LETTER pubs.acs.org/Langmuir
Quantification of a Single Aggregate Inner Porosity and Pore Accessibility Using Hard X-ray Phase-Contrast Nanotomography Pavel Trtik,*,† Miroslav Soos,*,‡ Beat M€unch,† Alexandros Lamprou,‡ Rajmund Mokso,§ and Marco Stampanoni§,|| †
Empa, Swiss Federal Laboratories for Materials Science and Technology, 8600 D€ubendorf, Switzerland Institute for Chemical and Bioengineering, Department of Chemistry and Applied Biosciences, ETH Zurich, 8093 Z€urich, Switzerland § Swiss Light Source, Paul Scherrer Institut, 5232 Villigen, Switzerland Institute for Biomedical Engineering, University of Zurich and ETH Zurich, 8092 Z€urich, Switzerland
)
‡
ABSTRACT: The 3D structure of three individual aggregates composed of 165 nm polystyrene primary particles is revealed nondestructively by hard X-ray phase-contrast synchrotron nanotomography. Three-dimensional image analysis allows us for the first time to obtain the complex inner porosity of the entire aggregate. It is demonstrated that despite their rather compact structure, characterized by a fractal dimension equal to 2.7, the produced aggregates are still porous, with porosity increasing with its size. Generated pores have diameters from 100 nm to 3 μm and are almost completely interconnected.
1. INTRODUCTION Porous materials (PMs) find application in many fields, including solid-phase synthesis, extraction, ion exchange, catalysis, pollutant adsorption, and chromatography. Independent of the application, there is a continuous need to optimize the properties of PMs, including the porosity, pore accessibility, pore size distribution (PSD), and size and shape of porous particles. Therefore, the characterization of the inner porosity of the PMs is the key issue in their development and optimization. Here, mercury intrusion porosimetry (MIP), N2 adsorption, or inverse size-exclusion chromatography can be used to provide some of this information. The disadvantage of these methods is that pore characteristics are obtained for a statistically large population of particles and they rely on indirect measurements of the internal structure and require models to interpret the PSD and pore space accessibility. Image-based techniques can provide detailed information on the 3D internal structure of porous materials. Among them, electron tomography (ET)1,2 was successfully applied to obtain nanometer-scale resolution of the particle internal structure. However, the volumes investigated by ET are about 1 μm3; therefore, the use of ET is restricted to macroscopically homogeneous samples. Theoretically, (cryo)focused ion beam nanotomography (FIB-nt)3,4 could provide measurements of larger samples of up to tens of micrometers in all three directions. However, because of an elaborate sample preparation and a long acquisition time for representative data sets, the use of FIB-nt (and ET as well) in high-throughput applications is not realistic. In principle, confocal laser scanning microscopy5 and (synchrotronbased) X-ray microtomography6 can be applied; however, their respective spatial resolutions are restricted to about 1 μm. r 2011 American Chemical Society
This letter represents a pilot project for measuring the 3D internal structure of aggregates composed of polymer primary particles utilizing Zernike phase-contrast X-ray nanotomography. Even though Zernike phase-contrast X-ray imaging was applied to the acquisition of 2D images,7 10 its application to determine the 3D internal structure of inorganic material such as zinc oxide particles11 and nanoporous gold12 or even cells13 is rather recent. The presented work represents the next step in the utilization of Zernike phase-contrast X-ray nanotomography toward the development of a high-throughput, high-resolution 3D technique for imaging very weakly absorbing materials.
