358 Chapter 8. Polarization igure 8.45, Two linear polarizers and two quarter wave plates. ‘which point it is circular, As it passes through C. another 90° retardance is added on, resulting once more in a linearly polar: ized wave. In effect. B and C together form a half-wave plate which merely flips the linear light from A through a spatial angle of 26, in this case 90°, Since the linear wave from C is parallel o the transmission axis of D, it passes through it and ‘out of the system. In this simple process we've actually proved something that is rather sublle. Ifthe circular polarizers 4 + Band C+ D are both left-handed, we've shown that lefi-cireular light entering a left-circular polarizer from the output side will be transmitted, Furthermore, it should be apparent, at least after some thought, that right-eircular light will produce a P-state perpendicular to the transmission axis of D and so will be absorbed. The converse is true as well; that is, of the ro cir ‘cular forms, only light in an SR-ssate will pass through a right circular polarizer having entered from the output side. A crumpled piece of celophane placed between two crossed Polarards shows 2raibow of colors. Depending ants thickness and the frequency ofthe light, the oelophane rotates the Efeld by afer: ent amounts, Rotating fether one of the Polaroxds wal shift the colors to ther complements 8.9 Polarization of Polychromatic Light 8.9.1 Bandwidth and Coherence Time of a Polychromatic Wave By its very nature purely monochromatic light, which is of course not a physical reality, must be polarized. The wo orthogonal components of such a wave have the same fre quency, and each has a constant amplitude. Ifthe amplitude of either sinusoidal component varied, it would be equivalent to the presence of other additional frequencies in the Fourier analyzed spectrum. Moreover, the two components have a constant relative phase difference; tha i, they are coherent. A monochromatic disturbance is an infinite wavetrain whose properties have been fixed for alltime; whether itis in an St YF... P-.or Estate, the wave is completely polarized ‘Actual light sources are polychromatic; they emit radiant having a range of frequencies. Let’s now examine what happens on a submicroscopic scale, paying particular attention to the polarization state of the emitted wave. Envision an clec- tron-oscillator that has been excited into vibration (possibly by collision) and thereupon radiates. Depending on its precise ‘motion, the oscillator will emit some form of polarized light As in Section 7.4.3, we picture the radiant energy from a single atom as @ wavetrain having a finite spatial extent Al, ‘Assume for the moment that its polarization state is essential: ly constant for a duration of the order of the coherence time At, (which, as you recall, corresponds to the temporal extent of the wavetrain, e.. Al,/e). A typical source generally con: sists of a large collection of such radiating atoms. which can be envisioned as oscillating with different phases at some dominant frequency B. Suppose then that we examine the light coming from a very small region of the source, such that the emitted rays arriving at a point of observation are essentially parallel. During a time that is short in comparison with the average coherence time, the amplitudes and phases of the ‘wavetrains from the individual atoms will be essentially com- stant, This means tha if we were to look toward the source in some direction, we would, at least for an instant, “see” a ‘coherent superposition of the waves emitted in that direction, We would "see" resultant wave having a given polarization state, That state would only last for an interval less than the ‘coherence time before it changed, but even so it would corre spond to a great many oscillations atthe frequency P. Clearly, if the bandwidth Avis broad, the coherence time (Ar, = 1/0)willbe small, and any polarization state willbe short-lived Evidently the concepts of polarization and coherence ave relaved in fundamental was Now consider a wave whose bandwidth is very small in comparison with its mean frequency, a quasimonochromatic ‘wave, It can be represented by two orthogonal harmonic ?- states, as in Eqs. (8-1) and (8.2), but here the amplitudes and iia! phase anges re funetions of time. Furthermore the tre- «quency and propagation nomber correspond tothe mea vals of the spectrum present inthe wave, namely, d and k. Thus (8.34a) wr + en) Een) = 1B .,(0) c08 1k: and E(t) = f(t) cos (8.34b) art en) The polarization state, and accordingly Eoy(1), Eoy(t).&(0)- and 6 (0), will vary slowly, remaining essentially constant over allarge number of oscillations, Keep in mind that the narrow bandwidth implies a relatively large coherence time. If we watch the wave during a much longer interval, the amplitudes and phase angles will vary somehow, either independently or in some correlated fashion. If the variations are completely uncorrelated, the polarization state will remain constant only foran interval that is small compared to the coherence time. In, other words, the ellipse describing the polarization state may change shape, orientation, and handedness. Since, speaking practically. no existing detector could discem any one partic: lar state lasting for so short a time, we would conclude that, the wave was unpolarized Antithetically, ifthe ratio Ey, (1)/Eo,(t) was constant even though both terms varied, and if e= «,(1) ~ ¢,(1) was constant as well, the wave would he polarized, Here the necessity for correlation among these different functions is obvious. Yet we can actually impress these conditions on the wave by merely passing