98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved.Ž .PII: S0030-4018 98 00501-X

( )N. Madamopoulos, N.A. RizarOptics Communications 157 1998 225–237226

experiment are based on holographic polymer dispersedŽ .liquid crystals PDLCs . The basic operation of a holo-

graphic PDLC device is as an active holographic elementwhose diffractive power can be electrically modulated.Interest in PDLC devices for delay lines has been recent as

w x w xproposed by us 17 , and others 18 . We have also recentlyshown how PDLC devices can act as both active polariza-tion switches and beam routers and combiners to form low

w xcomponent count PDLs 19 . In this paper, we do notelectrically modulate the PDLC diffractive power to forma polarization switch. Instead, the PDLC devices are usedas fixed PSHs to form beam routers and combiners. Thecurrent PDLC device limitations are discussed and ways ofimproving the overall PDL performance are presented. Afeature of our proposed PDL is that it uses 2-D liquidcrystal polarization switching arrays and single large areaPSHs. This makes the physical PDL implementation asomewhat simpler assembly operation. Note that currentliquid crystal technology can give high pixel count, lowcost, 2-D flat panel switching arrays. In addition, the largearea, single pixel polarization dependent opticalroutersrcombiners required in our PDL can be fabricatedusing low cost, batch production, standard holographictechniques. The rest of the paper discusses the detailsrelated to the theory and experimental verification of our

w xPDLC-based PDL 20 .

2. The polarization selective hologram-based switchedPDL

The proposed PSH-based PDL is shown in Fig. 1. Anarray of linearly polarized collimated laser beams enters

Ž .the PDL system. The polarization switch PS array con-Ž .trols the state of polarization SOP of the individual

beams. When the switch is set in its ‘on’ state it rotates theincident polarization by 908. On the other hand, when thePS is set in its ‘off’ state it leaves the incident polarizationunaffected.

Depending on the SOP of the beams, the PSH eitherdeflects the beam at a specific and predefined angle orleaves the incident beam to pass straight through. Hence,in the first case, the optical signal follows the delay path,

Ž .and after reflection from the two mirrors M it passesthrough the second PSH. Note that in any single bit PDLmodule, the two paths have to be combined before thebeam exits the module. This is because single bit modulesare cascaded to form the N-bit PDL. As seen in Fig. 1, weuse a PSH device for this important 2:1 recombinationoperation. In the second case, the light travels unaffectedthrough the two PSHs towards the output of the PDL bit.Imaging lenses can be used in the PDL system to minimize

w xthe interchannel crosstalk 21 .

3. Basic characteristics of the polymer dispersed liquidcrystal devices

As mentioned earlier, we use PDLC devices as PSHs.In this section, we will briefly describe the operation of thePDLC devices in general and then their use as PSHs.Researchers have recently been studying the electro-optic

w xcharacteristics and applications of PDLCs 22–28 . Thebasic operation of the PDLC devices is as an activeelement that can be electrically switched from a lightscattering state to a transparent state without polarizers and

Fig. 1. A single bit of the proposed photonic delay line based on polarization switching devices and polarization selective holograms.Dashed lines: delay path; Solid lines: non-delay path; PS: polarization switch; PSH: polarization selective hologram; M: mirror; L: lens.

( )N. Madamopoulos, N.A. RizarOptics Communications 157 1998 225–237 227

alignment layers. Hence, PDLC devices have been used asw xlight valves for displays 24 . Recently, holographic ele-

ment whose diffractive power can be electrically modu-w xlated have been demonstrated using PDLCs 29 . Thus,

PDLC-based switchable diffractive elements for opticalŽ .beam steering, holographic read only memories ROMs ,

and switchable focus lens applications can be realizedw x23,27 . The PDLC electrically switched gratings are based

Žon a photopolymer holographic recording material e.g.,.Polaroid, DMP-128 whose microstructure after processing

contains a distribution of sub-micron pores correspondingto a variation of refractive index. The porosity permitsbirefringent liquid crystals to be infused into the material,filling the empty spaces. By selecting the appropriateliquid crystal material, the right index modulation can beachieved which in conjunction with the proper gratingthickness and period, and incident linear polarization, re-sults in a thick grating that can strongly diffract light into

w xthe first order for the correct Bragg angle 30 . By reorient-ing the liquid crystal molecules in an electrical field, theeffective index of the nematic liquid crystal filled poreobtains a value that is close to the index of refraction ofthe adjacent polymer areas and thus the index modulationvanishes and therefore the diffraction grating vanishes.This applied electrical field is of the order of ;200–400

V across a 8 mm thick material. This is a rather highvoltage and it is currently limited due to the large surfaceinteraction of liquid crystal molecules in the sub-micron

w xpores 23 . Another important class of holographic PDLCgratings are based on a simple single step fabrication

Žtechnique, the polymerization-induced-phase separation or.PIPS process. In the PIPS process, the grating is formed

by phase separation of the liquid crystal during holo-graphic curing of the photopolymer. This is done using ahomogeneous mixture of a photopolymer and a liquid

w xcrystal 31 . The chemical potential of the system changesas the photopolymer cures. This increases the miscibility

w xgap between the liquid crystal and its host 32 . The liquidcrystal therefore separates as a microdroplet phase and anindex modulation between the polymer and the liquidcrystal droplets can be realized. Thus, a grating thatstrongly diffracts into the first order can be formed if theright combination of grating thickness, period, and state ofpolarization of the input light are selected.

