Scribd Logo
Upload

Welcome to Scribd!

  • 5

    Uploaded by

    Adit Wicaksono
    9 views20 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    9 views20 pages

    5

    Uploaded by

    Adit Wicaksono

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 20
    8.1 The Nature of Polarized ight Ithas already been established that light may be treated as @ transverse electromagnetic wave. Thus far we have considered only linearly polarized or plane-polarized light, that is, light for which the orientation of the electric field is constant, although its magnitude and sign vary in time (Fig. 3.14). In that case, the electric field o optical disturbance resides in ‘what is known as the plane-of-vibration. That fixed plane contains both E and k. the electric field vector and the propa- zation vector in the direction of motion, Imagine two harmonic, linearly polarized lightwaves of the same frequency, moving through the same region of space, in the same direction. If their electric field vectors are colinear, the superimposing disturbances will simply combine to form a resultant linearly polarized wave. Its amplitude and phase will, be examined in detail, under a diversity of conditions, in the next chapter, when we consider the phenomenon of interfer ence. On the other hand, if the two lightwaves are such that, their respective electricfield directions are mutually perpen- dicular, the resultant wave may or may not be linearly potar- iued. The exact form the light takes (i.e., its stare of polarization) and how we can observe it, produce it, change it, and make use of it isthe concern of this ehapter. 8.1.1 Linear Polarization ‘The two orthogonal optical disturbances that were considered. above can be represented as E,(z.1) = 1, cos (Az — wt) 1) a Jo, cos te - a +0) _ Polarization where e is the relative phase difference between the waves, both of which are traveling in the z-direction. Keep in mind from the start that because the phase isin the form (kz ~ «9, the addition of a positive e means that the cosine function in Eq, (8.2) will not attain the same value as the cosine in By (8.1) until a later time (e/w). Accordingly, E, lags E, by €> 0. Of course, if eis a negative quantity, E, leads E, by © < 0. ‘The resultant optical disturbane: twvo perpendicular waves: Ben 0+ Bye) 83) Ie is zero or an integral multiple of 27, the waves are said the vector sum of these Many animals can see variations in polarization jst 2s we see variations in ola. The pygmy octoous is one such creature. Te varying pattern of polarized light refected trom is surface suggests tight be “cor municating” with other pygmy octopuses, the way beds display coor (Peto cutesy Tra W Cova aR Sasa, Unverty f Nonne 326 Chapter 8 Polarization Figure 8.1 Linear ig. (2)The ld ieary polarized nthe fst and third quadrants. (by That same ‘ostiltng fed seen ead on. (c) Light Inearly polarized inthe second and fourth quaants to be in-phase. In that case Eq, (8.3) becomes E = GB. + Jo) 008 (kz ~ wn) 4) “The resoluant wave has a fixed ample equal 10 dE, + JE.) in other words, it 00 is linearly polarized (Fig. 8.1). The waves advance owarda plane of observation wher the elds Ivete be metsured. There one see a single resiltat ose iting, along a ited line cosinsoidall in ine (Fig. 8-10) The E-field progresses through one complete oscilaary cycle as the wave advances ong the z-axis through one Wave Tength This press canbe cared out equally well in reverse that is, we ean resolve any plane polarized wave into 160 onbopone component Suppose now that ¢ is an odd integer muliple of. The two wives ae 180" out-of phase, and E = (ik, ~ J.) ¢08 ks ~ wt) ws) This wave is again linearly polarized, but the plane-of-vibra- tion has been rotated (and not necessarily by 90°) from that of the previous condition, as indicated in Fig, 8.2 igure 8.2 Liner light osciating in the second and fourth quadrats, ae Ce Figure 8.3. Rightcrcular light. (a) Here the electric fd, wich has @ constant amplitude, rotates clockwise wih the same frequency wih hich Roseifates,(b) To perpendicular antemas ralating with 2 90° sae ference produce cicularly polerized electromagnetic waves, ‘The Nate of Polaized Light 327 8.1.2 Circular Polarization Another case of particular interest arises when both con- stituent waves have equal amplitudes (ie., Eo, Eo). and in addition, their relative phase difference e= —n/2 + 2mar, where m = 0, #1, *2,...In other words, e = ~2/2or any value increased or decreased from ~ 7/2 by whole-num- ber multiples of 2m, Accordingly By(z, 0) = bEgcos (ke wn) (8.6) and Ey(c.) = jE osin ke ~ on) aD The consequent wave is E = Eufivcos (kz ~ ot) +Jsinde- op) (8.8) (Fig, 83), Notice that now the scalar amplitude of F that is, (E*E)'/? = Eo, is a constant. But the direction of E is time- varying, and is not restricted, as before, toa single plane Figure 84 depicts what is happening at some arbitrary aon the axis, AUT = 0, E lies along the reference a Fig. 8a, and so E, = 1k, cos key and E, = [Eo sin key Ata later time, 1 = ka/t,E, = 18, B, = 0, and E is along the ans. The resultant electric feld vector E is rotating clockwise at an angular frequency of w, a8 seen by an obser Figure 8.4 Rotation ofthe electic vector in 8 rigttcrcular wave. Note that te rotation rate fs ‘wand ke = 1/4 328 Chapter 8. Polarization Figure 8.5. Rightcircuar ight Looking down the zaxis toward the or Bi, we se the electric field vector rotates clockmse asthe wave ‘advances toward the observer. cer toward whom the wave is moving (.., looking back atthe source), Such a wave is right-cireularly polarized (Fig. 8.5), and one generally simply refers to it as right-circular light. ‘The E-vector makes one complete rotation as the wave advances through one wavelength. In comparison, if = 77/2. Sa/2, 9m/2, and so on (ic., € = /2 + 2mm, where m = 0, £1, #2, #3,..), then E = Bali cos (kz — wr) — jsin (kz — ot] (89) The amplitude is unaffected, but EB now rotates counterclock- wise, and the wave is left-cireularly polarized. A linearly polatized wave can be synthesized from two ‘oppositely polarized circular waves of equal amplitude. In par- ticular, if we add the right-cireular wave of Eq. (8.8) to the lefi-circular wave of Eq, (8.9), we get = 264i cos (ke — 0) (8.10) which has a constant amplitude vector of 2Ff and is therefore linearly polarized 8.1.3 Elliptical Polarization As far as the mathematical description is concerned. both lin ear and circular light may be considered to be special cases of elliptically polarized light, or more simply elliptical