Laser Control of Chemical Reactions
6
A lthough the science of chemistry has made considerable progress in the past century, the main principles behind the industrial prac- tice of chemistry have remained basi- cally unchanged. Methods for breaking and re-forming chemical bonds still of- ten rely heavily on altering the temper- ature and pressure of the reaction or adding a catalyst. This approach is of- ten ineective because it takes no ac- count of our understanding of the mo- tions of molecules. As a result, bulk re- actions are often inecient, generating large quantities of useless by-products in addition to the desired materials. Recently investigators have devised new techniques, based on illuminating chemical compounds with lasers, that can potentially control the paths taken by reactions. These methods promise to alter the yields in selected ways by exploiting an essential feature of quan- tum mechanicsnamely, the wavelike properties of both light and matter. The latest calculations show that lasers can be enormously eective in steering a reaction down a preferred pathway. The idea of using lasers to drive reac- tions is not revolutionaryin fact, at- tempts to do so began shortly after the rst laser was invented about 35 years ago. These devices emit radiation of a precise frequency, or color, and thus can impart a well-dened parcel of en- ergy to a given target. Chemical bonds were seen as individual springs of dif- ferent strengths, each of which vibrated when given a certain amount of energy. The hope was that lasers could be tuned to attack a particular bond, weakening or breaking it and thereby encouraging one product to form in lieu of another. But this approach, called mode-selec- tive chemistry, has seen only limited success. In fact, it is doomed to fail for the vast majority of molecules because it assumes that chemical bonds are largely independent of one another. Only a few compounds have bonds that meet this requirement. Rather most bonds are strongly interdependent: en- ergy readily ows between them. As a consequence, the energy imparted by the laser is distributed throughout the molecule in a fashion similar to that re- sulting from traditional, and far cheap- er, sources. The latest approach, called coherent control, has grown out of research be- gun in the mid-1980s, when investiga- tors began to take a look at properties of laser light that had been ignored in earlier considerations of chemical con- trol. One such property is the coher- ence of laser light. Coherence relates to the way that atoms emit light. In con- ventional light sources, such as house- hold bulbs, electricity heats up the l- ament, exciting tungsten atoms. The atoms spontaneously lose this energy by giving o a bit of light. Each atom, however, emits light independently of the other atoms. The total light from the lament therefore consists of a jumble of waves from individual atoms. In a way, light from ordinary sources resem- bles a column of soldiers marching out of step with one another. Physicists say that such light is incoherent, or out of phase. In contrast, atoms in lasers act together. Consequently, all the waves that make up the entire laser beam match perfectly, behaving like soldiers marching in step. Coherent light readily displays an important propertyconstructive and destructive interference. The term in- terference does not necessarily mean that the waves disturb one anothers tra- jectories; it refers to the way all waves ( including those on water ) combine. In constructive interference, the crests or troughs of two (or more) waves meet so that the wave heights or dips add to- gether. The wave therefore becomes bigger and deeper ( in other words, the amplitude of the wave increases). In destructive interference, crest meets trough, and the wave is extinguished. Interfering Light Waves T he phenomena of constructive and destructive interference become ap- parent if we shine a laser beam on an opaque plate that has two narrow slits cut into it. Each slit acts as a new source of light waves. The waves emanating from each slit, which fan out, can then interfere with each other. The interfer- ence can be rendered visible if we place a viewing screen just beyond the slits. The screen will show a pattern of dark and light stripes, or fringes. The light areas correspond to constructive in- terference; the dark areas, destructive interference. The latest technique to control mole- cules takes advantage of the fact that light is not the only entity that displays interference. In accordance with a basic Laser Control of Chemical Reactions For years, chemists have sought to control reactions with lasers—and have mostly failed. Success may come from exploiting subtle quantum e›ects resulting from the interaction of light and matter by Paul Brumer and Moshe Shapiro PAUL BRUMER and MOSHE SHAPIRO have been research collaborators since they met at Harvard University more than 20 years ago. Brumer is a professor in the chemical physics theory group at the University of Toronto and is current- ly a Killam Research Fellow. He recently received the Chemical Institute of Cana- das Palladium Medal, one of the organ- izations highest awards. He earned his Ph.D. from Harvard. Shapiro obtained his Ph.D. from the Hebrew University of Je- rusalem. The Jacques Mimran Professor of Theoretical Chemical Physics at the Weizmann Institute of Science in Rehov- ot, Israel, he has also held several visit- ing appointments throughout the U.S., Europe and Canada. The authors grate- fully acknowledge funding and support from the U.S. Oce of Naval Research. 56 SCIENTIFIC AMERICAN March 1995 Copyright 1995 Scientific American, Inc.