Ž .The ordinary refractive index n of the liquid crystalo

is chosen to match the refractive index of the surroundingŽ . Ž .medium n . The extraordinary index n differs signifi-p e

w xcantly 32 . When the electrically controlled refractiveindex matches the index of refraction of the host polymer,the index modulation vanishes and thus the grating also

Ž .Fig. 2. The PDLC as a polarization selective hologram a horizontally polarized input ‘sees’ the grating and is deflected into the first order,Ž .b vertically polarized input does not ‘see’ the grating and passes through unaffected.


vanishes. Light then travels straight through the devicewithout any deflection. When the applied voltage is re-moved, the liquid crystal molecules reorient back to theirinitial state, again creating an effective index modulationthat significantly differs from the polymer host index ofrefraction. For a particular polarization and direction oflight incident at the Bragg angle, the PDLC device diffractslight into the first order for one of its states and leaves thelight unaffected for the other state.

As mentioned earlier, the thick gratings are in generalw xpolarization dependent 30 . In the case of PDLC gratings,

the polarization dependence is even more evident due tothe liquid crystal birefringence. This is possible when theliquid crystal is infused into the photopolymer pores withthe molecular director aligned such that with no electricfield applied across the PDLC device, the PDLC acts as agrating for one polarization, and as a glass plate for theother orthogonal polarization. This can be obtained if theliquid crystal molecular directors are aligned parallel to the

Ž .fringe direction Fig. 2 . Thus, light with polarizationŽ .parallel to the fringe direction i.e., horizontal will ‘see’

an index modulation created by the index of refraction ofŽ .the polymer host n and the extraordinary index ofp

Ž . Ž .refraction n of the crystal Fig. 2a . On the other hand,e

vertically polarized light does not ‘see’ any index modula-tion. This is because vertically polarized light ‘sees’ theordinary index of refraction of the liquid crystal whichalmost matches the index of refraction of the polymer hostŽ .Fig. 2b . In our experiments, we used two PDLC basedgratings fabricated by Foster–Miller. The polymer hostwas Polaroid DMP-128 and the liquid crystal was E7.With no electric field applied and for incident light at

Ž .Bragg angle ;308 and polarization along the gratingfringe direction, the PDLC devices diffract the light at an

Žangle of 608 with respect to the zero order beam for our.PDLC devices this happens for horizontal polarized light

Ž .see Fig. 2 . On the other hand, if the incident polarizationŽis perpendicular to the grating fringe direction i.e., verti-

.cal polarization , the light passes through the PDLC unaf-Ž .fected Fig. 2 . Note that due to the small mismatch

between n and n , there is some small vertically polar-p o

ized leakage in the first order. There is also a horizontallypolarized leakage in the zero order beam, due to thenon-optimized index difference between the polymer hostand the extraordinary index of refraction of the liquidcrystal.

4. Experimental demonstration of the polarization se-lective hologram-based photonic delay line

Fig. 3 depicts the experimental set-up for our proposedPSH-based PDL. A Lasertron model QLINK1-051 mi-

Ž .crowave fiber-optic transmitter ls1310 nm is used asthe laser source. The modulated light is coupled into the

Ž .PDL via a single mode fiber SMF connectorized to aŽ .gradient index GRIN lens. A mechanical fiber-optic

polarization controller is used to make the input polariza-Ž .tion horizontal. A high extinction ratio 10,000:1 polarizer

is also used to suppress any polarization leakage at theother orthogonal state. A FLC device is used as a polariza-tion switch. This switch acts as an electrically controlledhalf wave plate. When the switch is set ‘off’, the SOP ofthe light does not change. As mentioned earlier, the PDLCdevice diffracts the light when the incident polarization ishorizontal since the light ‘sees’ the grating. The firstdiffracted order is at 608 with respect to the zero orderbeam. After reflection from the mirrors, the light hits the

Fig. 3. The experimental set-up of the proposed PDL using a FLC device as a polarization switch and PDLC devices as polarizationselective holograms. Dashed lines: delay path; Solid lines: non-delay path; SMF: Single mode fiber; RF: radio-frequency.


second PDLC device with an angle of 608 with respect tothe non-delay path. The light is then diffracted towards theoutput of the PDL. On the other hand, when the FLCpolarization switch is set ‘on’, the polarization of the lightis rotated by 908. Hence, the light does not ‘see’ thegrating in PDLC1 device and passes straight through to-wards the PDLC2 device that also acts as a flat glass platefor the vertically polarized light. Thus, the input lightpropagates towards the output port of the PDL.

5. Leakage noise measurements and system improve-ments

An important issue of our PDL is the optical signal-to-Ž .leakage noise ratio SNR . The optical SNR is defined as

Ž .10 log signal powerrleakage noise power . As signal, wedefine the optical power in the optical beam that travelsthrough the desired delay or non-delay path of the module;all other optical power measured at the output is regardedas leakage noise optical power. Note that the optical SNRis a measure of the leakage noise in the system due to thenon-optimized PDLC devices diffraction efficiency as wellas leakage from the FLC switch.

Optical SNR measurements were obtained by indepen-dently measuring the signal and leakage noise of each PDLsetting at the output of the PDL bit. For example, for thedelay setting, the noise was measured by physically block-ing the signal traveling through the delay path, while thesignal was measured by physically blocking the noisetraveling through the non-delay path. The optical signalwas detected at the output of the PDL bit using a large

Ž .area detector 1 mm in diameter and a power meter. Table1 shows the optical SNR for the non-delay and delaysetting of our PDL.

