Measuring location, size, distribution, and loading of NiO crystallites in individual SBA-15 pores by electron tomography

Article abstract of DOI:10.1021/ja0728876

By the combination of electron tomography with image segmentation, the properties of 299 NiO crystallites contained in 6 SBA-15 pores were studied. A statistical analysis of the particle size showed that crystallites between 2 and 6 nm were present with a distribution maximum at 3 and 4 nm, for the number-weighted and volume-weighted curves, respectively. Interparticle distances between nearest neighbors were 1-3 nm with very few isolated crystallites. In the examined pores, a local loading twice the applied average of 24 wt % NiO was found. This suggests that a very high local loading combined with a high dispersion is achievable.

Full text of DOI:10.1021/ja0728876

Published on Web 07/27/2007
Measuring Location, Size, Distribution, and Loading of NiO
Crystallites in Individual SBA-15 Pores by Electron
Tomography
Heiner Friedrich,*,† Jelle R. A. Sietsma,† Petra E. de Jongh,† Arie J. Verkleij,‡ and
Krijn P. de Jong*,†
Contribution from Inorganic Chemistry and Catalysis, Utrecht UniVersity,
Sorbonnelaan 16, 3584 CA Utrecht, The Netherlands, and Molecular Cell Biology,
Utrecht UniVersity, Padualaan 8, 3584 CA Utrecht, The Netherlands
Received April 25, 2007; E-mail: h.friedrich@uu.nl; k.p.dejong@uu.nl
Abstract: By the combination of electron tomography with image segmentation, the properties of 299 NiO
crystallites contained in 6 SBA-15 pores were studied. A statistical analysis of the particle size showed
that crystallites between 2 and 6 nm were present with a distribution maximum at 3 and 4 nm, for the
number-weighted and volume-weighted curves, respectively. Interparticle distances between nearest
neighbors were 1-3 nm with very few isolated crystallites. In the examined pores, a local loading twice the
applied average of 24 wt % NiO was found. This suggests that a very high local loading combined with a
high dispersion is achievable.
prior research7-19 make catalysts prepared from SBA-15 ideal
model systems. For this study, a highly dispersed NiO/SBA-15
intermediate was chosen, since nickel is used as active
component in many industrially relevant processes.20-23
Structural characterization of OMM-based catalysts to study
the influence of various preparation treatments is mostly carried
out by bulk techniques, such as, N2-physisorption10-19 or X-ray
diffraction12-19 (XRD). From these measurements structural
parameters, like pore diameter, pore wall thickness, particle size,
and accessibility of the active phase, are derived. One major
1. Introduction
The intermediate size regime of nanostructured materials,
where the collective behavior of bulk materials emerges from
the discrete nature of molecular properties, provides the op-
portunity to observe unusual optical,1 magnetic,2 and catalytic3
behavior. One important factor for the design and production
of such novel materials, in general, and more efficient catalysts,
in particular, is the development of new characterization
procedures. In heterogeneous catalysis, size, structure, location,
distribution, and accessibility of the active particles within the
support determine the selectivity and activity of the catalyst.
Since commercial catalysts are highly complex, fundamental
studies are advantageously carried out on well-defined model
systems. In this respect ordered mesoporous materials (OMM)
have attracted much attention as model supports since 1992.4,5
This special interest can be attributed to their tunable but well-
defined pore geometry, bridging the gap between ordered
microporous zeolites and disordered meso- and macroporous
oxides. One of the most widely studied ordered mesoporous
silica is SBA-15, exposing a two-dimensional hexagonal sym-
metry of cylindrical pores.6 Its thermostability and the extensive
(7) Janssen, A. H.; Van Der Voort, P.; Koster, A. J.; de Jong, K. P. Chem.
Commun. 2002, 1632.
(8) Janssen, A. H.; Yang, C.-M.; Wang, Y.; Schu¨th, F.; Koster, A. J.; de Jong,
K. P. J. Phys. Chem. B 2003, 107, 10552.
(9) Ziese, U.; de Jong, K. P.; Koster, A. J. Appl. Catal., A 2004, 260, 71.
(10) Ravikovitch, P. I.; Neimark, A. V. J. Phys. Chem. B 2001, 105, 6817.
