Integrated Optics and Guided Waves- a Report of the Topical Meeting R. V. Pole, S. E. Miller, J. H.. Harris, and P. K. Tien A report of the first Topical Meeting on Integrated Optics-Guided Waves, Materials, and Devices has been prepared by four members of the Technical Program Committee. The meeting was held 7-10 February 1972 and was sponsored by the Lasers Technical Group of the Optical Society of America, the International Commission for Optics, and the International Union of Pure and Applied Physics. Foreword (P. K. Tien) Thanks to the earnest planning and preparation of R. V. Pole and J. W. Quinn, the first topical meeting on integrated optics-attended by 268 scientists from 11 countries-was held in Las Vegas on 7-10 February 1972. Fifty-four papers were presented concerning mainly fiber optics, thin-film passive and active optical devices,materials, and fabrication of smallstructures. Although the concept of processing optical signals in lightguides can be traced back to the literature of the early 1960's, the importance of integrated optics was not fully realized until 1968. Several papers' published since 1968 stressed the needs of thin film technology for fabrication and miniaturization of optical circuits. In 1969-1970, the prism-film and grating couplers 2 were developed, and they triggered many experiments involv- ing light propagation in thin films. Since then, the field has grown rapidly; about fifty papers were pub- lished in the past three years. The field of integrated optics, as we can visualize today, involves the development and research of various thin-film optical devices and circuits that may be used as integrated elements either to broaden the present capability of microelectronics or to process optical signals transmitted by glass fibers. It became apparent during the meeting that in this new and complex field, "so many deeds cry out to be done." The ideas and the experiments presented in this meeting marked only the beginning of integrated optics. We thought it would be important to record some of these initial efforts which will be the stepping stones to developments of tomorrow. R. V. Pole is with IBM Watson Research Center, P. 0. Box 218, Yorktown Heights, New York 10598; J. H. Harris is with the University of Washington, Department of Electrical Engi- neering, Seattle, Washington 98105; the other authors are with Bell Telephone Laboratories, Murray Hill, New Jersey 07974. Received 6 April 1972. Many participated in this report: R. V. Pole re- viewed the plenary session, materials and fabrication techniques, providing a comprehensive view for the present state of technology. S. E. Miller wrote the section on optical fibers, which is, by itself, an excellent paper. J. H. Harris reported various subjects of prime importance including passive and active optical devices. Many authors supplied the original photographs that brighten the article. This report is a snapshot of the growing field that may, one day, serve both the elec- tronics and communication industries. Plenary Session (R. V. Pole) Early Experiments of Light Guiding Opening the conference was an invited paper by Osterberg. 3 He described some of the early experi- ments in which he observed the transmission of light energy along the surface of thermally tempered glasses. These experiments, in which optical surface waves were obtained due to the higher material density near the surface of such glasses, marked the beginning of the field we now call integrated optics. Single-Crystal Films and a Novel Magnetooptic Switch Among the various subjects surveyed in the next (invited) paper by Tien, 4 the highlights were discussions concerning single-crystal films and his very recent ex- periment with a novel magnetooptic switch. Tien be- gan by emphasizing the needs for single-crystal films in the development of active integrated elements. He then continued: The only useful high-quality single- crystal films available today are Al.Gal-,As and garnet systems. Both systems are cubic and are made presently by liquid phase epitaxy. In integrated optics, we generally require the material of the film to be different from that of the substrate. This involves heteroepitaxy. It is known that for successful heteroepitaxy, the lattice constants of August 1972 / Vol. 11, No. 8 / APPLIED OPTICS 1675
Transcript
Integrated Optics and Guided Waves-a Report of the Topical Meeting
R. V. Pole, S. E. Miller, J. H.. Harris, and P. K. Tien
A report of the first Topical Meeting on Integrated Optics-Guided Waves, Materials, and Devices hasbeen prepared by four members of the Technical Program Committee. The meeting was held 7-10February 1972 and was sponsored by the Lasers Technical Group of the Optical Society of America,the International Commission for Optics, and the International Union of Pure and Applied Physics.
Foreword (P. K. Tien)
Thanks to the earnest planning and preparation of R.V. Pole and J. W. Quinn, the first topical meeting onintegrated optics-attended by 268 scientists from 11countries-was held in Las Vegas on 7-10 February1972. Fifty-four papers were presented concerningmainly fiber optics, thin-film passive and active opticaldevices, materials, and fabrication of small structures.
Although the concept of processing optical signals inlightguides can be traced back to the literature of theearly 1960's, the importance of integrated optics wasnot fully realized until 1968. Several papers' publishedsince 1968 stressed the needs of thin film technology forfabrication and miniaturization of optical circuits. In1969-1970, the prism-film and grating couplers2 weredeveloped, and they triggered many experiments involv-ing light propagation in thin films. Since then, thefield has grown rapidly; about fifty papers were pub-lished in the past three years.
The field of integrated optics, as we can visualizetoday, involves the development and research of variousthin-film optical devices and circuits that may be usedas integrated elements either to broaden the presentcapability of microelectronics or to process opticalsignals transmitted by glass fibers. It became apparentduring the meeting that in this new and complex field,"so many deeds cry out to be done." The ideas and theexperiments presented in this meeting marked only thebeginning of integrated optics. We thought it wouldbe important to record some of these initial effortswhich will be the stepping stones to developments oftomorrow.
R. V. Pole is with IBM Watson Research Center, P. 0. Box218, Yorktown Heights, New York 10598; J. H. Harris is withthe University of Washington, Department of Electrical Engi-neering, Seattle, Washington 98105; the other authors are withBell Telephone Laboratories, Murray Hill, New Jersey 07974.
Received 6 April 1972.
Many participated in this report: R. V. Pole re-viewed the plenary session, materials and fabricationtechniques, providing a comprehensive view for thepresent state of technology. S. E. Miller wrote thesection on optical fibers, which is, by itself, an excellentpaper. J. H. Harris reported various subjects of primeimportance including passive and active optical devices.Many authors supplied the original photographs thatbrighten the article. This report is a snapshot of thegrowing field that may, one day, serve both the elec-tronics and communication industries.
Plenary Session (R. V. Pole)
Early Experiments of Light GuidingOpening the conference was an invited paper by
Osterberg.3 He described some of the early experi-ments in which he observed the transmission of lightenergy along the surface of thermally tempered glasses.These experiments, in which optical surface waves wereobtained due to the higher material density near thesurface of such glasses, marked the beginning of thefield we now call integrated optics.