These optical SNR numbers are rather limited. Thereare two reasons for this limited PDL performance. Thefirst reason is that PDLC devices are not 100% efficient.For example, the PDLC1 device diffracts some part of thevertically polarized signal into the delay path, and some ofthe horizontally polarized light into the non-delay path.This ends up to be part of the leakage noise at the outputof our PDL. We will be referring to this noise as PDLC-based leakage noise. The second source of noise in oursystem is the FLC device and its limited onroff perfor-mance. We have seen that today’s FLC devices do notfully rotate the incident polarization by 908 when they are

Table 1ŽOptical SNR measurements for the two PDL settings without

.noise reduction

Ž .PDL setting Optical SNR dB

Non-delay 22.0Delay 16.7

w xset ‘on’ 10 . This means that when the FLC device is set‘on’, unwanted horizontal polarization leaks through thedelay path and contributes to the output leakage noise.Moreover, when the FLC device is set ‘off’ it does de-grade the SOP of the incident polarization and thus there isa vertical component which leaks through the non-delaypath. We will be referring to this noise as FLC switch-basedleakage noise. Optical SNR numbers of )30 dB arerequired for most PDL applications. Thus, a way of im-proving the system performance is necessary. Two differ-ent ways of improving the system performance were testedand are discussed in the following paragraphs.

5.1. PassiÕe leakage noise filter

As mentioned earlier, one source of the leakage noise isthe PDLC devices. This limitation is related with notgetting high enough diffraction efficiency from the PDLCdevices for horizontal polarized light, as well as the non-zero diffraction efficiency for vertically polarized light.This is due to the non-optimized index modulation. This isa fabrication process limitation and can be improved by

w xcareful fabrication techniques. Sutherland et. al. 33 , havesuggested that the diffraction efficiency and leakage noisecan be improved by appropriate PDLC device engineering.This can be accomplished by adjusting PDLC parameterssuch as droplet size, volume fraction of droplets, anddevice thickness by using material additives to the initialpolymer syrup .

For our PDLC1 device and for vertically polarizedinput incident light onto the device, all the light is ex-pected to pass through the device with no diffraction in thefirst order beam. Nevertheless, a 3% diffraction is ob-served into the first order beam, which translates intoPDLC-based leakage noise in our PDL. Note that thisleakage noise is vertically polarized. Thus, the use of ahorizontal polarizer can block the leakage traveling throughthe delay path. In a similar way, for horizontally polarizedlight incident onto PDLC1, a 100% diffraction efficiencyin the first order is expected. In this case, 89% of the inputlight is diffracted and 11% stays in the zero order beam.This rather large leakage noise travels through the non-de-lay path and significantly affects the PDL SNR perfor-mance. Note that the leakage noise is horizontal and thus avertical polarizer would block the leakage. Note also thatin both cases, the polarization of the signal travelingthrough the desired path is parallel to the axis of thepolarizer placed in the path and thus will not be affectedby the polarizers. Thus, a vertical and a horizontal polar-izer were positioned along the non-delay and the delaypaths, respectively. Table 2 shows the optical SNR mea-surements obtained using this passive noise filter. FromTable 2, we can conclude that the use of the passive noisefilter suppresses the noise leakage in the delay setting by;10 dB. In the non-delay setting, we do not observe such


Table 2Optical SNR for all the different leakage noise filtering approaches


Without noise Passive noise Active noise With passive andfilter filter filter active noise filter

Non-delay 22.0 23.3 33.8 44.0Delay 16.7 25.1 17.0 48.0

The SNR without any noise filter is also shown for comparison.

an improvement basically because the optical SNR numberis already )20 dB. Based on the measured PDLC diffrac-tion efficiency on the first order and the bypass efficiencyon the zero order beam, the theoretically expected opticalSNR, without taking into account the FLC switch-basedleakage noise, is 26 and 17 dB for the non-delay and thedelay settings, respectively. These numbers are close toour experimentally obtained ones. The passive noise filtergives optical SNR numbers of 23 and 25 dB, respectively.This is still not adequate for advanced phased array an-tenna applications or other PDL applications that requireoptical SNRs of )30 dB.

5.2. ActiÕe noise filter

The other source of leakage noise in our PDL is theFLC polarization switch. The FLC switch does not fullyrotate the polarization of the incident light when it is set

w x‘on’ 10 . This causes the horizontally polarized leakagenoise to be deflected by the PDLC1 device in the delaypath. Moreover, when the FLC switch is set ‘off’, the SOPof the light in not maintained at the high input polarizationextinction ratio. This vertically polarized FLC-based leak-age does not ‘see’ the grating on the PDLC1 device andtravels through the non-delay path unaffected. In bothcases, signal and FLC-switch based leakage noise are oforthogonal polarizations at the output of the PDL. Thus, byusing an additional FLC device and a vertical polarizer atthe output of the PDL, we can suppress this FLC switch-based leakage noise. For example, for the delay setting,horizontally polarized signal travels through the delaypath, and vertically polarized FLC switch-based leakagenoise travels through the non-delay path. The output FLCis set ‘on’ and thus rotates the polarization of both thesignal and the leakage noise. The vertical polarizer is thenused to block the leakage noise and pass through thesignal. In the non-delay setting, FLC2 is set ‘off’ and thusthe vertically polarized signal from the non-delay pathremains unaffected and passes through the polarizer, whilethe horizontally polarized FLC switch-based leakage-noiseis blocked. Note that the two FLC devices operate inopposite modes, i.e., when FLC1 is ‘on’, FLC2 is ‘off’ andvice versa. This is because the two states of the FLC

w xdevices do not perform equally well 10 . The SNR mea-surements of the PDL system using the active noise filter

are shown in Table 2. Using the active noise filter, weimproved the non-delay setting by )10 dB, but the delay

Ž .setting SNR remains at low levels e.g., -20 dB . Thislimited improvement for the delay setting is due to thevertically polarized PDLC-based leakage noise that eventu-ally passes through the active noise filter and contributesto a low SNR. The FLC2 switch also contributes to thisnoise since it is set in its ‘on’ state and there is some FLCswitch-based leakage.