Transcript of Laser Control of Chemical Reactions
Although the science of chemistry has made considerable progress
in the past century, the mainprinciples behind the industrial prac-tice of chemistry have remained basi-cally unchanged. Methods for breakingand re-forming chemical bonds still of-ten rely heavily on altering the temper-ature and pressure of the reaction oradding a catalyst. This approach is of-ten ineÝective because it takes no ac-count of our understanding of the mo-tions of molecules. As a result, bulk re-actions are often ineÛcient, generatinglarge quantities of useless by-productsin addition to the desired materials.
Recently investigators have devisednew techniques, based on illuminatingchemical compounds with lasers, thatcan potentially control the paths takenby reactions. These methods promiseto alter the yields in selected ways byexploiting an essential feature of quan-tum mechanicsÑnamely, the wavelikeproperties of both light and matter.The latest calculations show that lasers
can be enormously eÝective in steeringa reaction down a preferred pathway.
The idea of using lasers to drive reac-tions is not revolutionaryÑin fact, at-tempts to do so began shortly after theÞrst laser was invented about 35 yearsago. These devices emit radiation of aprecise frequency, or color, and thuscan impart a well-deÞned parcel of en-ergy to a given target. Chemical bondswere seen as individual springs of dif-ferent strengths, each of which vibratedwhen given a certain amount of energy.The hope was that lasers could be tunedto attack a particular bond, weakeningor breaking it and thereby encouragingone product to form in lieu of another.
But this approach, called mode-selec-tive chemistry, has seen only limitedsuccess. In fact, it is doomed to fail forthe vast majority of molecules becauseit assumes that chemical bonds arelargely independent of one another.Only a few compounds have bonds thatmeet this requirement. Rather mostbonds are strongly interdependent: en-ergy readily ßows between them. As aconsequence, the energy imparted bythe laser is distributed throughout themolecule in a fashion similar to that re-sulting from traditional, and far cheap-er, sources.
The latest approach, called coherentcontrol, has grown out of research be-gun in the mid-1980s, when investiga-tors began to take a look at propertiesof laser light that had been ignored inearlier considerations of chemical con-trol. One such property is the coher-ence of laser light. ÒCoherenceÓ relatesto the way that atoms emit light. In con-ventional light sources, such as house-hold bulbs, electricity heats up the Þl-ament, exciting tungsten atoms. Theatoms spontaneously lose this energyby giving oÝ a bit of light. Each atom,however, emits light independently ofthe other atoms. The total light from the
Þlament therefore consists of a jumbleof waves from individual atoms. In away, light from ordinary sources resem-bles a column of soldiers marching outof step with one another. Physicists saythat such light is incoherent, or out ofphase. In contrast, atoms in lasers acttogether. Consequently, all the wavesthat make up the entire laser beammatch perfectly, behaving like soldiersmarching in step.
Coherent light readily displays animportant propertyÑconstructive anddestructive interference. The term Òin-terferenceÓ does not necessarily meanthat the waves disturb one anotherÕs tra-jectories; it refers to the way all waves(including those on water) combine. Inconstructive interference, the crests ortroughs of two (or more) waves meetso that the wave heights or dips add to-gether. The wave therefore becomesbigger and deeper ( in other words, theamplitude of the wave increases). Indestructive interference, crest meetstrough, and the wave is extinguished.