Single-Crystal Films and a NovelMagnetooptic Switch
Among the various subjects surveyed in the next(invited) paper by Tien,4 the highlights were discussionsconcerning single-crystal films and his very recent ex-periment with a novel magnetooptic switch. Tien be-gan by emphasizing the needs for single-crystal films inthe development of active integrated elements. Hethen continued: The only useful high-quality single-crystal films available today are Al.Gal-,As and garnetsystems. Both systems are cubic and are made presentlyby liquid phase epitaxy. In integrated optics, we generallyrequire the material of the film to be different from that ofthe substrate. This involves heteroepitaxy. It is knownthat for successful heteroepitaxy, the lattice constants of
the film and the substrate must be matched within approxi-mately 0.01 A. Many systems satisfy this stringent con-dition. For example, the difference in lattice constantsbetween AlAs and GaAs is 0.009 A and the differencebetween AlP and GaP is almost none. He then sum-marized recent advances in AlxGal-.As system.
Next, Tien turned to magnetic films. The use ofgarnet films as magnetooptic waveguides has not beendiscussed previously; these films were developed formagnetic bubble devices. The lattice constants of thegarnets can be made continuously variable over a widerange by using different rare-earth ions, and conse-
Fig. 1. A photograph of 0.6328 ,um light propagation in a singlecrystal of europium gallium garnet film (P. K. Tien et al.
4).
DETECTOR
- .5cm-
.
DOUBLE ENDED PRISM
NITROBENZENE
SUBSTRATE
SUBSTRATE
Fig. 2. An interdigital thin film electrooptic modulator (J. N.Polley and J. H. Harris5).
quently, a perfect lattice match can always be obtainedby choosing a garnet of one composition as the film andthat of another composition as the substrate. In oneslide (Fig. 1), Tien showed light propagation in a single-crystal europium gallium garnet film on a gadoliniumscandium aluminum garnet substrate. In anotherslide, he showed all possible film-substrate combina-tions including their refractive indices, for optical wave-guides which can be made from the garnets.
In the second part of the paper, Tien reported anovel magnetooptic light switch4 which involved aniron garnet film as the waveguide. A light wave wasfed into the film as a TM mode. It was converted into aTE mode by means of magnetooptic effect. The con-version between TM and TE modes depended upon thedirection of magnetization which was controlled elec-tronically by a small integrated electrical circuit.
Two Dimensional Optics andElectrooptic Diffraction
In the third invited paper, Harris, discussed aspectsof two dimensional effects in films, the neglected dimen-sion being that normal to the waveguide surface.Through changes in film thickness or overlaying ofmaterials, equivalent mode refractive indices can bevaried in relatively arbitrary spatial fashion within thelimits imposed by scattering due to rapid changes in themodal configuration. Single mode structures can pro-vide index variations in the 0.1-0.5 range commonlyfound in natural and manmade optical systems. Heshowed slides of beam refraction that correlate with two-dimensional ray theory. When uniaxial layers arepresent, propagation exhibits angular cut-off propertiescorresponding to guided wave behavior in some direc-tions and leaky wave behavior in others.
Harris also discussed the importance of the use of thinfilms in permitting application of periodic electric fieldsto waveguides with periods comparable to the opticalwavelength. He described the results of experimentswith electrooptic Bragg diffraction using cross-guidefinger electrodes and interdigital structures (Fig. 2).The active medium was a film of nitrobenzene. Withthe interdigital structure a modulation depth of about50% was reported in the undiffracted beam using a pathlength of 5 mm. The finger separation and width was12 yu and the voltage required was of the order ofhundreds of volts. He indicated that with reductionin finger size a further decrease in required voltage andwide bandwidth operation should be achievable.
Materials and Fabrication Techniques (R. V. Pole)
Material Development andCircuit Manufacturability
The two materials sessions and a session on fabrica-tion techniques were highlighted by an invited papergiven by Chynoweth.6 The speaker began by notingthat integrated optics is roughly at the point integratedelectronics was some fifteen years ago; feasibilitydemonstrations have been made of a number of in-dividual components, both active and passive, but not
much has been done to stitch them together into com-plete operating functional circuits. Adopting the viewthat some form of hybrid rather than monolithic tech-nology is likely to form the basis for optical circuits, thespeaker outlined the four categories of methods of filmpreparation: by growth or deposition on a substrate;by chemical treatment; by radiation treatment of asubstrate; and finally those, such as purely mechanicalmethods, which fall into none of the above categories.
In the second part, the speaker concentrated on suchproblems as manufacturability and reliability of poten-tial integrated optical circuits considering in some detailvarious alternatives, tradeoffs, and constraints.
Electron Beam Lithography and MiniatureOptical Waveguides
Two papers dealt with the important problem ofapplying electron beam lithography and back-sputter-ing techniques to the fabrication of optical waveguides.Ostrowsky and DuBois7 reported~optical waveguidesseveral microns wide (Fig. 3) using polyvinylsiloxane(PVS) as a (negative) resist and a flying-spot scanner-controlled scanning electron microscope as a source ofexposure. Although they produced usable guides, theedge resolution was less than satisfactory due, primarily,to the limited resolution of the mask-flying-spot-scan-ner combination. Since PVS is convertible into silica,this method suggested the possibility of forming 5i02
directly in the masking layer. A refined process isapparently under development.
Goell8 obtained a good edge resolution (<500 A rmsdeviation) using sputter etching. He employed ananalog programmed e-beam to expose the pattern inpolymethylmethacrylate and a lift-off technique to forma manganese mask. Experiments with 0.6328-um lightprism-launched in these waveguides showed less than 3dB/cm loss (Fig. 4).
Optical Waveguides by Ion Implantation
Although ion implantation may be one of the moreimportant techniques of integrated optical devices,there was only one paper on this subject: by Garmire,Yariv and Hunsperger.9 They have successfully im-planted 300 keV protons into n type GaAs, forming aninsulating layer 3 ,m thick and 0.5 mm long. Since thequality of the resulting guides depends greatly on bal-ancing off conflicting requirements-higher dosagelevels and free carrier concentrations lead to higherindex changes but also greater radiation damage-agreat deal of work in this area still remains to be done.
Films and Methods of Deposition
There were several papers describing different guidematerials and their methods of deposition. Standleyand Rand'0 obtained low loss (0.4 dB/cm) films bychemical vapor deposition of- silicon oxynitride on afused quartz substrate. By controlling the molarratio of the reactants they were able to obtain highquality films with refractive indices in the 1.46-1.54range.
I ~
i
:IA'. k
! 2:
Fig. 3. Glass optical waveguide and coupler structures fabri-cated by electron beam masking and plasma etching (D. B.
Ostrowsky and J. C. Dubois7).
Fig. 4. A 3.5-a wide sputtered 7059 glass lightguide with a 3 mmradius of curvature that was delineated using an electron-litho-
Tien, Smolinsky and Martin" reported on their in-vestigation of various organic films obtained by the rfdischarge polymerization process of organic chemicalmonomers. Vinyltrimethylsilane (VTMS) and hexo-methyldisiloxane (HMDS) deposited both individuallyand as a mixture seem to have been their primarychoice. Since the two have somewhat different refrac-tive indices the resulting index can be made to varyeither from film to film or even within a given film,which suggested interesting possibilities in fabricationof lenslike optical elements.