5.3. Combination of the actiÕe and passiÕe noise filter

From the PDL optical SNR results obtained from thepassive and active noise filters, we see that each approachimproves the SNR for only one of the two settings. If wewere to use both of the filtering methods simultaneously,we would obtain higher optical SNR numbers for bothPDL settings. The experimental set-up showing the combi-nation of the two noise filters is depicted in Fig. 4. Thepassive noise filter suppresses the PDLC-based leakagenoise, while the active noise filter suppresses the FLCswitch-based leakage noise. Optical SNR measurements of)40 dB were obtained for both PDL settings. Theseoptical SNR numbers are highly desirable for PDL applica-tions such as phased array antennas.

5.4. ‘Orthogonal driÕe’ PDLC deÕice configuration

The current limitation in our PDL system is mainly thelow diffraction efficiency of our PDLC devices in the first

Ž .order e.g., ;89% . Note also that in any of the two PDLsettings, both of the PDLC devices either deflect the signalor bypass it. The difference between the bypass efficiency

Ž .in the zero-order beam ;97% and the diffraction effi-Ž .ciency ;89% in the first-order beam also leads to an

unbalanced SNR performance for the two PDL settings.w xWe propose an ‘orthogonal drive’ 34 , PDLC device

configuration to compensate for that unbalanced perfor-mance. This ‘orthogonal drive’ device configuration was

w xfirst proposed and demonstrated in Ref. 7 to compensatefor the poorer performance of the FLC devices when theyare set in their ‘on’ state compared with when they are setin their ‘off’ state. For our case, this ‘orthogonal drive’PDLC device configuration is obtained by setting one ofthe PDLC devices to diffract in the first order and the


Fig. 4. The PDL experimental set-up showing the passive and active noise filters. P: polarizer; GRIN: gradient index lens; SMF: singlemode fiber; L: lenses; LD: semiconductor laser.

other device to bypass in the zero order for each of thePDL settings. This can be accomplished by fabricating thetwo PDLC devices such that one device diffracts thevertical polarization and the other device diffracts thehorizontal polarization. Nevertheless, this approach willlead to increased expense and fabrication time, since twodifferent sets of PDLC devices will have to be fabricated.We choose to use a simpler approach, where we place ahalf wave plate in each of the PDL paths. Thus, thepolarization of the light in any of the paths is rotatedbefore it reaches the PDLC2 device. In this case, a hori-zontal polarization signal coming from PDLC1 devicethrough the delay path is rotated to vertically polarizedlight, and does not ‘see’ the grating in the PDLC2 device.Thus, it passes through the PDLC2 device unaffected.Similarly, vertically polarized light coming from PDLC1device through the non-delay path is rotated to horizontallypolarized light, and is deflected by the PDLC2 device.

Optical SNR measurements were obtained for this PDLthat makes use of the ‘orthogonal drive’ configuration andare shown in Table 3. Note that we actually use thisorthogonal drive configuration twice in our system; oncefor the PDLC devices and once for the FLC devices. From

Table 3, we can conclude that the ‘orthogonal drive’configuration improves the overall SNR performance forour PDL even when no noise filter technique is used. Notethat the SNR numbers of our PDL are still not high enoughwhen only one of the noise filter techniques is used. Thisis mainly due to the low diffraction efficiency into the firstorder beam of the PDLC devices. Note that if the perfor-mance of the PDLC devices is improved such that they cangive 99% diffraction in the first order for the horizontallypolarized light and a 99% bypass efficiency in the zeroorder for the vertically polarized light, then the theoreticalexpected SNR would approach 40 dB with only one of thenoise filters.

6. Time delay measurements

Time delay measurements were also obtained for thesingle bit PSH-based PDL. Fig. 5 shows oscilloscope

Žtraces of the non-delayed and the delayed signal bottom.traces . The top trace represents the reference signal from

the signal generator. A fiber pigtailed fast photodetector

Table 3Optical SNR for the PDL with the orthogonal drive configuration and for all the different leakage noise filter approaches


Without noise Passive noise Active noise With passive andfilter filter filter active noise filter

Non-delay 21.3 22.7 26.0 46.3Delay 22.7 26.4 24.9 48.0

The SNR without any noise filter is also shown for comparison.


Ž . Ž .Fig. 5. Oscilloscope traces showing a the non-delayed photodetected signal and b the delayed photodetected signal. Top traces: referencesignal from the oscilloscope. Bottom traces: the photodetected output signal.