(11) Van Der Voort, P.; Ravikovitch, P. I.; de Jong, K. P.; Neimark, A. V.;
Janssen, A. H.; Benjelloun, M.; Van Bavel, E.; Cool, P.; Weckhuysen, B.
M.; Vasant, E. F. Chem. Commun. 2002, 1010.
(12) Khodakov, A. Y.; Zholobenko, V. L.; Bechara, R.; Durand, D. Microporous
Mesoporous Mater. 2005, 79, 29.
(13) Esparza, J. M.; Ojeda, M. L.; Campero, A.; Dom´ınguez, A.; Kornhauser,
I.; Rojas, F.; Vidales, A. M.; Lo´pez, R. H.; Zgrablich, G. Colloids Surf., A
2004, 241, 35.
(14) Sietsma, J. R. A.; Van Dillen, A. J.; de Jongh, P. E.; de Jong, K. P. Stud.
Surf. Sci. Catal. 2006, 162, 95.
(15) Kantorovich, D.; Haviv, L.; Vradman, L.; Landau, M. V. Stud. Surf. Sci.
Catal. 2005, 156, 147.
† Inorganic Chemistry and Catalysis.
‡ Molecular Cell Biology.
(16) Vradman, L.; Landau, M. V.; Kantorovich, D.; Koltypin, Y.; Gedanken,
A. Microporous Mesoporous Mater. 2005, 79, 307.
(17) Cho, Y. S.; Park, J. C.; Lee, B.; Kim, Y.; Yi, J. Catal. Lett. 2002, 81, 89.
(18) Li, X.-K.; Ji, W.-J.; Zhao, J.; Wang, S.-J.; Au, C.-T. J. Catal. 2005, 236,
181.
(1) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993,
115, 8706.
(2) Bala, T.; Sankar, C. R.; Baidakova, M.; Osipov, V.; Enoki, T.; Joy, P. A.;
Prasad, P. L. V.; Sastry, M. Langmuir 2005, 21, 10638.
(3) Bezemer, G. L.; Bitter, J. H.; Kuipers, H. P. C. E.; Oosterbeek, H.; Holewijn,
J. H.; Xu, X.; Kaptein, F.; van Dillen, A. J.; de Jong, K. P. J. Am Chem.
Soc. 2006, 128, 3956.
(19) Park, Y.; Kang, T.; Kim, P.; Yi, J. J. Colloid Interface Sci. 2006, 295,
464.
(20) Satterfield, C. N. Heterogeneous Catalysis in Practice; McGraw-Hill: New
York, 1980.
(4) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S.
Nature 1992, 359, 710.
(5) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.;
Schmitt, K. D.; Chu, C. T.-W.; Olson, D. H.; Sheppard, E. W.; McCullen,
S. B.; Higgins, J. B.; Schlenkert, J. L. J. Am. Chem. Soc. 1992, 114, 10834.
(6) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B.
F.; Stucky, G. D. Science 1998, 279, 548.
(21) Bond, G. C. Heterogeneous Catalysis; Oxford University Press: Oxford,
U.K., 1988.
(22) Augustine, R. L. Heterogeneous Catalysis for the Synthetic Chemist; Marcel
Dekker, Inc: New York, 1996.
(23) van der Lee, M. K.; van Dillen, A. J.; Bitter, J. H.; de Jong, K. P. J. Am.
Chem. Soc. 2005, 127, 13573.
9
10.1021/ja0728876 CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 10249-10254
10249

A R T I C L E S
Friedrich et al.
XRD and N2-Physisorption. Powder XRD measurements were
performed with a Bruker AXS D8 Advance diffractometer using Co
KR1,2 radiation with a wavelength of 0.179 nm. Small-angle XRD
patterns of the SBA-15 material were recorded from 0.5-5° (2θ) at an
angular increment of 0.0086°. The diffraction pattern was corrected
for the KR2 component prior to further processing. To provide
reproducible results and minimize effects of XRD sample preparation,
the average over eight individual small angle XRD patterns was taken.
XRD measurements of the NiO/SBA-15 were obtained from 40 to 100°
(2θ) at an angular increment of 0.0086°. The NiO crystallite size was
calculated from the fwhm of the broadened diffraction peaks using the
Scherrer equation.
drawback of bulk techniques is that no direct information on
the actual location and distribution of the active particles within
the support is obtained. If additional, spatially resolved informa-
tion is required, transmission electron microscopy (TEM)14-19,24-27
and scanning electron microscopy (SEM)25-29 are applied.