Fujimori, Sasaki and Honda'2 described their in-vestigation of the tantalum pentoxide dielectric films.'3
The highlights of their paper were curves showing theoptimum oxidation temperatures that yield lowestlosses in a film of given thickness.
Hammer'4 discussed epitaxial ZnO films that weregrown on sapphire substrates. Laue x-ray patterns withwell-defined spots indicating single domain structure inthe film were shown. The top surface of the film waspolished before the light propagation experiment andscattering of 0.6328-,um light was found to lead to about8 dB/cm loss in the film.
The remaining papers dealt with infrared waveguidesin silicon," waveguide fabrication by solid state diffu-sion,' 6 improved photoresist films, 17 and embossedwaveguides.8
Optical Fibers (S. E. Miller)
Problems Characteristic of Optical Fibers:Transmission Loss and Dispersion
The session on optical fibers as long-distance trans-mission media was opened with two invited reviewpapers-one by Clarricoats" and the other by Marca-tili.20 Taken together these papers provided back-ground on the state of the art. Several forms of fibersare being studied. In one a core of one type glass issurrounded by a glass of slightly lower index with a stepchange in index at the boundary. A second fiber form,which has been given the name Selfoc by the Japanesegroup that originated it, has a continuously decreasingindex of refraction between the axis of the fiber and out-side fiber boundary, following a parabolic law. Thefiber with a discrete change in index between core andcladding may be dimensioned to have a single mode ofpropagation, or with a larger core diameter may carrya large number of modes. The field distributions asso-ciated with the various modes of propagation werenoted.
Transmission loss is a characteristic of prime im-portance and may be caused by (1) absorption due toimpurities (Fe, Cu, Co, etc.), (2) Rayleigh scatteringdue to frozen-in inhomogeneities in the glass-makingprocess, (3) Mie scattering (predominantly forwardangle) due to larger inhomogeneities, (4) waveguidescattering due, for example, to irregularities in the core-cladding interface and (5) nonlinear scattering due to
stimulated Raman or Brillouin interactions. Refer-ence was made to an earlier Corning Glass Works pub-lication describing a single-mode fiber having a mea-sured loss near 20 dB per kilometer at 0.6328 Am.
Dispersion is another characteristic of prime impor-tance and may be due to (1) intrinsic variation of glassindex with wavelength, or (2) waveguide mode dis-persion. In unimode fibers the predominant com-ponent of delay distortion is the intrinsic dispersioncharacteristic of the glasses from which the fiber ismade; these can be small permitting transmission ratesof 300 megabits or higher. In the multimode fiber thedifference in wave propagation velocities for the variousmodes all at the same wavelength can be the dominantdispersion factor and is appreciably larger than the dis-persion due to glass properties. For a multimode fiberwith 1% index difference between core and cladding thedelay difference can be as large as 30 nsec/km. In theparabolic index fiber less dispersion can be anticipated,with an associated higher radiation loss in sharplycurved sections.
In England the primary interest appears to be inintercity transmission by fiber bundles with a systemanalogous to coaxial carrier transmission systems.This interest is best served by the unimode fiber. AtBell Laboratories there is also strong interest in single-mode fibers for intercity transmission, but there isequally strong interest in intracity applications wherethe small size of the fiber and other physical and elec-trical fiber attributes may be advantageous. Here themultimode fiber in combination with light-emittingdiodes is of interest. The bandwidths needed in thelatter applications may range as low as a few MHz, i.e.,those already in use.
Unger2l presented an analytical paper on wavepropagation in monomode and multimode fibers withimperfections. He and his coauthors 2 ' reported exten-sive theoretical studies of the launching loss for multi-mode fibers and of the effects of various imperfectionsincluding random straightness variations in the com-pleted structure. For a fiber having a core radius 55wavelengths in diameter and with a fractional indexdifference between core and cladding of 3.2%, theoptimum width of a Guassian excitation beam wasfound to be approximately 0.64 core radius, and powerefficiency of 93% into the lowest order mode was in-dicated. Larger Gaussian beams excited increasingamounts of radiation, whereas smaller beams threwadditional power into the higher order guided modes.
Analysis was also indicated for bending losses for theparabolic index fiber, for the simple step-index-changeclad fiber, and for a hybrid of the two involving para-bolic index decrement for a certain radial distance fromthe axis and a step decrement to a lower cladding indexat a radius part way between the fiber axis and outsidesurface. The bending losses for these structures arelowest for the simple clad fiber, intermediate for thehybrid form of fiber, and largest when the parabolicindex decrement extends to the fiber wall. Alterna-tively, dispersion is lowest in the parabolic index fiber.A wealth of quantitative results were presented in the
talk, but a full assimilation must await the appearanceof this material in written form.
Loss Measurements
Kaiser et al.2 2 reported loss measurements of uncladoptical fibers drawn at Bell Laboratories. A non-destructive loss measurement was carried out by im-mersing the unclad fiber in a liquid with slightly higherrefractive index and intercepting the radiated energywith a silicon photodetector also immersed in the liquid.A broad band incoherent source was used in combina-tion with a series of bandpass filters which permittedobservations in 100-A bands every 200 A from 0.5 mto 1.1,4m. The measurements were automated so thatthe loss over the entire spectrum could be observed in ashort time-before atmospheric contaminants couldchange the characteristics of the unclad fiber. Aftercareful cleaning of the fiber surface, losses on lengthsfrom several tens to 100 m were recorded for severalgrades of commercial fused silica. In the "water-free" Suprasil W2 observed losses were in the 15-18dB/km region from 0.72 gm to 1.1 m, with an addi-tional 12 dB/km loss peak centered near 0.95 m dueto the second overtone of the OH resonance. In otherfused silica water absorption peaks at 0.95 Am ranged ashigh as 1000 dB/km, but loss minima as low as 8 dB/km were observed at selected wavelengths. The un-clad fiber measurement provides an effective means forevaluating impurity content and scattering in a singleglassy material and also provides a stepping stone to afiber with large index difference between core andcladding.
