An Appraisal of High Resolution Scanning Electron Microscopy Applied To Porous Materials

Introduction

In this report, we classify porous materialsbased on structural characteristics into the fol-lowing three categories: (i) Materials with aperiodic pore arrangement with uniform sizeeither in periodic framework structures such asmicroporous zeolites or in amorphous silica wallforming "cavity crystals" such as meso- andmacro-porous silica crystals; (ii) Materialswhich have irregular size of pores in a randomarrangement such as mesoporous silica withworm-like pores or random pores in size andarrangement; and (iii) Hollow nano spheres withor without nano-particles inside. Bothnanoporous (micro-, meso- and macro-porous)materials and their composites with nano-parti-cles have been prepared. All of these materialsare extremely important in a number of differentindustries. In (i), pores often provide acidic sitesand large surface areas for heterogenous cataly-sis to occur as is the case for (ii) where poresmay also be modified with transition metals orused to reversibly bind molecules for medicinalapplications. Owing to their porosity and struc-ture it is even possible to use them for both tis-sue engineering and drug delivery. In the case of(iii), hollow spheres with nanoparticles havealso been synthesized with the purpose of main-taining characteristic features of the nanoparti-cles whilst preventing their aggregation at cat-alytically active conditions. They also exhibitnew properties with their encapsulating hollowspheres. In order to maximize both the efficien-

cy and range of applications of these porouscrystals, an understanding of their structure andgrowth mechanism is therefore extremelyimportant for further control of growth process-es to tailor these materials for our specificneeds. Such properties we wish to controlinclude, but are not limited to, pore size, level ofexposure of pores to the crystal surface, orienta-tion of pores with respect to crystal surface, crys-tal size and shape, hydrophobicity/hydrophilici-ty and properties of the guest material.

With regards to porous crystals, scanning elec-tron microscopy (SEM) has long been used as amethod to rapidly gather information such as thesize and shape of crystals. This is because themicroscopes have a large depth of focus and arerelatively easy to operate with a minimum oftuition. As visualisation of crystals occurs atscan speeds and at a range of magnifications it istherefore possible to gather a statistically signifi-cant amount of data. This information is impor-tant when discussing crystal growth trends [1],but does not supply information to scientistsabout the crystal formation and structural chem-istry at the nanometer scale*. Such informationmay be found from observation of fine surfacestructure, for example: in zeolite LTA the heightand shape of surface terraces, observed by atom-ic force microscopy (AFM), provides evidenceof the existence of discrete growth units from

which the crystals form and the nature of thegrowth mechanism as a function of reactionconditions [2]. In scientific reports of thisnature, SEM images usually appear as littlemore than supporting information or visualexpatiation [3] to the scientific observations anddiscussions on crystal growth as gleaned fromother analytical techniques such as atomic forcemicroscopy (AFM) and also by powder X-raydiffraction (XRD), transmission electronmicroscopy (TEM) and others.

In the last decade there has been a quantumleap forward in the utility of SEM, which wenow dub high-resolution SEM (HRSEM).Particular advances include: Improvement ofobjective lens with smaller chromatic and spheri-cal aberration coefficients; precise control of thelanding energy of impact electrons and thereforea dramatic reduction in charging (the main con-tributor to loss of information in observation ofinsulating materials) whilst retaining a suitablysmall probe size; and an ability to obtain selec-tive information by tuning proper electron ener-gy ranges and collecting angles for detection.Important crystal features, such as mesoporouschannel openings, terminations and curvature [4]and other surface features such as twinning andgrowth fronts in zeolites [5] have been reported.

We begin by comparing the level of detail andvarious other merits between HRSEM and AFM,then further quantify the measurable level of sur-face topography and show that the HRSEM isable to identify nanometer scale crystal surfacefeatures through an elegant example, whilst atthe same time providing information on largercrystal details - both with a high level of fidelity.We then show that through new sample prepara-tion techniques, the detail HRSEM can reach is

Sam M. Stevens†,††, Kjell Jansson†, Changhong Xiao†, Shunsuke Asahina†††,Miia Klingstedt†, Daniel Grüner†, Yasuhiro Sakamoto†, KeiichiMiyasaka†,††††, Pablo Cubillas††, Rhea Brent††, Lu Han†,†††††, Shunai Che†††††,Ryong Ryoo††††, Dongyuan Zhao††††††, Mike Anderson††, Ferdi Schüth†††††††,and Osamu Terasaki†,††††*