2. MATERIAL AND METHODS The primary particles used in this study were prepared by miniemulsion polymerization14 using styrene as a monomer, divinylbenzene as a cross-linker, 2,2-azobis-2-methylpropionitrile as an initiator, sodium dodecyl sulfate as a surfactant, and hexadecane as a hydrophobe. These primary particles were swollen by monomer addition and consequently aggregated in a controlled fashion by a gradient of salt addition in the presence of shear.15 To freeze the aggregate internal structure, the last step in the aggregates preparation was the postpolymerization of the monomer previously added to the primary particles.16 The aggregates prepared in this way exhibited substantially higher mechanical stability compared to those composed of the primary particles assembled by van der Waals forces only. A Mastersizer 2000 device (Malvern, U.K.) was used to measure the ensemble-average radius of gyration of the aggregate population, ÆRgæ, and their fractal dimension, df. Received: September 1, 2011 Revised: September 21, 2011 Published: September 22, 2011 12788
dx.doi.org/10.1021/la203432v | Langmuir 2011, 27, 12788–12791
Langmuir
LETTER
Figure 1. Example of a structure factor measured by static light scattering. The line indicates the power-law scaling used to determine the fractal dimension of a population of aggregates, which for this case is equal to 2.7. To prepare the sample for both MIP and nanotomography experiments, the aggregate slurry was first diluted and subsequently dried on a glass surface. It is worth noting that as confirmed by the light-scattering analyses of the rewetted aggregates, the drying has no significant influence on ÆRgæ or df. The pore size distribution (PSD) and sample porosity were measured using a Porosimeter 2000 (Fisons, Italy). The obtained mercury intrusion pressure was used to evaluate the pore diameter through the Washburn equation17 using a Hg surface tension of 0.48 N/m and a contact angle of 130. The sample porosity was evaluated under the assumption of cylindrical pores. Zernike phase-contrast hard X-ray nanotomography in the full-field mode of image acquisition was applied to assess the inner porosity of a single aggregate composed of polystyrene primary particles with a diameter of 165 nm. Samples for nanotomography were prepared by transferring a few polystyrene aggregates inside a tapering quartz microcapillary that served as a sample holder. The aggregates remained stably attached to the inside of the microcapillary by surface forces, and no additional material was used for the fixation. The nanotomography experiments were performed on an individual polystyrene aggregate located in the narrow part of the tapering microcapillary using a full-field X-ray microscope18,19 at the TOMCAT beamline20 of the Swiss Light Source. Because the polystyrene aggregate absorbs only a fraction of a percent of the incoming X-ray beam (estimation based on X-ray transmission in a solid using http://www-cxro.lbl.gov), the modulation of intensity was achieved by a Zernike phase-contrast optical element that shifted the phase of the undiffracted beam by π/2. The Rayleigh resolution limit of the optical arrangement was equal to about 85 nm. A CCD device placed downstream from the test arrangement collected the raw projections. A binning factor of 2 was employed with the projection acquisition, leading to the isometric pixel size of projections equal to 50 nm. The tomographic experiment consisted of the acquisition of 361 raw projections in the equiangular positions over 180 of rotation. The acquisition of the sample raw projections preceded and was completed with 20 flat-field and 5 dark-field projections. The exposure time for each projection was 0.8 s. The 3D image was then reconstructed using a filtered back-projection algorithm with ring artifact suppression.21
3. RESULTS AND DISCUSSION An example of S(q) measured by static light scattering (SLS) is presented in Figure 1. As can be seen from the Guinier region (bending part) of the S(q),22 the aggregates investigated in this work have ÆRgæ = 40 μm and are very compact as indicated by df = 2.7. This is supported by an SEM picture of an aggregate population as well as by details of the single aggregate surface presented in Figure 2a,b. Despite rather high df values, several
Figure 2. (a) Detailed SEM image of the aggregate population together with (b) the detail of a single aggregate surface.
Table 1. Summary of Data for Three Aggregates aggregate Vagg (μm3) Ds (μm) εtotal (%) εdead (%) Finterconnected (%) large
3593.1
24.8
33.5
0.2
99.6
medium
1900.5
20.2
19.5
0.2
98.5
106.4
7.8
7.2
0.5
93.7
small
pore entrances in Figure 2b are clearly visible, indicating that the produced aggregates are still porous. To access information about the porosity of individual aggregates, hard X-ray nanotomography was applied on three individual aggregates having sizes of approximately 8 to 25 μm (Table 1). Figure 3a c shows vertical cross sections of the investigated aggregates located in tapering capillaries. All three investigated aggregates have an elongated shape with an aspect ratio of approximately 1.5, which is in agreement with the SEM image presented in Figure 2a as well as with our previous 12789
dx.doi.org/10.1021/la203432v |Langmuir 2011, 27, 12788–12791
Langmuir
LETTER
Figure 3. Vertical cross sections of the investigated polystyrene aggregates enclosed in quartz tapering microcapillaries: (a) small, (b) medium, and (c) large. The scale bars correspond to 10 μm.