Ž .New Focus, Model: 1414-50 was used to detect themodulated optical signal. The fiber used was a multi-mode

Ž .fiber 50 mm core diameter connectorized to a GRIN lens.Fig. 5a shows the non-delayed signal, with a relative delayfrom the reference signal of 36.28 ns. The time markershave been positioned at the on-set of the traces. As on-settime, we define the time when the pulse gets to 10% of itsmaximum value. Fig. 5b shows the delayed signal, with arelative delay from the reference signal of 40.34 ns. Therelative time delay between the two photodetected signalsis the time delay obtained from our PDL bit and iscalculated to be 40.34 nsy36.28 nss4.08 ns. The ex-pected time delay can be estimated from the optical pathlength difference between the two paths. The PDLC-to-PDLC distance for the non-delay path is 29 cm. ThePDLC-to-PDLC distance for the delay path is 152 cm.Thus, the expected time delay can be found from thefollowing equation:

Delay Path y NonyDelay PathŽ . Ž .Dts

c

153=10y2 y29=10y2 mŽ .s s4.10 ns. 1Ž .83=10 mrs

The 0.02 ns difference between the expected and themeasured time delay is due to measurement errors in thepath lengths as well as the tolerance in the position of thetime markers on the oscilloscope screen.

7. Insertion loss of the PDLC-based photonic delay line

An important issue of the PDL is the system insertionloss. Recently, we have discussed insertion loss issues forour previously proposed PDL systems due to the bulkoptical components in PDL structures as well as the effect

w xof the FLC devices in the loss 35 . The main contributor

to the insertion loss is the FLC devices that are currentlyŽlimited to an average of 85% transmission efficiency or

.15% optical insertion loss for operation at 1300 nm. Thenew element in the PSH-based PDL is the PDLC devices.The device insertion losses are due to reflection, absorp-tion and scattering. Reflection losses can be minimizedusing anti-reflection coatings on the glass substrates. Ab-sorption loss is rather minimal and the primary contributor

w xto the PDLC loss is scattering 33 . Studies of the liquidcrystal droplet size have shown that scattering can be

Žreduced using small droplets e.g., droplet diameter 0.04. w xmm 33 .Our PDLC1 and PDLC2 devices had insertion loss of

Ž . Ž .0.8 dB or 17% and 0.7 dB or 15% , respectively. Theoverall insertion loss of the PDL when both noise filtersare used was measured at 2.9 dB. The use of the activenoise filter contributes a 0.7 dB to the total PDL insertionloss, while the passive noise filter contributes only a 0.07dB to the PDL insertion loss. However, the great improve-ment obtained for the optical SNR when using the noisefilters justifies the use of the noise filters eventhough thePDL shows an increased insertion loss. Nevertheless, asnew FLC materials are developed and studied, reducedinsertion loss numbers for the FLC switches are expected.Improved insertion loss is also expected from the PDLCdevices as the PDLC technology matures.

8. A compact photonic delay line architecture based onpolarization selective holograms

8.1. Experimental set-up

The overall size of the PSH-based PDL bit can bereduced if a reflective design is used. Fig. 6 shows such areflective design. A mirror is positioned in each of the twoPDL paths. These mirrors reflect back the light onto the


Fig. 6. The compact reflective photonic delay line architecture based on polarization selective holograms.

PDLC device through exactly the same path. A quarterŽ .wave plate QWP is also used in each path so that the

polarization of the light is rotated by 908 after passingthrough the QWP twice. Thus, light is directed towards theoutput of the PDL bit. This reflective design uses half theoptical path compared with the transmissive design dis-cussed earlier. It also uses fewer optical components since

Žit reuses the lenses and the PSH in our case the PDLC.device .

The compact reflective PSH-based PDL works as fol-lows. When PS1 is set ‘on’, it rotates the incident horizon-tal polarization by 908. This vertically polarized lightpasses through the PSH unaffected. After going throughQWP1 twice, it is rotated back to horizontally polarizedlight and this time ‘sees’ the grating in the PSH. Thus, it isdiffracted towards the output port of the PDL. When thePS1 is set in its ‘off’ state it leaves the incident polariza-tion unaffected. The horizontally polarized light is thendiffracted in the first order beam and after passing throughthe QWP2 twice, it changes to vertically polarized light.This vertically polarized light does not ‘see’ the grating inthe PSH and thus passes through it unaffected towards theoutput of the PDL. Note that the ‘orthogonal drive’ config-uration comes naturally in the reflective PDL design be-cause it is needed to separate the output from the inputport.

8.2. Leakage noise measurements

A limitation of this reflective PSH-based PDL architec-ture is that no passive noise filter can be used since the

polarization of the signal traveling in any of the two pathsexists in both orthogonal polarizations. This limits theoptical SNR performance for our PDL using the PDLCdevices available in our lab. Nevertheless as mentionedearlier, higher diffraction efficiency PDLC devices can beobtained and thus better SNR numbers can be reached.However, the active noise filter can be used to furthersuppress the FLC1 switch-based leakage noise.

Optical SNR measurements were obtained for this com-pact reflective PDLC-based PDL and are shown in Table4. Two sets of measurements were obtained, one withoutany noise filter and one with the active noise filter. OpticalSNRs )27 dB are obtained for both settings. This closeto 30 dB numbers can potentially exceed the 30 dB levelby using improved PDLC devices with higher diffractionefficiency.