Interfering Light Waves
The phenomena of constructive anddestructive interference become ap-
parent if we shine a laser beam on anopaque plate that has two narrow slitscut into it. Each slit acts as a new sourceof light waves. The waves emanatingfrom each slit, which fan out, can theninterfere with each other. The interfer-ence can be rendered visible if we placea viewing screen just beyond the slits.The screen will show a pattern of darkand light stripes, or fringes. The lightareas correspond to constructive in-terference; the dark areas, destructive interference.
The latest technique to control mole-cules takes advantage of the fact thatlight is not the only entity that displaysinterference. In accordance with a basic
Laser Control of Chemical Reactions
For years, chemists have sought to control reactions with lasers—andhave mostly failed. Success may come from exploiting subtle
quantum e›ects resulting from the interaction of light and matter
by Paul Brumer and Moshe Shapiro
PAUL BRUMER and MOSHE SHAPIROhave been research collaborators sincethey met at Harvard University morethan 20 years ago. Brumer is a professorin the chemical physics theory group atthe University of Toronto and is current-ly a Killam Research Fellow. He recentlyreceived the Chemical Institute of Cana-daÕs Palladium Medal, one of the organ-izationÕs highest awards. He earned hisPh.D. from Harvard. Shapiro obtained hisPh.D. from the Hebrew University of Je-rusalem. The Jacques Mimran Professorof Theoretical Chemical Physics at theWeizmann Institute of Science in Rehov-ot, Israel, he has also held several visit-ing appointments throughout the U.S.,Europe and Canada. The authors grate-fully acknowledge funding and supportfrom the U.S. OÛce of Naval Research.
56 SCIENTIFIC AMERICAN March 1995 Copyright 1995 Scientific American, Inc.
principle of quantum mechanics, parti-cles such as electrons, atoms and mole-cules can also behave as waves that caninterfere with one another [see ÒTheDuality in Matter and Light,Ó by Bert-hold-Georg Englert, Marlan O. Scully andHerbert Walther; SCIENTIFIC AMERICAN,December 1994]. Experiments have ful-ly conÞrmed the existence of such mat-ter waves and of the associated inter-ference of particles.
How can chemists exploit interfer-ence phenomena to control reactions?Numerous techniques have been devel-oped, but the simplest one trains twodiÝerent lasers beams on molecules.Each beam excites the wavelike aspectof the molecules in a particular way.These two matter waves can then inter-fere with each other. Constructive inter-ference, in turn, may result in the for-mation of a particular product. Destruc-
tive interference would result in the ab-sence of that product or the enhancedformation of another. It turns out thatwe can control the interference pat-ternÑthat is, the yield of the reactionproductsÑby adjusting the coherenceand intensity properties of the two la-ser beams.
A more speciÞc explanation incorpo-rates an abstraction known as a wavefunction, which physicists use to de-
SCIENTIFIC AMERICAN March 1995 57
LASER BEAMS Þred into a vessel Þlled with paired sodiumatoms, or dimers, can control the breakup of the molecules.The beams, each of a diÝerent frequency (appearing as yel-low and red), place the molecules into quantum states that
interfere with one another. Each molecule subsequentlybreaks up into one sodium atom in its normal state and an-other sodium atom in any of several excited states. The par-ticular state formed is governed by the degree of interference.
RIC
KI R
OS
EN
SA
BA
Copyright 1995 Scientific American, Inc.
scribe atomic and molecular systems.Consider a molecule made of three com-ponents connected in a lineÑcall themA, B and C. Because particles behave aswaves, physicists can describe the ini-tial state of the molecule with a mathe-matical entityÑa wave function. It em-bodies all available information aboutthe state of a particle and its motions,or dynamics. For example, the squareof the wave function gives the proba-bility of Þnding the molecule in a par-ticular geometry. Wave functions serveas the basic descriptive tool of quan-tum mechanics.