it through a polarizer, thereby removing any undesired constituents. The time interval over which the wave thereafter maintains its polarization state is no longer dependent on the bandwidth because the wave's components have been appro- priately correlated. The light could be polychromatic (even ‘hte), yet completely polarized, t will behave very much like the idealized monochromatic waves treated in Section 8.1 Between the two extremes of completely polarized and unpolarized light is the condition of pastial polarization. In fact, t can be shown that any quasimonochromatie wave can, be represented as the sum of a polarized and an unpolarized wave, where the two are independent and either may be zero, 18.9 Polarization of Plyehromatic Light 358 8.9.2 Interference Colors Insert a crumpled sheet of cellophane between two Polaroids illuminated by white light. Alternatively, take an ordinary plastic bag (polyethylene), which shows nothing special between crossed Polaroids. and stretch it. That will align its ‘molecules, making it birefringent. Now crumple it up and. examine it again. The resulting pattern will be a profusion of multicolored regions, which vary in hue as either Polaroid rotates. These interference colors arise from the wavelength dependence of the retardation. The usual variegated nature of the patterns is due to local variations in thickness, bireftin- gence, or both, ‘The uppearance of interference colors is commonplace and ‘can easily be observed in any number of substances. For ‘example, the effect can be seen with a piece of multilayered mica. a chip of ice, a stretched plastic bag, of finely crushed particles of an ordinary white (quartz) pebble. To appreciate how the phenomenon occurs, examine Fig. 8.46, A narrow bheam of monochromatic linear light is schematically shown passing through some smuall region of a birefringent plate ¥. ‘Over that area the birefringence and thickness are both assumed to be constant. The transmitted light is generally elliptical. Equivalently, envision the light emerging from & as composed of two orthogonal linear waves (Le., the x- and y- components of the total E-field), which have a relative phase difference Ag, determined by Eg, (8.32). Only the compo- nents of these two disturbances, which are in the direction of the transmission axis of the analyzer, will pass through it and (on (0 the observer. Now these components, which also have a phase difference of Ag. are coplanar and can thus interfere. When Ag = =, 37, Figure 8.46 the ongn of interference colors360 Chapter 8 Polarization Sm... they are completely out-of-phase and tend (o cancel each other. When Ag = 0, 2m, 4... the waves ae in-phase and reinforce each other. Suppose then that the retardance arising at some point P, on 3 for blue light (Ay = 435 nm) is 4m. In that case blue will be strongly. transmitted. It follows from Eq. (8.32) that Ay = 2nd, ~ m4) is essentially a constant determined by the thickness and the birefringence. At the point in question, therefore, Ay Sg = 17407 forall wavelengths. If we now change to incident yel low light (Ay = 580 nm), Sy ~ 3zzand the light from P, is completely canceled. Under white-light illumination that par lar point on Swill seem as if t had removed yellow com pletely, passing on all the other colors, but none as strongly’ as blue. Another way of saying this i that the blue light emerg- ing from the repion about P; is linear Ag = 4) and parallel to the analyzer's transmission axis, In contrast, the yellow light is linear (Ag = 317) and along the extinction axis; the other colors are elliptical The region about P behaves like @ half-wave plate for yellow and ful-wave plate for blue. Ifthe analyzer were rotated 90°, the yellow would be transmitted, and the blue extinguished By definition two colors are said to be complementary ‘when their combination yields white light. Thus when the ana- lyzer is rotated through 90° it will alternately transmit or absorb complementary colors. In much the same way there right be a point Ps somewhere else on © where Ay = 4m for red (Ag = 650 nm), Then, Aug = 26007, whereupon bluish- green light (lo = 520 nm) will have a retardance of Sz and be extinguished. Clearly, if the etardance varies from one region to the next over the specimen. so t00 will the color ofthe ight transmitted by the analyzer. 8.10 Optical Activity ‘The manner in which light interacts with material substances. can yield a great deal of valuable information about their mol- cecular structures. The process to be examined next, although (of specific interest in the study of Optics, has had and is con- tnuing to have far-reaching effects in the sciences of chem- istry and biology. In 1811 the French physicist Dominique F. J. Arago first observed the rather fascinating phenomenon now known as ‘optical activity. It was then that he discovered that the plane of vibration of a beam of linear light underwent a continuous rotation as it propagated along the optic axis of a quartz plate (Fig. 8.47). At about the same time Jean Baptiste Biot Quan Figu 17 Optical activity displayed by quart (1774-1862) saw this same effect while using both the ‘aporous and liguid forms of various natural substances like turpentine. Any material that causes the E-field ofan incident linear plane wave to appear to rotate ix said tobe optically active, Moreover a8 Bio found, one must isinguishhetween Tights and left-handed rotation. I while looking in the dies tion ofthe source, the plane-of vibration appears to have revolved clockwise, the substance is referred 10 as desire tory, ot d-rotatory (rom the Latin desto, meaning righ). Alternatively, if E appears to