light ‘This means that, in general the resultant electric-field vector E will rotate and change its magnitude, a5 well In such cs es the endpoint of wil race out an ellipse na fixed pace perpendicular to Kas the wave sweeps by, We can see ts Dever by ataly wring an expression forthe euve traverse by the tp of E. To tht end real that, E, = Ba,c0s ke ~ 0) uy and B= bcos ks — wr +0) 12) ‘The equation of the curve we ate looking for should not bea funetion of either positon or time; in other words, we should be able t0 get rid of the (Az ~ wn) dependence, Expand the expression for E,into E,/ Ey = 005 (kz ~ a1) 08 €~ sin (ke ~ a sin € and combine it with E, /. to yield Ee Be Fo Fa Ie follows from Eg. (8.11) that in (ks ~ of) sine @.13) sin (ke = wt) = [1 = Bx/ Egg)" 0 Bi. (8.13) leads to Finally. on rearranging terms, we have By -(B Bjorn ‘This is the equation of an ellipse making an angle a with the . E,)-coordinate system (Fig. 8.6) such that oko, COS Ba Boy Equation (8.14) might be a bit more recognizable ifthe princi pal axes of the ellipse were aligned with the coordinate axes, thatis, a= O or equivalently e= 2/2, =3x/2, #5x/2, in which ease we have the famitiar form 8.15) 8.16) 8.17) 8.1. The Netue of Polanzed Light 329 ‘These are both straight lines having slopes of 2 Eo, /Eo.: in other words, we have linear light Figure 8.7 diagrammatically summarizes most of these conclusions. This very important diagram is labeled across the bottom “E, leads E, by: 0, 7/4, m/2, 3/4..." where these are the positive values of to be used in Eq, (8.2). The same set of curves will occur if“E, leads E, by: 2m, 77/4, 31/2, Sa7/4....” and that happens when ¢ equals ~27, ~77/4, ~ 3/2, ~577/4, and so forth. Figure 8.7b illustrates how Fy leading £, by 2/2 is equivalent to E, leading E, by 3/2 Figure 8.6 Eliotica ight. Te endpoint ofthe electric field vector (where the sum of these two angles equals 27). This will be of sneeps out an elipse at rotates once around. ‘continuing concer as we go on to shift the relative phases of the (wo orthogonal components making up the wave, We are now in a position to refer to a particular lightwave which, in agreement with our previous results, isa circle, Ie in terms of its specific state of polarization. We shall say that isan even multiple of 7, Eq. (8.14) yields linearly polarized or plane-polarized light is in a -state, and right-of left-circular light is in an 2. or state, respectively Similarly, the condition of elliptical polarization corresponds to an &-sate. We've already seen that a P-state can be repre- sented 28 a superposition of 2%- and #states [Eg. (8.10)], and the same is true for an &-state. In this case, as shown in Fig. 88, the amplitudes of the two circular waves are different. (An (8,19) analytical treatment is lef for Problem 8.3.) (8.18) an m moo Baw ft 0 Teta Figure 8.7 (2) Various polarization configurations. The ligt would be cular with © = 1/2 or 32/2 f Eo, ~ Ey, but here for he sake of gener ity E, was taken to be lager than E. (DE leads E, (rE, les E,) by & 17/2, or ateratvely, E, leads E. (or E. lags 6, by 37/2 330 Chapter 8. Poisiztion ure 8.8 Eloticl it as the superpositon of an ot and stat, 8.1.4 Natural Light An ordinary light source consists of a very large number of | randomly oriented atomic emitters. Each excited atom radiates ‘a polarized wavetrain for roughly 10" s, All emissions hav. ing the same frequency will combine to form a single resultant polarized wave, which persists forno longer than 10". New ‘wavetrains are constantly emitted, and the overall polarization changes in a completely unpredictable fashion, If these changes take place at so rapid a rate as to render any single resultant polarization state indiscernible, the wave is referred to as natural light. It is also known as unpolarized light, but this is a misnomer, since in actuality the light is composed of a rapidly varying succession of the different polarization states. Randomly polarized is probably a better Way to speak of it We can mathematically represent natural light in terms of ‘ovo arbitrary, incoherent, orthogonal, linearly polarized waves of equal amplitude (ie., waves for which the relative phase 4ifference varies rapidly and randomly). Keep in mind that an idealized monochromatic plane wave ‘must be depicted as an infinite wavetrain. If this disturbance is resolved into two orthogonal components perpendicular to the direction of propagation, they, in tum, must have the same frequency, be infinite in extent, and therefore be mutually coherent (ie., ¢ = constant). In other words, a perfectly ‘monochromatic plane wave is always polarized. In Fact. Es, (8.1) and (8.2) are just the Cartesian components of a trans- verse (E. = 0) harmonic plane wave. Whether natural in origin or artificial, light is generally neither completely polarized nor completely unpolarized: both cases are extremes, More often, the electric-field vector varies in a way that is neither totally regular nor totally iregu- Jar, and such an optical disturbance is partially polarized One useful way of describing this behavior isto envision it as the result of the superposition of specific amounts of natural and polarized light 8.1.5 Angular Momentum and the Photon Picture We have sready seen that an electromagnetic wave impinging con an object ean impart both energy and linear momenta © that body, Moreover ithe incident plane wave is circularly polarized: we can expeet electrons within the material to be se into circular motion in response tothe force generated bythe rotating E-field. Alternatively, we might picture the field as being composed of two orthogonal Pstates that are 90 uta phase These simultaneously drive the electron in two perpendicular directions with a m/2 phase difference. The resting motion i again circular. nee, the trgue exerted by the Bild averages to zero over an orbits and the Feld hives the electton with an angular velocity equal tthe ie quency ofthe electromagnetic wave, Angolr momentum wil thus be imparted by the wave tothe substance in which the electrons are imbedded and to which the) are bound. We can treat the problem ater simply without actually going int the details ofthe dynamics. The power delivered tothe system the energy transfered pec uni ime, d/l. Furthermore the poower generated by atorgue Facing on aotating bay is us (oT (which is analogous to OF for linear motion) 50 ae dt or (8.20) Since the torque is equal to the time rate-of-change of the angular momentum L, it follows that on the average ae 821 dt dt ow A charge that