Dispersion Measurements
Gloge et al.23 '2 4 described experimental techniques forevaluating dispersion in low-loss multimode fibers.Two distinct approaches were outlined: (1) the use of asequence of short pulses from a mode-locked laser witha pulse width observation before and after traversal ofthe fiber length; (2) measurement of the baseband fre-quency response of the fiber section by comparing themodulated light source spectrum with the modulationafter traversal of the fiber.24 In the short pulse tech-nique, pulses about 0.2 nsec long were utilized in a fiberwith provision for exciting all the modes (approximately600) in a fiber 20 m long, 56-Mum core diameter, and witha numerical aperture of 0.13. The maximum delaydifference between the fiber modes in this sample wasabout 0.4 nsec. The experimentally observed outputagreed well with that expected from a convolution of a0.2-nsec input pulse with a 0.4-nsec rectangular impulsefunction. The alternate measurement technique-direct determination of the baseband frequency re-sponse-compared the detected beat spectrum of lightfrom a free-running ion laser before and after trans-mission through the fiber. Each fiber mode is an in-dependent transmission channel; consequently, thebaseband amplitudes (not light amplitudes) added atthe detector and yielded increasingly out-of-phase addi-tion of output currents from the various modes as the
baseband modulation frequency increased. A spectrumanalyzer placed at the detector output (when suitablyaveraged) showed a flat spectrum out to approximately2.5 GHz from the free-running laser, and after a 30-m-long section of the fiber described above the spectrumanalyzer showed decreasing output above 1 GHz withthe first minimum in the vicinity of 1.6 GHz. The ad-vantages of the baseband technique compared with theshort pulse technique is simplicity for the laboratoryapparatus and greater measurement sensitivity due tothe intrinsically narrow-band type of observation.
Liquid-Core Fibers
Stone2" reported on liquid-core hollow fibers havingexceptionally low losses. The hollow fibers were drawnfrom fused silica tubing to approximately 5 mils outsidediameter in a oxyhydrogen burner. A number ofliquids were explored as the lightguiding core, the bestbeing tetrachlorethylene. The observed losses in theregion between 0.5-1.1 m approximately followed the1/X 4 Rayleigh scattering law with a small loss peak(approximately 10 dB/km) in the vicinity of 0.95 mdue to the second overtone of the OH resonance. Aminimum loss in the vicinity of 13.5 dB/km in the regionnear 1.06 m was observed for a length of 450 m.This is the lowest-loss clad fiber yet reported.
Selfoc Fibers
Koizumi26 gave an invited post-deadline review ofdevelopments related to optical fibers and integratedoptical circuits. He indicated the following new de-velopments in what the authors call Selfoc fibers, whichare fibers having a parabolic decrement in index of re-fraction from the axis to the side wall of the fiber:(1) miniature lenses, 1-mm diameter by 5 mm long;(2) a Selfoc guide for use in a Nippon Electric Com-
pany photocopying device, the fiber being 20 0 -judiameter and 60 cm long;
(3) continuous wave glass lasers using the Selfoc fiberstructure have been demonstrated providing out-put power of 3.5 W with an input of 4 K•W,
(4) long fibers using the Selfoc structure and intendedfor long-distance communication have been ad-vanced, the best achievement to date being a 300-m fiber having 55 dB/km loss near 1.06 Am.
The structures have been fabricated using an ion ex-change process in which potassium replaces thalium andsodium in a lead-soda-silicate glass during a high tem-perature soaking cycle in potassium nitrate. He indi-cated that the ionic replacement process had also beenused to create guiding layers in sheets of glass for use inintegrated optics. A thin-film type of guide with lossas low as 0.37 dB/cm was quoted which indicated avery high quality surface in the completed structure.
Fiber Coupler
Krumpholz27 addressed the problem of a practicalcoupler for single-mode fiber waveguides. Using fiberswith core and cladding indexes of 1.578 and 1.516, acore diameter between 0.5 and 1.0 m was utilized toassure single-mode operation. A theory for the launch-
Fig. 5. Experimentally observed quantum efficiency vs lightwavelength of a new kind of Si-avalanche photodiode for use in
connection with lightguides (S. Maslowski2 8).
ing loss from a Gaussian laser beam to the fiber wasgiven, and a generalization to the case of intercouplingloss between two fibers at a joint was also reported. Amanually adjustable structure for aligning the opticalbeams at the juncture of two single-mode fibers wasachieved by mounting the fibers eccentrically in rotata-ble members. Intercoupling efficiencies of 90% wereobtained experimentally.
Detector for Glass Fiber Waveguides
Maslowski28 reported a new kind of detector for usein communication systems with glass fiber waveguides.A silicon p-n junction is employed with the light to bedetected propagating parallel to the junction in itsimmediate vicinity. With this mode of optical propa-gation, large penetration depth of the energy at infra-red wavelengths still results in a short path to the ex-ternal circuit for the hole-electron pairs generated.The transverse beamwidth of the light wave must bemaintained within a few micrometers, but this is not aserious limitation for a detector to be used with single-mode fibers. Figure 5 shows the experimentally ob-served quantum efficiency vs wavelength. Quantumefficiencies were reported to be higher than for conven-tional planar photodiodes at visible wavelengths andappreciably higher in the region at 1 and above.Avalanche multiplication of the photocarriers may beused. With a photocurrent gain of 70, a demodulationcutoff frequency of 1 GHz was reported for a 6 Am-diamimpinging light beam at 0.632 8-Am wavelength. Bydisplacing the point of impingement, gain and band-width could be traded within the limits on gain roughly100 and on demodulation bandwidth 2-3 GHz.
Passive Optical Devices-Theory andExperiment (J. H. Harris)
Passive Integrated Elements
In a review paper, Li29 was concerned with passive
integrated circuits consisting of three-dimensional di-electric waveguide structures embedded in planar sub-strates. He reviewed requirements for low-loss wave-guides, radiation loss due to bends, and coupling betweenadjacent guides. He then discussed the use of glass,organic and crystal films and methods of making wave-guides including (1) electron beam exposure of resistand etching or sputtering etching, (2) direct writing ofpolymers using a uv laser beam, (3) embossed tech-niques, and (4) field induced methods.
Optical Waveguides of Periodic Layers
The generation of waveguides with the unique prop-erty of lying on substrates and having cutoff behaviorin air was discussed by Arnaud and Saleh.'0 The wave-guide would consist of stacks of quarter wave alternatehigh and low index layers. The field penetration inquarter wave periodic structures is exponential so thatthe guided wave would have exponential properties inboth directions from the air-dielectric interface. Thenumber of layers required is dependent on the indexdifference between adjacent layers. With differences of0.2, sixteen layers are required to reduce the field in thelayers to one third for TE modes. As was pointed out,this reduction may not be sufficient. For the smallindex differences that can be achieved by direct opticalexposure techniques, the number of layers may be ex-cessive.
Scattering and Power Distribution inMultimode Waveguides
Several papers that followed discussed aspects ofmode and beam coupling in plane structures withuniform properties in the direction transverse to thedirection of propagation. Marcuse" reviewed thenature of light propagation along linear multimodewaveguides in which there are significant scatteringeffects. Over distances Az that are large compared withthe correlation length of the scatterers, the powerscattered from, say, the nth to the th mode is ap-proximately proportional to •z and to the power in thenth mode, Pn. Thus, the power distribution in a guidecontaining N modes may be found by solving the set oflinear first order differential equations:
dP. IN \ Nd = _n- ( A hn) P + E hnP,. (1)
The power per unit length scattered out of the guide,into other modes, and back into the nth mode are therespective terms in Eq. (1). Except in special casesinvolving degeneracy, these equations have exponentialsolutions of the form
NP, = E ciB(') exp[-c(i)z], (2)
i=i
where a(') is the N eigenvalues obtained by setting Eq.(2) into Eq. (1), Bn(i) is the eigenvector for each i ob-tained in the same way and normalized to Bl() = 1, andcf are the coefficients determined by the relative modalexcitation at the guide input.