† Department of Physical, Inorganic & Structural Chemistry, Arrhenius Laboratory, Stockholm University

†† The University of Manchester ††† JEOL Europe

†††† Korea Advanced Institute of Science and Technology ††††† Shanghai Jiao Tong University

†††††† Fudan University ††††††† Max Planck Institute Mülheim

Nanoporous materials such as zeolites and mesoporous silica crystals have attracted a lot of attention in recent years. In particular,the incorporation of various materials such as organic molecules, or metal nanoparticles and other inorganic compounds within theirpores which give rise to fascinating new functions. For such materials, it is essential to determine their structure, composition andmechanisms of growth in order to maximize their utility in future applications.

Recent progress in the performance of SEM is enormous, especially in low energy imaging where we can now directly observe finesurface structures of porous materials even those that are electrical insulators. Furthermore, by precise filtration and detection of emit-ted electrons by their energy, we can selectively obtain different types of information such as material composition, location of parti-cles inside or outside the pores etc. The physical processes and technologies behind this precise tuning of landing and detection ener-gies for both impact and emitted electrons, respectively, are explained and illustrated using a number of porous materials includingzeolite LTA, SBA-15, SBA-16, zeolite LTL, FDU-16 and Au@TiO2 ' rattle spheres,' along with comparisons with other techniques suchas atomic force microscopy (AFM) and transmission electron microscopy (TEM). We conclude that, by using extremely low landingenergies, advanced sample preparation techniques and through a thorough understanding of the physical processes involved, HRSEMis providing new and unique information and perspectives on these industrially important materials.

†SE-10691 Stockholm, Sweden, terasaki@struc.su.se

††††Daejeon 305-701, Korea, terasaki@kaist.ac.kr

*That said, SEM is also able to determinethe chemical composition of points on a crys-tal by measuring the energy / wavelength ofthe characteristic X-ray in energy / wave-length dispersive spectroscopy (EDS / WDS),respectively but the elemental information isan average over the range of a micrometer.

JEOL News Vol. 44 No.1 17 (2009) (17)

not limited to the surface of the crystal butincludes internal pore structure by cross sectionanalyses - the three dimension structural infor-mation of which differs from the projectedimages observed in TEM. As most porous crys-tals are insulators, they are therefore subject tocharging which reduces image information.Using extremely low landing energies* by adeceleration of highly unstigmated electronsleaving the objective lens called gentle beam(also known as stage bias or retarding field), weare able to remove charging whilst maintaining afinely focused electron probe as electrons passthrough the objective lens with the column ener-gy. We then explore the other types of detectablecontrast produced from electron irradiated sam-ples by careful filtering and detection of emittedelectrons based on their energy using an energyfilter for secondary electrons called r-Filter andwe show that compositional contrast is extreme-ly useful in understanding the nature of thesenew and exciting materials.

Experimental, Results andDiscussion, etc.:

Instrument & Electron Interaction Overview

An SEM is composed of a number of cham-bers. The first is a column filled with an elec-

tron source, apertures, scanning coils and lens-es, down which an electron beam is first gener-ated, accelerated, demagnified to a small elec-tron probe, and then deflected in a raster fash-ion over the area to be imaged in the mainchamber. The interaction between the impactelectrons and the sample generates a numberof different electrons such as back scatteredelectrons (BE), secondary electrons (SE) andAuger electrons and other emissions such ascharacteristic and Bremsstrahlung X-rays, visi-ble light through elastic and inelastic scatteringprocesses. Inelastically generated and emittedSEs are classified into two types called SE1and SE2 which are produced within materialseither directly from an incident (primary) elec-trons or from internally scattered electrons,respectively. Therefore SE1 produces a con-trast highly dependant upon sample surfacegeometry and gives topographic informationselectively, while SE2 comes from largerdepth and volume than those of SE1, the con-tribution of SE2 to the image reduces topo-graphic information. It is noteworthy thatreduction of SE2 contribution to an image canbe made only by reducing the landing energy.Characteristic X-rays, which are used for EDSand WDS measurements, give chemical infor-mation. Rutherford scattering of electronsinvolves high scattering angles of impact elec-trons, a detectable proportion of which arescattered through a large enough angle suchthat their resultant trajectory allows them toexit through the sample surface as BE. As