absorbs a quantity of energy @ from the incident, cireular wave will simultaneously absorb an amount of angu- Jay momentum L such that, a 62) Ifthe incident wave is in an estate, its E-vector rotates clock- wise, looking foward the source. This isthe direction in which 1 positive charge in the absorbing medium would rotate, and ‘he angular momentum vectors therefore taken to point in the divetion opposite to the propagation direction," as shown in Fig. 89. ‘According fo the quantum-mechanical description, an elee- twomagnetic wave transfers energy in quantized packets or photons such that = hv. Thus & = hwo (where f= h/2m), and the intrinsic or spin angular momentum of a photon is either ~h or +, where the signs indicate right- or left-hand: edness respectively. Notice thatthe angular momentum of a photon is completely independent of is energy. Whenever a charged particle emits or absorbs electromagnetic radiation, slong with changes in its energy’ and linear momentum, it will undergo a change of hin its angular momentum ‘The eneray iransferred (oa target by an incident mono- chromatic electromagnetic wave ean he envisaged as being transporied in the form of a stream of identical photons. We can anticipate a corresponding quantized transport of angular nnomentum. A purely left-cireulaly polarized plane wave will Fgure 8.9. Anguiar momestum ofa photon. “This choice of terminology is admittedly bit award, Yet is use i esis farly well established, eventhough itis completely anthetc to themore reasonable convention adopted in elementary particle physics. "hs rater important et spe example, conse the hyrogen dt. tis composed ofa proton and an electron, each hav Spin of ‘v2. The atom hs sii mare energy when the spins of bth particles are inthe same drei. tis posse, however, tat once n 2 very To tre, roughly 10” years, one ofthe soins wil ip over ard be axipra tothe other, The change in angular momentum of the tom Ste, and tis imparted to an emit photon which cris off the sis excess in energy 25 wo. This isthe origin of the 21m meron emission, uhich s So Significant in ado astronomy. 12 Polariers 331 impart angular momentum tothe target as if al the constituent photons in the beam had their spins aligned in the direction of propagation. Changing the light to right circular reverses the spin orientation of the photons, as well as the torque exerted by them on the target. In 1935, using an extremely sensitive torsion pendulum, Richard A. Beth was actually able to per- form such measurement.* ‘Thus far we've had no difficulty in desesibing purely right- and left-creular ight in the photon picture; but what is linear- ly or ellipically polarized light? Classically, light in a Pstate ‘can be synthesized by the coherent superposition of equal ‘amounts of light in R-and F-staes (with an appropriate phase Lifference). Any single photon whose angular momentum is somehow measured will be found 10 have its spin either total ly parallel or antipaalle! to K. A beam of linear ight will iter- act with matter as iFit were composed, at that instant, of equal numbers of right-and left-handed photons. There i a subtle ‘point that has to be made here. We cannot say thatthe beam is actually made up of precisely equal amounts of well-defined right- and left-handed photons; the photons are all identical Rather. each individual photon exists in either spin state with equal likelihood. If we measured the angular momentum of the constituent photons. ~h would result as often as +f. This isall we can observe, We are not privy to what the photon is doing before the measurement (if indeed it exists before the measurement). As a whole, a linearly polarized light beam will impart no total angular momentum to a target. In contrast, ifeach photon docs not occupy both spin states with the same probability. one angular momentum, say +h. will be found to occur somewhat more often than the other, “hs In this instance, a net positive angular momentum will therefore be imparted to the target. The result en masse is cliiptically polarized light, that is, a superposition of unequal amounts of #- and flight bearing a particular phase relationship. 8.2 Polarizers Now that we have some idea of what polarized light is, the ‘next logical step is to develop an understanding of the tech- niques used to generate, change, and manipulate ito fit our needs. An optical device whose input is natural light and “rycnard A. Beth, “Mechanical Detecton and Measurement af the Angular Momentum of ight,” Phys. Rev. 0, 118 (1936). 332 Chapter 8 Polarization whose output is some form of polarized light is « polarizer, For example, recall that one possible representation of unpo- larized light isthe superposition of two equal-amplitude, inco- herent, orthogonal states. An instrument that separates these two components, discarding one and passing on the other, is, known as a linear polarizer, Depending on the form of the output, we could also have circular or elliptical polarisers. All these devices vary in effectiveness down to what might be called leaky or paral polarizers. Polarizers come in many different configurations, but they are all based on one of four fundamental physical mecha nisms: dichroism. or selective absorption; reflection: scatter- ing: and birefringence, or double refraction. There is. however, one underlying property that they all share: there ‘must be some form of asymmetry associated with the process. ‘This is certainly understandable, since the polarizer must somehow select a particular polarization state and discard all others. In truth. the asymmetry may be a subtle one related to the incident or viewing angle, but usually it is an obvious, anisotropy in the material ofthe polarizer itself 8.2.1 Malus's Law ‘One matter needs to be settled before we go on: how do we determine experimentally whether or not a device is actually 2 linear polarizer? By definition, if natura light is incident on an ideal linear polarizer, as in Fig. 8.10, only light in a state will be trans- ‘mitted. That state will have an orientation parallel to a spe~ cific direction called the transmission axis of the polarizer Only the component of the optical field parallel to the trans- ‘mission axis will pass through the device essentially unaffect- ed. Ifthe polarizer in Fig 8.10 is rotated about the <-axis, the | Bas Naot mits, oe Figure 8.10. Natural ight incident ona near polarize tited at an angle # wth respect to the vertical, reading of the detector (¢,g., a photocell) will be unchanged because af the complete symmetry of