Now imagine that we can break theABC molecule apart in two ways. Eitherthe bond between A and B can break,or the one between B and C. Thus, twopairs of compounds are possible: A andBC or AB and C. To produce either pos-sibility, we must add energy to the mol-ecule. We can irradiate it with a photonof a particular frequency (and thus im-part a known amount of energy to it),which the molecule absorbs. If the pho-ton is suÛciently energetic, it brings theABC molecule to the Þnal stateÑthat is,to the energy level at which ABC dissoci-ates into two products. The wave func-tion describing the ABC molecule in thisstate incorporates two kinds of infor-mation. One is the wave nature of thelaser light impinging on the molecule;the second is the quantum-mechanicalwave nature of the molecule itself.
But what if we irradiate the moleculeat one third the frequency (that is, withlight carrying one third the energy)? Inthis case, the molecule would have to
absorb three photons instead of one inorder to dissociate. The molecular wavefunction at the point of dissociationwould then be diÝerent, reßecting thefact that the molecule absorbed threephotons rather than just one.
We can control molecular motion ifwe simultaneously irradiate ABC withboth light Þelds. Doing so produces twodistinct wave functions at the energy ofdissociation. The two modes of excita-tion interfere, just as coherent lightdoes when it passes through two slits[see box on page 62 ]. It may seemstrange that two reaction pathways caninterfere, but such interference is theessence of quantum mechanics.
Fortunately for chemists seeking tocontrol molecules, the mathematicalterm describing the interference diÝersfor the two possible outcomes of thereaction. We can create predominantlyA and BC or mostly AB and C by adjust-ing the interference term. The interfer-ence, and hence the amount of eachproduct, depends on the relative ampli-tude and phase of the two original la-ser beams, and so it can be altered byadjusting these characteristics. Note
that because control relies on interfer-ence, the method does not require theuse of intense lasers. In other words,weak light can have a substantial eÝecton the dynamics of molecules.
Working with Chi K. Chan, then atthe University of Toronto, we studiedthe dissociation of the diatomic mole-cule iodine monobromine ( IBr). Thismolecule can separate into I + Br orI +Br*, where the asterisk indicates thatthe bromine atom has excess energy.Our calculations predict that varying theintensities and relative phase of two la-
60 SCIENTIFIC AMERICAN March 1995
PHASE COHERENCE refers to the wayatoms emit radiant energy. In ordinarysources such as lightbulbs, excitedatoms spontaneously give oÝ light in-dependently of one another. Hence, thelight waves from each do not Òline up,Óand the total output over time consistsof random emissions (left ). In laser light(right ), all the atoms emit in phase, sothe output is coherent.
CONTROLLED CHEMICAL DISSOCIATION can be accomplishedif a compound is irradiated simultaneously by two types oflaser photons, one of which is more energetic than the other.For instance, the molecule ABC can dissociate if it absorbs aphoton of a certain energy, or three photons, each with one
third the energy. The wave functions associated with the ex-citation by each type of photon quantum-mechanically inter-fere with one another. The extent of the interferenceÑandhence, the yieldÑis controlled by adjusting the relative ampli-tude (the height of the wave) and the phase of the laser light.
ORDINARY LIGHT
TOTAL OUTPUT
ATOM
AB C
AB C+
+
LASER PHOTONS
PHASE-SHIFTED PHOTON
AMPLITUDE-SHIFTED PHOTON
MOLECULE YIELD
YIELD
90 PERCENT
90 PERCENT
10 PERCENT
10 PERCENT
AB C
AB C+
+
AB C
AB C
DIM
ITR
Y S
CH
IDLO
VS
KY
Copyright 1995 Scientific American, Inc.
ser beams would provide an enormousrange of control over the reaction. Wewould be able to vary the amount ofenergetic bromine produced to ensurethat it accounted for between 25 and95 percent of the total product formed.This degree of control far exceeds thehopes of conventional industrial chem-istry, which seeks to improve reactionselectivity by about 10 percent. Experi-ments by Daniel S. Elliott of PurdueUniversity have demonstrated the suc-cess of this approach for controlling theionization of atoms. Robert J. Gordon
of the University of Illinois at Chicagohas also obtained beautiful results inthe control of diatomic and polyatomicmolecules.
