Printed by Jouve, 75001 PARIS (FR)

(19)E

P2

722

865

A1

TEPZZ 7  865A_T(11) EP 2 722 865 A1

(12) EUROPEAN PATENT APPLICATION

(43) Date of publication: 23.04.2014 Bulletin 2014/17

(21) Application number: 12189369.7

(22) Date of filing: 22.10.2012

(51) Int Cl.:H01J 37/04 (2006.01) H01J 37/28 (2006.01)

(84) Designated Contracting States: AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TRDesignated Extension States: BA ME

(71) Applicant: FEI COMPANYHillsboro, Oregon 97124-5793 (US)

(72) Inventors: • Kieft, Erik

5611 CG Eindhoven (NL)• Kiewiet, Fred

5622 BR Eindhoven (NL)

• Lassise, Adam3532 ES Utrecht (NL)

• Luiten, Otger5614 AH Eindhoven (NL)

• Mutsaers, Peter5663 HT Geldrop (NL)

• Vredenbregt, Edgar5627 GK Eindhoven (NL)

• Henstra, Alexander3544 PW Utrecht (NL)

(74) Representative: Bakker, HendrikFEI Company Patent Department P.O. Box 17455602 BS Eindhoven (NL)

(54) Beam pulsing device for use in charged-particle microscopy

(57) A charged-particle microscope comprising:- A charged-particle source, for producing a beam ofcharged particles that propagates along a particle-opticalaxis;- A sample holder, for holding and positioning a sample;- A charged-particle lens system, for directing said beamonto a sample held on the sample holder;- A detector, for detecting radiation emanating from thesample as a result of its interaction with the beam;- A beam pulsing device, for causing the beam to repeat-edly switch on and off so as to produce a pulsed beam,

wherein the beam pulsing device comprises a unitaryresonant cavity disposed about said particle-optical axisand having an entrance aperture and an exit aperture forthe beam, which resonant cavity is embodied to simulta-neously produce a first oscillatory deflection of the beamat a first frequency in a first direction and a second os-cillatory deflection of the beam at a second, different fre-quency in a second, different direction. The resonant cav-ity may have an elongated (e.g. rectangular or elliptical)cross-section, with a long axis parallel to said first direc-tion and a short axis parallel to said second direction.

EP 2 722 865 A1

2

5

10

15

20

25

30

35

40

45

50

55

Description

[0001] The invention relates to a charged-particle mi-croscope comprising:

- A charged-particle source, for producing a beam ofcharged particles that propagates along a particle-optical axis;

- A sample holder, for holding and positioning a sam-ple;

- A charged-particle lens system, for directing saidbeam onto a sample held on the sample holder;

- A detector, for detecting radiation emanating fromthe sample as a result of its interaction with the beam;

- A beam pulsing device, for causing the beam to re-peatedly switch on and off so as to produce a pulsedbeam.

[0002] The invention also relates to a method of usingsuch a charged-particle microscope.[0003] As used throughout this text, the ensuing termsshould be interpreted as follows:

- The term "charged particle" encompasses an elec-tron or ion (generally a positive ion, such as a Galliumion or Helium ion, for example, though a negativeion is also possible). It may also be a proton, forexample.

- The term "charged-particle microscope" (CPM) re-fers to an apparatus that uses a charged-particlebeam to create a magnified image of an object, fea-ture or component that is generally too small to beseen in satisfactory detail with the naked human eye.In addition to having an imaging functionality, suchan apparatus may also have a machining function-ality; for example, it may be used to locally modify asample by removing material therefrom ("milling" or"ablation") or adding material thereto ("deposition").Said imaging functionality and machining function-ality may be provided by the same type of chargedparticle, or may be provided by different types ofcharged particle; for example, a Focused Ion Beam(FIB) microscope may employ a (focused) ion beamfor machining purposes and an electron beam forimaging purposes (a so-called "dual beam" micro-scope), or it may perform machining with a relativelyhigh-energy ion beam and perform imaging with arelatively low-energy ion beam.

- The term "sample holder" refers to any type of table,platform, arm, etc., upon which a sample can bemounted and held in place. Generally, such a sampleholder will be comprised in a stage assembly, withwhich it can be accurately positioned in several de-grees of freedom, e.g. with the aid of electrical actu-ators.

- The term "charged-particle lens system" refers to asystem of one or more electrostatic and/or magnetic

lenses that can be used to manipulate a charged-particle beam, serving to provide it with a certain fo-cus or deflection, for example, and/or to mitigate oneor more aberrations therein. In addition to (varioustypes of) conventional lens elements, the charged-particle lens system (particle-optical column) may al-so comprise elements such as deflectors, stigma-tors, multipoles, aperture (pupil) plates, etc.

- The phrase "radiation emanating from the sample"is intended to encompass any radiation that ema-nates from the sample as a result of its irradiation bythe charged-particle beam. Such radiation may beparticulate and/or photonic in nature. Examples in-clude secondary electrons, backscattered electrons,X-rays, visible fluorescence light, and combinationsof these. Said radiation may also simply be a portionof the incoming beam that is transmitted through orreflected from the sample, or it may be produced byeffects such as scattering or ionization, for example.

- The term "detector" should be broadly interpreted asencompassing any detection set-up used to register(one or more types of) radiation emanating from thesample. Such a detector may be unitary, or it maybe compound in nature and comprise a plurality ofsub-detectors, e.g. as in the case of a spatial distri-bution of detector units about a sample holder, or apixelated detector. The detector may be used in im-age formation and/or for spectroscopic investigation(e.g. as in the case of techniques such as EDX orWDX (Energy- or Wavelength-Dispersive X-raySpectroscopy), EELS (Electron Energy-Loss Spec-troscopy) and EFTEM (Energy-Filtered Transmis-sion Electron Microscopy)).

[0004] In what follows, the invention will - by way ofexample - often be set forth in the specific context ofelectron microscopes. However, such simplification is in-tended solely for clarity/illustrative purposes, and shouldnot be interpreted as limiting.[0005] Electron microscopy is a well-known techniquefor imaging microscopic objects. The basic genus of elec-tron microscope has undergone evolution into a numberof well-known apparatus species, such as the Transmis-sion Electron Microscope (TEM), Scanning Electron Mi-croscope (SEM), and Scanning Transmission ElectronMicroscope (STEM), and also into various sub-species,such as so-called "dual-beam" tools (e.g. a FIB-SEM),which additionally employ a "machining" beam of ions,allowing supportive activities such as ion-beam millingor ion-beam-induced deposition, for example.[0006] In traditional electron microscopes, the imagingbeam is continuously "on" during a given imaging ses-sion. However, in recent years, the possibility of beingable to perform "time-resolved" microscopy using apulsed electron beam has attracted interest. In such mi-croscopy, the use of a pulsed input beam allows outputfrom the employed detector to be discretized into a tem-poral train of timestamped components (e.g. images or