Although TEM is a very powerful technique, the resulting
images are in first approximation a two-dimensional (2D)
projection of a three-dimensional (3D) structure.30 This implies
that for complex 3D objects, like the examined NiO/SBA-15,
the actual 3D location and distribution of the NiO phase cannot
be unambiguously determined from single TEM images.
N2-physisorption measurements were performed at -196 °C (Mi-
cromeritics, TriStar 3000 V6.01 A) on the SBA-15 and the NiO/SBA-
15 intermediate. From the adsorption isotherms, the pore volume and
pore size distribution were deduced using NL-DFT theory37-39 imple-
mented in the DataMaster 4.01 software (Micromeritics).
Complex 3D structures can be imaged using electron tomog-
raphy (ET), sometimes referred to as 3D-TEM, which combines
multiple TEM images from different directions.7-9,24-27,31-35
First applications of ET to heterogeneous catalysts focused on
the qualitative description of the observed nanoscale structures.7-9
More recent applications involve a posteriori image segmenta-
tion to quantify catalyst properties, like, porosity,33 localization
of the active phase,34 and loading of active components.35 As
for any data derived from electron microscopy, care has to be
taken that the imaged structures are representative for the bulk;
thus, bulk characterization is mandatory to provide a basis to
evaluate the relevance of nanoscale observations. In this study,
ET in combination with posteriori image segmentation is applied
to measure the 3D location and size of single NiO crystallites
in individual SBA-15 pores. On the basis of the segmentation
results, the NiO particle size distribution, 3D nearest neighbor
distances, and a loading model of individual SBA-15 pores are
derived. Obtained results are verified and discussed in context
of the bulk characterization. To our knowledge, this is the first
application of ET to quantify the particle size distribution, 3D
nearest neighbor distances, and the local loading in individual
pores of the support material.
Electron Tomography (ET). TEM tomography grids were prepared
by applying 10 nm sized colloidal gold particles (Aurion, PAG
conjugated) from aqueous suspension to a Quantifoil R2/1 carbon film
supported on a parallel bar Cu TEM grid. After drying in air, a small
amount of the NiO/SBA-15 powder was deposited directly onto the
grid. Excess catalyst was shook-off before inserting the TEM grid into
a Fischione, model 2020, advanced tomography holder. ET was
performed in bright-field TEM mode using a Tecnai 20 electron
microscope (FEI) operating at 200 kV acceleration voltage, equipped
with a LaB6 electron source and a TWIN objective lens. A series of
TEM images (tilt-series) were acquired by rotating the specimen inside
the microscope over a tilt range of at least +70 to -70° at increments
of 1°. Images were recorded at each tilt step on a 2048 pixel × 2048
pixel TVIPS CCD camera using Xplore3D software (FEI). Diffraction
effects caused by a combination of the microscope aberrations and
strongly excited high-order beams, resulting in nonsymmetric white
halos around crystalline material, were excluded by using an objective
aperture. The defocus was set between -130 and -200 nm depending
on the thickness of the examined SBA-15 particle to keep the entire
object in underfocus. The nominal magnification was 29 000×, which
corresponds to a pixel size of 0.27 nm. Images of the tilt-series were
aligned with respect to a common origin and rotation axis using the 10
nm gold particles as fiducial markers. Finally the aligned tilt-series
was binned to 0.54 nm pixel size prior to reconstruction of the three-
dimensional volume. Alignment and reconstruction by filtered back-
projection was performed in IMOD.40 Preliminary results using the
sequential iterative reconstruction technique (SIRT) showed good
contrast in the reconstruction but did not improve the segmentation
described in the next paragraph.
Assessment of the NiO phase within the SBA-15 mesopores, a
process called segmentation, was carried out in Amira 3.1 (TGS-Visual
Concepts/Mercury Computer Systems, Inc.). Segmentation involved in
a first step the manual tracing of the NiO crystallite contours throughout
the volume of the examined mesopores. Since the NiO crystallites are
three-dimensional, three perpendicular cutting planes were used to
identify connected regions with dark contrast. As practical limit for
the smallest detectable particle, a connected region of at least 3 × 3 ×
2 pixels (∼3 nm3) had to be present. This practical limit exceeds the
rule-of-thumb for the theoretically predicted resolution which equals
three times the thickness of the sample divided by the number of images.