Rutherford scattering is highly dependant uponthe electrostatic potential of the sample atomnucleus, and therefore the atomic number, BEtherefore provides compositional contrast .The electrons generated are collected by ascintillator or solid state detector and the signalis multiplied and delivered as a grayscalevalue, depending upon the intensity, to a view-ing screen that is progressively scanning insynchronisation with the scanning beam. Theresult is an image of the crystal that may besaved to film or digital image file. All HRSEMimages in this article were taken using a JEOLhigh-resolution scanning electron microscopefitted with in-lens detector, gentle beam, coldfield emission electron source and EDS detec-tor (JSM-7000F and JSM-7401F) installed atStockholm University, Sweden.

More information about the type of interac-tions produced can be found both in reference[6] and throughout this article. We begin byexamining topographic contrast.

Level of Information/Resolution &Comparison With Atomic Force Microscopy

An elegant example of the power of imagingin HRSEM compared to AFM is shown inFigure 1. Part A shows a twinned crystal ofzeolite LTA where each twin is of similar sizeand shape and considerably overlapped.Visualisation of this twin would not be possi-ble in AFM owing to the steep sides of thecrystal. In parts B and C, we increase the mag-

(18) JEOL News Vol. 44 No.1 18 (2009)

*Landing voltage =accelerating (column) voltage - stage-bias voltage

Fig.1 HRSEM images of twinned crystals of zeolite LTA. (B) and (C) show the interface between the two twins located in the boxes on (A). 1 keVlanding energy, 2.5 keV column energy. In-lens detector, Sb mode (See Figure 8-1, D).

Fig.2 (A) composite HRSEM (left,grayscale) and AFM (right, red-scale) image of the exact same(100) surface of the exact samecrystal of zeolite LTA. (B) Crosssection is of AFM height meas-urements corresponding to thepurple line with a zoom in (C).(D) whole (100) surface of zeo-lite LTA with a magnification ofthe position of prevailing screwdislocation in (E). HRSEM: 1keV landing energy, 2.5 keV col-umn energy, In-lens detector, Sbmode. AFM: Constant ForceMode.

JEOL News Vol. 44 No.1 19 (2009) (19)

nification to two areas where the twins inter-sect each other. The AFM would not be able toprobe into the sharp angle between these twosurfaces owing to hindrance from the tip shapenor would it be able to image the macroporosi-ty along the twin intersection. Surface steps areclearly visible in Figure 1B and C implying ahigh level of vertical resolution - the size ofsuch features is quantified below.

As discussed in the introduction, SEM has,until recently, been unable to image surface ter-races of nanoporous crystals which are essentialto further the understanding of their growth withthis information only reliably attained by AFM.The topographic sensitivity is so high that sur-face terraces on zeolite LTA of 1.2 nm are clear-ly observable in HRSEM images. We have con-firmed this by imaging the surface of zeoliteLTA ibidem (in the same place) using both AFMand HRSEM and overlaying the two images asshown in Figure 2A. It can be seen that the twoimages overlap with extremely good correlationand all the terraces visible in AFM are observ-able in the HRSEM image. Height cross-sectionsfrom the AFM (Figure 2.B and C) show that thestep heights are all 1.2 nm and therefore thenominal vertical sensitivity is higher than theguaranteed lateral resolution of the microscopeat this beam condition (1.5 nm).

HRSEM images can be collected much morerapidly and from a larger area by HRSEM thanAFM, thus making the technique particularly

efficient. Images may also be obtained at lowermagnifications than that of the AFM and still theterrace information is conserved (Figure 2D andE). In AFM, the area to be scanned is limited bythe range of the piezoelectric crystal used tomove either the cantilever or the stage (depend-ing on the type of AFM) which delineates, andtherefore distorts the image at high displacement.The distortion in HRSEM is so slight that it canbe neglected: in Figure 3, the left hand image iscomposed of four individual images of zeoliteLTL with detailed surface terraces which matchup well enough so that they join together seam-lessly. The surface information is still conservedas in the three digital zooms shown in Figure 3A2 to 3A4, across the length of the crystal.