1 2

EP 2 722 865 A1

3

5

10

15

20

25

30

35

40

45

50

55

spectra). The principle behind such microscopy can becompared to the principle underlying stroboscopic pho-tography, where the use of a high-speed flash allowscontinuous motion of a photographed object to be cap-tured as a temporal train of freeze-frame exposures -facilitating accurate analysis of subtle differences be-tween the captured frames. By using a pulsed (or"flashed") charged-particle beam in a CPM, it becomespossible to accurately investigate dynamic processes ina sample, such as phase transitions, mechanical vibra-tions, heat dissipation, chemical reactions, biological celldivision, fluid flow, electrical response, radioactive decayprocesses, etc.[0007] Of particular interest as regards the techniqueelucidated in the previous paragraph is the ability to pro-duce ultra-short-duration charged-particle pulses, sincethese can in turn be used to investigate ultra-fast dynamicprocesses. One known way of producing such pulses isto embody the charged-particle source to comprise aphoto-electric emitter (e.g. a heated LaB6 crystal), andembody the beam pulsing device as a pulsed laser beamthat irradiates said emitter with ultra-short bursts of light(e.g. with a duration of the order of picoseconds (ps) orfemtoseconds (fs)); such a technique is set forth, for ex-ample, in US 2005/0253069 A1. Since it is relatively easyto acquire and use lasers that are capable of producingextremely short light pulses, this particular method wouldappear to lend itself to ultra-fast time-resolved analysisof samples. However, a significant drawback of thisknown technique is that the employed photo-electricemitter produces a much lower electron brightness thanis typically available from a conventional (e.g. Schottky)source in an electron microscope, which severely limitsthe practical usefulness of this approach.[0008] In an alternative solution, one could elect to usea conventional electron source (such as the aforemen-tioned Schottky source), and embody the beam pulsingdevice as a beam "chopper" or "blanker" that inter-rupts/passes the electron beam in an oscillatory manner.For example, the beam pulsing device might employ abeam deflector in which electrodes generate an oscilla-tory electric field that periodically laterally deflects theelectron beam away from a nominal particle-optical axis,thus effectively creating a pulsed beam further down-stream. In a more sophisticated variant, one could use aresonant cavity to produce said beam deflection. Al-though this method has the advantage of using a high-brightness electron source, it is generally of limited flex-ibility as regards its application to practical beam chop-ping. In particular, this method does not lend itself to theproduction of short pulses (e.g. with a pulse lengths inthe fs-ps range) at relatively low frequencies (e.g. of theorder of 100 MHz), since the former aspect (short pulses)requires a relatively high deflection frequency whereasthe latter aspect (relatively long period between pulses)requires a relatively low deflection frequency, and thesetwo different demands are difficult to mutually reconcile.For an example of this approach, reference is made to

the article by K. Ura et al., "Picosecond Pulse Strobo-scopic Scanning Electron Microscope", J. Electron Mi-crosc., Vol. 27, No. 4 (1978), pp. 247-252.[0009] In a variant of the approach set forth in the pre-vious paragraph, one could attempt to embody the beampulsing device as a series configuration of two "crossed"deflectors with intermediate drift space; here, the term"crossed" is used to indicate that the deflection direction(e.g. along an x-axis) of one deflector is perpendicular tothe deflection direction (e.g. along a y-axis) of the otherdeflector. The idea here is that the input to the seconddeflector is pulsed by the first deflector, so that the seconddeflector produces "a pulse of a pulse" or, in effect, abeat. Such a configuration would allow more flexibility,in that there are now different frequencies that can beadjusted (deflection frequencies of first and second de-flections, and beat frequency of the superimposed de-flections) so as to allow more independent variation ofthe pulse length and pulse frequency of the output (re-sultant) pulse. However, it should be noted that use ofsuch a set-up in a CPM beam pulsing device would tendto significantly complicate the employed particle-opticalcolumn between source and sample. This is because, inorder to work satisfactorily, each deflector is ideally sit-uated at the focal point of a lens (located upstream in theparticle-optical column), e.g. a condenser/objective lens.If one employs an x-deflector at position Z1 along theparticle-optical axis and a y-deflector at position z2 alongthe same axis, then a first stigmator will have to be usedupstream of the deflectors so as to deliberately introduceenough astigmatism to give each deflector its respectivefocus, and a second stigmator will have to be used down-stream of the deflectors in order to subsequently mitigatethis deliberately introduced astigmatism. Because thetwo deflectors are in series arrangement and have anintermediate drift space, the distance Δz = |Z2 - Z1| willbe relatively large (e.g. of the order of a few cm), thusrequiring relatively large and powerful stigmators - whichtends to be a significant disadvantage in a (typically)cramped particle-optical column.[0010] It is an object of the invention to address theseissues. More specifically, it is an object of the inventionto provide a CPM in which ultra-fast time-resolved micro-scopy can be satisfactorily performed. In particular, it isan object of the invention that such a CPM should employa charged-particle beam pulsing device with which it rel-atively easy to (independently) adjust the obtained pulselength and pulse frequency, without introducing exces-sive astigmatism. More specifically, it is an object of theinvention that such a beam pulsing device be capable ofproducing ultra-short beam pulses (e.g. with ps or fs pulselengths) at relatively low frequencies (e.g. of the order of100 MHz). Moreover, it is an object of the invention thatthe charged particle beam in such a CPM should havesatisfactory brightness.[0011] These and other objects are achieved in acharged-particle microscope as set forth in the openingparagraph, characterized in that the beam pulsing de-

3 4

EP 2 722 865 A1

4

5

10

15

20

25

30

35

40

45

50

55

vice comprises a unitary resonant cavity disposed aboutsaid particle-optical axis and having an entrance apertureand an exit aperture for the beam, which resonant cavityis embodied to simultaneously produce a first oscillatorydeflection of the beam at a first frequency in a first direc-tion and a second oscillatory deflection of the beam at asecond, different frequency in a second, different direc-tion.[0012] The beam pulsing device in the CPM accordingto the present invention simultaneously produces two dif-ferent deflections in a single (unitary) resonant cavity;consequently, there is no longer a need to use the pow-erful stigmators referred to above, because the above-mentioned substantial focus separation Δz is no longerpresent. The absence of such large stigmators, coupledwith the much more compact size of the inventive beampulsing device parallel to the particle-optical axis direc-tion (single cavity rather than dual cavities with interme-diate drift space), results in very significantly reduced de-mands on available space in the particle-optical columnof the CPM. The beam pulsing device according to thepresent invention is thus more space-saving and produc-es fewer aberration issues.[0013] In a particular embodiment of the invention, saidunitary resonant cavity is:

- Substantially cylindrical in form, with a cylindrical ax-is that is substantially collinear with said particle-op-tical axis (z-axis);

- Embodied to be excited in TM110 resonant mode.