unpolarized lat. Keep in mind that we are dealing With waves, but Beeause of he very high frequency of light, our detector will measure only the incident radiance. Sine the radiance is proportional the square of the amplitude of the electric field (Eq. (3.44)}, ‘se need only concer ourseves with that ampli ow suppose tht we intradice a second identical ie in ca polarizer, or analyzer. whose transmission axis svete (Fig. 8.11 rte amplitude of the electric ld transmitted the firs polarizer is En, ony its component, Ecos pa lel wo the transmission axis ofthe analyzer will be passed on the detector (assuming no absorption). According to F (3.44). the irradiance reaching the detectors then given by 110) = iy cost ax) “The maximum ance, (0) = cep By/2 = occas when te angle Obeten the wansmisson ates ofthe analyzer and pla Figure 8.11 Alnear polarizer and ano Iyzer—Walus's Law. Natural ight of rac ance Ji incident on linear polarizer tited at an angle # wth respect to the vertical. The wadance leaving the frst linear oolcizers 1, = Hoy. The iradince leaving the second ear polanzer (ich makes an ane o ith te Fst 1 izeris zero, Equation (8.23) can he rewritten as 110) = 1(0} co This is known as Malus’s Law, having first been publ 1809 by Etienne Malus, military engineer and captain in the army of Napoleon Keep in mind that /(0) i the irradiance arriving on the ana- lyzer. Thus, if 1000 W/m? of natural light impinges on the first linear polarizer in Fig, 8.11, assuming that polarizer ideal, it wil pass 500 W/m? of linear light on to the analyzer; that’s (0). Depending on 8, we can use Eq, (8.24) to calculate the transmitted irradiance 1/8). Alternatively, suppose the inci- dent beam is 1000 W/m" of linear light parallel to the trans- mission axis of the first polarized. In that case 1/0) = 1000 W/m. Observe that 1/90°) = 0. This arises from the fact that the electric field that passed through the polarizer is perpendicular to the transmission axis of the analyzer (the two devices so amanged are said to be erossed). The field is therefore parallel to what is called the extinction axis ofthe analyzer and has no component along the transmission axis. We can use the Setup of Fig. 8.11 along with Malus's Law to determine whether a particular device isa linear polarizer. ‘As we'll see presently, the most common kind of finear polarizer used today is the Polaroid filter. And although you certainly can confirm Malus’s Law with two ordinary Polaroids. you'll have to be careful to use light in the range from =450 nm to =650 nm. Ordinary Polaroids are not very 00d at polarizing IR, 8.3 Dichroism {nits broadest sense, the term dichroism refers to the selective absorption of one of the two orthogonal ?-state components, ofan incident beam. The dichroic polarizer itself is physically anisotropic, producing a strong asymmetric or preferential absorption of one field component while being essentially transparent to the other 8.3.1 The Wire-Grid Polarizer ‘The simplest device of this son is a grid of parallel conducting, wires, as shown in Fig. 8.12. Imagine that an unpolarized elec- tromagnetic wave impinges on the grid from the right. The clectrie field can be resolved into the usual two orthogonal 8.3 Dictwoism 333 ‘components, in this case, one chosen to be parallel tothe wires and the other perpendicular to them, The y-componeat of the field drives the conduction electrons along the length of each wire, thus generating a current, The electrons in turn collide with lattice atoms, imparting energy to them and thereby heat- ing the wires (joule heat). In this manner energy is transferred from the field to the grid. In addition, electrons accelerating along the y-axis radiate in both the forward and backward directions. As should be expected, the incident wave tends to be canceled by the wave reradiated in the forward direction resulting in little oF no transmission of the y-component of the field, The radiation propagating in the backward direction simply appears as a reflected wave. In contrast, the electrons, are not free to move very far in the x-direction, and the corre- sponding field component of the wave is essentially unaltered as it propagates through the grid. The transmission axis of the grid is perpendicular to the wires. It is a common error to assume naively that the y-component of the field somehow slips through the spaces between the wires. ‘One ean easily confirm our conclusions using microwaves and a grid made of ordinary electrical wire. I is not so easy a ‘matter, however, to fabricate a grid that will polarize light, but it has been done! In 1960 George R. Bird and Maxfield Par- fish, Jr. constructed a grid having an incredible 2160 wires per ‘mm.* Their feat was accomplished by evaporating a stream of | {gold (ort other times aluminum) atoms at nearly grazing inei- ence onto a plastic diffraction grating replica (see Section 10.2.7). The metal accumulated along the edges of each step in the grating to form thin microscopic “wires” whose width and spacing were less than one wavelength across. Figure 8.12 A weed polarizer. The grid eliminates the vertical component ive, the one paral tothe wres) ofthe Efe and passes the horizontal component. "6. R. Bird and M. Parish, J. "The Wie Grd as a Nesrhnfrared Polarizer” J. Opt. Soc. Am. 50, 886 (1960) 334 Chapter 8. Polarization Although the wire grid is useful, particularly in the infrared, tis mentioned here more for pedagogical than prac tical reasons. The underlying principle is shared by other, ‘more common, dichroic polarizers. 8.3.2 Dichroic Crystals Cerin matrals re inherently dichroic because of an anisotropy in heir respective erytalline structures. Probably thebest known these the natal occuring mineral malin, asemipreious stone often used in jewelry Actually there af several ourmalines, which are boron silicates of di. fering chemical composition fg, NaFe,B AlSiOsy OH), ] For this substance there ia specifi direction within the ery tal known the principal or api xi, which is determined by its atomic configuration. "The electri ield component of an incident lightwave thats perpendicular tthe principal ais is Stomaly absorbed bythe sample. The ticker the cyst the rior complete the absorption (Fig. 8.13) A plate cut froma tourmaline crystal parallel toi repel ais and several ail limeters thik wil serve a linear polarizer. In this nlance the crystals principal wus becomes the polarze's tans sion axis But the useflness of tural i rather ited by the fat tat its erystals are comparatively smal. Moreover even te ransnied