have been displaced counter: clockwise, the materials levorotaars or Frotatory (rom the Latin fevo, meaning lf, tn 1822 the English astronomer Sir John F. W. Herschel (1792-1871) recognized that d-rolatory and rotatory bhav- iorin quar setualy correspond to twa different ental raphie structures. Although the molecules are identical (SiO), erystal quartz cane either right- or lefthandes, depending onthe atngement of those molecules. As shown in Fig. $48, the extemal appearances ofthese to forms are the same inal respects, excep that one isthe miro image of the other: they are sid tobe enanromorphs ofeach other. All transparent enantiomorphie substances are optically active. Farthermore, molten quart and fase quar, neither of which is crystalline, are not optically active. Evidently, in quartz optical activity is sociated with the structural distbution of the molecules asa whole. There are many substances, both organic and inorganic (-. benzil and NaBrOs, respectively ‘chick ke grt exhibit optical sctivity only in rt form In contrast, many naturally occuring organic compounds,Figure 8.48 fight and lefthanded quartz crystal. such as sugar, tartare acid, and turpentine. are optically active insoluton or inthe iquid state. Here the roratory power, as it is often referred to, is evidently an atribute of the individual molecules. There are also more complicated substances for which optical activity is associated with both the molecules themselves and their arrangement within the various crystal An example is rubidium tartrate, A d-rotatory solution ofthat compound will change to -rotatory when crystallized, In 1825 Fresnel, without addressing the actual mechanism involved, proposed a simple phenomenological description of| optical activity. Since the incident linear wave can be repre- sented a a superposition of -and stats, he suggested that these two forms of circular light propagate at different speeds. An active material shows cireular birefringence; thts, i pos sesses two indices of refraction, one for P-states (n,) and one for Fstates(n,).In traversing an optically active specimen, the wo circular waves would get out-of-phase, andthe resul tan linear wave would appear to have rotated. We can see how thisis possible analytically by ecurning to Eqs. (8.8) and (8.9), 8.10 Optical Actuty 361 which described monochromatic right- and left-circular light propagating in the z-direction. It was seen in Eq. (8.10) that the sum of these two waves is indeed linearly polarized. We now alter these expressions slightly in order to remove the factor of ‘00 in the amplitude of Eg, (8.10), in which ease 2 ficos thas ~ wn) + jsin hae ~ ew) 8.350) Bo fhcos thy ~ at) — jin ys on] (8.350) represent the right-and left-handed constituent waves. Since w is constant, ky = ky and ky = Kym. The resultant distur- bance is given by E = Ey + Ey, and after a bit of trigono- ‘metric manipulation, it becomes EB = By cos [ky + ky)e/2 — eat} Ee0s thy — kyie/2 + [sim (ky — ky)e/2) (8.36) Atthe position where the wave enters the medium (= = 0) itis linearly polarized along the x-axis, as shown in Fig. 8.49: that is, E = icosor (8.37) Notice that at any point along the path, the v0 components, have the same time dependence and are therefore in-phase. This just means that anywhere along the =-axis the resultant is, linearly polarized (Fig. 8.50), although its orientation is cer- tainly afunction of z. Moreover, if n> my or equivalently ky > ky. E will rotate counterclockwise, whereas ik > Ky the rotation is clockwise (looking toward the source). Traditional ly, the angle B through which E rotates is defined as positive Figure 849. the superposition of an ea an sale atz=0.362 Chapter 8 Poisizaton woo 1 when itis clockwise. Keeping this sign convention in ming, it should be clear from Eq, (8.36) that the field at point = makes. an angle of B = (ky ~ ky)z/2 with respect to its original orientation. Ifthe medium has a thickness d the angle through which the plane-of-vibration rotates is then ad = te nl 3 Baa 3 (8.38) Where my > my is d-rotatory and my > my is rotatory (Fig. 851), Fresnel was actually able 10 separate the constituent Jt- Figure 8.51 The superposition ofan Jt and an state at 2 ~ d Uy kA ky) and states ofa linear beam using the composite prism of Fig. 8.52. 1 consists of a umber of right- and left-handed «quartz segments cut with their optic axes as shown. The Sate propagates more rapidly in the first prism thn in the se- ond and is thus refracted toward the normal to the oblique boundary, The opposites tre forthe state, and the two ci cular waves inerease in angular separation at each interface In sodium light the specif rotatory power. whichis defined 38 B/dis found to be 21.7°/mm for quartz Ke follows that js = me = 7.1% 10 "for light propagating along te optic axis In that particular direction ordinary double refraction vanishes However, with the incident lisht propagating normal to the optic axis (ass frequently the casein polarizing prisms, wave plates. and compensators), quartz behaves like any optically inactive. positive, uniaxial crystal. There are other birefringent, optically active crystals, both uniaxial and biaxial such a cinnabar, HgS (n, = 2.854, n. = 3.201), which has & rotatory power of 32.5°/mm. fn contrast, the substance NaCIOs is opt cally active (3.1°/mm) but not birefringent. The rotatory pow- er of liquids, in comparison, isso relatively small that it "usually specified in terms of 10-em path lengths: for example, in the case of turpentine (CjgH,) it is only —37°/10 em (10°C with Ay = 589.3 am), The rotstory power of solutions varies withthe concentration. This facts particularly helpful in deter mining, for example, the amount of sugar present in a urine sample or commercial sugtr syrup You can observe optical activity rather easily using color Jess corm syrup, the kind availabe in any grocery store. You won't need much of it, since B/d is roughly +30°/inch. Put about an inch of syrup ina glass container between crossed Polaroids an iluminate i with a ashlight. The beautiful coloptic mes Figure 8.52 the Fresne! composite prism, ors that appear as the analyzer is rotated arise from the fact that Bisa function of Ay, an effect known as rorarory dispersion, Using a filter to get roughly monochromatic light, you can readily determine the rotatory power of the syrup.