However, the practical achievable resolution has been reported to be
considerably better than the calculated value.9,31 In the last segmentation
step, the volume and the center-of-mass of each NiO crystallite were
calculated.
2. Materials and Methods
Preparation of NiO/SBA-15. SBA-15 (pore volume 0.73 cm3‚g-1
)
was synthesized according to a procedure described by Zhao.6 Deposi-
tion of the NiO precursor was carried out by incipient wetness
impregnation of SBA-15 using a 4.2 M aqueous Ni(NO3)2 solution (Ni-
(NO3)2·6H2O (Acros) in demineralized water) followed by drying at
120 °C for 12 h. Conversion of the dried precursor was done by heating
the sample at 1 °C‚min-1 to 450 °C in a flow of 1 vol % NO in He.
The sample was kept at 450 °C for 4 h. Following this procedure, highly
dispersed NiO/SBA-15 at a loading of 24 wt % NiO was obtained.36
(24) Meurig Thomas, J.; Midgley, P. A. Chem. Commun. 2004, 1253.
(25) Datye, A. K. J. Catal. 2003, 216, 144.
(26) Gai, P. L.; Boyes E. D. Electron microscopy in heterogeneous catalysis;
Institute of Physics Publishing: Bristol, U.K., 2003.
(27) Anderson, M. W.; Ohsuna, T.; Sakamoto, Y.; Liu, Z.; Carlsson, A.; Terasaki,
O. Chem. Commun. 2004, 907.
(28) Smith, D. J.; Yao, M. H.; Allard, L. F.; Datye, A. K. Catal. Lett. 1995, 31,
57.
(29) Bore, M. T.; Pham, H. N.; Switzer, E. E.; Ward, T. L.; Fukuoka, A.; Datye,
A. K. J. Phys. Chem. B 2005, 109, 2873.
(30) Hawkes, P. W. The Electron Microscope as a Structure Projector. In
Electron Tomography: Three-Dimensional Imaging with the Transmission
Electron Microscope; Frank, J., Ed.; Plenum Press: New York, 1992.
(31) Weyland, M.; Midgley, P. A.; Meurig Thomas, J. J. Phys. Chem. B 2001,
105, 7882.
(32) de Jong, K. P.; Koster, A. J. ChemPhysChem 2002, 3, 776.
(33) Ziese, U.; Gommes, C. J.; Blacher, S.; Janssen, A. J.; Koster, A. J.; de
Jong, K. P. Stud. Surf. Sci. Catal. 2005, 158, 633.
(34) Gommes, C. J.; de Jong, K. P.; Pirard, J. P.; Blacher, S. Langmuir 2005,
21, 12378.
(35) Midgley, P. A.; Meurig Thomas, J.; Laffont, L.; Weyland, M.; Raja, R.;
Johnson, B. F. G.; Khimyak, T. J. Phys Chem. B 2004, 108, 4590.
(36) Sietsma, J. R. A.; van Dillen, A. J.; de Jongh, P. E.; de Jong, K. P. Angew.
Chem., Int. Ed. 2007, 46, 4547.
(37) Tarazona, P. Phys. ReV. A 1985, 31, 2672.
(38) Tarazona, P. Phys. ReV. A 1985, 32, 3148.
(39) Tarazona, P.; Marini Bettolo Marconi, U.; Evans, R. Mol. Phys. 1987, 60,
573.
(40) Kremer, J. R.; Mastronarde, D. N.; McIntosh, J. R. J. Struct. Biol. 1996,
116, 71.
9
10250 J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007

NiO Crystallites in Individual SBA-15 Pores
A R T I C L E S
Figure 1. Geometrical model of the hexagonal mesopore ordering in SBA-
15 (a) and low-angle XRD pattern (b).
Figure 2. N2 adsorption and desorption isotherms of SBA-15 and NiO/
SBA-15 (a) and pore size distributions derived by NL-DFT (b).