ight suffers &cenain amount of absorp tion. To complicate matters, this undesirable absorption ts strongly wavelength dependent, and the specimen will there fore be Colored. A tourmaline esta held upto natural bite light might appear green (ey come in other colors a8 well) then viewed normal tothe principal axis and nearly black ten viewed along that axis Wher al te E-felds are perpen: dicular to H ergo the term dichroic, meaning nwo colors) There are several oer substances that display similar char acters. Aerytal ofthe miner! hyperthene, a feromag- nesium silicate, might look preen under white ght polarized imone tection and piak fora differen polarization dnection ‘We can get a qualitative picture ofthe mechanism that ives ise to ryt dichroism by considering the microscopic Structure ofthe sample. You might want to take another look a Section 3.5) Recall that the atoms within a crystal ate Stronaly hound together by short-range foes to for pert oc latce The electrons, which are esponsibe forthe opi cal properties, canbe envisioned as elastically ed to ci Tetpecilveeqllbium positions. Blesironsasociated with & given atom te also node te feuence of the surrounding hear sows, which themselves may not be ymaerielly Figure 8.13 A dichroic crystal. The E'ld paral othe optic ais is transmitted without ary deanton The natualy occurring ridges evdere inthe photograoh of the tourmaline crystals correspond tothe apt Bs. Pat ov distributed. AS a result, the elastic binding forces on the elec trons will he different in different directions. Consequently their response to the harmonic electric field of an incident electromagnetic wave will vary with the direction of E. IFin addition to being anisotropic the material is absorbing, a detailed analysis would have fo inelude an orientation-depes- dent conductivity, Currents will exist, and energy from the wave will be converted into joule heat. The attenuation, in addition to varying in direction, may be dependent on fre- ‘quency as well. This means that ifthe incoming white light is in a Pstat, the crystal will appear colored, and the color will depend on the orientation of E, Substances that display two or even three different colors are said to be dichroic or trichroi. respectively. 8.3.3 Polaroid In 1928 Edwin Herbert Land, then a 19-year-old undergradu- ale at Harvard College, invented the first dichroic sheet polar- izer, known commercially as Polaroid J-sheet. I incorporated a synthetic dichroic substance called herapathite, or quinine sulfate periodide.” Land's own retrospective account of his, early work js rather informative and makes fascinating read- ing. I is particularly interesting (0 follow the sometimes, whimsical origins of what is now, no doubt, the most widely used group of polarizers. The following is an excerpt from Land's remarks: In the literature there are a few pertinent high spots in the ‘development of polarizers particularly the work of William Bird Herapath, a physician in Bristol, England, whose pupil. a Mr. Phelps, had found that when he dropped iodine into the urine of a dog that had been fed quinine, litle scintillating sreen crystals formed inthe reaction Figu. Phelps went. is, teacher, and Herapath then di something which Land] think was curious under the circumstances; he looked atthe erystals under a microscope and noticed that in some places they were light where they overlapped and in some places they were lark. He was shrewd enough to recognize that here was 2 remarkable phenomenon, a new polarizing material [now known as herapathit] Herapath's work eaught the attention of Sir David Brow ster, who vas working in those happy day’ on the kaleido= scope... Brewsie, who invented the kaleidoscope, wrote a ‘ook aout and in that book he mentioned that he would ike touse heraputhite crystals forthe eyepiece. When Iwas rea ing this book, back in 1926 and 1927, | eame across his refer ence to these remarkable crystals, and that started my’ interest ‘in eeapatit. “Wore wil be said about these pracesses late on when we consider bietingence. Suffice to sey now that for crystals classified 2 uniaxial bere are to distinct drectons, and terefare two colors may be ds dyed by absorbng specimens. n biaxial crystals there are tree ds inet drectons and the possbilly of tree coors. LH Land, “Some Aspects ofthe Development of Sheet Polarzers,” J Ot Soc, Am, 41, 957 1951) 8.3 Dichrowm 335 -Apair of crossed polaoids. Each polarod appears gray becouse i absorbs raughy hal the incident ht. Pato by EA Land’s initial approach to creating @ new form of linear polarizer was to grind herapathite into millions of submicro- scopic erystals, which were naturally needie-shaped. Theit smal siz lessened the problem of the scattering of light. In his earliest experiments, the erystals were aligned nearly parallel to each other by means of magnetic or electric fields. Later, Land found that they would be mechanically aligned when a viscous colloidal suspension of the herapathite needles was extruded through a long narrow slit, The resulting J-sheet was effectively a large flat dichroic crystal. The individual submi croscopic crystals still scattered light a bit, and as a result, J- sheet was somewhat hazy. In 1938 Land invented H-sheer, which is now probably the most widely used linear polarizer. It does not contain dichroic tals but is instead a molecular analogue of the wire grid. A sheet of clear polyvinyl alcohol is heated and stretched in a given direction, its long hydrocarbon molecules becoming aligned in the process. The sheet is then dipped into an ink fin iodine. The iodine impregnates the plastic and attaches to the straight long-chain polymeric molecules, effec- tively forming a chain of its own. The conduction electrons associated with the iodine can move along the chains as if they ‘were long thin wites, The component of E in an incident wave that is parallel to the molecules drives the electrons, does work fon them, and is strongly absorbed. The transmission axis of the polarizer is therefore perpendicular to the direction in which the film was stretched. Each separate miniscule dichroic entity is known as a dichromophore. In H-sheet the dichromophores are of molee- ular dimensions, so seattering represents no problem, H-sheet isa very effective polarizer across the entire visible spectrum, but is somewhat less so atthe blue end, When a bright white light is viewed through a pair of crossed H-sheet Polaroids the solution 336 Chapter 8. Polarization extinction color will be a deep blue as a result of this leakage. HN-50 would be the designation of a hypothetical, ideal H- sheet having a neutral color (N) and transmitting 50% of the incident natural light while absorbing the other 50%, which is the undesired polarization component. In practice, however, bout 4% of the incoming light will be reflected back at each surface (antireflection coatings are not generally used), leav ing about 92%. Half of this is presumably absorbed, and thus ‘we might contemplate an HIN-46 Polaroid. Actually, lange ‘quantities of HN-38, HN-32, and HN-22, each differing by the amount of iodine present, are produced commercially and are readily available (Problem 8,10), Many other forms of Polaroid have been developed.