© ‘The first great scientific contribution made by Louis Pas- teur (1822-1895) came in 1848 and was associated with his doctoral research. He showed that racemic acid, which is an ‘optically inactive form of tartaric acid, is actually composed of | mixture containing equal quantities of right- and left-handed constituents. Substances of this sort, which have the same ‘molecular formulas but differ somehow in structure. are called isomers, He was able to crystallize racemic acid and then sep- arate the two different types of mirror-image crystals (enan- tiomorphs) that resulted. When dissolved separately in water, they formed d-rotatory and -rotatory solutions. This implied the existence of molecules that, although chemically the same, were themselves mirror images of each other; such molecules are now known as optical stereoisomers. These ideas were the basis for the development of the stereochemistry of organic and inorganic compounds, where one is concerned with the three-dimensional spatial distribution of atoms within & given molecule, 8.10.1 A Useful Model ‘The phenomenon of optical activity is extremely complicated, and although it can be treated in terms of classical Electro~ ‘magnetic Theory, it actually requires a quantum-mechanical “A geiabn fier works wel, but apiece of colored celloohane wil also a rcely. Just remember that the celophane wil act as a wave plate ste Section 8.7.1), so dont putt between te polaoids unless you alg its principal ates appropriately 8.10 Optical Achuty 363, solution,” Despite this, we will consider a simplified model, ‘which will yield a qualitative. yet plausible, description of the process. Recall that we represented an optically isotropic ‘medium by a homogeneous distribution of isotropic electron- ‘oseilltors that vibrated parallel tothe E-field ofan incident wave, An optically anisotropic medium was similarly depict ed as adistibution of anisotropie oscillators that vibrated a Some angle tothe driving E-field. We now imagine thatthe electrons in opially active substances are constrained 10 ‘move along twisting paths tat for simplicity are assumed 0 be helical Such a molecule Is pictured much as fit were a conducting helix. The silicon and oxygen atoms ina quart crystal are known tobe aranged in ether ight-o left-handed Spirals about the opte axis, a8 indicated in Fig. 8.53. In he presen representation ths crystal would corespond toa pa- allel aray of helices. In comparison, an active sugar solution ‘would be analogous (0a distribution of randomly oriented helices, each having the same handedness In quartz we might anticipate thatthe incoming wave would interaet differently withthe specimen, depending on whether tsa" right or lef-handed helices. Thus we could expect different indices fo the #-and components of the wave. The detailed treatment ofthe provess that leads to cir cular biteingence in crystals is by no means simple, but a least the necessary asymmetty is evident. How, then, can a random aray of helices, coresponding oa solution, produce ical activity? Let us examine one such molecule in hist plified representation, for example, one whose axis happens (0 be parallel tothe harmonic E-field of the elecromagnetic wave, That field wil drive charges up and down along the length ofthe molecule, effectively producing a time-varying elecre dipole moment (1), parallel to the ais, In adklition, "The review aticle “Optical Acuity and Molecular Dissymmety” by 8 F Mason, Contemp. Phys, 9, 239 (1968), contams 2 fay extensive Ist of references for futher reading. "In ation to these sold and aul states, theresa thd clasiicaton of substances, which s useful Because ofits remarkable optical proper tis, Its known asthe mesormorphic o Fguid crystal state, Ligud crys- tals are organic compounds that can fow and yet mati thei charac teristic moleculr orientation. in parbcuar, cholestere au crystals have a cal structure and therefore exhibit extremely large rotatory overs, ofthe order of 40000°/mm. The pc ofthe screwke molec lar arrangement is considerably smal than that of quart364 Chapter 8. Polsization g % ‘we now have a current associated with the spiraling motion o' the electrons. This in turn generates an oscillating magnetic dipole moment m/t), which is also along the helix axis (Fig, 8.54), In contrast. ifthe molecule was parallel to the B-field o the wave, there would be a time-varying flux and thus an induced electron current circulating around the molecule. This would again yield oscillating axial electric and magnetic dipole moments. In either case #1) and m/s) will be parallel or antiparallel to each other. depending on the sense of the particular molecular helix. Clearly, energy has been removed from the field, and both oscillating dipoles will scatter (i. reradiate) electromagnetic waves. The electric field E,, emitted ina given direction by an electric dipole is perpendicular to the electric field E,, emitted by magneti¢ dipole, The sum of these, which