Photoelectrons measuring the phase of light

Europhysics News, Jul 2018

Péter Dombi, Alexander Apolonski, Ferenc Krausz

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

Photoelectrons measuring the phase of light

the phase of light 0 PéterDombi , AlexanderApolonski1andFerencKrausz1-2 1Institutfür Photonik , Technische Universität Wien, Austria 2Max-Planck-Institutfür Quantenoptik , Garching , Germany - ltrashort, visible laser pulses have gained immense impor tance inthe past decade. The newtechnologybasedmaini­ onTi:sapphirelasershas achievedseveralbreakthroughs inrespect ofincreasingthe magnificationofbyfarthebest temporal magni fying glass mankind has ever had. With this tool even those extremelyfast processes can be resolvedin time that would nor mally appear as smeared out just like the images taken with : camera with insufficient shutter speed. These advanced light sources have not onlyrevolutionisedfemtochemistrybyaffording an insight into ultrafast chemical processes, earningAhmed H. Zewail the Nobel Prize for Chemistry in 1999, but have also alloweda huge further stepto be madebythe production of iso lated attosecond (1 as = 10~18s) pulses in2001 [ 1 ]. Interaction of femtosecondpulses with gases results inX-raygenerationwhich under certaincircumstances, cangiverisetoasingleX-rayburst a short as 250 as - an order of magnitude shorter than could have been dreamed of fiveyears ago. Apart fromthe unprecedented temporal resolution that they provide, these laser-driven X-ray sourcesaffordpromiseofcompact,coherentX-raydiagnostictool that are somuch desiredbythe medical industry. Several further intriguing questions are raised, such as, what happens whenthe lengthofsuchavisiblepulse (the current stateof-the-art is around 3.5fs) becomes comparableto the oscillatior cycle (2.7 fs at the typical wavelength of 800 nm of Ti:sapphir< lasers). Does the momentary amplitude of the oscillations (the envelope of the pulse - seeupper part ofFigure 1) remain a cru­ cial parameter in determiningthe interaction ofthese pulses with matter orhas onerather to taketheevolutionofthe actual electro­ magneticwaveformintoaccount? Intuition, simulations and, most recently, experimentshaveallindicatedthat thelatteristhecase,for example, when one assesses the photoelectronyieldfroma metal surface inducedbycontrolled, few-cycle optical waveforms, aswe ►Fig.1:Atrainofchirped laser pulseswithacentral wavelength of750nm, producedbyaspecialTksapphireoscillatorwitha repetition ratefr=24MHz,iscompressedby12reflectionsoff speciallycoated (chirped) mirrors (middlepart offigure) after whichthe pulsesattainadurationof4fs.Thepulsetrain hasalso beencarrier-envelope phasestabilised (inasetupnot shown in thefigure).This resultsinawaveformevolutionsimilartothat depicted inthe upper part ofthefigure;i.e.everynthlaser pulse hasthe sameelectromagneticwaveform(inthefigure,n=4for thesakeofsimplicity;inthe experiment n=24was realised).The Dulsetrain interacts withagoldsurfacethat emitselectrons ina multi-photon process.Thesearethen amplifiedinamultiplier tube,the output signal ofwhichclearlyshowsafrequency tomponent at fr/n,this beingaclearindicationthat the emission process isdirectlysensitivetothe actual optical waveform, not just tothe amplitude envelope.Bymeans ofsimulation resultsthe measurement can becalibratedsothat wecantell at anarbitrary timethe shape ofthewaveforms interactingwiththe surface. have recentlyshown [ 2 ], Ifthis is true, the standard approach to light-matter interactions aimingonlyat control overthe evolution of the temporal amplitude envelope of such laser pulses has become out-of-date. Even though there is a wide spectrum of technologies that allowalmost arbitraryshaping ofthe envelope oflaser pulses, the control of the actual waveformwithin the envelope remained a challenge until 2000. It then became possible to gainaccesstothis last final parameter oflight in such awaythat the endresult was