3. Results and Discussion
Bulk Characterization. A schematic model of the SBA-15
support is shown in Figure 1a. The straight cylindrical pores
are viewed along the pore direction and the 2D hexagonal
symmetry (p6mm) of the pore ordering is apparent. The small-
angle XRD pattern of the SBA-15 (Figure 1b) shows distinct
peaks that are associated with the hexagonal symmetry.6 From
the position of the first three peaks, the unit cell size (a0) was
calculated to 10.5 ( 0.2 nm.
The corresponding SBA-15 adsorption and desorption iso-
therms with their typical hysteresis loop are displayed in Figure
2a. Analysis by NL-DFT calculations37-39 gave a porosity of
0.73 cm3‚g-1 and a very narrow pore size distribution with a
maximum at 9 nm (Figure 2b). These results confirm a uniform
SBA-15 support with 1-2 nm thick pore walls.
Figure 3. XRD pattern of the NiO crystallites within SBA-15 after
impregnation, drying and calcination.
Upon addition of the NiO phase to the SBA-15 support by
impregnation, drying, and calcination, the accessible pore
volume decreased (Figure 2a). The desorption branch of the
isotherm did not indicate a significant mesopore blocking.8,11
NL-DFT analysis yielded a porosity of 0.54 cm3‚g-1 and a pore
size distribution similar to the one observed for the blank SBA-
15 with the maximum again at 9 nm (Figure 2b). The XRD
pattern of NiO/SBA-15 (Figure 3) displays three distinct peaks
associated with the NiO crystal structure. The width of the
diffraction peaks confirmed that only nanometer-sized NiO
crystallites were created upon calcination of the precursor. All
three peaks had a fwhm that led via the Scherrer equation to a
NiO crystallite size of 4 nm. The formation of small NiO
crystallites in combination with the observed loss in pore
volume suggests that the NiO phase was located inside the
SBA-15 pores.
Characterization by ET. TEM observations, as in Figure
4a, confirm that the NiO phase was assembled inside the SBA-
15 pores. Using ET, the actual 3D structure is resolved. Figure
4b shows a 0.54 nm thick numerical cross section through the
reconstruction of the support particle A. It is immediately
apparent that not all pores of the support contained NiO
crystallites. In this particular structure, approximately 45% of
the pores were filled. This phenomenon, of filling only a fraction
9
J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007 10251

A R T I C L E S
Friedrich et al.
Figure 4. TEM image of NiO/SBA-15 structure designated particle A. The dark spots are 10 nm gold particles used for alignment of the tilt series (a). Also
shown is a 0.54 nm thick numerical cross section through the same object (b), with the magnified part of the cross section with a colored outline of the
segmented NiO crystals. Different colors have been assigned to crystals in different pores (c). Additionally shown is a surface view of segmented NiO
particles (d).
of the pores, has previously been reported for ZrO2/SBA-158
and MnOx/SBA-1541 and might be related to an inhomogeneous
concentration of surface hydroxyl groups.42,43 The extent of pore
filling by NiO throughout the whole SBA-15 structure can be
best observed by consulting the supporting movie slices.mpg.
For a quantitative characterization, image segmentation was
applied. In Figure 4c, a part of the numerical cross section has
been magnified. The manually traced contours of the NiO are
marked in different colors. Crystallites belonging to different
pores have been color coded green, blue, yellow, pink, and red.
A 3D surface model of the NiO is displayed in Figure 4d. This
example shows how ET in combination with a posteriori image
processing provides the volume and the 3D location of the NiO
crystallites in individual SBA-15 pores. For further analysis a
total of 299 NiO crystallites, contained in six filled pores (three
in support particles A and B), were segmented. The six filled
pores were chosen from different regions of the two SBA-15
structures to ensure proper statistical sampling. The relative
location of the three pores in support particle A is indicated in
the supporting movie slices.mpg.
the range of tilt angles, e.g., the finite thickness of the specimen
holder and restrictions of the microscope stage.44-46 The
elongation factor, e, which occurs in the direction of the electron
beam, can be calculated from the maximum tilt angle (R):47
R + sin R cos R
e )
(1)
xR - sin R cos R
Measured volumes of the NiO crystals have to be corrected for
elongation by dividing by the elongation factor. Simple division
can be applied, since elongation only occurs along one axis of
the three-dimensional structure.