This form of coherent control is notrestricted to the use of one photon andthree photons at one third the energy.Rather quantum mechanics permitscontrol over the probability that a par-ticular reaction will occur if each of thetwo paths corresponds to the absorp-tion of either an even number or an oddnumber of photons. For example, An-dre Bandrauk of the University of Sher-brooke in Quebec has computationallydisplayed extensive control over thephotodissociation of a chlorine mole-cule into energetic and nonenergeticatoms when the molecule absorbs twophotons from one laser beam and fourphotons, each of half the energy, fromthe second laser beam. Similarly, wehave shown that using one photon inconjunction with two photons, each ofhalf the energy, can be used to controlthe direction that molecules take whenthey leave the reaction region. Such anability may render the separation ofproducts easier and thereby boost theeÛciency of the reaction.
Current Limitations
Although the concepts behind con-trolling reactions with coherent
light apply to a wide range of isolatedmolecules, there are at least two obsta-cles to their immediate broad applica-tion. One is that the eÛciency of lasercontrol drops substantially when the
light waves or the molecular wave func-tions have ill-deÞned phases. Loss ofphase deÞnition occurs mainly becauseof collisions between molecules, whichincrease with higher temperature andpressure and which commonly occur inindustrial environments. Further workis necessary to be able to introduce co-herent control into modern commer-cial settings. Consequently, at present,the schemes must be applied in clever-ly designed environments or must berestricted to a limited class of reactions.For instance, current technology workswell on dilute gases, where the mole-cules are far apart and so collide lessfrequently.
The second major obstacle, now be-ginning to seem surmountable, involvesthe phase of the laser light. Given twoarbitrary laser sources, we generally donot know the extent to which the lightfrom one will be in phase with the oth-er. In addition, the phase diÝerence isaÝected by any instabilities in the equip-ment. An unstable phase diÝerence be-tween the two lasers reduces the degreeof interference and control.
Sophisticated optical techniques havethe potential to eliminate such phaseproblems. For example, photons maybe generated by passing light at onefrequency through a particular materialthat is thereby induced to emit light atanother frequency. This process yieldstwo light Þelds whose phase relationsare well deÞned. We can further con-trol the phase diÝerences between thetwo light sources by temporarily slow-ing one light beam relative to the other.
SCIENTIFIC AMERICAN March 1995 61
TWO-PULSE LASER STRATEGY can also govern the yield of achemical reaction. The Þrst pulse places the molecule in a su-perposition stateÑin other words, it sets the molecule vibrat-ing and rotating in a particular way. The motion depends on
the characteristics of the molecule and the laser. The secondpulse causes the molecule to break up. The diÝerences inyield are controlled by varying the time between pulses andby changing the frequencies that make up the pulses.
TOTAL OUTPUT
LASER LIGHT
MOLECULE
PULSE 1 PULSE 2
+
+
A
B
CA
B
C
A
B
C
A
B
C
DIM
ITR
Y S
CH
IDLO
VS
KY
DIM
ITR
Y S
CH
IDLO
VS
KY
Copyright 1995 Scientific American, Inc.
Another way to control phase prob-lems relies on intense laser beams, thefocus of many recent investigations.With Zhidang Chen of the University ofToronto, we have shown that such la-sers make it possible to bypass the needfor radiation having well-deÞned andcarefully controlled phases. In addition,such strong-Þeld methods have the po-tential to increase substantially the ab-solute yield of the reaction and to over-come unwanted collisional eÝects. Aparticular scheme has just been usedexperimentally at the Weizmann Insti-tute of Science in Israel by Irit Sofer,Alexander Shnitman, Ilya Golub, Am-non Yogev and us to demonstrate con-
trol over the various products formedin the dissociation of paired sodiummolecules.