AFM uses a physical probe, rather than abeam of electrons, that scans in raster fashionacross the surface of the area to be observed.The tip is therefore prone to probe-tip dilationas can be seen in Figure 3B where the surfacecrater of zeolite LTL is imaged in both AFM(Figure 3B1) and SEM (Figure 3B2 to 3B3)and the appearance of terrace forking, the gapbetween each prong, is undetectable to thebulky AFM probe. The latter shows muchmore information about the crater detail. AFMdoes not perform well when the sample sur-face is far from orthogonal to the probe, suchcases include when crystals have rounded sur-faces as in silicalite-2 [7] or upon examiningsurfaces that intersect at obtuse angles.

As HRSEM is unable to measure verticalheights directly, AFM was used to confirm verti-cal resolution illustrating one advantage of AFMover SEM [8]. In terms of lateral resolution,however, HRSEM is the optimum technique. Wenow illustrate how this level of resolution andcontrast is not just limited to surface features butalso applicable to the internal structure of materi-als by first using a new sample preparation tech-nique called cross-section polishing.

Cross-Section Polishing & ComparisonWith Transmission Electron Microscopy

Cross-section polishing uses a beam of acceler-ated argon ions to polish the material of interest.By placing a shield that is resistant to the argonion beam on top of the material, this shield pro-tects a portion of the material from the abrasiveargon ion beam, and polishing occurs as shownin Figure 4. There is a high dependancy of attri-tion rate upon the angle between material anddirection of the argon ion beam. As the rate ofattrition is effectively zero when the surface ofthe material is parallel to the argon ion beam, avery flat cross section is therefore formed. It isfree from contamination and damage associatedwith other types of polishing such as mechanicalor chemical etching [9]. In the case of porouscrystals, they are first embedded in a (preferablyconductive) medium such as silver loadedepoxy or carbon glue before being polished. In

Fig.3 Zeolite LTL images. (A1) is acomposite of four SEM images,the areas highlighted at the top(magenta - A2), middle (cyan -A3) and bottom (green - A4) ofthe crystal are 2x digital zoom ofthose images. Right hand crystalis shown in both AFM (B1), SEM(B2) and HRSEM (B3) images,notice the probe-tip dilation effectin the AFM reducing the amountof information around the smallcrater towards the bottom of thecrystal. HRSEM: 0.8 keV landingenergy, 2.3 keV column energy.In-lens detector, Sb mode. AFM:Constant Force Mode.

Fig.4 Schematic of the Cross Section Polisher as viewed from the side (top-right) and isometrically (bottom-left).

CP blockSample

The argon ion beam is parallel to across section surface.

Example condition:6kV, 120uA, 3hr

Example Mehod:Back side Ar ion irradiation CP Ar ion

Si

Shield Plate

Fig.5 Mesoporous silica SBA-15.(A) and (B) are HRSEMimages of the unaltered crystalsclearly displaying surfacechannels and terminations. (C)and (D) are HRSEM of thepowder after cross section pol-ishing, notice surface profile isvisible along with grain bound-aries and change of channelorientation. (e) and (f) are ofcrystals with spherical mor-phology viewed both as a crosssection (HRSEM) and projec-tion (HRTEM), respectively.

order to reduce imperfections in the etchingprocess, the resin is heated or left under vacuumwhen drying in order to remove air bubbles, asthese small cavities create eddies and fluctua-tions in the otherwise homogenous stream of theargon ion beam. It is noteworthy that much ofthe atrophied material produced in the polishingprocess presents enough of a volume to con-tribute to contamination contrast. To reduce thisproblem, the supports are cleaned with ethanolbefore sample preparation and heated to 250°Cprior to introduction into the exchange chamberof the HRSEM.

In Figure 5, HRSEM images of SBA-15are shown where the mesoporous channelsterminate in corrugations on the side wallsurface or as U-turns (Figure 5A) or open(Figure 5B) hexagonally distributed pores onthe front wall. By encapsulating these crys-tals in a solid medium,it is possible to polishthese materials as shown in Figure 5C to 5D.Once the internal structure is exposed it canbe seen in the HRSEM that the polishingprocess is gentle enough to preserve both themesoporosity and the interface between thedifferent grains of SBA-15, the latter is invis-ible to TEM where the size of the crystalwould be too thick for transmission and thelack of periodicity of the interface. Similarly,the change in orientation of the pore acrossan individual crystal can be followed. Thispoint is reiterated in the spherical counter-parts shown in both TEM (Figure 5F) and bycross section polishing (Figure 5E) where the

contrast changes from lines (channels runperpendicular to the optic axis) to spots(channels run parallel). Cross-section polish-ing also allows a profile view of surface ter-minations that are visible in TEM but diffi-cult to distingusih and disprove to be fracturelines of a crystal fragment caused by thecompulsory crushing required to accommo-date crystals of this size upon a TEM grid.