[0014] It should be noted that the term "cylindrical" isused here in a strict mathematical sense, and thus en-compasses cylinders that do not have a circular cross-section. According to standard usage in the field of elec-tromagnetism, the symbol "TM" indicates a TransverseMagnetic field, i.e. an electromagnetic field that has nolongitudinal magnetic component (so that B = 0 along thez-axis). The triplet of subscripts "110" denotes integereigenvalues of a wave vector k needed to satisfy bound-ary conditions pertaining to Maxwell’s equations in thecavity. Without going into further mathematical detail, aTM110 mode is a dipole mode with a strong lateral mag-netic field at radius r = 0 (measured outward from the z-axis) and zero electric field at r = 0. Such a mode can,for example, be excited in the cavity with the aid of aHertzian dipole loop antenna placed close to the wall ofthe cavity (distal from the z-axis). An antenna of this typecan, for example, be achieved by:

- Creating a small bore in a wall of the cavity;- Feeding the inner conductor of a coaxial cable

through this bore to the interior of the cavity, in sucha way that said inner conductor does not touch said(conducting) wall;

- Creating a loop in said inner conductor proximal tosaid wall;

- Orienting the loop appropriately (e.g. so that its plane

is normal to the y-axis, to excite a magnetic field par-allel to y);

- Connecting said coaxial cable to an oscillating RadioFrequency (RF) power supply.

[0015] The vibrational behavior of the cavity can beadjusted in various ways. For example, the frequency ofsaid oscillating power supply can be altered. Alternative-ly, a small conducting (e.g. metallic) or dielectric "plunger"(tuning element) can be partially inserted into the cavity,e.g. through a small bore opposite the above-mentionedantenna; the extent of insertion of such a plunger willthen influence the resonant frequency of the cavity, be-cause:

- Insertion of a conducting plunger will locally de-crease the effective radius of the cavity, with an at-tendant increase in resonant frequency;

- Insertion of a dielectric plunger will increase the ef-fective dielectric constant of the cavity, with an at-tendant decrease in resonant frequency.

[0016] Needless to say, when the cavity is excited on-resonance (i.e. the frequency of the oscillating powersupply is matched to the resonant frequency of the cav-ity), the resulting electromagnetic fields in the cavity willbe at their largest. The skilled artisan in the field of elec-tromagnetism will be familiar with such concepts, and willbe able to implement and optimize them according to thedetails/requirements of a particular configuration. In par-ticular, he will realize that other types and/or locations ofantenna (or other means of excitation) can be employed,as well as other types and/or locations of tuning element/plunger. He will also understand that he is not limitedper se to a TM110 resonance mode, and that, in principle,other types of TM, TE (Transverse Electric) and/or Trans-verse Electro-Magnetic modes may be equally or bettersuited to a given set-up.[0017] In order to simultaneously excite two differentresonances of mutually different frequency in the samecavity, one can, for example, concurrently use two differ-ent excitement antennae (of a type as described above,or similar), each antenna working in unison with its ownplunger / tuning element (again of a type as describedabove, or similar). In such a set-up, one antenna/plungerpair can be aligned so as to produce an oscillatory de-flection along said first direction, and the other plung-er/antenna pair can be aligned so as to produce an os-cillatory deflection along said second direction. The po-sitions of the plungers and/or the driving frequency of theantennae can then be adjusted to as to give said oscil-latory deflections the desired frequencies. However,there are alternatives to such a set-up. For example, onecould attempt to use a single excitement antenna in con-junction with two different plungers - though such an ar-rangement will generally be less flexible than a two-an-tenna approach. Yet another alternative is set forth in thenext paragraph.

5 6

EP 2 722 865 A1

5

5

10

15

20

25

30

35

40

45

50

55

[0018] In a noteworthy embodiment of a CPM accord-ing to the current invention, the resonant cavity - whenviewed in a direction normal to said particle-optical axis(z axis) - has an elongated cross-section, with a long axisparallel to said first direction and a short axis parallel tosaid second direction. In a particular such embodiment,said long and short axes are substantially perpendicular(though this is not strictly necessary). Because its cross-section has two different characteristic dimensions, sucha resonant cavity can simultaneously support two differ-ent resonances - one along each said dimension (as ex-plained above, resonant frequency depends inter alia onthe (effective) internal dimension of the cavity). In manypractical applications of the invention, only a relativelysmall frequency difference will be required between saidtwo resonances, so that the difference between said twodimensions may be correspondingly small; nevertheless,the current embodiment will generally allow larger fre-quency differences to be achieved (if desired) than theset-up described in the previous paragraph. Examplesof cross-sectional forms as alluded to here include rec-tangles and quasi-rectangular forms such as "racetracks"(in which the two opposing straight sides of a rectangleare replaced by curved sides).[0019] In a particular example of an embodiment asset forth in the previous paragraph, the resonant cavity’scross-section is (substantially) an ellipse, with a majorand a minor axis that correspond respectively to said firstand second directions. Such a geometrical configurationis advantageous in that:

- An ellipse doesn’t contain any discontinuities in itsform (e.g. corners, angles), thus simplifying the elec-tromagnetic field configurations produced inside anelliptical cavity;

- An ellipse is a relatively good approximation to a cir-cular cross-section, with its various symmetry-asso-ciated benefits. This is particularly the case for anellipse of relatively mild eccentricity;

- Even relatively mild eccentricity of an ellipse pro-vides enough scope to slightly vary the effective in-ternal dimensioning of the cavity in a particular di-rection, for the purpose of resonant frequency ad-justment.

In the current context, an oval or quasi-oval may be re-garded as an approximation to an ellipse.[0020] To make an embodiment as set forth in eitherof the preceding two paragraphs tunable, one could makeuse of tuning elements / plungers as set forth above.However, as an alternative/supplement to such an ap-proach, one can also embody the resonant cavity to bemechanically deformable so as to adjust a ratio of thelengths of said long and short axes (major and minoraxes). For example, one could conceive a scenariowhereby the cavity walls are (at least partially) made ofa pliable material (such as plastic) that is coated with afilm of metal or another conducting material; such walls