Using Pulses
Because interference of molecularpathways is the key to governing re-
actions, any laser scenario that inducessuch interference may serve as a meansof controlling reactions. Instead of shin-ing two steady beams on a target, onemight use ultrashort pulses of laserlight. Modern lasers can generate burstsas short as 10Ð14 second. Unlike contin-uous-wave radiation, a light pulse ismade up of a collection of distinct fre-
quencies and, hence, of a collection ofphotons with diÝerent energies. Suchlight also has a perhaps counterintuitiveproperty. The briefer the pulse, thebroader the range of energies within it.
This property plays a major role inpulsed-laser methods for controllingthe outcomes of chemical reactions. Bydelivering a range of energies, a pulsecan induce motion (such as vibrationor rotation) in a molecule, which in turnaÝects the way it interacts with otherlight pulses. Ordinarily a molecule suchas ABC exists at a speciÞc (that is, quan-tized) energy value. A system at one ofthese Þxed energies resides in a so-calledstationary state and does not move over
62 SCIENTIFIC AMERICAN March 1995
Perhaps the most unusual aspect of quantum mechan-ics is that, under certain circumstances, matter be-
haves just as waves do. In particular, it displays interfer-ence. Because this property is essential to the ability oflasers to control chemical reactions, some of the mathe-matics and more technical quantum-mechanical conceptsunderlying the phenomenon may be of interest.
Interference arises from the way waves add together.The rule for combining many different waves is first tosum their amplitudes (the height of the wave) and thensquare the result. Consider the interference of two waveswhose amplitudes at a given position and time are a andb. The intensity of eachwave is a2 and b2. Thecombined amplitude, c,is the sum c = a + b,and the combined in-tensity is c2 = (a + b)2 =a2+ b2 + 2ab.
Note that the com-bined intensity is notmerely the sum of theintensities of each wave(which would be a2+b2), but that an addi-tional interference term,2ab, contributes. If aand b are both positiveor both negative, thisinterference term ispositive. The resultingintensity, c2, is thus greater than the simple sum of inten-sities of each wave. The interference in this case is said tobe constructive. If a is positive and b is negative (or viceversa), the interference term is negative, and the resultingintensity is smaller than the simple sum of the individualintensities. Interference in this case is said to be destructive.
The variation in intensity produced by interfering lightwaves can be seen in the famous double-slit experiment,in which a beam of coherent light passes through two slitsand onto a screen (diagram ). The bright regions on thescreen arise from the constructive interference betweenthe two beams of light passing through the two slits; darkregions result from the destructive interference. Intensi-ties in between these two extremes arise from the combi-
nations of maxima and minima meeting at the screen.Now consider replacing the light with particles such as
electrons, atoms or molecules. Classical intuition wouldsuggest that the emergent pattern on the screen would re-semble two nearly rectangular blobs, resulting from thedirect passage of particles through the slits. In fact—andthis is the essential feature of quantum mechanics—theobserved pattern can clearly display an interference pat-tern. That is, the particles can show behavior characteris-tic of waves (as long as the particles are themselves “co-herent”—that is, prepared so that they have well-definedwave functions). There is simply no way to explain these
observations on the ba-sis of a classical theoryof particles.
For this reason, theproper description ofparticle dynamics inquantum mechanics en-tails a wave functionthat, as in the case forwaves, is described byboth an amplitude anda phase. In the double-slit experiment appliedto particles, quantumwave functions arrivingat the screen via theslits interfere with oneanother. Most signifi-cant is another unusual
feature of the quantum world: interference arises becausewe do not know through which of the slits the particlespassed. If we do record such information, the interferencepattern disappears.
The two-slit experiment embodies the fundamental quan-tum principle that two (or more) phase-preserving routesthat a molecule can take to some final state can be madeto interfere. In the case of the one-photon plus three-pho-ton experiment mentioned in the main body of the article,we do not know which of the two possible excitationroutes leads to the observed final state. Hence, the tworoutes interfere. This optically induced interference formsthe foundation for coherent radiative control of molecularprocesses.