Fig. 6. Power distribution among the individual waveguidemodes in the slowest attenuating coupled mode of a waveguide
with random wall distortion (D. Marcuse").
Figure 6 illustrates computed results for the co-efficients of the coupled modes in a two-dimensionalconfiguration involving a film of thickness 2d and indexni embedded in a medium of index n and having aGaussian random wall distortion of correlation lengthD. Only the slowest attenuating term i 1 of Eq. (2)is shown. Note that as the correlation length D de-creases, thus providing increased scattering out of theguide, the attenuation of higher order mode com-ponents decreases ever more rapidly. Marcuse alsodiscussed aspects of pulse transmission along guides.He pointed out that the spread of a pulse in a multi-mode guide due to differing group velocities is onlyproportional to distance in the absence of scattering.In analog to the random walk problem, the spread isproportional to the square root of distance when thereis extensive energy interchange among the modes.Miyazaki"2 also discussed aspects of mode conversionand pulse delay and dispersion in waveguides.
Passive Mode Conversion
Shah, Crow, and Wang3 discussed passive modeconversion techniques with potential electronic modula-tion aspects using a uniaxial crystal.' 4 They have ob-served TE-TM mode conversion on the order of half thetransmitted power with a quartz crystal 2.6 mm inlength clamped to a glass waveguide. Figure 7 showsthe light coupled out of the waveguide as seen through a,polarizer. The appearance of two lines indicates modeconversion and corresponds to the case in which theoptic axis of the crystal lies in a plane transverse to thedirection of propagation (longitudinal). A single lineshows no conversion and corresponds to the case inwhich the optic axis lies in a plane containing one linecolinear with the direction of propagation (equatorial).
Beam-Film Coupling
Aspects of light wave couplers2 were reviewed byTamir and Bertoni,"5 who pointed out that couplingwith the aid of a prism or a periodic structure consistsof a modified version of the Goos-Hanchen effect. Thiseffect describes the spatial displacement of a beam onreflection from a surface under conditions of total reflec-tion. They also pointed out that improved efficiencycan be obtained with a periodic coupler if the couplingis tapered in the manner of a prism with a variabletunneling region.
Ash, Seaford, and Pennington 6 discussed the applica-tion of holography to beam coupling to waveguides aswell as to coupling between waveguides. The pro-cedure of roughing a film surface or otherwise couplinglight out of a guide and making a holographic recordingby intercepting the scattered light can provide a usefulcoupling technique. Reconstruction of the originalwavefront leads to a coupling efficiency that, in prin-ciple, is limited only by the diffraction efficiency of thehologram and the success in intercepting the scatteredlight.
Active Integrated Elements-ElectronicModulation (J. H. Harris)
Acoustooptic ModulationThe demonstration of active electronic modulation
effects on solid films is clearly in an early stage of de-velopment. The exception to this is work with acousto-optic modulation3 7 that was discussed by Kuhn' inan invited paper. He described aspects of both colinearand transverse modulation, employing surface acousticwaves as indicated in Fig. 8. In the former case, modeconversion from first to third order TE as well as TMmodes was achieved with conversion efficiencies as highas 55% over an 8-mm path length. Colinear inter-action requires phase matching (or momentum con-
: : LONGITUDINALI
TE
EQUATORIAL
INPUT
(a) (c)
TM INPUT
(b) (d)
Fig. 7. Output mode lines viewed through a polarizer showingmode conversion for longitudinal orientation of a quartz sub-
SURFACE ACOUSTIC e. Io I INTERDIGITALTRANSDUCER I GLASS FILM
METALFILM
LiNbO,SUBSTRATE
ZOPTICAL GRATINGCOUPLERS
(b)
Fig. 8. Transverse and longitudinal acoustooptic modulatorconfigurations (L. Kuhn' 8).
servation) of the modal wave numbers and the acousticwave number in the form
k,, + k = kw3,
The interaction is therefore dependent on the acousticfrequency. It was observed that mode conversion hada narrow frequency band sin(w - coo)/( - coo) behaviorwith an acoustic bandwidth of 850 kHz at a center fre-quency of 320 MHz. A deflection efficiency of 66% wasachieved with the transverse modulation.
Electrooptic ModulationUse of GaAs epitaxial layers to achieve electrooptic
modulation was discussed by several people. Previouswork with end excited junction structures' 9 was citedby Yariv40 who also emphasized the need for structurescompatible with integrated optical systems. Cheo4ldescribed experiments with GaAs films of reducedcarrier concentration (10'2 -10"/cc) vapor deposited onhigh carrier concentration substrates. The index dif-ference between film and substrate was estimated to be0.3 at 10.6 where the experiments were carried out.Multimode guides 2 cm in length and ranging from 18A to 43 in thickness were fabricated. The various
modes were successfully excited using a Ge prism, butattempts at electrooptic phase modulation were pro-hibited by a voltage maximum of about 20 V that couldbe applied to the Schottky barrier electrode.
Nonlinear Interaction (P. K. Tien)
Nonlinear Thin-Film DevicesWhen a light wave propagates in a film, the light
energy is concentrated in the film. If the film is verythin, the field intensity inside the film can be very largeeven at a moderate laser power level. Although thislarge field intensity is important in nonlinear optics,progress has been slow in the development of thin-filmnonlinear devices. The main difficulties involved are42
lack of single-crystal films that have large nonlinearoptical coefficients, and difficulties in controlling thethickness and the refractive index of the film within thetolerances required for a long coherence length. In aninvited paper, Suematsu4 ' discussed his calculation ofphase-match conditions for various parametric inter-actions in lightguides. In another paper, Yariv40
described a method of phase-match that involves athin-film waveguide with a corrugated top surface.The surface corrugation introduces spatial harmonicswith new propagating constants that can be phasematched. Figure 9 is an electronmicrograph of acorrugated grating on the surface of a thin GaAsepitaxial film. The grating was fabricated by ion mill-ing.
Free-Standing GaAs Optical Waveguides
Anderson4 4 described his experiments of parametricinteraction using Xe and C02 laser emission lines.The optical waveguides used were free-standing polishedsingle crystals that had cross sections of 3 X 10-100,pm and lengths of 1-4 mm. Phase-match was accom-plished by selection of a specific set of waveguide modes.Some experiments demonstrated a conversion efficiencyof 1% and other experiments showed second harmonicgeneration over a wide bandwidth using a pump powerof less than 1 W.