can then be locally nudged/squeezed/moved by appro-priately placed actuators so as to change their form/di-mensioning/position, thus (locally) altering the effectiveinternal dimensions of the cavity and, accordingly, its res-onant behavior.[0021] For good order, it should be noted that, in a cy-lindrical cavity with a perfectly circular cross-section, andwith perfectly smooth walls that are uninterrupted bybores or protrusions, excitation of a given resonancemode can, in fact, concurrently produce more than onedegenerate "versions" of the same mode; for example,one could obtain two degenerate TM110 modes - one witha magnetic field oriented along the x-axis and anotherwith a magnetic field oriented along the y-axis. However,the moment an imperfection is introduced into such acavity (e.g. by sliding in an antenna or plunger, or byintroducing a deformation of the circular cross-section),such degeneracy is broken, and one of said "versions"will become dominant. The embodiments in the preced-ing paragraphs exploit this effect.[0022] A resonant cavity as used in the current inven-tion may, in a relatively simple embodiment, contain avacuum through which the charged-particle beam prop-agates. However, in an alternative embodiment, the in-ventive resonant cavity comprises a dielectric materialthat is disposed about said particle-optical axis and thatcontains a substantially axially symmetric void throughwhich the beam can pass. Use of a dielectric material inthis manner allows a higher field amplitude and frequencyto be achieved for the same input power, or, equivalent-ly/alternatively, results in lower power requirements togenerate the same field amplitude and frequency (ascompared to a vacuum-filled cavity). Examples of candi-date dielectrics in this context include materials such asZrTiO4, sapphire, fused quartz, alumina, PTFE (Poly-TetraFluoroEthylene) and ZrO2, for instance. Said voidis preferably substantially axially symmetric with respectto the particle-optical axis - e.g. (quasi-)conical or bell-shaped, or cylindrical - so as to allow the charged-particlebeam (in its various states of deflection and non-deflec-tion) to continue to traverse the cavity without (signifi-cantly) intercepting the dielectric material.[0023] In a refinement of an embodiment as set forthin the previous paragraph, an interface between said voidand said dielectric material is at least partially coated bya film of electrically conductive material. Put another way,the inward-facing surface of the dielectric body that de-limits said void is (at least partially) metallized or coatedwith a (thin) film of other conductive material, such asITO (Indium Tin Oxide). In this way, the accumulation ofunwanted/parasitic electric charge at the interface be-tween the void and dielectric is advantageously mitigated- since the presence of such charge would tend to disturbthe operation / deflection performance of the cavity. How-ever, so as not to disturb the intended electromagneticfield distribution inside the cavity, the employed conduc-tive film should be relatively thin. More specifically, inorder to make the film essentially transparent to said field

7 8

EP 2 722 865 A1

6

5

10

15

20

25

30

35

40

45

50

55

at the resonance frequency of the cavity, the film thick-ness should be substantially smaller than the so-calledskin depth δ at that frequency, which is given by the ex-pression:

where ρ and m are, respectively, the resistivity and mag-netic permeability of the film material, and ω = 2πf is theangular frequency corresponding to a linear frequency f.As an example, for copper (ρ = 1.71 3 10-8Ω · m, m ≈ m0= 4π 3 10-7N·A-2) at a resonance frequency of 3 GHz, δ= 1.2 mm. For (non-magnetic) film materials with higherresistivity, the skin depth will be larger. Needless to say,if the film is relatively thin, then its DC conductance willbe accordingly low; however, even a relatively poor con-ductor at the surface of the dielectric will be capable ofpreventing a gradual build-up of electric charge.[0024] The discussion above has been largely struc-tural in nature, but it is also possible to give a more func-tional description of the operation of the invention. In thisrespect, it should be remembered that:

- When two waveforms of (slightly) different frequencyare superimposed, one obtains a train of beats, andthe frequency of these beats is (highly) sensitive tothe frequency difference between said superim-posed waveforms. Similarly, the frequency of thebeam pulses observed at sample level in a CPM ac-cording to the present invention can be tuned by al-tering the frequency difference between the two res-onances that are simultaneously produced in the uni-tary resonant cavity of the inventive beam pulsingdevice.

- If the frequency of a deflection is kept constant andits amplitude is increased, then the linear speed ofthe deflection must also necessarily increase, andvice versa. Accordingly, if one increases the ampli-tude of the output of the oscillating power supply thatis driving (at least one of the resonances in) the in-ventive resonant cavity, the speed with which thecharged-particle beam is deflected back and forthwill also increase; a point situated on the deflectionpath of the beam will thus experience a shorter pulseduration as the beam passes by. And vice versa.

- By adjusting the relative phase between the two os-cillatory deflections in the inventive resonant cavity,one can influence the location of a beat at a givenmoment in time. This effect can, if desired, be ex-ploited to influence spatial alignment and/or synchro-nization aspects of the pulsed charged-particlebeam, as set forth in more detail below.

[0025] In further continuance of this discussion, an as-

pect of the current invention ischaracterized in that:

- In operation, said resonant cavity causes thecharged-particle beam to trace out a composite ge-ometrical figure (e.g. a so-called Lissajous figure) ona masking plane perpendicular to the particle-opticalaxis;

- The microscope comprises a masking plate locatedin said masking plane and having an opening thatcan be positioned so as to intersect said compositegeometrical figure, thus serving to admit a pulse ofcharged particles as the beam traces across saidopening.

[0026] As set forth above, a resonant cavity as speci-fied by the current invention is capable of simultaneouslyexciting two different oscillatory deflections (of a charged-particle beam) in two different lateral directions and atmutually different frequencies. The net effect /resultantof these two deflections will be to cause the beam to"dance" out a composite (and periodic) geometrical figureon a plane normal to the (nominal) particle-optical axis.More specifically, in the case of sinusoidal oscillationsalong the x- and y-axes (which may differ in amplitude,frequency and phase), said figure will be a Lissajous fig-ure - which term, in fact, encompasses a whole family offigures whose shapes are sensitive to the aforemen-tioned amplitude, frequency and phase values. If, whiletracing out such a composite geometrical figure, thebeam traverses an opening (e.g. relatively small hole orslit) in a front side of a masking plate, it will momentarilypoke through the opening and cause a beam pulse to beobserved at the back side of the plate. For a given geo-metrical figure, if said opening doesn’t initially lie at somepoint along the course of the figure, it can be made to doso in various ways, e.g. by changing the position of theopening (moving the plate or certain parts thereof) and/oradjusting the aforementioned relative phase between thetwo oscillatory deflections.[0027] Even if the oscillating power supply drives theinventive cavity with a non-sinusoidal waveform (such asa sawtooth or block wave), it should be remembered thatsuch waveforms can nevertheless be decomposed intoa sum of sinusoids (Fourier decomposition), whence thediscussion above remains pertinent.[0028] For the sake of clarity, it should be explicitly re-membered that pulse length is not necessarily directlyrelated to pulse frequency, and that a pulse train maycomprise extremely short pulses (e.g. a few picosecondsor femtoseconds per pulse) at a relatively "low" frequency(e.g. in the Megahertz range), simply because there is arelatively long "dead time" between successive pulses.The current invention allows independent adjustment ofpulse length and pulse frequency - an aspect that is interalia of importance in the context of the next paragraph.[0029] The dynamic processes investigated using theCPM/method of the current invention may, if desired, be