8.3. Time delay measurements

Time delay measurements were also obtained for thecompact reflective single bit PSH-based PDL. Fig. 7 shows

Table 4Optical SNR for the compact reflective PDL with and without theactive noise filter

PDL setting Optical SNR

Without noise Active noisefilter filter

Non-delay 19.4 28.7Delay 20.1 27.4


Ž . Ž .Fig. 7. Oscilloscope traces showing a the non-delayed photodetected signal and b the delayed photodetected signal for the compactreflective PDLC-based PDL. Top traces: reference signal from the oscilloscope. Bottom traces: the photodetected output signal.

oscilloscope traces of the non-delayed and the delayedŽ .signal bottom traces . The same technique as in the case

of the transmissive PSH-based PDL was used. Fig. 7ashows the non-delayed signal, with a relative delay fromthe reference signal of 34.28 ns. Fig. 7b shows the delayedsignal, with a relative delay from the reference signal of38.36 ns. The relative time delay between the two photode-tected signals is the time delay obtained from our PDL bitand is calculated to be 38.36 nsy34.28 nss4.06 ns. Theexpected time delay is again 4.10 ns.

9. Alternative polarization selective hologram designs

A unique feature of our PSH-based PDL is that it usesthe mature, low cost liquid crystal technology for theimplementation of the polarization switches and standardholographic techniques for the PSH devices. In our experi-mental demonstration, we used PDLC devices as PSH. Analternative PSH technique is based on computer generated

Ž . w xholograms CGH 14 . The maximum possible diffractionefficiency from a CGH depends on the accuracy withwhich the theoretically continuous phase values are actu-ally fabricated. Theoretically, diffraction efficiencies of98.7% can be obtained for a 16 level phase CGH, and

w x99.7% for a 32-level phase CGH 14 . Standard litho-graphic techniques used for microelectronic fabrication canbe used to create an accurate multilevel phase profile withhigh spatial resolution. These CGHs are generally polariza-tion independent but a technique of making polarizationdependent CGH has been reported and demonstrated usinga birefringent substrate to make a four level birefringent

w xCGH with a diffraction efficiency of 60% 14 . Sixteen or32 level phase birefringent CGHs that show higher diffrac-

tion efficiencies can also be used to implement a PSH-basedPDL.

As mentioned earlier, thick gratings are polarizationw xsensitive 30 . Thus, thick gratings can also be used as

PSH. For a transmission type of phase volume hologram, ithas been shown that there exists a condition where the

Ž .diffraction efficiency of one polarization s or p is 100%and the diffraction efficiency of the orthogonal polariza-

Ž . w xtion p or s is 0% 36 . It has been reported that apolarization selective hologram has been recorded inDMP-128 photopolymer with a normalized diffraction effi-ciency of 99% for s-polarized light and 1% for p-polarized

w xlight 37 . The diffraction efficiency of the volume holo-grams depends on the maximum index modulation that canbe achieved for a material, the wavelength, the gratingperiod and the thickness. Usually the thickness is materialdependent and the wavelength is set by the system require-ments and use. Thus, the two factors important for a PSHare the index modulation that can be controlled by therecording condition, and the possible Bragg angle anddiffraction angles.

Another technique for implementing polarization selec-tive beam routing elements is based on the mature nematicliquid crystal technology. Due to the capability of NLCmaterials to be controlled by low electrical voltages, recon-figurable beam routing elements can also be formed. Liq-uid crystal based phase gratings for high efficiency light

w xvalves has been proposed as early as 1979 38 . Pro-grammable liquid crystal devices for variable focal-length

w x w xlenses 15,16 , and for adaptive optical interconnects 39 ,w xand alignment 40 , applications have been proposed and

demonstrated. For our PDL application, theseŽ .birefringent-mode BM NLC devices can be set such that

the liquid crystal nematic director is along one of the twoorthogonal polarizations. Hence, the index of refraction


Fig. 8. A PSH via a birefringent mode nematic liquid crystal based device used as a polarization dependent diffraction grating.

has a specific phase profile for one of the polarizations andŽ .a uniform index of refraction no index modulation for the

other polarization. Thus, only one of the incident linearpolarization will be deflected to a predefined direction.These BM NLC-based beam routing elements can beeither active or passive.

Fig. 8 shows a BM-NLC device that acts as a polariza-tion dependent grating that strongly diffracts light in thefirst order for one polarization and does not diffract thelight for the other orthogonal polarization. The liquidcrystal molecules are oriented in layers such that horizon-tally polarized light ‘sees’ a maximum index modulationn yn while vertically polarized light ‘sees’ a uniforme o

Ž .index of refraction n . Transparent indium tin oxideo

Ž .ITO electrodes are properly spaced to give the requiredgrating spacing for maximum diffraction efficiency.

An alternative BM-NLC based routing element will beone based on a ramp like index of refraction profile andhence an induced ramp like optical phase shift to theincident beam. This approach can make use of a thin

w xfilm-resistor network 41 , on the device substrate layer tocontrol the voltages of the independent electrodes by theuse of only one external driver. This technique gives a nearcontinuous index pertubation to match the prism like phase

w xpertubation required for the beam deflection 41 . Fig. 9shows a possible configuration for the this BM NLCdevice using on-chip resistor-based control electronics. Agradual tilt of the liquid crystal molecules gives the re-

Fig. 9. Top view of a PSH formed with a thin-film-resistor network based NLC deflector.


quired index of refraction profile to the device and thus forhorizontal light the devices act as a beam deflector.

10. Conclusion

We have proposed and experimentally demonstrated aphotonic delay line based on FLC devices for polarizationswitching and PDLC devices as polarization selective opti-cal path routing components. Extensive investigation of theleakage noise in the system was performed, and twodifferent leakage noise filters were investigated for im-proving the SNR numbers. Improved SNR performanceŽ .)45 dB was obtained by combining both noise filtersand using the ‘orthogonal drive’ configuration. Time delaymeasurements were also performed for our single bit sin-gle channel PDL. An alternative reflective architecture wasproposed. This reflective architecture uses fewer numberof optical components. In addition, the propagation of lighttwice through the two PDL paths also gives a reduced sizefor our reflective PDL architecture. PSH based on birefrin-gent CGHs, phase volume holograms, and BM-NLC de-vices for PDL applications were also proposed. Futurework relates with experiments using improved diffractionefficiency PDLC devices. PDL architectures based onalternative polarization selective holograms, such as com-puter generated holograms, or NLC-based routing elementswill also be investigated.