a trainofultrashort laser pulses inwhichthe relativephasebetween thecarrierwaveandtheenvelopeofthepulse (thecarrier-envelope phase, CEP) shifts fromone pulse to the next by aknown, con­ trolled and stabilised amount (as depicted in the upper part of Figure 1) [ 3 ], Eventhoughthis didnot implythat the actual carri­ er-envelopephasevalueofanindividual pulse couldbemeasured, itwas atrue revolution in another sense. Namely, when onelooks at suchaphase-stabilisedpulsetrain in the frequencydomainone findsthat it iscomposedofacomb ofequidistant frequencieswith afixed and stable offset fromzero frequency, i.e. atrulyunprece­ dentedreference inthe visible and near-infrared optical domains [ 4 ]. This optical frequencyruler canthen be usedformetrological applications in a domain that was only accessible before with a dozensynchronisedoscillators typicallyfillingan industrial-scale facility. The improvement was immense and this branch of research has been intensivelypursued since then. The measuring accuracyhas improvedbyseveral orders ofmagnitudethankssoley to this technology. In this way such exciting aspects as, for example, themeasurement ofthe much-debatedtime drift offun­ damental constants have been put within reach. Apart fromthis most important spin-off ofcarrier-envelope phase stabilisation, it has opened anewerain the investigation oflight-matter interac:ions. In 2003 several opticallyinduced phenomena were shown to dependon the carrier-envelope phase. One ofthemis ofparticular interest, since it allows unambiguous measurement of the phase oflow-energylaser pulses. It is based on the well-known photoelectric effect, the researchhistoryofwhich spans fromthe ate nineteenth century until the present and in which eminent physicistshaveplayedaprominent role. In 1886-1888 Hertz and Hallwachs observed the emission of electrons fromametal surfacewhenlight of sufficientlyhighfre­ quencyimpinges onit. This process isknown as the photoelectric effect.Theso-called“light electricity”was explainedbyEinsteinin 1905, for whichhe receivedthe Nobel prize in 1921. The photon a p p ro a c h o f E in ste in allo w ed th e early e x p e rim e n ts to be explained: the photoelectric effect takes place w hen the photon energy exceeds the threshold for freeing the electrons from the metal. If it occurs, the num ber o f electrons emitted depends on the light power (i.e. the num ber of photons) and its polarisation. In addition to these parameters, we have found that the photoelectric effect (photoelectron emission from a gold surface) also depends on the phase of the light (or, m ore precisely, the carrier-envelope phase) in the case of few-optical-cycle laser pulses. The mathematical expression for the electric field of a short light pulse (EL(t) in Figure 1) includes its temporal amplitude evolution (A (t)), central carrier frequency ( ω l) and the carrier-envelope phase (ϕ ). It has not been possible to measure the last parameter, the phase, until now, but the photoelectric effect has allowed us to access this quantity, w ith th e indispensable help o f colleagues from the National Institute of Advanced Industrial Science, Japan, and Max-Planck-Institut für Quantenoptik, Germany [ 2 ]. By util­ ising the fact that the carrier-envelope phase evolves in the pulse train periodically with a fixed frequency, we have presented exper­ imental p roof of the phase sensitivity o f photoelectron emission from a metal cathode by dem onstrating that the output signal of the multiplier has the same periodic variation (see Figure 1). The known theoretical prediction o f the maximum of the photocurrent Physics in daily life: Drag'n' Roll L.J.F. Hermans, Leiden University, The Netherlands W h e t h e r we ride our bike or drive our car: there is resistance to be overcome, even on a flat road; that much we know. But when it comes to the details, it’s not that trivial. Both components of the resis­ tance—rolling resistance