is the resultant field E, scattered by a helix, will not be parallel to the incident field E, along the direction of propagation, (The same is of course true for the magnetic fields.) The plane-of-vibration of the resultant transmitted light (E, + E,) will thus be rotated in a direction determined by the sense of the helix. The amount of the rotation will vary -ach molecule, but it will always be in the same direction for helices of the same sense. Although this discussion of optically active molecules as helical conductors is admittedly superficial, the analogy is well worth keeping in mind, In fact, if we direct a linear 3-cm microwave beam onto a box filled with a large number of identical copper helices (e.g., 1 em long by 0.5 em in diameter @ ( Optic axis A F CCC Figure 8.54 The radiation rom heica molecues.and insulated from each other), the transmitted wave will undergo a rotation ofits plane-of vibration.” 8.10.2 Optically Active Biological ‘Substances Among the most fascinating observations associated with optical activity are those in biology. Whenever organic mole: cules are synthesized in the laboratory, an equal number of d= and isomers are produced, with the effect that the compound. isoptically inactive, One might thea expect that if they exist at all, equal amounts of d- and L-optical stereoisomers will be ‘ound in natural organic substances. Thisis by no means the case Natural sugar (sucrose, CyH2,0),), no matter Where itis, ‘grown, whether extracted from sugar cane or sugar beets, is always d-rotatory. Moreover, the simple Sugar dextrose or d- slucose (CHj20,), which as its name implies is d-rotatory, is the most important carbohydrate in human metabolism. Evi- denily, living things can somehow distinguish between optical Al proteins are fabricated of compounds known as amino acids. These in tm are combinations of carbon, hydrogen, ‘oxygen, and nitrogen. There are twenty-odd amino acids, and all of them (with the exception of the simplest one, glycine, which is not enantiomorphic) are generally [-rotatory. This, ‘means that if we break up a protein molecule, whether it comes from an egg or an eggplant, a beetle or a Beatle, the constituent amino acids will be Lrotatory. One important exception isthe group of antibiotics, such as penicilin, which, do contain some dextro amino acids. In fact, this may well account for the toxie effect penicillin has on bacteria. kis intriguing to speculate about the possible origins of life on this and other planets. For example, did life on Earth origi- nally consist of both mirror-image forms? Five amino acids were found in a meteorite that fell in Victoria, Australia, on September 28, 1969, and analysis has revealed the existence of roughly equal amounts of the optically right- and left-handed forms, This is in marked contrast to the overwhelming pre- dominance of the left-handed form found in terrestrial rocks, Tre implications are many and marvelous.” “LTioco and M. . Freeman, “The Optical Activity of Oriented Copper 'see Physics Today, Feb. 1971, p. 17, for addtional discussion and ref crenes fr further reading 8.11 Induced Optical Efects—Opbical Modustors 365 8.11 Induced Optical Effects— Optical Modulators A number of different physical effects involving polarized Tight all share the single common feature of somehow being. externally induced. In these instances, one exerts an external influence (e.g., a mechanical force, a magnetic or electric field) on the optical medium, thereby changing the manner in ‘which it transmits Tight 8.11.1 Photoelasticity In 1816 Sir David Brewster discovered that normally trans parent isotropic substances could be made optically anisotrop: ic by the application of mechanical stress. The phenomenon is. known as mechanical birefringence, photoelasticity. or stress birefringence. Under compression or tension, the material takes on the properties of a negative or positive uniaxial ery tal, respectively. In either case, the effective optic axis is inthe direction of the stress, and the induced birefringence is pro- portional to the siress. Ifthe stress is not uniform over the sam. ple, neither isthe birefringence or the retardance imposed on a transmitted wave (Bq. (8.32)] Photoelasticity serves as the basis of a technique for study- ing the stresses in both transparent and opaque mechanical structures (see photo}, Improperly annealed or carelessly mounted glass, whether serving as an automobile windshield ‘ora telescope lens, will develop internal stresses that can ea ily be detected. Information concerning the surface strain on ‘opaque objects can be obtained by bonding photoelastic coat-366 Chapter 8. Poiarizaton ings to the parts under study. More commonly, a transparent, scale model of the partis made out of a material optically sen sitive to stress, such as epoxy, glyptol, or modified polyester resins. The model is then subjected to the forces thatthe actu al component would experience in use. Since the birefringence varies from point to point over the surface of the model, when itis placed between crossed polarizers, a complicated varie sated fringe pattern will reveal the internal stresses, Examine almost any piece of clear plastic or even a block of unilavored gelatin between two Polaroids; try stressing it further and watch the pattem change accordingly (see photos). The retardance at any point on the sample is proportional to the principal stress difference; that is, (7, ~ 72), where the sign the orthogonal principal stresses. For example, if the sample were a plate under vertical tension. 