Figure 4. (a) Cross-sectional cut of the medium aggregate. (b) Threedimensional rendering of a part of the polystyrene aggregate showing the segmentation of polystyrene (yellow), the largest connected region of the aggregate inner porosity (red), and the inner porosity unconnected to the outer space (blue). (c) Skeletonized view of the largest connected region of the aggregate inner porosity. The thickness of the skeleton tubes is scaled down by a factor of 2. The scale bar represents the pore diameter.
work.15,23 As illustrated by dark areas inside the investigated aggregates (Figure 3a c), they contain substantial intra-aggregate porosity, which is the focus of the further analysis. To reconstruct the 3D shape of an aggregate, a segmented mask of solid and porous regions was obtained by applying a threshold to the image histogram. Consequently, a sequence of morphological operations (dilation, filling, and erosion) was performed on the solid mask of a single aggregate to provide realistic mask discrimination between the intra-aggregate porosity (red and blue areas in Figure 4b) and the outer space. For comparison, the original cross-sectional view through the aggregates is presented in Figure 4a. An example of a skeletonized view of the largest connected region of the aggregate inner porosity is presented in Figure 4c. On the basis of this mask, the total volume of the aggregates, together with their intra-aggregate porosity and pore interconnectivity, was determined. It was found that the porosity is a strong function of the aggregate size (Table 1), with a negligible portion of dead pores (blue areas in Figure 4b). Interestingly, the detected pores are almost completely interconnected without the formation of isolated loops as indicated by the fraction of fully interconnected pores (Finterconnected in Table 1 close to 100%). This would imply good mass transport properties of the produced PMs. To obtain the PSD of individual aggregates, two numerical approaches were employed, in particular, a 3D continuous concept24 and a simulated MIP intrusion process.21 The obtained results are summarized in Figure 5a. The pore sizes cover the range from 100 nm to approximately 3 μm. Even though both approaches presented in Figure 5a represent the PSD, the MIP simulation is constrained by the occurrence and size of bottlenecks that are acting as restrictions for the intrusion process. Consequently, the
Figure 5. (a) Comparison of the cumulative distribution function of intraaggregate porosity obtained for three investigated aggregates based on the 3D continuous approach (dashed line) and a 3D simulation of the MIP (solid line). (b) Pore accessibility ratio of the intra-aggregate pore space for the spherical entities of given radii obtained for small (solid line), medium (shortdashed line), and large aggregates (dash-dotted line). (c) Comparison of the pore size distribution obtained from nanotomography for small (solid line), medium (short-dashed line), and large aggregates (dash-dotted line) together with the PSD measured by MIP using the original aggregate population (dashed line). The true spatial resolution of the technique is about 150 nm. As a result, the values below this limit should be perceived with this fact in mind.
accessibility ratio of the intra-aggregate pore space can be interpreted as the ratio between the constrained PSD obtained from the simulated MIP process and the one obtained from the unconstrained 3D continuous approach (Figure 5b). Although data obtained for medium and large aggregates are closely comparable showing that 10, 50, 90% of the volume of the intra-aggregate porosity is accessible from the outer space by spherical entities with radii of 800, 500, 300 nm, respectively, a much higher accessibility was found for small aggregates. This is due to the low porosity of the small aggregate combined with its location mostly at the aggregate periphery (Figure 3a). The last quantity evaluated from nanotomography data is the PSD (Figure 5c). All three aggregates show comparable PSDs covering the range from 100 nm to approximate 3 μm. However, because the spatial resolution for this measurement was about 150 nm,25 the results obtained below this value may contain a significant error. When comparing the obtained PSDs with that measured by conventional MIP (dashed line in Figure 5c), it can be seen that whereas in the range of small pores good agreement 12790
dx.doi.org/10.1021/la203432v |Langmuir 2011, 27, 12788–12791
Langmuir between both methods was found, a discrepancy exists in the region of large pores. Although the conventional MIP experiment was realized on the complete sample of the aggregate population having sizes from approximately 5 to 200 μm, because of field-of-view restrictions the nanotomography measurement can be realized only on aggregates with sizes smaller than approximately 35 μm. Aggregates larger than this limit that may contain larger pores are excluded from nanotomography analysis. Despite this difference, the obtained information is very valuable for the study of mass transport within single aggregates and its contribution to mass transport in the entire porous medium. In summary, we demonstrate that the 3D structure of the single polystyrene aggregate can be obtained by Zernike phasecontrast nanotomography. The entire single aggregate analysis allows the distinction of the intra-aggregate porosity from the outer space, thus providing valuable insight into the pore size distribution and its connectivity and accessibility. Moreover, we show that for samples of heterogeneous internal structure, X-ray nanotomography is a very valuable technique in obtaining information about large and, in our case, entire samples that cannot be measured with other methods (e.g., ET).2 In particular, we conclude that despite their rather compact structure the aggregates contain a network of micrometer-sized, highly interconnected pores that are to a large extent accessible to spherical particles of large radii.