A statistical analysis of the 299 NiO crystallite volumes
(Figure 5a) showed that a large number of small NiO crystallites
(<30 nm3) were present. The measured volumes are, in contrast
to those resulting from techniques like TEM or XRD, not based
on a sphere model. The volume statistics could be fitted well
to a log-normal distribution as predicted by Granqvist to be
observed independently of the prevalent growth mechanism.48,49
For comparison with XRD, the size of each NiO crystallite
can be estimated by its equivalent diameter, i.e., the diameter
of a sphere having the same volume. The resulting size
distribution and log-normal fit are given in Figure 5b. Despite
the fact that the crystallites did not appear to be spherical, a
close match is observed. When volume weighting was applied
to the log-normal fit, the curve maximum shifted from 3 to 4
nm, which corresponds very well with the NiO crystallite
diameter obtained by XRD. The deviation at a diameter value
of 5 nm was caused by poor statistical sampling as seen in the
For accurate volume measurements reconstruction artifacts
have to be considered. Dense objects will appear elongated in
the tomographic reconstruction, an effect which is known as
the missing wedge problem.44-46 The missing wedge of
information in Fourier space is caused by the inherent limit to
(41) Imperor-Clerc, M.; Bazin, D.; Appay, M.-D.; Beaunier, P.; Davidson, A.
Chem. Mater. 2004, 16, 1813.
(42) Moreno, M. S.; Weyland, M.; Midgley, P. A.; Bengoa, J. F.; Cagnoli, M.
V.; Gallegos, N. G.; Alvarez, A. M.; Marchetti, S. G. Micron 2006, 37,
52.
(43) Jentys, A.; Kleestorfer, K.; Vinek, H. Microporous and Mesoporous Mater.
1999, 27, 321.
(47) Radermacher, M.; Hoppe, W. Proceedings of the SeVenth European
Congress on Electron Microscopy, Den Haag, The Netherlands; Institute
of Physics: London, 1980; p 132.
(48) Granqvist, C. G.; Buhrman, R. A. J. Catal. 1976, 42, 477.
(49) Datye, A. K.; Xu, Q.; Kharas, K. C.; McCarty, J. M. Catal. Today 2006,
111, 59.
(44) Penczek, P.; Marko, M.; Buttle, K.; Frank, J. Ultramicroscopy 1995, 60,
393.
(45) Mastronarde, D. N. J. Struct. Biol. 1997, 120, 343.
(46) Midgley, P. A.; Weyland, M. Ultramicroscopy 2003, 96, 413.
9
10252 J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007

NiO Crystallites in Individual SBA-15 Pores
A R T I C L E S
Table 1. Properties of NiO-Filled SBA-15 Pores Derived from the
Reconstructions of Particles A and B
pore
length (nm)
NiO vol (nm3)
SiO2 vol (nm3)
NiO loading (wt %)
A1
A2
A3
B1
B2
B3
all
134
211
198
94
111
139
887
1650
1998
1711
939
1081
916
5588
8799
8257
3920
4629
5797
36991
47
41
39
42
42
32
41
8295
and very few isolated crystallites. In conjunction with the
measured particle size distribution, neighbors were separated
by only 1-3 nm. To our knowledge, no other technique can
provide this local nanoscale information.
Nanoscale Loading Model. The loading of an individual
SBA-15 pore can be calculated if the NiO and the SiO2 volumes
associated with that particular pore are known. The NiO volume/
filled pore equals the sum over the individual crystallite volumes,
while the SiO2 volume can be calculated from a geometrical
model:
x
3
2
a0
2
VSiO
)
‚Lpore
(2)
2
1 + FSiO (Vpores + VNiO
)
2
with a0 ) 10.5 nm the unit cell size, FSiO ) 2.2 g‚cm-3 the
2
density of silica, Vpores ) 0.54 cm3‚g-1 the pore volume
measured by N2-physisorption, VNiO ) 0.047 cm3‚g-1 the
volume of impregnated NiO/g of silica, and Lpore in nm the pore
length. The pore length is approximated by the center-of-mass-
distance between the first and the last segmented particle in
the pore.
Figure 5. Volume statistics over 299 NiO particles (a) and size distributions
of NiO particles expressed in diameter of sphere having the same volume
(b).