Gentle Beam Helps With Charging InInsulating Porous Materials

An unfortunate occurrence when imaginginsulating specimens in HRSEM is that theyoften experience a phenomenon known ascharging. This results in the masking of imageinformation by areas of extremely high or lowcontrast or by a distortion in the image. This isdue to a build-up of unstable and in-homoge-nous electric fields generated by either a netdeficit or surplus of electrons provided by theimpact electrons of the electron source whichthen leave the sample either by conduction toearth or by emission. These electric fields dis-rupt the trajectory and generation of emittedelectrons. As most porous materials are com-posed of silica, they are virtually non-conduc-tive and so balancing of electrons within thespecimen requires tight control of the ratio ofemitted to impact electrons, known as the elec-tron yield, σ, preferably keeping it as close aspossible to unity for a given beam condition.The electron yield varies greatly with both sur-

face angle (thus giving rise to topographic con-trast, as discussed above) and acceleratingenergy, or more precisely, the landing energywhich is the energy electrons impact the sur-face of the sample. The critical energy and thegeneral dependency of electron yield differsfrom material to material but follows a generalcurve with a negative gradient*. The landingenergy may be controlled to find the criticalenergy which is the point below/above whichthe electron yield is respectively greater/lessthan unity [10]. In siliceous materials, this isusually between 500 eV and 1.5 keV. To high-light this we have imaged both SBA-16 andFDU-16; one silica based and the other carbonbased, both of which possess the same structur-al symmetry and approximately the same poredensity and shape. In Figure 6, where the con-trast of each image is normalised against thehighly ordered pyrolytic graphite (HOPG) sub-strate. As HOPG is very conductive and there-fore has effectively a constant electron yield of1 over all landing energy used. It can be seen inthe case of SBA-16 that the critical energy isbetween 1.4 and 1.7 keV, whereas in FDU-16,the range is much larger and at relatively higherenergies. Unfortunately for SBA-16, at the

*The function is actually hump shaped with asecond, lower critical energy at a lower landingenergy. As this side of the maxima (with a posi-tive gradient) falls below the practical range ofcurrent microscopes, it is removed from thescope of this paper for pragmatic reasons.

Fig.6 Mesoporous silica SBA-16 (top row) and reciprocal carbon analogue FDU-16 (bottom row) imaged at different accelerating voltages. Contrast nor-malized to that of the HOPG substrate. No surface bias. In-lens detector, Sb mode.

(20) JEOL News Vol. 44 No.1 20 (2009)

range of desirable accelerating energies wherecharging is minimised, the diffraction error andchromatic aberration dramatically limit theminimum possible probe diameter and resolu-tion drops below that required to image theimportant surface features [11].

To overcome the reduction in resolvingpower of the SEM at accelerating energiesbelow 5 keV, the use of beam retardation, oth-erwise known as gentle beam mode, has beenadopted. This is where the accelerating energy,which determines the minimum probe size andtherefore the limit of the resolution when pass-ing through the column (column energy), isretarded by a negatively charged stage bias to alower, landing energy equal to the columnenergy minus the stage bias. The landing ener-gy used is significantly lower than that usedbefore in obtaining images of such high levelsof resolution. Using this method it is possible toretain the advantage of high resolution imagingby having a high column energy but also reduc-ing charging effects by operating at a lowerlanding energy.

A further benefit of gentle beam is that theelectric field employed is orders of magnitudestronger, more homogenous and more stablethan that of the small electric fields discussedabove arising from sample surfaces with steepgradients or thin areas (i.e. tips and edges). Thismeans that the stage bias supersedes the smallfields and removes streaking and localised areasof charging, thus cancelling the disruption to thetrajectory and energy of the emitted electrons.This can be seen in Figure 7 where a hierarchi-cal porous material [12] templated from poly-styrene spheres (Figure 7A) is crushed producingmany tips and edges. When imaged without andwith a relatively small stage bias of 500 eV(Figure 7C1 to 7C2, respectively) the differenceto the improvement of the contrast is striking.There is also an increased resolution that can beseen as the mesopores within the walls of themacropore owing to the higher column energyand the mesoporosity in the macroporous wallare now observable (Figure 7D).