* K-sheet, which is humidity: and heat-resistant, has as its dichromophore the straight-chain hydrocarbon poly vinylene, A combination of the ingredients of Hand K-shects leads to HR-sheet, a near-infrared polarizer. And there are commer- cially available dichroic sheet linear polarizers that function in the ultraviolet from ~300 nm to =400 am. ‘Remember that sheet dichroic polarizers are designed for a specific wavelength range. A pair of crossed sheet linear polarizers intended to block the visible will leak substantially below =450 nm and above ~650 nm, Polaroid veetograph is a commercial material at one time designed to be incorporated in a process for making three- dimensional photographs. The stuff never was successful at its intended purpose, but it ean be used to produce some rather thought-provoking, if not mystifying, demonstrations. Vecto- graph film is a water-clear plastic laminate of two sheets of | polyvinyl alcohol arranged so that their stretch directions are ‘at right angles to each other. In this form there are no condue- tion electrons available, and the film is not a polarizer. Using ‘an iodine solution, imagine that we draw an X on one side of the film and a ¥ overlapping it on the other. Under natural ilu- ‘mination the light passing through the X will be in a Pstate perpendicular to the state light coming from the ¥. In other ‘words, the painted regions form two crossed polarizers. They «will be seen superimposed on each other. Now, if the vecto~ ‘graph is viewed through a linear polarizer that can be rotated, either the X, the Y, or both will be seen. Obviously, more imaginative drawings can be made. (One need only remember ‘o make the one on the far side backward.) “See Polarized Light: Producton and Use, by Suc, arts more read- able ite brother, Polarized Light, by Sturclff and Balad 8.4 Birefringence Many cnytilline substances (i.e, slkls whose atoms ae am©anged in some sort of regular repetitive ara are opicaly anisotropic. Tei oplial properties are not the same nal direction within any given sample. The dichroic ens of the previous section are but one special subgroup. We saw there that ifthe crystals lticeafoms were not completely symmetrically arayed, the binding forces on te electrons oul be anisotopi,Ealer, in Fig 3.386 we represented the isotropic osilar using the simple mechanical model ofa spherical charged sell bound by identical springs to fixed point. This was ine for opialy isotopic substances (amor Dhous solids, such as glas and pes, are ull. but ma tiways, isotopic Figure 8.14 shows another charged shel this one bound by spring of differing stfnes i, having diferent spring constant). An leton thats pled from ultram long direction parallel to one sto springs” wal evidently oscillate with ferent characterise tequen Ey than it would were it displaced in some other direction AAs was pointed ou previously. ligh propagates though a transparent substance by excing ie atoms within the mei tm. Phe electrons ae driven bythe Efe, and they read these secondary wavelets recombine. andthe sulanrefact ed wave moves on. The speed of the wave, and therefore he index of retraction, i determined bythe difference between the frequency of the E-field and the natural frequency of the Figure 8.14 Mechanical model depicting a negatively charged shel bound to positive nucleus by pars of springs having different sifess. Figure 8.18. Retractve index versus frequency along two ates in @ tal. Relons where c/a < 0 correspond to absortion bands. som. An anisotropy in the binding force will be manifest in anisotropy in the refractive index. For example, if P-state light was to move through some hypothetical crystal so that it ‘encountered electrons that could he represented by Fig. 8.14, its speed would be governed by the orientation of E, IF E was parallel to the stiff springs, that is, in a direction of strong binding, here along the x-axis, the electron's natural frequen- ¢y Would be high (proportional to the square root ofthe spring constant. In contrast, with E along the y-axis, where the bind= ing force is weaker, the natural frequency Would be somewhat lower. Keeping in mind our earlier discussion of dispersion and the n(w) curve of Fig, 3.41, the appropriate indices of ‘eftaction might look like those in Fig, 8.15, A material of this Sor, which display’s two different indices of refraction, is said lobe birefringent * If the erystal is such that the frequency of the incident light appears in the vicinity of @,, in Fig. 8.15. it resides in the ahsomption band of n/a), A crystal so illuminated will be strongly absorbing for one polarization direction (y) and trans- parent for the other (x). A birefringent material that absorbs “The word refingence used to be used instead of our presently term retacton, comes from the Lata reractus by way ofan etymological rate beginning mth ranger, meaning to break. 