9 10

EP 2 722 865 A1

7

5

10

15

20

25

30

35

40

45

50

55

deliberately precipitated/maintained/steered by applyingan appropriate external stimulus (e.g. radiative, thermal,electrical, chemical, acoustic and/or mechanical) to thesample under investigation, whereby the application ofsuch stimulus may, if desired/required, be matched tothe timing/phase of the pulsing behavior of the employedcharged-particle beam. For example, the sample underinvestigation may have some property that can be influ-enced by the light and/or heat that is locally delivered bya focused laser beam. The laser in such an instance willtypically deliver relatively short light pulses that are sep-arated by dead periods in which the lasing cavity "re-loads", e.g. resulting in picosecond pulses at a frequencyof, say, 75 MHz. It can be highly advantageous if a CPMthat is used to obtain imagery and/or spectroscopic in-formation from such a sample is capable of producingcharged-particle beam pulses that are matched (in termsof duration, frequency and phase) to those of the laser.In this way, innovative analysis techniques - such as high-brightness FEELS (Femtosecond Electron Energy-LossSpectroscopy) - can be exploited for sample analysis pur-poses. It should be noted that the laser-based examplegiven here is not limiting. For example, one could applyan electrical stimulus (e.g. inductively, or using a contactprobe), or a mechanical stimulus (using a vibrating mem-brane), etc.[0030] The invention will now be elucidated in moredetail on the basis of exemplary embodiments and theaccompanying schematic drawings, in which:

Figure 1A renders a transverse cross-section (in planview) of a unitary resonant cavity comprised in abeam pulsing device for use in a charged particlemicroscope (CPM) according to the present inven-tion.Figure 1B shows a longitudinal cross-section (in el-evation) of the subject of Figure 1A.Figure 1C shows magnetic and electrical field con-figurations in the subject of Figures 1A and 1B for aTM110 resonance mode.Figure 1D illustrates a Lissajous figure that can betraced out by a charged-particle beam traversing aresonant cavity as depicted in Figures 1A and 1B, inoperation.Figure 2 shows a longitudinal cross-sectional viewof a particular type of CPM (in this case, a TEM) inwhich the present invention can be implemented.Figure 3 renders a longitudinal cross-sectional viewof another type of CPM (in this case, a SEM) in whichthe present invention can be employed.Figure 4 depicts a longitudinal cross-sectional viewof yet another type of CPM (in this case, a FIB mi-croscope) in which the present invention can be putto use.

[0031] In the Figures, corresponding parts may be in-dicated using corresponding reference symbols.

Embodiment 1

[0032] Figures 1A and 1B render various views of aunitary resonant cavity 201 comprised in a beam pulsingdevice for use in a charged particle microscope (CPM)according to the present invention. More particularly:

- Figure 1A renders a lateral cross-sectional view ofthe resonant cavity 201, observed along the z axis.When located in the CPM, the cavity 201 will beplaced/aligned so that the CPM’s particle-optical ax-is 219 extends along this z-axis;

- Figure 1B renders a longitudinal cross-sectionalview of the same resonant cavity 201, in which thez axis is now vertical.

Also illustrated are x and y axes, which form a Cartesiancoordinate system together with said z axis.[0033] As depicted in Figure 1A, the cavity 201 has asubstantially elliptical transverse cross-section, with amajor axis parallel to the x-axis (first direction) and a mi-nor axis parallel to the y-axis (second direction). Thedashed circle 205 in Figure 1A is drawn centered on thecavity 201, to act as a reference to more clearly revealthe elliptical form of the depicted cross-section. As is ev-ident from Figure 1B, the cavity 201 is cylindrical in shape,and its cylindrical axis is substantially collinear with thedepicted optical axis 219. Entrance aperture 221a andexit aperture 221b allow a charged-particle beam prop-agating along the optical axis 219 to enter and leave theinterior of the cavity 201, respectively. A cavity of suchform is sometimes referred to as a "pillbox cavity".[0034] In practice, entrance aperture 221a and exit ap-erture 221b may be quite small, e.g. of the order of abouta millimeter wide (whereby it should be noted that a typ-ical width of the charged-particle beam will be of the orderof about a few tens of microns); if desired, exit aperture221b may be somewhat wider, to allow for greater de-flection amplitudes of the charged-particle beam from theparticle-optical axis 219. In this respect, it should be notedthat apertures 221a and 221b are not drawn to scale.[0035] Cavity wall 203 is made of conducting material,e.g. copper sheet with a thickness of a few mm. Providedin this wall 203 at successive angular intervals of 90°about the z-axis and in a common plane normal to the z-axis are small bores 207a, 207b, 207c, 207d, such thatbores 207a, 207c oppose one another along the y-axisand bores 207b, 207d oppose one another along the x-axis. Through these bores 207a, 207b, 207c, 207d pro-trude respective rods 209a, 209b, 209c, 209d. Two ofthese rods - 209a, 209b - carry respective excitementantennae 211, 213, whereas the other two rods - 209c,209d - carry respective tuning elements 215, 217. Thesetuning elements 215, 217 take the form of plungers(disks) that can be moved laterally into and out of thecavity 201, thus allowing adjustment of the effective in-ternal width of the cavity 201 along the y- and x-axes,respectively (or serving to alter the net dielectric constant

11 12

EP 2 722 865 A1

8

5

10

15

20

25

30

35

40

45

50

55

of the cavity interior). The excitement antennae 211, 213are located proximal the wall 203, as far as practicablefrom the particle-optical axis 219.[0036] The excitement antennae 211, 213 are con-nected via coaxial cables to respective (RF) Gigahertzoscillating power supplies (not shown); in fact, as set forthabove, one possible embodiment of the antennae 211,213 simply takes the form of a loop in the inner conductor(core) of each such coaxial cable. The output of said pow-er supplies is adjusted so as to produce the desired si-multaneous resonances of the cavity 201 - in the currentexample, two TM110 resonance modes that are mutuallyperpendicular. Because of the elliptical cross-section ofthe cavity 201, with the attendant difference in its internalwidth along the x- and y-axes, these two resonances willhave (slightly) different resonant frequencies. The exactresonance frequency values can be tuned by (slightly)sliding either or both of the tuning elements 215, 217 into/ out of the cavity 201; alternatively/supplementally, if thewall 203 is pliable, the eccentricity of the cavity 201 canbe slightly altered, e.g. with the aid of (undepicted) actu-ators (and/or by manual adjustment).[0037] In a specific example, it is elected to have theresonant frequencies of the cavity 201 at or close to avalue of 3GHz. This is not a limiting value: it is merelyconveniently achievable, and has an additional advan-tage of being the 40th harmonic of 75MHz. This latterfrequency is the pulse frequency of many commerciallyavailable lasers, and such lasers can, if desired, be em-ployed to apply an external stimulus to a sample duringexamination in a CPM (see above). It is thus relativelyeasy to phase-lock a 3GHz signal from an oscillating pow-er supply and a 75MHz output from a laser. It is alsorelatively easy to tune the two resonant frequencies ofthe cavity 201 to, for example, 3GHz along the ellipticalcavity’s major axis (x-axis) and 3.075GHz along its minoraxis (y-axis) - leading to a frequency difference of 75MHz,and thus allowing excellent synchronization to light puls-es produced with such a laser. For the frequency value(s)in question, the lateral dimensions of the cavity 201 willdepend inter alia on the dielectric medium present withinthe wall 203. For example:

- If the dielectric is vacuum, one obtains a minor-axislength (ay) of ca. 122 mm and a major-axis length(ax) of ca. 134 mm. The power loss (P) for such acavity is ca. 393 W, and its so-called Quality factor(Q) has a value of ca. 11100, assuming a magneticfield value of B = 3 mT (milliTesla) in the cavity.