Charging is just one way in which emittedelectrons may lower the amount of informationdetected in an HRSEM image. By careful tuningof both the landing energy and the stage bias, itis possible to produce electrons with the maxi-mum information desired by the user. We shallshow in the next section how it is also possible to

selectively tune the energy of emitted electronsand the corresponding information to be gleanedfrom such electrons.

r-Filter Used To Separate DifferentTypes of Contrast

As discussed in the overview section, elec-trons with different energies, intensities andtrajectories are generated during SEM depen-dant upon different factors. Fortunately, theenergy difference between secondary andbackscattered electrons is very large andSEMs fitted with an r-Filter are capable ofindependently detecting these different elec-trons. This is done by placing a detector notat the side of the SEM chamber but inside theobjective lens. These in-lens detectors notonly have a higher signal-to-noise ratio (asthey are catching high-angle scattered elec-trons which are more prevalent when usinggentle beam) but also facilitate the incorpora-tion of electrodes placed in the objective lenswhich are then able to gate out/in electrons ofdifferent energies.

This can be seen in Figure 8, whereAu@TiO2 rattlespheres [13] are observed inSb mode (predominantly secondary electrons,Figure 8.1A) where the information is topo-graphic in nature, all electrons (Figure 8.1B)and Bs mode (predominantly back scatteredelectrons, Figure 8.1C) is almost all composi-tional contrast. By using both modes we areable to observe gold nanoparticles bothencapsulated inside and outside the macrop-orous TiO2 hollow spheres. The gold appearsas bright spots of contrast in Bs mode. As canbe seen in the images, both the short penetra-tion of impact electrons with low landingenergies and large attenuity of SE electronsallow for observation of gold particles on thesurface of the spheres (Figure 8.2A to 8.2B)but is unable to distinguish other gold parti-cles that are visible in the Bs mode appearingas very bright spots of contrast below the sur-face, encapsulated within the TiO2 shell. Thecomposition of both the shell and thenanoparticles with smaller than 20 nm indiameter are confirmed by now highly spatial-ly resolved EDS measurements in spectra 2and 1, respectively in Figure 8.3 - the samearea is chemically mapped in Figure 8.4 forboth titanium and gold.

Conclusion:

We have shown that HRSEM is a techniquecapable of extracting nanoscopic informationboth from the surface and inside nano-, meso-and macro porous materials. Such informationis not readily accessible using other techniquessuch TEM and AFM. The information is maxi-mized using a combination of technologiesincluding; cross-section polishing; gentle beamand r-Filter. By using ultra-low landing ener-gies whilst maintaining a highly stigmatedelectron probe we were able to produce imagesof extremely high resolution without chargingmasking the information. A more detailedaccount may be found in our upcoming reviewarticle [14].

Acknowledgments

The authors wish to acknowledge The Knut& Alice Wallenberg Foundation, SwedishResearch Council, Japan Science andTechnology Agency, EPSRC and ExxonMobilResearch and Engineering for their financialsupport, JEOL SAS (Europe) and JEOL Ltd.,for their technical support and MohammadJalil, Peter Alberius and L. Itzel Meza for sup-plying their samples. OT & YS acknowledgeProf. Kazuyuki Kuroda (Waseda University)for his continued encouragement and support.

References

[1] Yang et al. Revision of Charnell's proceduretowards the synthesis of large and uniformcrystals of zeolites A and X. Microporousand Mesoporous Materials vol. 90 (1-3)53-61. (2006)

[2] Agger et al. Crystallization in Zeolite AStudied by Atomic Force Microscopy. J.Am. Chem. Soc. vol. 120 (41) 10754-10759.(1998) ; Sugiyama et al. AFM observationof double 4-rings on zeolite LTA crystalssurface. Microporous and MesoporousMaterials vol. 28 (1) 1-7 (1999); Meza etal. Differentiating fundamental structuralunits during the dissolution of zeolite A.Chemical Communications (Cambridge,United Kingdom) (24) 2473-2475. (2007)

[3] Smaihi. Investigation of the CrystallizationStages of LTA-Type Zeolite by

JEOL News Vol. 44 No.1 21 (2009) (21)

Fig.7 Hierarchical Porous Material. (A) polysterene spheres used to form macroporous template, (B) free powder, (C1) and (C2) taken without and with a 500eV stage bias of the free powder lightly crushed exposing mesoporous network highlighted in(D). 3 keV column energy. In-lens detector, Sb mode.