8.4 Bretrngence 337 one of the orthogonal states, passing on the other. is dichroic. Frthermore. suppose that the crystal symmetry i Such thatthe binding ores the and zdrectons are iden Tel in ther words, each of these springs ha the sare natu alfrequeney and they ae equally ss. The axis no defines the direction of the optic axis. Inasmuch asa erytl ean be represented by an ray ofthese oriented ansoropic charged sella the opi axis is actualy a diretion and nol mere: tra single ine, The model works rather nicely fr dichoig erysials, since i light was to propagate along th ope axis (E Inthe plane), it would be strongly absorbed. nd if moved normal that ait would emerge linet polarized (Often the natural requencies of bireingent crystals re above the optical range, and they appear colores. Ths rep- fesened by Fig 815 where the incident light snow coms ered 1o have frequencies in he region of wy. Two diferent indices are apparent, but absorption fa either polarization is negligible Equation (3.71) shows that mas) varies inversely with he matral frequency This means that lage effective Spring constant (ie. strong binding) corresponds toa tow Polarzabliy, alow dielectric constant anda low refractive index "We wl eostuc, foal pictorial, a linear polarizer a- lizingbireingence by causing he two orthogonal states to follow differen paths and separate. Even more fascinating things canbe doe with birefringent crystals a5 ve shal ee inter 8.4.1 Calcite Let's spend a moment relating the above ideas to a typical birefringent crystal, caleite. Calcite or calcium carbonate (CaCOs) is a common naturally occurring substance. Both ‘marble and limestone are made up of many small calcite crys- tals bonded together. OF particular interest are the beautiful large single crystals, which, although they are becoming rare, ‘can still be found, particularly in India, Mexico, and South Arica. Calcite is the most common material for making linear polarizers for use with high-power lasers. Figure 8.16 shows the distribution of carbon, caleium, and ‘oxygen within the caleite structure; Fig. 8.17 isa view from above. looking down along what has. in anticipation, been labeled the optic axis in Fig. 8.16, Each CO, group forms a tri- angular cluster whose plane is perpendicular to the optic axis. If Fig. 8.17 is rotated about a line normal to and passing 338 Chapter 8. Polarization through the center of any one of the erbonate groups, the same enact confgiration of tons would appear Ire tes dhorng each revolution. The diection designated a the optic ‘xis corresponds ton special crystallographic orientation in that isan axis of 34d symmetry. The lrg birefringence dleplayed by alte aries from the fact thatthe carbonate groups ar all in plans norma othe optic ans. Te behavior Of tir electrons, or rather the mutual interaction of the induced oxygen dipoles. markedly diferent when Eis eter im or noxmal to those planes (Problem 8.25). In any even the asymmetry is clea enough Calcite samples can readily be split forming smooth sur foces known as cleavage planes. The crystal i esenially made to come apart tween specific planes of atoms where the interatomie, bonding fs relatively Weak All cleavage planes nea (Fg. 8-17) are norma to hee efferent dre: tions Asa crystal grows, sloms ae aldd layer upon layer. following the same pate. But more raw material may be available othe growth process on one side than on another, resulting i a exysal with an externally complicated shape. Even so, the cleavage lanes are dependent onthe atom eon. Higuraion and if one cuts sample so that each sutface is & cleavage plane. its orm wil be related tothe Base arange- | Optic Caubon catsum © ove ©) | Figure 8.16. Arrangerent of atoms in cle Figure 8.17 Alomic arrangement for caeite looking down the optical ans ‘ment of its atoms. Such a specimen is referred t0 as a cleay- age form. In the case of calcite it is a thombohedron, with each face a parallelogram whose angles are 78° 5’ and 101° 55’ (Fig. 8.18). ‘There are only (wo blunt comers where the surface planes meet to form three obtuse angles. A Tine passing through the vertex of either of the blunt comers, oriented so that it makes ‘equal angles with each face (45.5°) and each edge (63.8°), i clearly an axis of 3-fold symmetry. (This would be a bit more obvious if we cut the rhomb to have edges of equal length) Figure 8.18. Calcte cleavage form, Evidently, such a ine must correspond to the optic axis. Wha ever the natural shape of a particular calcite specimen, you need only find a blunt corner and you have the optic axis, In 1669 Erasmus Bartholinus (1625-1692), doctor of med: icine and professor of mathematics atthe University of Copen: hagen (and incidentally, the father-in-law of Ole Romer, the man who in 1679 first measured the speed of light), came upon a new and remarkable optical phenomenon in calcite, Which he called double refraction. Calcite had been discovered not long before, near Eskifjordur in Iecland, and was then known as Iceland spar. In the words of Bartholinus:® Greatly prized by all men is the diamond, and many are the joys which similar treasures bring, such as precious stones and pears... bute, who, on the ther hand, prefers the knowledge ‘of nasual phenomena to these delights, he will I hope. have noless joy in anew sort of body. namely, transparent crystal, recently brought ro us from Iceland, which perhaps is one of the greatest wonders that nature has produced. [As my investigation of this erytal proceeded there showed itself a wonderful and extraordinary phenomenon: objects ‘which are looked at through the erystal do nat show. a in the «ase of other transparent bodies, a single refracted image, but ‘ey appear double. ‘The double image referred to by Bartholinus is quite evi dent in the accompanying photograph. If we send a narrow beam of natural light into a caleite crystal normal to a cleavage plane, it wil split and emerge as two parallel beams. To see the same effect quite simply, we need only place a black dot “W.F, Magi, A Source Book in Pysies. 8.4 Bretnngence 339 Double image formed by calcite crystal (not cleavage form) ‘on a piece of paper and then cover it with a caleite rhomb. The image will now consist of two gray dots (black where they overlap). Rotating the crystal will cause one of the dots to remain stationary while the other appears to move in a circle about it, following the motion of the crystal. The rays forming the fixed dot, which is the one invariably closer to the upper blunt corner, behave as if they had merely passed through a plate of glass. In accord with a suggestion made by Bartholi- rus, they are known as the ordinary rays, or o-rays. The rays coming from the other dot, which fashion, are Known as the extraordinary rays, ot e-rays. If the crystal is examined through an analyzer, it will be found that the ordinary and extraordinary images are Finearly potar- ized (see photo). Moreover, the two emerging Pstates are ‘orthogonal ehave in such an unusual 2 erystal (Bunt corer on the bottom. The transmission axes of the two polarizers are paral to thee short edges. Where the image is doubled the lower, undetected one is the ordinary image. Take along Took: there's aot inthis one, matby Ex 340 Chapter 8. Poltzation Any number of planes can be drawn through the rhomb so sto coniain the optic axis, and these are all called principal planes. More specifically ifthe principal plane is also normal oa pair of opposite surfaces of the cleavage form, it slices the crystal across a