7, would be the maximum principal stress in the vertical direction and « \would be the minimum principal stress inthis case zero, hor izontally. In more complicated situations, the principal stress es, as well as their differences, will vary from one region to the next. Under white-light illumination, the loci ofall points on the specimen for which (2; ~ 3) is constant are known as isochromatic regions, and each such region corresponds to a particular colo. Superimposed on these colored fringes will tea separate sytem of black bands. Atany point where the E> field of the incident Tinea igh is parallel to ether local prin cipal stress axis, the wave will pass through the sample unaffected, regardless of wavelength, With erosed polarizers, that light will be absorbed by the analyzer, yielding a black region known as an isoctinie band (Problem 8.49). In addition to being beautiful to look at, the fringes also provide both a {qualitative map of the stress pattern ancl a basis for quantita tive calculations 8.11.2 The Faraday Effect Michael Faraday in 1845 discovered that the manner in which Tight propagated through a material medium could be influ enced by the application of an external magnetic field. In par ticular, he found that the plane-of-vibration of linear light incident on a piece of glass rotated when a strong magnetic, field was applied in the propagation direction. The Faraday Effect was one of the earliest indications of the interrelation: ship between electromagnetism and light. Although it is rem niscent of optical activity, there is an important distinct ‘The angle f (measured in minutes of arc) through which the plane-of-vibration rotates is given by the empirically determined expression B= Bd (8.39) ‘where B is the static magnetic flux density (usually in gauss), dis the length of medium traversed (in em), and 1s a factor of proportionality known as the Verdet constant. The Verdet, constant for a particular medium varies with both frequency olastic between crossed Poarods, (a) A permanently stressed piece of cle tem changes wih the application of 2 The finge pat(@ropping off rapidly as v decreases) and temperature. Its roughly of the order of 10” min of are gauss! cm’ for gas es and 10°? min of arc gauss”! em” for solids and liquids (Gee Table 8.2). You can get a better feling forthe meaning of these numbers by imagining for example a I-cm-loag sample ‘of HO in the moderately large field of 10° gauss. (The Earth's field sabout one half gauss.) In that panicle case, a rotation of 2°11’ would reslt since ¥"= 0.0131 By convention, a pestve Verde constan corresponds 104 (eiamagnetie) material for which the Faraday Effet Fro4e- tory when the light moves parallel tothe applied Bfeld and ttrotarory when it propagates antiparallel 10 B. No sich reversal of handedness occurs in the case of natural optical activity Fora convenient miemonic, imagine the Bild to be generated by solenoidal coll wound about the sample. The plane-of-vibration, when 1” positive, rotates in the same Giretion asthe curent in the col, regardless ofthe Bean's propagation direction along is axis, Consequently. the effect Can be amplified by reflecting the light back and forth afew times though the sample ‘The theoretical treatment ofthe Faraday Effect involves the quantur-meckanical theory of dispersion, including. the effets of B on the atomie or molecular energy levels. til suffice here merely to outline the imited classical argument for onmagnetic material. Suppose te incident light tobe circular and monochromat- ic-Anelasticlly bound electron wil ake on a steady-state cir cular obit being driven by the rotating E-feld of the wave (Meeffect ofthe wave's B field is negligible.) The intexdue fiom a large constant applied magnet ld pependicular to TABLE 8.2 Verdet Constants for Some Selected Substances Material Temperature (°C) -V (min of are gauss 'em=!) gh int pass 6 oon7 Waer 20 00131 act 6 00359 uuu 2 0166 NEFe(SO,) 12H, 26 ~po0nss ee o 627 x 10° cos o 939% 10% '578 nm and 760 mH Nore extensive listings are gven inthe usual handbooks. 8.11 Induced Ootica Efects—Optcal Modulators 367 the plane of the orbit will result in a radial force Fay on the electron. That force can point either toward or away from the circle’s center, depending on the handedness of the light and the direction ofthe constant B-field. The total radial force (Fy plus the elastic restoring force) can therefore have two differ- ent values, and so too can the radius of the orbit. Consequent- ly, fora given magnetic field there will be two possible values of the electric dipole moment, the polarization, and the per- mitivity, as well as two values of the index of refraction, nx and n. The discussion can then proceed in precisely the same fashion as that of Fresnel’s treatment of optical activity. As before, one speaks of two normal modes of propagation of electromagnetic waves through the medium, the 2% and ‘F states. For ferromagnetie substances things are somewhat more complicated, In the ease of a magnetized material is propor- tional to the component of the magnetization inthe direction of propagation rather than the component of the applied de field. ‘There are a numberof practical applications ofthe Faraday Effect. It can be used to analyze mixtures of hydrocarbons, since each constituent has a characteristic magnetic rotation When utilized in spectroscopic studies, it yields information about the properties of energy states above the ground level. Interestingly, the Faraday Effect has been used to make opti- cal modulators. An infrared version, constructed by R. C. LeCraw, utilized the synthetic magnetic crystal yttriumiron ‘garnet (YIG), to which has been added a quantity of gallium. YYIG has a structure similar o that