(22) JEOL News Vol. 44 No.1 22 (2009)

Complementary Characterization Techniques.European Journal of Inorganic Chemistryvol. 2003 (24) 4370-4377. (2003)

[4] Che et al. Direct observation of 3D meso-porous structure by scanning electronmicroscopy (SEM): SBA-15 silica andCMK-5 carbon. Angewandte Chemie(International ed. in English) vol. 42 (19)2182-5 (2003); Terasaki et al. Structuralstudy of meso-porous materials by elecrronmicroscopy, "Mesoporous Crystals andRelated Nano-Structured Materials",Studies in Surface Science and CatalysisVol. 148, 261-288, Elsevier (2004);.Wakihara et al. Investigation of the surfacestructure of zeolite A. Physical ChemistryChemical Physics vol. 7 (19) 3416-3418(2005); Tueysuez et al. Direct imaging ofsurface topology and pore system ofordered mesoporous silica (MCM-41,SBA-15, and KIT-6) and nanocast metaloxides by high resolution scanning electronmicroscopy. Journal of the AmericanChemical Society vol. 130 (34) 11510-11517

[5] Terasaki. Electron Microscopy Studies in

Molecular Sieve Science. Molecular Sieves- Science and Technology vol. 2 71-112(1999)

[6] Reimer. Scanning Electron Microscopy.Berlin: Springer-Verlag (1998)

[7] Stevens et al. High-Resolution scanningelectron and atomic force microscopies:observation of nanometer features on zeo-lite Surfaces. Studies in Surface Science andCatalysis vol. 174 (2) 775-780 (2008)

[8] Anderson Michael et al. Modernmicroscopy methods for the structural studyof porous materials. ChemicalCommunications (8) 907-16 (2004)

[9] Erdman et al. Precise SEM cross sectionpolishing via argon beam milling.Microscopy Today (2006) vol. May pp. 22-25; Takahashi et al. A New Method ofSurface Preparation for High SpatialResolution EPMA/SEM with an Argon IonBeam. Microchimica Acta vol. 155 (1-2)295-300 (2006)

[10] Cazaux. e-Induced secondary electronemission yield of insulators and chargingeffects. Nuclear Instruments & Methods inPhysics Research, Section B: Beam

Interactions with Materials and Atoms vol.244 (2) 307-322 (2006); Rau et al. Secondcrossover energy of insulating materialsusing stationary electron beam under nor-mal incidence. Nuclear Instruments andMethods in Physics Research Section B:Beam Interactions with Materials andAtoms vol. 266 (5) 719-729 (2008)

[11] Reimer. Image Formation in Low-VoltageScanning Electron Microscopy. SPIE:Optical Engineering Press (1993);Mullerova and Frank. Scanning Low-Energy Electron Microscopy. Advances InImaging And Electron Physics vol. 128309-443 (2003)

[12] Loiola et al. Synthesis and characterizationof hierarchical porous materials incorporat-ing a cubic mesoporous phase. Journal ofMaterials Chemistry vol. 18 (41) 4985-4993

[13] Schüth et al. To be published; Pablo M.Arnal et al. High-Temperature-StableCatalysts by Hollow Sphere Encapsulation.Angew. Chem. Int. Ed. 45, 8224-8227(2006)

[14] Jansson et al. To be submitted.

Fig.8.1 Au@TiO2; Titanium nanoparticles creating interconnected macrospherical cavities, each encapsulating a gold nanoparticle. Images were takenat 3.0 kV with different ratios of SE (secondary) and BE (backscattered) electrons. Images constructed of only SE electrons (A), of a mixture ofSE and BE electrons (Sb mode) using the r-Filter (B) and of only BE electrons (C).

8.2 A reduced landing energy of 500 eV (column energy 1.5 keV) produces topographical images (A and B) showing clearly the encapsulated gold(see as a compositional contrast change in top right image (C)- the position of the gold).

8.3 Confirmation, by EDS chemical analysis, of position of gold nanoparticle as found in Figure 8.2.8.4 EDS mapping of spheres (A) showing titanium K-lines (B) and gold M-lines (C).