LETTER
(14) Antonietti, M.; Landfester, K. Prog. Polym. Sci. 2002, 27, 689. (15) Soos, M.; Moussa, A.; Ehrl, L.; Sefcik, J.; Wu, H.; Morbidelli, M. J. Colloid Interface Sci. 2008, 319, 577. (16) Marti, N.; Quattrini, F.; Butte, A.; Morbidelli, M. Macromol. Mater. Eng. 2005, 290, 221. (17) Washburn, E. W. Phys. Rev. 1921, 17, 273–283. (18) Stampanoni, M.; Marone, F.; Mikuljan, G.; Jefimovs, K.; Trtik, P.; Vila-Comamala, J.; David, C.; Abela, R. Dev. X-ray Tomogr. 2008, 7078, V780. (19) Stampanoni, M.; Marone, F.; Mikuljan, G.; Jefimovs, K.; Trtik, P.; Vila-Comamala, J.; David, C.; Abela, R. J. Phys.: Conf. Ser. 2009, 186, 012018. (20) Stampanoni, M.; Groso, A.; Isenegger, A.; Mikuljan, G.; Chen, Q.; Meister, D.; Lange, M.; Betemps, R.; Henein, S.; Abela, R. Synch. Radiat. Instrum. 2007, 879, 848. (21) M€unch, B.; Trtik, P.; Marone, F.; Stampanoni, M. Opt. Express 2009, 17, 8567. (22) Sorensen, C. M. Aerosol Sci. Technol. 2001, 35, 648–687. (23) Ehrl, L.; Soos, M.; Morbidelli, M. Langmuir 2008, 24, 3070. (24) M€unch, B.; Holzer, L. J. Am. Ceram. Soc. 2008, 91, 4059. (25) Modregger, P.; Lubbert, D.; Schafer, P.; Kohler, R. Phys. Status Solidi A 2007, 204, 2746–2752.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected];
[email protected].
’ ACKNOWLEDGMENT We thank Dr. Thomas Suter, Mr. Erwin Pieper (both at Empa), and the entire staff of the TOMCAT beamline for assistance with and the support of the experiment. ’ REFERENCES (1) Frank, J. Electron Tomography: Three-Dimensional Imaging with the Transmission Electron Tomography; Plenum Press: New York, 1992. (2) Yao, Y.; Czymmek, K. J.; Pazhianur, R.; Lenhoff, A. M. Langmuir 2006, 22, 11148. (3) Holzer, L.; Indutnyi, F.; Gasser, P. H.; M€unch, B.; Wegmann, M. J. Microsc. (Oxford, U.K.) 2004, 216, 84. (4) Holzer, L.; M€unch, B.; Rizzi, M.; Wepf, R.; Marschall, P.; Graule, T. Appl. Clay Sci. 2010, 47, 330. (5) Lu, P. J.; Conrad, J. C.; Wyss, H. M.; Schofield, A. B.; Weitz, D. A. Phys. Rev. Lett. 2006, 96, 4. (6) Flannery, B. P.; Deckman, H. W.; Roberge, W. G.; Damico, K. L. Science 1987, 237, 1439. (7) Neuhausler, U.; Schneider, G.; Ludwig, W.; Meyer, M. A.; Zschech, E.; Hambach, D. J. Phys. D: Appl. Phys. 2003, 36, A79. (8) Youn, H. S.; Jung, S. W. J. Microsc. (Oxford, U.K.) 2006, 223, 53. (9) Youn, H. S.; Shin, T. J. J. Microsc. (Oxford, U.K.) 2007, 228, 107. (10) Zschech, E.; Geisler, H.; Rinderknecht, J.; Schneider, G.; Spolenak, R.; Schmeisser, D. Curr. Nanosci. 2008, 4, 256. (11) Li, W. J.; Wang, N.; Chen, J.; Liu, G.; Pan, Z. Y.; Guan, Y.; Yang, Y. H.; Wu, W. Q.; Tian, J. P.; Wei, S. Q.; Wu, Z. Y.; Tian, Y. C.; Guo, L. Appl. Phys. Lett. 2009, 95, 053108. (12) Chen, Y.-C. K.; Chu, Y. S.; Yi, J.; McNulty, I.; Shen, Q.; Voorhees, P. W.; Dunand, D. C. Appl. Phys. Lett. 2010, 96, 043122. (13) Stampanoni, M.; Mokso, R.; Marone, F.; Vila-Comamala, J.; Gorelick, S.; Trtik, P.; Jefimovs, K.; David, C. Phys. Rev. B 2010, 81, 140105. 12791
dx.doi.org/10.1021/la203432v |Langmuir 2011, 27, 12788–12791