Interference and Coherent Control
INTERFERENCE FRINGES are made if coherent light passes throughtwo slits. Particles such as molecules also interfere in this way.
SCREEN
SLITS
INTERFERENCEPATTERN
LASER
DIM
ITR
Y S
CH
IDLO
VS
KY
Copyright 1995 Scientific American, Inc.
time. For a molecule to undergo the dy-namics, it must live in several energylevels at once. Such an assemblage ofenergy levels is called a superpositionstate. The wave function describing thesuperposition state is the sum of wavefunctions representing stationary statesof diÝerent energies. To construct it, re-searchers shine a pulse of coherent la-ser light on the molecule. The way themolecule then moves depends on thenature of the light pulse and its interac-tion with the molecule. Thus, we can ef-fect dynamic changes in the molecule byshifting the relative contribution of thefrequencies that compose the pulseÑthat is, by shaping the pulse.
Several researchers have developedthese ideas. They include Stuart A. Riceof the University of Chicago, David J.Tannor of Notre Dame University, Her-schel Rabitz of Princeton University,Ronnie KosloÝ of the Hebrew Universityof Jerusalem and Kent R. Wilson of theUniversity of California at San Diego.Their results show that pulses built outof a complicated mixture of frequenciesare required to control molecular dy-namics optimally, but simple approxi-mations often suÛce to break apartmolecules in a controlled way.
Although a single light pulse can al-ter the dynamics of a molecule, it doesnot by itself aÝord an active means ofcontrolling the yield of a chemical reac-tion. Rather an idea originally intro-duced by Rice and Tannor and subse-quently extended by us in collabora-tion with Tamar Seideman, now at theNational Research Council of Canada,does allow for control with pulses. Spe-ciÞcally, one needs to employ a two-pulse sequence. The Þrst pulse placesthe molecule in a superposition statethat dictates how that molecule will lat-er respond to the follow-up pulse. Thesecond pulse breaks up the moleculeinto diÝerent products.
Although not apparent, this scenariois similar to that using continuous-wavelasers in that quantum interference be-tween wave functions is responsible forthe control. Interference between mo-lecular wave functions, however, is nowcreated by the various frequencies with-in the two light pulses incident on themolecule. The interference, and hencethe yield of the products formed, canbe altered easily by varying the intervalbetween the two pulses and the frequen-cies that make up the Þrst laser pulse.Hence, unlike the continuous-wave laserexperiments in which a steady streamof products is formed, pulsed lasers letus take advantage of time as an experi-mental variable.
Indeed, computational studies by Iz-hak Levy, formerly at the Weizmann In-
stitute, and us show that the range ofcontrol is potentially quite extensive. Inan analysis of the dissociation of di-atomic molecules, we showed that theyield can be varied so that the desiredproduct accounts for between 3 percentof the total yield and 95 percent, de-pending on the laser settings selected.One can basically turn a process com-pletely on or oÝ. For polyatomic mole-cules, control is not as extensive but isconsiderable nonetheless. We have, forinstance, successfully applied this ap-proach to the dissociation of monodeu-terated water (HOD), made of hydrogen,oxygen and deuterium, to produce con-trollable amounts of H+ OD or D +OH.
Help for Pharmaceuticals
One of the Þrst real-world applica-tions of laser control may well be
enjoyed by the pharmaceutical indus-try. Currently chemists must take careto ensure that reaction products adopta speciÞc conformation. Often the samemolecule can exist in two forms, knownas enantiomers. Like our right and lefthands, enantiomers are mirror imagesof each other. Indeed, such moleculesare often referred to as right- or left-handed. Drug companies expend con-siderable eÝort to form compoundswith the correct handedness becauseoften one enantiomer is biologically ac-tive, and the other is either inactive orharmful.