- In an alternative case, the cavity 201 is largely filledwith a ceramic dielectric comprising ZrTiO4 dopedwith <20% SnTiO4. The employed body of dielectricis circle-cylindrical in form, fills the reference circle205, and comprises a central 3mm-wide shaft to al-low passage of the electron beam. The above-men-tioned values then become (approximately) ay ≈ 20mm, ax ≈ 22 mm, P ≈ 45 W and Q ≈ 2600. The valuesof ax and ay scale according to √εr, where εr is the

relative permittivity of the employed dielectric (rela-tive to vacuum), with εr ≈ 37 in the current case.

[0038] The above example assumes a length of thecavity 201 (parallel to the z-axis) of ca. 17 mm (thoughother values are, of course, possible).[0039] Figure 1C schematically illustrates the field ge-ometry of one of the TM110 modes in the subject of Fig-ures 1A and 1B, as follows:

- Left drawing: The magnetic field (B) lies purely in aplane parallel to the xy-plane. Close to the z-axis, itis oriented substantially along the y-axis (front-to-back in Figure 1C). Distal from the z-axis, it demon-strates a whirlpool form.

- Right drawing: The electric field (E) is zero in the x-and y-directions, and also zero along the z-axis. Toeither side of the z-axis, it demonstrates a clear di-pole form, with field lines oriented up/down parallelto the z-axis.

The other (simultaneously excited) TM110 mode in thecavity 201 will be basically the same, but will be laterallyrotated (in the xy-plane) through an angle of 90°, so thatthe magnetic field (B) lines close to the z-axis point left-right instead of front-back. In practice, the TM110 B-fieldexcited in the cavity will not be confined to a single plane:the situation illustrated on the left of Figure 1C will existin any plane taken normal to the z-axis (within the cavity).[0040] The effect of the abovementioned simultaneousresonances will be to superimpose an oscillatory x-de-flection and an oscillatory y-deflection on a charged-par-ticle beam propagating along the particle-optical axis219. The resultant oscillation will cause said beam totrace out a composite geometrical figure (a Lissajous fig-ure) on a (non-depicted) plane located downstream ofthe cavity 201 (beneath the aperture 221b in Figure 1B).An example of such a geometrical figure is illustrated inFigure 1D. If a (non-depicted) masking plate containinga small opening (such as a hole or slit) is located in saiddownstream plane, and the plate’s opening is positionedso as to be intercepted by the geometrical figure, then,as the beam traverses the opening, it will produce acharged-particle pulse downstream of the masking plate.[0041] In a specific configuration using the frequencyvalues quoted above, the inventors observed a 150 MHzrepetition rate for the figure depicted in Figure 1D. Whenviewed along a given lateral direction, each such repeti-tion will involve an outward and a homeward motion ofthe beam; if only one of these is selected, a 75MHz pulsefrequency will be obtained. Using a masking plate havingan opening (hole) width of ca. 10 mm and placed ca. 10cm downstream of the exit aperture 221b, one can createelectron pulses with a pulse length (duration) of ca. 100fs for the abovementioned magnetic field value of B ≈ 3mT.

13 14

EP 2 722 865 A1

9

5

10

15

20

25

30

35

40

45

50

55

Embodiment 2

[0042] Figure 2 renders a highly schematic longitudinalcross-sectional view of a particular embodiment of a CPMin which the current invention can be applied. In thepresent instance, the CPM is a TEM.[0043] The depicted TEM comprises a vacuum hous-ing 120 that is evacuated via tube 121 connected to avacuum pump 122. A charged-particle source in the formof an electron gun 101 produces a beam of electronsalong a particle-optical axis (imaging axis) 100. The elec-tron source 101 can, for example, be a field emitter gun,a Schottky emitter, or a thermionic electron emitter. Theelectrons produced by the source 101 are acceleratedto an adjustable energy of typically 80 - 300 keV (althoughTEMs using electrons with an adjustable energy of 50 -500 keV, for example, are also known). The acceleratedelectron beam then passes through a beam limiting ap-erture / diaphragm 103 provided in a platinum sheet. Toalign the electron beam properly to the aperture 103, thebeam can be shifted and tilted with the aid of deflectors102, so that the central part of the beam passes throughthe aperture 103 along axis 100. Focusing of the beamis achieved using magnetic lenses 104 of a condensersystem, together with (part of the) final condenser lens105. Deflectors (not depicted) are used to center thebeam on a region of interest on a sample, and/or to scanthe beam over the surface of the sample. In this sche-matic, functional depiction, the deflectors 102 are shownrelatively high up in the CPM, and final condenser lens105 is shown as being relatively small; however, theskilled artisan will appreciate that deflectors 102 may bemuch lower in the CPM (e.g. nested within the lens 105),and that item 105 may be much larger than depicted.[0044] The sample to be examined is held by a sampleholder 112 in such a manner that it can be positioned inthe object plane 111 of projection system 106 (whoseuppermost lens element is conventionally referred to asan objective lens). The sample holder 112 may offer var-ious positional/motional degrees of freedom (one or moreof translation(s), pitch, roll and yaw), and may also havetemperature control functionality (heating or cryogenic).It may be a conventional type of sample holder for holdinga static sample in a containment plane; alternatively, thesample holder 112 can be of a special type that accom-modates a moving sample in a flow plane/channel thatcan contain a stream of liquid water or other solution, forexample.[0045] The sample is imaged by projection system(projection lens system, projection column) 106 onto flu-orescent screen 107, and can be viewed through a win-dow 108. The enlarged image formed on the screen typ-ically has a magnification in the range 103x-106x, andmay show details as small as 0.1 nm or less, for example.The fluorescent screen 107 is connected to a hinge 109,and can be retracted / folded away such that the imageformed by the projection system 106 impinges upon im-age detector 151. It is noted that, in such an instance,