Raman Gain in Optical Fiber
Ippen and Stolen4 ' observed Raman gain coefficient of4 X 10"11 cm/W in a single-mode cladded fiber made byCorning Glass Works using a 0.526-pm argon laser as thepump. The core diameter of the fiber was 4 pm andthe length was 10 m. Although the gain coefficient issmall, large amplification can be achieved over longlengths of low-loss fibers and would thus restrict thepower that can be transmitted in such fibers.
Integrated Sources and Amplifiers (P. K. Tien)
Thin-Film Dye Lasers and Amplifiers
Contrary to the slow pace of the development of thin-film nonlinear optics, significant progress has beenmade in the area of thin-film dye lasers and amplifiers.Early in 1971, Kogelnik and Shank46 reported laseroscillation in a dielectric grating made of gelatin doped
with rhodamine 6G. Periodic variation of the refrac-tive index provided. the feedback necessary for theoscillation. Later, Weber and Ulrich4 6 reported aring dye laser that had a doped polyurethane filmcoated on the surface of a cylindrical rod. The feed-back for laser oscillation was established around thecircumference of the rod. In these and many laterexperiments, the dye commonly used has been rhod-amine 6G. This dye has an absorption band near 0.53pm and can thus be pumped effectively either by a 0.515-pm argon laser or by the second harmonic of a Nd :YAGlaser. It can also be pumped less efficiently by uvlight. The absorption of the pump light is of the orderof 20 dB/cm for a doping level of 5 X 10' molecules/cm', or 8 X 106 iM/l. The dye can be dissolved inliquids or mixed in solid matrices. Stimulated emissionhas been observed in doped solid matrices at wave-lengths of 0.552-0.630 m with pump power densitytypically at 1 MW/cm2 . The chief disadvantage ofthe dye laser is, of course, its short lifetime. Thenumber of photons that a dye molecule can absorb be-fore it becomes useless is called the bleaching number.This number ranges between 1.4 X 105 and 25 X 10' asmeasured by E. P. Ippen in liquids and I. P. Kaminowin solid matrices. With this background in mind, wenow discuss the papers presented to this meeting.
The first paper in this session was given by Bjorkholmand Shank,4 7 who discussed a dye laser using a light-guiding polyurethane film doped with rhodamine 6G,which was similar to the film used by Weber and Ulrich.However, instead of forming a grating structure in thegain medium as in the laser reported by Kogelnik andShank, they optically pumped the film with interferencefringes formed by first splitting and then recombiningthe 0.35-pm light beam from a doubled ruby laser.Since the spacing between the interference fringes
Fig. 9. Electron-microscope enlargement of a corrugationgrating on the surface of a thin GaAs film. The grating period is1.4 um and the depth is 0.15 ,um. The grating is made by ionmilling through a mask prepared holographically using a photo-resist layer. The gratings are used to obtain phase matching innonlinear optical experiments in thin films and as input andoutput light couplers in various GaAs active components (E.Gamire, S. Somekh, H. Stoll, A. Yariv,40 H. Garvin, and E.
Wolfe).
should be half of the wavelength of oscillation, thewavelength of the dye laser was tunable by adjusting thespacing of the fringes. Stimulated emission typicallywith a linewidth of a few tenths of an angstrom wasobserved.
In another paper, Chang et al.4 8 described a dye ampli-fier. They had used a polyurethane film doped withrhodamine B that was optically pumped by a nitrogenlaser. By coupling a 0.632 8-,pm He-Ne laser beaminto the active film, they observed amplification of thelight wave in the film and measured a net gain on theorder of 20 dB/cm.
Ippen et al.4 9 reported another dye amplifier in whichboth the pump and the laser light propagated in thesame light-guiding film. The film was covered by a 3X 10-3 M/l solution of rhodamine 6G. The authorsnoted that, when the refractive index of the film wasclose to that of the dye solution, the evanescent field ofthe waveguide mode penetrated deep into the solutionand could be used effectively to pump the dye. Withthe above experimental arrangement, they observedsuperradiance at a pump power level of 1 kW. Thepump used was a 0.532-Am doubled Nd: YAG laser.
Thin-Film Neodymium Glass AmpliferFinally, we discuss a thin-film Nd-glass amplifier
reported by Yajima et al.50 The amplifier was operatedat 1.06 pm. A thin film of neodymium glass, 2.7 mthick, was sputtered on a piece of Corning 7059 glasssubstrate at a rate of 1500 A/hr; the film was thencovered with a layer of Corning glass for protection.The Nd-glass film had a refractive index of 1.55 and aloss of the order of 0.05 dB/cm. When it was pumpedby a xenon flash lamp, a net gain of 16% over a pathlength of 3 cm with a pump energy of 130 J was mea-sured.
This report was edited by P. K. Tien.The Technical Program Committee for this OSA
Topical Meeting on Integrated Optics-Guided Waves,Materials and Devices has the following members:R. V. Pole, Chairman (IBM Watson Research Center),E. Ash (University College London), A. J. DeMaria(United Aircraft Research Laboratories), J. H. Harris(University of Washington), D. R. Herriott (BTL),Lawrence Kuhn (IBM Watson Research Center), S. E.Miller (BTL), Heinrich Nassenstein (FarbenfabrikenBayer AG), P. K. Tien (BTL), and Yashuharu Sue-matsu (Tokyo Institute of Technology).
References1. R. Shubert and J. H. Harris, IEEE Trans. MMT-16, 1048
(1968); S. E. Miller, Bell System Tech. J. 48, 2059 (1969);P. K. Tien, Appl. Opt. 10, 2395 (1971).
2. J. H. Harris, R. Shubert, URSI Conf. Abs., Apr. 1969; P. K.Tien, R. Ulrich and R. J. Martin, Appl. Phys. Lett. 14, 291(1969); M. L. Dakss, L. Kuhn, P. F. Heidrich, and B. A.Scott, Appl. Phys. Lett. 16, 523 (1970); H. Kogelnik and T.Sosnowski, Bell Syst. Tech. J. 49, 1602 (1970); J. H. Harris,R. Shubert, and J. N. Polky, J. Opt. Soc. Am. 60, 1007(1970); P. K. Tien and R. Ulrich, J. Opt. Soc. Am. 60,1325(1970); R. Ulrich, J. Opt. Soc. Am. 60, 1337 (1970); J. E.
Midwinter, IEEE J. Quant. Electron. QE-6, 583 (1970);J. H. H-arris and R. Shubert, IEEE Trans. MTT-19, 269(1971); R. Ulrich, J. Opt. Soc. Am. 61, 1467 (1971); T.Tamir and H. L. Bertoni, J. Opt. Soc. Am. 61, 1397 (1971);P. K. Tien and R. J. Martin, Appl. Phys. Lett. 18, 398(1971).
