Marcatili E A J. Dielectric rectangular waveguide and directional coupler for integrated optics. Bell Syst. Tech. J. 48:2071-102, 1969. [Bell Laboratories, Holmdel, NJ]
The transmission properties of a guide by the multiplicity of dielectrics and
consisting of a dielectric rod of rec- the cornered cross-section was enough tangular cross-section surrounded by to discourage people from tackling the dielectrics of smaller refractive indices problem. However, I was lucky enough are determined. This guide is the basic to recognize that the liability of the component in a new technology called rectangular cross-section could be integrated optical circuitry. The direc- transformed into a very useful asset. In tional coupler, a particularly useful fact, it became obvious that the cross- device, made of two of those guides section of the dielectrics surrounding closely spaced is also analyzed. [The the core could be divided into a few SCI® indicates that this paper has been discrete regions and that the elec- cited over 145 times since 1969.] tromagnetic field in some of these regions could be neglected. E. A. J. Marcatili "This approximation simplified the Bell Laboratories problem to such a degree that the Crawford Hill Laboratory mathematics needed to solve the trans- Holmdel, NJ 07733 mission problems became trivial, in fact, so trivial that even the more com- July 25, 1978 plicated problem of determining the "In 1969 my friend and boss, S. E. properties of a directional coupler con- Miller, came across the wonderful idea sisting of two such guides parallel to of fabricating the many components of each other was solved in the same optical transmitters and receivers on a paper. single substrate using highly sophisti- "Why has this paper been widely cated but well known photolithographic quoted? I believe there are three techniques. These devices were very reasons: timeliness, simplicity of the small (a few microns across), precise, treatment and results, and introduction insensitive to mechanical and thermal of the new t r ick described above. Let transients and, if fabricated in large me expand these three points. My numbers, economical. Integrated op- paper appeared in the same issue of the tical circuits were supposed to do for Bell System Technical Journal where S. optics what integrated circuits did for E. Miller introduced the basic ideas of electronics. integrated optics. 1 This coincidence "It was immediately recognized that was no accident, and my calculations the basic component for this new tech- were filling a need created by his ideas. nology was a waveguide consisting of a Had my article been an exact treat- dielectric rod of approximately rec- ment it probably would have been so tangular cross-section, surrounded by complicated that readers would have other dielectrics of lower refractive in- been 'turned off.' Finally, the trick of dices. In those days the analytical solu- neglecting the field in some parts of tion of such a problem was next to im- the guide cross-section was neat and possible. Even the much simpler ques- has been used in many subsequent ar- tions of propagation in a dielectric rod ticles. Thus, my paper had not only an of circular cross-section surrounded by intrinsic value in itself but also provid- a single dielectric were extremely dif- ed dividends in the form of a simple ficult to answer. The complexity added mathematical tool useful elsewhere."
1. Miller S E. Integrated optics: an introduction. Bell Syst. Tech. J. 48:2059-69, 1969.
Hawkins Electrical Guide, Number One
Questions, Answers, & Illustrations, A Progressive Course
of Study for Engineers, Electricians, Students and Those
Desiring to acquire a Working Knowledge of Electricity and
its Applications