the projection system 106 may need to be (slightly) re-focused so as to form the image on the image detector151 instead of on the fluorescent screen 107. It is furthernoted that the projection system 106 may additionallyform intermediate images at intermediate image planes(not depicted).[0046] The image detector 151 may, for example, com-prise a Charge-Coupled Device (CCD) or a Complemen-tary Metal Oxide Semiconductor (CMOS) device, both ofwhich can be used to detect impinging electrons. As analternative to electron detection, one can also use aCCD/CMOS that detects light - such as the light emittedby a Yttrium Aluminium Garnet (YAG) crystal (for exam-ple) that is bonded to the CCD/CMOS, or connectedthereto by optical fibers (for example). In such an indirectdetector, the YAG crystal emits a number of photonswhen an electron hits the crystal, and a portion of thesephotons is detected by the CCD/CMOS; in direct detec-tors, electrons impinge on the semiconductor chip of theCCD/CMOS and generate electron/hole pairs, therebyforming the charge to be detected by the CCD/CMOS.The detector 151 is connected to a processing apparatus(controller) and display unit [not depicted].[0047] The image formed on the fluorescent screen107 and on the image detector 151 is generally aberrateddue (for example) to imperfections produced in the pro-jection system 106. To correct such aberrations, variousmultipoles can be deployed in/near the projection system106. Such multipoles are not depicted in Figure 2, so asto avoid cluttering the drawing, but the skilled artisan willbe familiar with their design, positioning and implemen-tation.[0048] Where the imaging beam impinges on the sam-ple 111, "stimulated radiation" is generated in the formof secondary electrons, visible (fluorescence) light, X-rays, etc. Detection and analysis of this radiation canprovide useful information about the sample 111. Toachieve such detection, Figure 2 shows a supplementarydetector 130, which is connected to a voltage source 132.As here depicted, the detector 130 is positioned at theside of the sample plane 111 proximal the gun 101; how-ever, this is a matter of design choice, and a detector 130may alternatively be positioned at the side of the sampleplane 111 distal the gun 101, for example. The "detector"alluded to in the appended claims may be either or bothof detectors 151 or 130, or another (undepicted) detector.[0049] In addition to the detectors 151/130, the depict-ed apparatus may also be equipped with EELS or EFTEMfunctionality, for example. In this context:

- For EELS: Deflectors 152 can be used to deflecttransmitted electrons (traversing the sample) in a di-rection away from the optical axis 100 and towardan off-axis EELS detector, which is not shown in Fig-ure 2.

- For EFTEM: Use can be made of an energy "filter",whose purpose is to select which energy range ofelectrons will be admitted to the detector 151 at any

15 16

EP 2 722 865 A1

10

5

10

15

20

25

30

35

40

45

50

55

given time. Such filter functionality can be fulfilled bythe deflection coils 152, which will "pass" certainelectron energies while deflecting others aside.

[0050] It should be noted that Figure 2 only shows aschematic rendition of a (simplified) TEM, and that, inreality, a TEM will generally comprise many more deflec-tors, apertures, etc.[0051] In the context of the present invention, it is de-sirable to be able to pulse/chop the electron beam beforeit impinges on the sample being investigated. To this end,a (non-depicted) beam pulsing device according to thepresent invention is disposed about the particle-opticalaxis 100 at some point between the source 101 and thesample holder 112, preferably at a crossover point, e.g.a focal point of the penultimate condenser lens 104. Thispulsing device will comprise a unitary resonant cavity asset forth above, e.g. similar to that in Figures 1A/1B andEmbodiment 1.

Embodiment 3

[0052] Figure 3 renders a schematic longitudinalcross-sectional view of another embodiment of a CPMin which the current invention can be applied. In thepresent instance, the CPM is a SEM.[0053] In Figure 3, a SEM 400 is equipped with an elec-tron source 412 and a SEM column (particle-optical col-umn) 402. This SEM column 402 uses electromagneticlenses 414, 416 to focus electrons onto a sample 410,and also employs a deflection unit 418, ultimately pro-ducing an electron beam (imaging beam) 404. The SEMcolumn 402 is mounted onto a vacuum chamber 406 thatcomprises a sample stage 408 for holding a sample 410and that is evacuated with the aid of vacuum pumps (notdepicted). The sample stage 408, or at least the sample410, may be set to an electrical potential with respect toground, using voltage source 422.[0054] The apparatus is further equipped with a detec-tor 420, for detecting secondary electrons that emanatefrom the sample 410 as a result of its irradiation by theimaging beam 404. In addition to the detector 420, thisparticular set-up (optionally) comprises a detector 430,which here takes the form of a plate provided with a cen-tral aperture 432 through which imaging beam 404 canpass. The apparatus further comprises a controller 424for controlling inter alia the deflection unit 418, the lenses414, 416, the detectors 420 and 430, and displaying ob-tained information on a display unit 426.[0055] As a result of scanning the imaging beam 404over the sample 410, output radiation, such as secondaryelectrons and backscattered electrons, emanates fromthe sample 410. In the depicted set-up, secondary elec-trons are captured and registered by the detector 420,whereas backscattered electrons are detected by detec-tor 430. As the emanated output radiation is position-sensitive (due to said scanning motion), the obtained (de-tected/sensed) information is also position-dependent.

The signals from the detectors 420 and 430, either sev-erally or jointly, are processed by the controller 424 anddisplayed. Such processing may include combining, in-tegrating, subtracting, false coloring, edge enhancing,and other processing known to the skilled artisan. In ad-dition, automated recognition processes, such as usedin particle analysis, for example, may be included in suchprocessing.[0056] In an alternative arrangement, voltage source422 may be used to apply an electrical potential to thesample 410 with respect to the particle-optical column402, whence secondary electrons will be accelerated to-wards the detector 430 with sufficient energy to be de-tected by it; in such a scenario, detector 420 can be maderedundant. Alternatively, by substituting one or more ofthe detectors 420 for the detector 430, these detectors420 can assume the role of detecting backscattered elec-trons, in which case the use of a dedicated detector 430can be obviated. In light of such variants, the "detector"alluded to in the appended claims may be either or bothof detectors 420 or 430, or another (undepicted) detector.[0057] If desired, one can realize a controlled environ-ment (other than vacuum) at the sample 410. For exam-ple, one can create a pressure of several mbar, as usedin a so-called Environmental SEM (ESEM), and/or onecan deliberately admit gases - such as etching or precur-sor gasses - to the vicinity of the sample 410. It shouldbe noted that similar considerations apply to the case ofa TEM, e.g. as set forth in Embodiment 2 above, wherebya so-called ETEM (Environmental TEM) can be realized,if desired.[0058] Once again, in the context of the present inven-tion, it is desirable to be able to pulse/chop the electronbeam 404 before it impinges on the sample 410 beinginvestigated. To this end, a (non-depicted) beam pulsingdevice according to the present invention is disposedabout the particle-optical axis of the SEM 400 at somepoint between the source 412 and the sample holder 408,preferably at a crossover point, e.g. a focal point of thelens 414. This pulsing device will comprise a unitary res-onant cavity as set forth above, e.g. similar to that inFigures 1A/1B and Embodiment 1.