3. MAI Invited Paper Early Work on Optical SurfaceWaves H. Osterberg and L. W. Smith, American OpticalCorporation, Southbridge, Massachusetts 01550.
4. MA2 Invited Paper Development of Thin-Film Optoelec-tronics in the Past Two Years P. K. Tien, Bell TelephoneLaboratories, Holmdel, New Jersey 07733; P. K. Tien,R. J. Mar-tin, S. L. Blank, S. H. Wemple, and L. J. Varnerin,"Optical Waveguides of Single-Crystal Garnet Films";P. K. Tien, R. J. Martin, R. Wolfe, R. C. LeCraw and S. L.Blank, "Switching and Modulation of Light in Magneto-Optic Waveguides of Garnet Films," (both papers to bepublished).
5. MA3 Invited Paper Two-Dimensional Optics J. H.Harris, University of Washington, Seattle, Washington98105.
6. TuAl Invited Paper Materials Aspects of Thin-FilmOptical Circuits A. G. Chynoweth, Bell Telephone Labora-tories, Murray Hill, New Jersey 07974.
7. MB7 Contributed Paper Electron-Beam Masking forOptical-Waveguide Fabrication D. B. Ostrowsky and J. C.DuBois, Thomson/CSF-LCR/DR1, 91 Orsay, France.
8. MB8 Contributed Paper Electron-Resist Fabrication ofIntegrated Optical Circuits J. E. Goell, Bell TelephoneLaboratories, Holmdel, New Jersey 07733.
9. MB10 Contributed Paper Waveguiding in Proton-Im-planted GaAs E. Garmire, A. Yariv, California Institute ofTechnology, Pasadena, California 91109; and R. G. Hun-sperger, Hughes Research Laboratories, Malibu, California90265.
10. TuA3 Contributed Paper Guided-Wave Propagation inMixed Silicon Oxide-Silicon Nitride Dielectric Films R. D.Standley, Bell Telephone Laboratories, Holmdel, NewJersey 07733, and M. Rand, Bell Telephone Laboratories,Allentown, Pennsylvania 18103.
11. TuA7 Contributed Paper Thin Organosilicon Films forIntegrated Optics P. K. Tien, G. Smolinsky, and R. J.Martin, Bell Telephone Laboratories, Holmdel and MurrayHill, New Jersey, 07974.
12. TuA2 Contributed Paper Optical Transmission Charac-teristics in Ta2l,0 Film M. Fujimori, M. Sasaki, and M.Honda, Fujitsu Laboratories, Kawasaki 211, Japan.
13. D. -l. Hensler, J. 1). Cuthbert, R. J. Martin and P. K. Tien,Appl. Opt. 10, 1037 (1971).
14. Post-Deadline Paper Single Crystal Epitaxial ZnO onSapphire Optical Waveguide J. M. Hammer, J. P. Wittke,D. J. Chamin, and M. T. Duffy, RCA Laboratories, Prince-ton, New Jersey 08540.
15. TuA6 Contributed Paper Infrared Waveguides in SiliconD. Vincent, University College London, London, WCIE 2JE,England.
16. TuA5 Contributed Paper Fabrication of Integrated OpticalCircuits by Solid-State Diffusion H. F. Taylor, W. E.Martin, V. N. Smiley, and D. B. Hall, Naval ElectronicsLaboratory Center, San Diego, California 92152.
17. TuA8 Contributed Paper Light-Guiding Structures ofPhotoresist Films H. P. Weber, R. Ulrich, E. A. Chandross,W. J. Tomlinson, Bell Telephone Laboratories, Holmdel andMurray Hill, New Jersey 07974.
18. MB9 Contributed Paper Embossed Optical WaveguidesR. Ulrich, H. P. Weber, E. A. Chandross, W. J. Tomlinson,
and E. A. Franke, Bell Telephone Laboratories, Holmdel,New Jersey 07733.
19. WAl Invited Paper Optical Fiber Waveguides P. J. B.Clarricoats, Queen Mary College, University of London,London, England.
20. WA2 Invited Paper Fiber Optics for Long-DistanceCommunication E. A. J. Marcatili, Bell Telephone Labora-tories, Holmdel, New Jersey 07733.
21. WA3 Contributed Paper Analysis of Optical WaveLaunching and Propagation in Monomode and MultimodeFibers with Imperfections H. J. Heyke, H. Kirchhoff, andH. G. Unger, Technische Universitat Braunschweig, Braun-schweig, Germany.
22. WA4 Contributed Paper Loss Measurements of UncladOptical Fibers P. Kaiser, A. R. Tynes, A. M. Cherin, andA. D. Pearson, Bell Telephone Laboratories, Holmdel,New Jersey 07733; WA6 Contributed Paper Measure-ment of the Angular Distribution of Light Scattered from aGlass Fiber Optical Waveguide E. G. Rawson, Bell Tele-phone Laboratories, Murray Hill, New Jersey 07974.
23. WA7 Contributed Paper Dispersion in a Low-Loss,Multimode Fiber Measured at Three Wavelengths-D.Gloge, R. D. Standley, W. S. Holden, E. L. Chinnock, andT. S. Kinsel, Bell Telephone Laboratories, Holmdel, NewJersey 07733.
24. WA8 Contributed Paper Direct Measurement of the(Baseband) Frequency Response of Multimode Fibers D.Gloge, E. L. Chinnock, and D. H. Ring, Bell TelephoneLaboratories, Holmdel, New Jersey 07733.
25. WA5 Contributed Paper Optical Transmission Loss inLiquid-Core, Hollow Fibers J. Stone, Bell TelephoneLaboratories, Holmdel, New Jersey 07733.
26. Invited Post-Deadline Paper Recent Development ofSeifoc Fibers K. Koizumi, Nippon Electric Company,Japan.
27. WB5 Contributed Paper Optical-Coupling Problems inCommunication Systems with Glass Fiber Waveguides 0.Krumpholz, AEG-Telefunken, Ulm, Germany.
28. WB6 Contributed Paper New Kind of Detector for Usein Communication Systems with Glass Fiber WaveguidesS. Maslowski, AEG-Telefunken, Ulm, Germany.
29. MA4 Invited Paper Passive Integrated Optical CircuitsT. Li, Bell Telephone Laboratories, Holmdel, New Jersey07733.
30. MA5 Contributed Paper Guidance of Surface Waves byPeriodic Layers J. A. Arnaud and A. A. M. Saleh, BellTelephone Laboratories, Holmdel, New Jersey 07733.
31. MB1 Contributed Paper Radiation Losses in Multi-mode, Dielectric Slab Waveguides D. Marcuse, BellTelephone Laboratories, Holmdel, New Jersey 07733, D.Marcuse, Bell Syst. Tech. J. 48, 3187 (1969); A. W. Snyder,IEEE Trans. MTT-17, 1138 (1969); J. H. Harris, D. P. GiaRusso, and R. Shubert, Proc. IEEE 59, 1123 (1971).