of natural gem garmets. The device is depicted schematically in Fig. 8.55. A linear infrared laserbeam enters the crystal from the lefi. A transverse de magnetic field saturates the magnetization of the YIG crystal in tha direction. The total magnetization vector (rising from the constant field and the field of the coil) can vary in direc- tion, being tilted toward the axis ofthe crystal by an amount proportional fo the modulating current in the coil, Since the Faraday rotation depends on the axial component of the mat netization, the coil current controls B. The analyzer then con- verts this polarization modulation to amplitude modulation by way of Malus’s Law (Eq, (8.24)| In short, the signal to be transmitted is introduced across the coil 3s @ modulating volt- age, and the emerging laserbeam carries that information in the form of amplitude variations. ‘There are actually several other magneto-optic effects. We shall consider only two of these, and rather succinctly at that The Voige and Cotton-Mouton Eiffects both arise when a con- stant magnetic field is applied toa transparent medium per-368 Chapter 8 Poiarizaion ppendicular to the direction of propagation of the incident light ‘beam, The former occurs in vapors, whereas the latter. which is considerably stronger, occurs in liquids. In either case the ‘medium displays birefringence similar to that of a uniaxial cerystal whose optic axis isin the direction of the de magnetic field, that is, normal to the light beam (Eq. (8.2)]. The two indices of refraction now correspond tothe situations in which the plane-of-vibration of the wave is either normal or paralle| to the constant magnetic field. Their difference An (ie. the birefringence) is proportional to the square ofthe applied mag- netic field. It arses in liquids from an aligning of the optically and magnetically anisotropic molecules of the medium with that field. I the incoming light propagates at some angle to the static field other than O or 7/2, the Faraday and Cotton-Mou- ton Effects occur concurrently, with the former generally being much the larger of the two. The Cotton-Mouton is the ‘magnetic analogue of the Kerr (electro-optic) Effect, to be ‘considered next 8.11.3 The Kerr and Pockels Effects ‘The first electro-optic effect was discovered by the Scottish physicist Joh Kerr (1824-1907) in 1875. He found that an isotropic transparent substance becomes birefringent when placed in an electric field EF, The medium takes on the charac- {eristies of a uniaxial crystal whose optic axis corresponds to the direction of the applied field. The two indices, mand 1, are associated with the two orientations of the plane-of-vibra- Figure 8.55 A Foracay effect ‘mavilator. tion of the wave, namely, parallel and perpendicular to the applied electric field, respectively. Their difference, An, isthe birefringence, and it is found to be An = AgKE? (8.40) where K is the Kerr constant. When K is positive, as it most fofien is, An, which can be thought of as n.~ nis positive, and the substance behaves like a positive uniaxial crystal. Values of the Kerr constant (Table 8.3) are often listed in clectrostat- ic units, so that one must remember to enter E in Eq. (8:40) in statvolts per em (one statvolt ~ 300 V). Observe that, as with the Cotton-Mouton Effect, the Kerr Effect is proportional i the square of the field and is often referred to as the quadratic, electro-optic effect. The phenomenon in liquids is atributed partial alignment of anisotropic molecules by the E-field In solids the situation is considerably more complicated, Figure 8.56 depicts an arrangement known as a Kerr shut ter or optical modulator. It consists of a glass cell containing ‘ovo electrodes, which is filled with a polar liquid. This Kerr cell, a8 iti called, is positioned between crossed linear pola- izers whose transmission axes are at +45° to the applied E- field. With zero voltage across the plates, no Tight will be {ransmitte; the shutter is closed. The application of a modu- lating voltage generates a field, causing the cell to function as, variable wave plate and thus opening the shutter proporton- ately. The great value of such a device Ties in the fact that it, ‘can respond effectively to Irequencies roughly as high as 10"° Hz. Kerr cells, usually containing nitrobenzene or carbonTABLE 8.3 Kerr Constants for Some Selected Liquids (20°C, Ao = 589.3 nm) Substance K Gin units of 10-7 em statvolt*) Benzene cate 08 Carbon dso cs a2 loro CEC, a5 Water Ho. a7 icone HNO, a Niobenzene GaANOS 20 disulfide, have been used! for a number of years in a variety of applications, They serve as shutters in high-speed photogea- phy and as light-beam choppers to replace rotating toothed ‘heels. AS such, they have been utilized in measurements of the speed of light, Kerr cells are also used as Q-switches in pulsed laser systems. the plates functioning as the electrodes have an effective length of cm and are separated by a distance d, the retarda- tion is given by Ag = 2nkeve/d? (an) swhere V is the applied voltage. Thus a nitrobenzene cell in which dis one em and is several em will require a rather large voltage, roughly 3 X 10* V, in order to respond asa al = wave plate. This is a characteristic quantity known as the half wave voltage, V,j2. Another drawback is that nitrobenzene is both poisonous and explosive. Transparent solid substances, such as the mixed crystal potassium tantalate ntobate (KTag4sNbo.3503), KTN for short, or barium titanate (BaTiO,), which show a Kerr Effect, are therefore of interest aselectro-optical modulators. There is another very important electro-optical effect known as the Pockels Effect, alter the German physicist Friedrich Carl Alwin Pockels (1865-1913), who studied it ‘extensively in 1893. Iisa linear electro-optical effect, inas- much as the induced birefringence is proportional tothe first power of the applied E-field and therefore the applied voltage The Pockels Effect exists only in certain erystal that lack a center of symmetry —in other words, crystals having no cen- tral point through which every atom can be reflected into an identical atom. There are 32 crystal symmetry classes. 