Acknowledgements

The authors would like to acknowledge Dr. Larry Do-mash and Conrad Gozewski of Foster Miller for providingthe PDLC devices and for their useful and helpful discus-sions. This work is partially supported by grantaN000149510988 from the Office of Naval Research.

References

w x Ž .1 N.A. Riza Ed. , Selected Papers on Photonic Control Sys-tems for Phased Array Antennas, SPIE Milestone Series,Vol. MS 136, June, 1997.

w x2 X.S. Yao, L. Maleki, A novel 2-D programmable photonictime-delay device for millimeter-wave signal processing ap-

Ž . Ž .plications, IEEE Photonic Technology Letters 6 12 19941463–1465.

w x3 D. Dolfi, P. Joffre, J. Antoine, J.-P. Huignard, D. Philippet,P. Granger, Experimental demonstration of a phased arrayantenna optically controlled with phase and time delays,

Ž . Ž .Applied Optics 35 26 1996 5293–5300.w x4 Y. Chang, B. Tsap, H.R. Fetterman, P.A. Cohen, A.F.J. Levi,

I.L. Newberg, Optically controlled serially fed phased arrayŽ .transmitter, IEEE Microwave Guided Wave Letters 7 3

Ž .1997 69–71.

w x5 N.A. Riza, Multichannel variable optical control systems forŽ .large coherent optical arrays, in: G.W. Kamerman Ed. ,

Laser Radar Technology and Applications, Vol. 2748, April,1996, pp. 333–339.

w x6 N.A. Riza, Advanced novel photonic instrumentation foradaptive and interferometric astronomy, in: P.Y. Bely, J.B.

Ž .Breckinridge Eds. , Space Telescopes and Instruments IV,Vol. 2807, SPIE Proc., pp. 335–342.

w x7 N.A. Riza, Photonically controlled ultrasonic arrays: scenar-ios and systems, in: M. Levy, S.C. Schneider, B.R. McAvoyŽ .Eds. , IEEE Ultrasonics Symposium Proceedings, Vol. 2,1996, pp. 1545–1550.

w x8 J. Fu, M. Schamschula, H.J. Caulfield, Modular solid opticŽ .time delay system, Optics Communication 121 1995 8–12.

w x9 N.A. Riza, Liquid crystal-based optical time delay units forphased array antennas, IEEErOSA Journal of Lightwave

Ž .Technology 12 1994 1440–1447.w x10 N.A. Riza, N. Madamopoulos, Characterization of a ferro-

electric liquid crystal-based time delay unit for phased arrayantennas, IEEErOSA Journal of Lightwave Technology 15Ž .1997 1088–1094.

w x11 T. Todorov, L. Nikolova, K. Stoyanova, N. Tomova, Polar-ization holography: III. Some applications of polarization

Ž .holographic recording, Applied Optics 24 1985 785–788.w x12 R. Kostuk, M. Kato, Y.T. Huang, Polarization properties of

substrate-mode holographic interconnects, Applied Optics 29Ž .1990 3848–3854.

w x13 Q.W. Song, M.C. Lee, P.J. Talbot, E. Tam, Polarizationsensitivity of birefringent photorefractive holograms and its

Ž .applications to binary switching, Applied Optics 31 19926240–6246.

w x14 F. Xu, J. Ford, Y. Fainman, Polarization-selective computer-generated holograms: design, fabrication, and applications,

Ž .Applied Optics 34 1995 256–366.w x15 S. Sato, Liquid-crystal lens-cells with variable focal length,

Ž . Ž .Japanese Journal of Applied Physics 18 9 1979 1676–1684.

w x16 N.A. Riza, M.C. DeJule, A novel programmable liquid crys-tal lens device for adaptive optical interconnect and beam-forming applications, Optical Computing International Con-ference, Inst. Phys. Conf. No. 139, Part II, Edinburgh, Scot-land, August, 1994, pp. 231–234.

w x17 N.A. Riza, Patent Disclosure, 1996.w x18 T.W. Stone, M.S. Malcuit, J.A. Kleinfeild, J. Kralik, Micro-

optic photonic time shifters based on switched gratings, in:Ž .A.P. Goutzoulis Ed. , Optical Technology for Microwave

Applications VIII, Proc. SPIE, Vol. 3160, 1997, pp. 17–26.w x19 N.A. Riza, N. Madamopoulos, Photonic delay line using

electrically switched gratings in polymer dispersed liquidŽ . Ž .crystals, Optical Engineering 37 11 1998 3061–3065.

w x20 N. Madamopoulos, N.A. Riza, Photonic delay lines usingpolarization selective holograms for optical signal routing, in:

Ž .D.R. Pape Ed. , Advances in Optical Information ProcessingVIII, SPIE Proceedings, Vol. 3388, No. 16, April 1998.

w x21 N.A. Riza, 25 Channel nematic liquid crystal optical time-de-lay unit characterization, IEEE Photonics Technology LettersŽ . Ž .7 11 1995 1285–1287.