32. MA6 Contributed Paper Propagation Properties of OpticalSignal Waves in Perturbed Dielectric Waveguides by Con-formal Mapping, Technique Y. Miyazaki, Nagoya Uni-versity, Nagoya, Japan.
33. MB2 Contributed Paper Optical-Waveguide Mode-Con-version Experiments and Further Development of the Theoryof Propagation in Waveguides with Gyrotropic and Aniso-tropic Substrates M. Shah, J. D. Crow, and S. Wang,University of California, Berkeley, California 94720.
34. F. K. Reinhart, 1). F. Nelson, and J. McKanne, Phys. Rev.177, 1208 (1969); S. Wang, J. D. Crow, M. Shah, Appl.Phys. Lett. 19, 187 (1971); S. Wang, M. Shah, and J. D.Crow, IEEE Trans. QE-8, 212 (1972); D. P. Gia Russo, Ph.D. Diss., Univ. of Wash. (1971).
35. MB3 Contributed Paper Unified Theory of Optical BeamCouplers T. Tamir and H. L. Bertoni, Polytechnic Instituteof Brooklyn, Nwe York, New York 11201.
36. MB4 Contributed Paper Holographic Couplers for In-tegrated Optical Circuits E. A. Ash, E. Seaford, UniversityCollege London, London, WCIE 2JE, England; K. Penning-ton, IBM Watson Research Center, Yorktown Heights,New York 10598.
37. L. Kuhn, M. L. Dakss, P. F. Heidrich, and B. A. Scott,Appl. Phys. Lett. 17, 265 (1970); L. Kuhn, P. F. Heidrich,and E. G. Lean, Appl. Phys. Lett. 19, 428 (1971); W. S. C.Chang, IEEE Trans. QE 7, 167 (1971).
38. TuB2 Invited Paper Interactions Between Optical GuidedWaves and Surface Acoustic Waves L. Kuhn, IBM WatsonResearch Center, Yorktown Heights, New York 10598.
39. D. F. Nelson and F. K. Reinhart, Appl. Phys. Lett. 5, 148(1964); F. K. Reinhart, J. Appl. Phys. 39, 3426 (1968);D. Hall, A. Yariv, and E. Garmire, Opt. Comm. 1, 403(1970) and Appl. Phys. Lett. 17, 127 (1970).
40. TuB1 Invited Paper Active Integrated Optics A. Yariv,California Institute of Technology, Pasadena, California91109.
41. TuB3 Contributed Paper Excitation and Modulation of10.6 Micrometer Guided Waves in GaAs Epitaxial ThinFilms P. K. Cheo, United Aircraft Research Laboratories,East Hartford, Connecticut 06108.
42. P. K. Tien, Appl. Opt. 10, 2395 (1971); P. K. Tien, R.Ulrich, and R. J. Martin, Appl. Phys. Lett. 17, 447 (1970);D. B. Anderson and J. T. Boyd, Appl. Phys. Lett. 19, 266,1971.
43. TuB4 Invited Paper Some Problems in Guided-WaveParametric Interactions Y. Suematsu, Tokyo Institute ofTechnology, Tokyo, Japan 152.
44. TuB5 Invited Paper Waveguide Phase-Matched Para-metric Interactions in GaAs D. B. Anderson, AutoneticsDivision of North American Rockwell, Anaheim, California92803.
45. ThA8 Contributed Paper Stimulated Raman Gain inOptical Waveguides E. P. Ippen and R. H. Stolen, BellTelephone Laboratories, Holmdel, New Jersey 07733.
46. H. Kogelnik and C. V. Shank, Appl. Phys. Lett. 18, 152(1971); H. P. Weber and R. Ulrich, Appl. Phys. Lett. 19, 38(1971).
47. ThAi Contributed Paper Distributed-Feedback Lasersin Thin-Film Optical Waveguides J. E. Bjorkholm andC. V. Shank, Bell Telephone Laboratories, Holmdel, NewJersey 07733.
48. ThA3 Contributed Paper Light AmplificationFilm M. S. Chang, P. Burlamacchi, C. Hu,Whinnery, University of California, Berkeley,94720.
49. ThA4Lasertories,
in a Thinand J. R.California
Contributed Paper Evanescent-Field-Pumped DyeE. P. Ippen and C. V. Shank, Bell Telephone Labora-
Holmdel, New Jersey 07733.
50. ThA9 Contributed Paper Amplification of 1.06 microm-eters Using Nd-Glass Thin-Film Waveguide H. Yajima,S. Kawase, and Y. Sekimoto, Electrotechnical Laboratory,Tanashi-shi, Tokyo, Japan.
IMPORTANT JOURNAL INFORMATION
At its October 1971 meeting, the Boardof Directors considered the problem ofever increasing production costs forthe Society journals. it is the in-tention of the Board that the Society'sjournals be available to members andnonmembers at a reasonable cost andthat all material judged acceptable bythe editors be printed. The Societyrelies upon several sources of incomein order to meet these two aims; dues,subscriptions, page charges, and adver-tising are the primary sources.A considerable drop in the percentageof institutions that agreed to honorthe payment of page charges occurredduring 1971. This decrease in incomefrom page charges will cause a drasticrevision in dues and subscription pricesif allowed to continue. In order toprevent the occurrence of this unde-sirable revision, the Board took severalactions. First, the printing methods ofboth the JOURNAL OF THE OPTICAL SOCIETYof AMERICA and APPLIED OPTICS were re-vised in order to effect considerablesavings at little loss in quality.Second, the Board established a pagebudget for papers that appear in thetwo journals.Under this new page budget, the totalnumber of pages for which the pagecharge is not honored is limited tonot more than 10% of the total numberof pages.Thus, authors of such papers may findthat the publication of their papersis delayed by this policy. We encour-age you to avoid this delay by impress-ing upon the responsible members ofyour institution the importance of pagecharges for successful publication ofthe journals. Authors whose researchis supported by federal agencies arereminded that the Federal Council forScience and Technology has approvedthe page charge as a proper researchcost.The page charge was established tocover the costs of editorial prepara-tion and composition. Printing anddistribution were to be borne by incomefrom subscriptions and dues. Pagecharge rates have not kept pace withthe rising costs of composition. There-fore, it is most important that thepage charge be honored on a largefraction of the published papers.
The Society has made a special effortduring the past six months to encouragethe advertising of optical products,materials, and services in APPLIEDOPTICS. We believe that the inclusionof this material is of great benefit tothe optical scientists and engineerswho comprise our readership. We hopethat you will take full advantage ofthe reader service card every month.In doing this, you not only help sus-tain our advertising support but youhelp make industry more responsive toyour product and service needs. Thissupport from the optical industry isvital to our publication of 4500 pagesof contributed papers and letters eachyear.