Detailed values for occupied pore length, segmented NiO
volume, calculated SiO2 volume, and the resulting loading in
individual pores are given in Table 1. To our surprise a local
loading much higher than the applied 24 wt % was found. The
loading average of 41 wt % corresponds to a NiO content in
the examined pores that is more than twice the expected value.
This observation is in line with the large percentage of unfilled
pores that arguably compensate for the high local loading.
Variations along the pores were studied by calculating the
loading in 10 consecutive pore segments as displayed for pore
A3 in Figure 7. Information on crystallite volume vs position
and a surface view complement the results. In general a
homogeneous distribution of the number of NiO particles is seen,
leading to few isolated crystallites as derived from the nearest-
neighbor distances. A preference toward higher loading at the
pore mouth was not found. Although small and large particles
were mostly distributed at random, the local loading decreased
in regions with mainly small particles.
Any conclusions to be drawn from this nanoscale pore loading
model have to be seen with regard to its limitations. Limiting
factors are the pore length approximation, the segmentation
accuracy, and the fact that NiO particles below 3 nm3 (∼1.8
nm sphere diameter) have not been detected. A 10% error in
both NiO and SiO2 volumes translated into 5 wt % difference
in local loading. Effects of the reconstruction point-spread
function beyond the elongation factor were not considered and
might further decrease the segmentation accuracy and, conse-
quently, the size determination.
Figure 6. Statistical analysis of the 3D nearest- and 3D second-nearest-
neighbor distances from 299 NiO crystallites.
region between 60 and 80 nm3 in Figure 5a. The observation
that particles of 6 nm and above were rare was also supported
by the N2-physisorption measurements and the inferred absence
of mesopore blocking.
In addition to the volume, the locations of all 299 NiO
crystallites were determined from their centers-of-mass. The
results of a statistical analysis of the nearest- and second-nearest-
neighbor distances between NiO crystallites can be seen in
Figure 6. Distances were measured in 3D between the centers-
of-mass of adjacent NiO crystallites belonging to the same pore.
Both curves are narrow indicating a homogeneous distribution
9
J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007 10253

A R T I C L E S
Friedrich et al.
Figure 7. Properties of pore A3 of NiO/SBA-15: position and corresponding volume of individual NiO crystallites (a); surface view of NiO crystallites (b);
local NiO loading calculated for 10 successive segments of the SBA-15 pore (c).
Electron tomograms contain large amounts of potentially
quantitative information, but the problem remains of how to
extract it. Manual segmentation, although subjective and its error
undetermined, is one rather straightforward solution to the
problem. On the basis of a total of 299 examined NiO
crystallites, a sufficient statistical sampling was given. Volume
and particle size distributions corresponded well with numerical
fits of a commonly assumed log-normal distribution.49 The close
match of the ET derived particle size to XRD results strongly
suggests that the studied crystallites were representative for the
bulk.
variations existed between and along the pores. The high local
loading was compensated by a large percentage of unfilled pores.
This suggests that a very high local loading combined with a
high dispersion is achievable, if the filling of the pores during
impregnation and drying is improved. ET has provided new
quantitative information on the 3D structure of NiO/SBA-15
and is expected to play a major role in the characterization of
nanostructured materials, in general, and heterogeneous catalysts,
in particular.
Acknowledgment. H.F. and K.P.d.J. acknowledge support
by the National Research School Combination Catalysis.
Microscopy was performed at the Electron Microscopy Group
Utrecht (EMU). We thank A. J. Koster, W. J. C. Geerts, J. D.
Meeldijk, T. P. van der Krift, M. N. Lebbink, A. Yakushevska,
and M. Barcena for their support.
4. Conclusions
ET in combination with a posteriori image processing was
successfully applied to study the size distribution and 3D
location of single NiO crystallites in individual SBA-15 pores.
Statistical analysis confirms a commonly assumed log-normal
size distribution with a volume-weighted maximum at 4 nm,
which was in excellent agreement with the crystallite size
determined by XRD. Interparticle distances between nearest-
neighboring crystallites were 1-3 nm with very few isolated
crystallites. The local loading in filled pores was almost twice
the applied overall loading of the support, and significant
Supporting Information Available: Slices.mpg, a movie of
consecutive numerical cross sections through reconstruction A
of NiO/SBA-15 structure, with the relative location of the three
segmented SBA-15 pores indicated. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA0728876
9
10254 J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007