Laser control could be a solution forachieving the correct outcome. We ex-amined the dissociation of a compoundthat can break up into either right- orleft-handed formsÑcall the substanceABA′, where A and A′ are enantiomers.The reaction can produce A and BA′ orA′ and BA. Because ABA′ is highly sym-metrical, traditional dissociation ofABA′ by absorption of light does notpush the reaction in any particular di-rection; the result is an equal yield of Aand A′. But our studies show that un-der certain conditions ( in particular, inthe presence of a weak magnetic Þeld)the two-pulse scheme can be used tocontrol enantiomeric yield so that wecan produce A rather than A′.
Procedures based on quantum inter-ference can do more than control chem-ical reactions. They can be used to pro-duce entirely novel kinds of technology.The methods can allow workers to se-lect the particular energy state of theproducts of a chemical reaction. Theseproducts could in turn generate laserlight at frequencies not obtainable withcurrent equipment. Even more interest-ing, Paul B. Corkum of the National Re-search Council of Canada has proposedusing interference eÝects to build lasers
that emit supershort bursts of light, onthe timescale of 10Ð16 second. This in-terval is about one tenth the length ofpulses from the best lasers in use today.
With Gershon Kurizki of the Weiz-mann Institute, we have proposed us-ing quantum interference to regulatethe ßow of electrons in semiconductors.One can design two pathways by whicha donor atom loses an electron on ab-sorption of light. These paths can bemade to interfere. Controlling this in-terference means that the direction ofthe ejected electronsÑand hence the di-rection of the electric currentÑcan beregulated. The result would be a fastoptical switch, perhaps on the order of10Ð12 second, many times faster thanpresent-day switches. Experimental ev-idence for such directional control hasnow been obtained by Boris Zeldovich,now at the University of Central Florida,in photoelectric detectors, by Corkumin semiconductor devices and by Elliottin atoms that are ionized by light.
The advent of quantum mechanicsintroduced new concepts in the under-standing of nature. But we are now mov-ing past the role of passive observer.As we approach the 21st century, it isclear that we can extend quantum-me-chanical ideas to open up unprecedent-ed possibilities for gaining further con-trol over atomic, molecular and elec-tronic processes.
SCIENTIFIC AMERICAN March 1995 63
FURTHER READING
COHERENT PULSE SEQUENCE CONTROL OFPRODUCT FORMATION IN CHEMICAL RE-ACTIONS. D. J. Tannor and S. A. Rice inAdvances in Chemical Physics, Vol. 70,Part 1, pages 441Ð523; 1988.
INTERFERENCE BETWEEN OPTICAL TRAN-SITIONS. C. Chen, Y.-Y. Yin and D. S. El-liott in Physical Review Letters, Vol. 64,No. 5, pages 507Ð510; January 29, 1990.
CONTROLLED PHOTON INDUCED SYMME-TRY BREAKING� CHIRAL MOLECULARPRODUCTS FROM ACHIRAL PRECURSORS.M. Shapiro and P. Brumer in Journal ofChemical Physics, Vol. 95, No. 11, pages8658Ð8661; December 1, 1991.
COHERENT LASER CONTROL OF BOUND-TO-BOUND TRANSITIONS OF HCL ANDCO. S.-P. Lu, S. M. Park, Y. Xie and R. J.Gordon in Journal of Chemical Physics,Vol. 96, No. 9, pages 6613Ð6620; May 1,1992.
COHERENT AND INCOHERENT LASER CON-TROL OF PHOTO CHEMICAL REACTIONS.M. Shapiro and P. Brumer in Interna-tional Reviews in Physical Chemistry,Vol. 13, No. 2, pages 187Ð229; Septem-ber 1994.
COHERENCE CHEMISTRY: CONTROLLINGCHEMICAL REACTIONS WITH LASERS. P.Brumer and M. Shapiro in Accounts ofChemical Research, Vol. 22, No. 12,pages 407Ð413; December 1994.
Copyright 1995 Scientific American, Inc.