and drag—deserve a closer look. Let us first remember the main cause of the rolling resistance. It’s not friction in the ball bearings, provided they are well greased and in good shape. It’s the tires, getting deformed by the road. In a way, that may be surprising: the deformation seems elastic, it’s not permanent. But there is a catch here: the forces for compression are not compensated for by those for expansion of the rubber (there is some hysteresis, if you wish). The net work done shows up as heat. The corresponding rolling resistance is, to a reasonable approxima­ tion, independent of speed (which will become obvious below). It is proportional to the weight of the car, and is therefore written: Froll = Cr mg, with Crthe appropriate coefficient. Now we can make an educated guess as to the value of Cr. Could it be 0,1? No way: this would mean that it would take a slope of 10% to get our car moving. We know from experience that a 1% slope would be a . better guess. as a fu n c tio n o f th e c a r rie r- e n v e lo p e p hase [ 5 ] se rv e d as an absolute calibration for carrier-envelope phase determ ination in the experim ent. This finding, p ro v id e d by a compact, solid-state detector, opens the door for th e experim ental characterisation of the complete waveform o f light p u lses and the optim isation of a huge variety o f nonlinear experim ents in optics on femtosecond and attosecond tim e scales. This paper is a sum m ary o f an article published in NJP (see ref. [ 2 ]). Right! For most tires inflat­ ed to th e recom m ended pressure, Cr = 0,01 is a stan­ dard value. By the way: for bicycle tires, with pressures about twice as high, Cr can get as low as 0,005. The conclusion is that, for a 1000 kg car, the rolling resistance is about 100N. What about the drag? In view of the Reynolds num­ bers involved (Re = 106) forget about Stokes. Instead, we should expect the drag Fd to be p ro p o rtio n al to 1/2pv2, as already suggested by B ernoulli’s law. On a Fig. 1 vehicle with frontal area A, one can write Fd = CdA-1/2pv2. Now, CDis a complicated function of speed, but for the relevant v-range we may take CDconstant. For most cars, the value is between 0,3 and 0,4. The total resistance is now shown in the figure, for a mid-size model car (m=1000 kg, Cr = 0.01, Cd= 0,4 an d A=2 m2). It is funny to realize that the vertical scale immediately tells us the energy consumption. Since 1 N is also 1 J/m, we find at 100 km/h approximately 500 kfJ/km for this car. Assuming an engine efficiency of 20%, this corresponds to about 7 liters of gasoline per 100 km. At still higher speeds, the figure suggests a dramatic increase in the fuel consumption. Fortunately, it’s not that bad, since the engine effi­ ciency goes up, compensating part of the increase. What about the engine power? Since P = F.v, we find at 100 km /h about 15 kW is needed. That’s a moderate value. But note that, at high speed where drag is dominant, the power increases almost as v3! Should we want to drive at 200 km /h, the engine would have to deliver the 8-fold power, or 120 kW. That’s no longer moderate, I would say, and I’m sure the police would agree.... [1] For reviews in these fields see, for example, A.H. Zewail , J. Phys. Chem. A 104 , 5660 - 5694 ( 2000 ) and T. Brabec , F. Krausz , Rev. Mod. Phys. , 72 , 545 - 592 ( 2000 ). Newer results can be found in, for example, M. Hentschel et al, Nature 414 , 509 ( 2001 ) and R. Kienberger et al, Nature 427 , 817 ( 2004 ). [2] P. Dombi et al, New J. Phys. 6 , 39 ( 2004 ). [3] D.J. Jones et al, Science 288 , 635 ( 2000 ) and A . Poppe et al, Appl. Phys. B , 72 , 373 ( 2001 ). [4] S.T. Cundiff , J. Phys . D 35 , R43 - R59 ( 2002 ). [5] Ch. Lemell et al, Phys.Rev.Lett . 90 , 076403 ( 2003 ).

This is a preview of a remote PDF:

Péter Dombi, Alexander Apolonski, Ferenc Krausz. Photoelectrons measuring the phase of light, Europhysics News, 129-130, DOI: 10.1051/epn:2004407