20 of which may show the Pockels Effect. Incidentally, these same 8.11 Induced Optical Efects Optical Modulators 369 20 classes are also piezoelectric. Thus, many crystals and all, liquids are excluded from displaying a linear electro-optic clfect. ‘The first practical Pockels cell, which could perform as a shutter or modulator, was not made until the 1940s, when suit- able crystals were finally developed. The operating principle for such a device is one we've already discussed. In bref, the birefringence is varied electronically by means of a controlled applied electric field, The retardance can be altered us desired, thereby changing the state of polarization ofthe incident linear ‘wave. In this way, the system functions as a polarization mod ulator. Early devices were made of ammonium dihydrogen phosphate (NH,H3PO,), or ADP, and potassium dihydrogen phosphate (KH:PO,), known as KDP; both are still in use. A. ‘great improvement was provided by the introduction of single erystals of potassium dideuterium phosphate (KD;PO,), or KD*P, whieh yields the same retardation with voltages less than half of those needed for KDP. This process of infusing crystals with deuterium is accomplished by growing them in 3 solution of heavy water. Cells made with KD*P or CD*A. (cesium dideuterium arsenate) have been produced commer cially for some time. A Pockels cell is simply an appropriate noncentrosymmet- tic, oriented, single crystal immersed in a controllable electric field, Such devices can usually be operated at fairly low volt ‘ages (roughly 5 to 10 times less than that ofan equivalent Kerr cell); they’ are linear, and of course there is no problem with toxic liquids. The response time of KDP is quite short, typi cally less than 10:ns, and it can modulate alight beam at up to about 25 GHz (ie., 25 % 10° Hz). Figure 8.56 A kerr cel370 Chapter 8 Polarization There are wo common cell configraions, referred ts transverse a longitudinal, depending on whether the aplied Eels perpendicular or parallel the direction of propass- ton, respectively. The longitudinal type illustrate, ints tnos basi form, in Fig. 887. Since the beam traverses the electrodes, these are usally made of transparent meta-onie coatings (eg. $90, nO, or CAO), thin mea films, gids, oF Fins. heer iself is generally nissan the absence of am applied field and its aligned such that its optic ox is along the beans propagation deetion, For such an atange- men the eardance is given by Ag = 2anbreV/Ay 8.42) Where ros is the electro-optic constant in m/V, nis the ordi nary index of refraction, Vis the potential difference in volts, and Ap is the vacuum wavelength in meters.* Since the crystals. are anisotropic, their properties vary in different directions, and they must be described by a group of terms referred to col lectively as the second-rank electro-optic tensor 7. Fortunate ly, we need only concern ourselves here with one of its ‘components, namely, roa, values of which are given in Table 84. The half-wave voltage corresponds to value of Ag in which case v Ag (8.43) Transparent ceetrode vote olaraer Figure 8.57 A Pociels cel “Ths exoression, along withthe appropriate one for the transverse ‘mode, is derived rater ricey in Yariv, Quantum Electronics. en so, the treatment is soptisteated and not recommended for casual reading TABLE 8.4 Electro-optic Constants (Room Temperature, Ao = 546.1 nm) 3 ny Vare Material (units of 10 m/V)_(approx.) (in kV) ADP (NHH,POD 8s 192 92 DP /K#.PO.) 106 151 1% KDA (KHAO =130 1s) 62 KO*P (KD,PO}) 23 12 =a and from Eq, (8.42) Ao. ‘n= Fae 8.44 a Der Lo) As an example, for KDP. 743 = 10.6 x 10-!? m/V, ny = 1.51, and we obtain Vaj2 ~ 7.6 X 10° V at Ay = 546.1 am. Pockels cells have been used as ultra-fast shutters, Q- switehes for lasers, and de to 30-GHz light modulators." 8.12 Liquid Crystals In 1888 the Austrian botanist Friedrich Reintzer observed that cholesteryl benzoate seemed to have two distinct transition points, one at which the crystal changed into a cloudy liquid and another where it became transparent. Known today as lig: uid erystal, he had discovered a new phase of matter that pos: sessed physical properties between those of ordinary liquids and solids, Liquid crystals (LCs) have long cigar-shaped mol- ecules that can move about, and consequently, like ordinary liquids, they lack positional order. Nonetheless, like crystals, their molecules strongly interact to sustain a large-scale orien- tational order. There are three types of liquid erystal distin- guished by the ways in which their molecules align. We'll focus on the nematic variety where the molecules tend t0 be ‘more or less parallel, even though their positions are fairly ran- dom (Fig. 8.58), To prepare a parallel nematic cell, we first coat one face of ceach of two pieces of flat glass with a transparent electrically conducting metalic film, such as indium tin oxide (which has ‘maximum transmission from 450 to 1800 nm). These two win- “The reader interested in ight modulation in general shoul consult D. F Nelson, "The Modulation of Laser Light." Scientific American June 1968), Also see Chaoter 14, Vol of Handbook of Optics (1995).