w x22 R.T. Ingwall, M. Troll, Mechanism of hologram formation inŽ . Ž .DMP-128 photopolymer, Optical Engineering 28 6 1989

586–591.w x23 L. Domash, C. Gozewski, A. Nelson, J. Schwartz, Pro-


grammable beamlet generator, dynamic lens, and opticalmemory using electrically switched holographic devices, in:

Ž .J. Lottorner, B. Javidi, S.T. Kowel, W. Miceli Eds. , Pho-tonics for Processors, Neural Networks, and Memories, Proc.SPIE, Vol. 2026c, 1993, pp. 642–652.

w x24 K. Takizawa, H. Kikuchi, H. Fujikake, Y. Namikawa, K.Tada, Polymer-dispersed liquid crystal light valves for pro-

Ž .jection display, Optical Engineering 32 1993 1781–1791.w x25 S.J. Klosowicz, J. Zmija, Optics and electro-optics of poly-

mer-dispersed liquid crystals: physics, technology, and appli-Ž .cation, Optical Engineering 34 1995 3440–3450.

w x26 V.P. Tondiglia, L.V. Natarajan, R.L. Sutherland, T.J. Bun-ning, W.W. Adams, Volume holographic image storage andelectro-optical readout in a polymer-dispersed liquid–crystal

Ž . Ž .film, Optics Letters 20 11 1995 1325–1327.w x27 L.H. Domash, Y.-M. Chen, B. Gomatam, C. Gozewski, R.L.

Sutherland, L.V. Natarajan, V.P. Tondiglia, T.J. Bunning,W.W. Adams, Switchable-focus lenses in holographic poly-

Ž .mer-dispersed liquid crystal, in: I. Cindrich, S.H. Lee Eds. ,Diffractive and Holographic Optics Technology III, SPIEProc., Vol. 2689, 1996, pp. 188–194.

w x28 S.-X. Cheng, R.-K. Bai, Y.-F. Zou, C.-Y. Pan, Electro-opti-cal properties of polymer dispersed liquid crystal materials,

Ž .Journal of Applied Physics 80 1996 1991–1995.w x29 D. Whitney, R.T. Ingwall, Fabrication and properties of

composite holograms recorded in DMP-128 photopolymer,Ž .in: R.A. Lessard Ed. , Photopolymer Device Physics, Chem-

istry, and Applications, Proceedings SPIE, Vol. 1213, No.18, 1990, pp. 18–26.

w x30 H. Kogelnik, Coupled wave theory for thick hologram grat-Ž . Ž .ings, The Bell System Technical Journal 48 9 1969

2909–2947.w x31 W. Doane, N.A. Vaz, B.-G. Wu, S. Zumer, Field controlled

light scattering from nematic microdroplets, Applied PhysicsŽ . Ž .Letters 48 4 1986 269–271.

w x32 R.L. Sutherland, V.T. Tondiglia, L.V. Natarajan, T.J. Bun-ning, W.W. Adams, Electrically switchable volume holo-

graphic gratings in polymer-dispersed liquid crystals, Ap-Ž . Ž .plied Physics Letters 64 9 1994 1074–1076.

w x33 L. Sutherland, L.V. Natarajan, V.P. Tondiglia, T.J. Bunning,B.L. Epling, D.M. Brandelik, Relation of electro-opticalcharacteristics to materials properties and morphology inpolymer-dispersed liquid crystal holographic gratings, in: I.

Ž .Cindrich, S.H. Lee, Eds. , Diffractive and Holographic De-vice Technologies and Applications IV, SPIE Proceedings,Vol. 3010, 1997, pp. 142–149.

w x34 N.A. Riza, J. Chen, Ultra-high y47 dB optical drop rejec-tion multi-wavelength add-drop filter using spatial filteringand dual bulk acousto-optic tunable filters, Optics Letters 23Ž . Ž .12 1998 945–947.

w x35 N. Madamopoulos, N.A. Riza, Directly modulated semicon-ductor laser fed photonic delay line using ferroelectric liquid

Ž . Ž .crystals, Applied Optics 37 8 1998 1407–1416.w x36 Y.-T. Huang, Polarization-selective volume holograms: gen-

Ž . Ž .eral design, Applied Optics 33 11 1994 2115–2120.w x37 J.J. Butler, M.S. Malcuit, T.W. Stone, Polarization-sensitive

holograms formed with use of DMP-128 photopolymer, Op-toelectronic Interconnect, Switching, Processing and Storage,CLEO’97 Proceedings, CWF53, Baltimore, May, 1997, pp.257–258.

w x38 Y. Hori, K. Asai, M. Fukai, Field controllable liquid–crystalphase grating, IEEE Transactions on Electronic Devices ED26Ž . Ž .11 1979 1734–1737.

w x39 N.A. Riza, S. Yuan, Robust packaging of photonic rf mod-ules using ultrathin adaptive optical interconnect devices, in:

Ž .A.P. Goutzoulis Ed. , Optical Technology for MicrowaveApplications VIII, Proc. SPIE Vol. 3160, 1997, pp. 170–177.

w x40 N.A. Riza, S. Yuan, Demonstration of a liquid crystal adap-tive alignment tweeker for high speed infrared band fiber-fed

Ž . Ž .free-space systems, Optical Engineering 37 6 1998 1876–1880.

w x41 N.A. Riza, M.C. DeJule, Three-terminal adaptive nematicŽ . Ž .liquid crystal lens device, Optics Letters 19 14 1994

1013–1015.

polarization selective hologram-based photonic delay lines

Documents