prineipal section. Evidently, three of these pass through any one point; each is a parallelogram having angles of 109° and 71°. Figure 8.19 is a diagrammatic repre- sentation of an initially unpolarized beams traversing a princi- pal section of a caleite rhomb. The filled-in citles and arrows drawn along the rays indicate that the o-ray has its electric- field vector normal tothe principal section, and the field ofthe e-ray is parallel tothe principal section. To simplify matters a bit, let Ein the incident plane wave be linearly polarized perpendicular t© the optic avis, as shown in Fig. 8.20. The wave strikes the surface of the erys- tal, thereupon driving electrons into oscillation, and they in turn reradiate secondary wavelets. The wavelets superimpose and recombine to form the refracted wave, and the process is repeated over and over again until the wave emerges from the crystal, This represents a cogent physical argument for applying the ideas of scattering via Huygens’s Principle, Huygens himself, though without benefit of electromagnetic theory, used his construction to explain many aspects of dou- Le opi ase Figure 8.19 A ight beam wit two orthogonal field companents tr vetsing a cakete pncpal section. Figure 8.20 An incident plane wave polarized perpendicsr tothe prcipal section ble refraction in calcite a8 long ago a8 1690, It shouldbe made lear from the oulet, however, that his treatment incomplete in which form itis nppealinly, though decep tively simple, Inasmuch asthe Eel i perpendicular othe opie axis, one assumes thatthe wavelron simulates countless toms on thesurTae, which ten acta soutesof spherical wavelets al of which sve in-phase. Presumably. as longus the fl ofthe traveler ds everwhere normal the opie ct, they wl fxpand nto te erystal in all directions with speed, a they would inan sotropie medium, (Keepin und thatthe speed isa function of fequeney,) Since the o-vave display no anomalous behavior. this assumption seems reasonable, The envelope ofthe waveles is essentially a portion ofa plane wave, which intrn stimulates a distribution of secondary omic point sources. The process continues, and the wave moves saight across heer st In comast, consider the incident wave in Fig. 8.21 whose E-field is parallel tothe principal section. Notice that E now has a component normal tothe optc axis, a6 well as 3 component parallel to it. Since the medium is birefringent, Tight ofa given frequency polarized parallel tthe optic axis propagates with speed where o,# 7. In priula for faleite and sodium yellow light = $89 nm), 1.4860, = 16580, ~ e What kindof Huygenss wavelets can we expect “A, Sommerfeld, Opts, p. 148. Figure 8.21. An incident plave wave polarized paral 0 the perp secton. ‘ow? AL the risk of oversimplifying matters, we represent each e-wavelet, for the moment at least, as a small sphere (Fig, 8.22), But v, > 0, so that the wavelet will elongate in all, DL (Me=n,) 18 ~0.172, and it is negative uniaxial. In com parison, there are other erystals, such as quartz (crystallized silicon dioxide) and ice, for which v, > v,, Consequently. the ellipsoidal e-wavelets are enclosed within the spherical o- wavelets, as shown in Fig. 8.25, (Quattz is optically active and therefore actually a bit more complicated.) In that ease, (0, ~ n,)is positive, and the crystal i positive uniaxial The remaining crystallographic systems, namely orthorhombic, monoclinic, and triclinic, have two optic axes and ate said tobe biaxial. Such substances, for example, mica IKHAI(Si0,)x]. have three different principal indices of refraction, Each set of springs inthe oscillator model would then be diferent. The birefringence of biaxial erystas is mea- sured as the numerical difference between the largest and smallest ofthese indices. Ae fe UV ew u Figure 8.24 Wavelets in a negative uniaxial crystal. The arrows and dos reoresent the elds f the extraordinary and ordinary waves, respectvaly. The Efe of te owave i everyanere perpendicular tothe optic as. 8.4.3 Birefringent Polarizers Iewill now be an easy matter, at least conceptually, to make Some sort of linear birefringent polarizer. Any number of schemes for separating the o- and e-waves have been employed, all of them relying on the fact that n, # Mt The most renowned birefringent polarizer was introduced in 1828 by the Scottish physicist William Nicol (1768-1851). The Nicol prism is now mainly of historical interest, having TABLE 8.1 Refractive Indices of Some Uniaxial Birefringent Crystals (Ao = 589.3 nm) Crystal ny ne ‘Tourmaline 169 1938 Caleie see aso use 15H uss Sodium ni sas 1349 ee 1.09 1313 tile (1103) 2616 2908 8.4 Bietringence 343 Figure 8.25. Wavelets in» poste ural crystal. The artows and ots represent the Efiels of the extraordinary and ordinary waves, Fespectwely, The Elie of the owave is everywhere perpendicular to the optic axis long been superseded by other, more effective polarizers Putting i rather succinctly. the device is made by first erind- ing and polishing the ends rom 71° to 6; se Fig. 820) of suitaly Tong, now esleiterhombohedron: ter citing the Thomb diagonally. the wo pieces are polished and cemented back together with Canada balsam (Fig. 8.26), The balsam cements transparent and has an index of 1.55 almost midway between n and, The incident beam enters the “pris,” The oan ery are refracted they separate and strike the alsa layer. The critical angle a the euleite-balsum interface forthe ‘o-ay is about 69° (Problem 8.38) The -ray (entering within A narow cone of roughly 2°) will be tall Intemally reflect fed and thereafter absorbed by a layer of black paint on the sides ofthe rhomb. The eay emerges laterally displaced but otherwise essentially unscathed, a least inthe opiea region ofthe spect, (Canada balsam absorbs inthe ultavilet) The Glan-Foucault polarizer (Fig. 8.27) 's constructed of nothing other than cal, whichis transparent from roughly 5000 nm in the infrared to about 230 nm i the wlio, I therefore can be used over abroad spectral range. The incon ing ray stikesthe surface normally and canbe resolved into component hat are citer completely parallel or perpendicu Tart the optic axis. The two rays traverse the fist ealcite sec 344 Chapter 8. Polarization Optic ane Figure 8.26 The Nicol prism. The litle fat onthe blunt corner lacates the optic ais. ato by Edt tion without any deviation. (We'll come back to this point lat er on when we talk about retarders.) Ifthe angle-of-incidence ‘on the calcite-air interface is @, one need only arrange things so that n, < L/sin

    You might also like