Embodiment 4

[0059] Figure 4 renders a schematic longitudinalcross-sectional view of yet another embodiment of aCPM in which the current invention can be applied. Inthe present instance, the CPM is a FIB microscope.[0060] Figure 4 shows a FIB tool 500, which comprisesa vacuum chamber 502, an ion source 512 for producinga beam of ions along a particle-optical axis 514, and aFIB column (particle-optical column) 510. The FIB col-umn includes electromagnetic (e.g. electrostatic) lenses516a and 516b, and a deflector 518, and it serves toproduce a focused ion beam (imaging beam) 508.[0061] A workpiece (sample) 504 is placed on a work-piece holder (sample holder) 506. The workpiece holder

17 18

EP 2 722 865 A1

11

5

10

15

20

25

30

35

40

45

50

55

506 is embodied to be able to position the workpiece 504with respect to the focused ion beam 508 produced bythe FIB column 502.[0062] The FIB apparatus 500 is further equipped witha Gas Injection System (GIS) 520. The GIS 520 compris-es a capillary 522 though which a gas may be directedto the workpiece 504, and a reservoir 524 containing thegas (or a precursor substance used to produce the gas).A valve 526 can regulate the amount of gas directed tothe workpiece 504. Such a gas may be used in depositinga (protective) layer on the workpiece 504, or to enhancea milling operation performed on the workpiece 504, forexample. If desired, multiple GIS devices 520 may beemployed, so as to supply multiple gases according tochoice/requirement.[0063] The FIB tool 500 is further equipped with a de-tector 530, which, as here embodied, is used to detectsecondary radiation emanating from the sample 504 asa result of its irradiation by the ion beam 508. The signalfrom the detector 530 is fed to a controller 532. This con-troller 532 is equipped with a computer memory for stor-ing the data derived from this signal. The controller 532also controls other parts of the FIB, such as the lenses516a/b, the deflector 518, the workpiece holder 506, theflow of the GIS 520 and the vacuum pumps (not depicted)serving to evacuate the chamber 502. In any case, thecontroller 532 is embodied to accurately position the ionbeam 508 on the workpiece 504; if desired, the controller532 may form an image/spectrum of detected/processeddata on monitor 524.[0064] Again, in the context of the present invention, itis desirable to be able to pulse/chop the ion beam beforeit impinges on the sample 504 being investigated. To thisend, a (non-depicted) beam pulsing device according tothe present invention is disposed about the particle-op-tical axis 514 at some point between the source 512 andthe sample holder 506, preferably at a crossover point,e.g. a focal point of the lens 516a. This pulsing devicewill comprise a unitary resonant cavity as set forth above,e.g. similar to that in Figures 1A/1B and Embodiment 1.

Claims

1. A charged-particle microscope comprising:

- A charged-particle source, for producing abeam of charged particles that propagates alonga particle-optical axis;- A sample holder, for holding and positioning asample;- A charged-particle lens system, for directingsaid beam onto a sample held on the sampleholder;- A detector, for detecting radiation emanatingfrom the sample as a result of its interaction withthe beam;- A beam pulsing device, for causing the beam

to repeatedly switch on and off so as to producea pulsed beam,characterized in that the beam pulsing devicecomprises a unitary resonant cavity disposedabout said particle-optical axis and having anentrance aperture and an exit aperture for thebeam, which resonant cavity is embodied to si-multaneously produce a first oscillatory deflec-tion of the beam at a first frequency in a firstdirection and a second oscillatory deflection ofthe beam at a second, different frequency in asecond, different direction.

2. A charged-particle microscope according to claim 1,wherein said resonant cavity:

- Is substantially cylindrical in form, with a cylin-drical axis that is substantially collinear with saidparticle-optical axis;- Is embodied to be excited in TM110 resonantmode.

3. A charged-particle microscope according to claim 1or 2, wherein, when viewed in a direction normal tosaid particle-optical axis, the resonant cavity has anelongated cross-section, with a long axis parallel tosaid first direction and a short axis parallel to saidsecond direction.

4. A charged-particle microscope according to claim 3,wherein said cross-section is substantially an ellipsewhose major and a minor axis correspond respec-tively to said first and second directions.

5. A charged-particle microscope according to claim 3or 4, wherein the resonant cavity is mechanically de-formable so as to adjust a ratio of the lengths of saidlong and short axes.

6. A charged-particle microscope according to any pre-ceding claim, wherein said resonant cavity compris-es a dielectric material that is disposed about saidparticle-optical axis and that contains a substantiallyaxially symmetric void through which the beam canpass.

7. A charged-particle microscope according to claim 6,wherein an interface between said void and said di-electric material is at least partially coated by a filmof electrically conductive material.

8. A charged-particle microscope as claimed in anypreceding claim, wherein:

- In operation, said resonant cavity is embodiedto cause the beam to trace out a composite ge-ometrical figure on a masking plane perpendic-ular to the particle-optical axis;

19 20

EP 2 722 865 A1

12

5

10

15

20

25

30

35

40

45

50

55

- The microscope comprises a masking platelocated in said masking plane and having anopening that can be positioned so as to intersectsaid composite geometrical figure, thus servingto admit a pulse of charged particles as the beamtraces across said opening.

9. A charged-particle microscope according to claim 8,wherein said composite geometrical figure is a Lis-sajous figure.

10. A charged-particle microscope as claimed in anypreceding claim, additionally comprising apparatusfor applying a stimulus to the sample, which stimuluscan be synchronized to the output of the beam puls-ing device.

11. A method of examining a sample using a charged-particle microscope, comprising the following steps:

- Providing the sample on a sample holder;- Using a charged-particle source to produce abeam of charged particles that propagates alonga particle-optical axis;- Using a beam pulsing device to repeatedlyswitch the beam, thus producing a pulsed beam;- Using a charged-particle lens system to directthe pulsed beam onto the sample;- Using a detector to detect radiation emanatingfrom the sample as a result of its interaction withthe beam,characterized by the following steps:- Employing a unitary resonant cavity as part ofthe beam pulsing device, disposing this cavityabout the particle-optical axis, and passing thebeam through the cavity via entrance and exitapertures;- Exciting the resonant cavity to simultaneouslyproduce a first oscillatory deflection of the beamat a first frequency in a first direction and a sec-ond oscillatory deflection of the beam at a sec-ond, different frequency in a second, differentdirection.

21 22

EP 2 722 865 A1

13

EP 2 722 865 A1

14

EP 2 722 865 A1

15

EP 2 722 865 A1

16

EP 2 722 865 A1

17

EP 2 722 865 A1

18

EP 2 722 865 A1

19

EP 2 722 865 A1

20

EP 2 722 865 A1

21

REFERENCES CITED IN THE DESCRIPTION

This list of references cited by the applicant is for the reader’s convenience only. It does not form part of the Europeanpatent document. Even though great care has been taken in compiling the references, errors or omissions cannot beexcluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description

• US 20050253069 A1 [0007]

Non-patent literature cited in the description

• K. URA et al. Picosecond Pulse Stroboscopic Scan-ning Electron Microscope. J. Electron Microsc., 1978,vol. 27 (4), 247-252 [0008]