Morphology of Layered Silicate- (NanoClay-) Polymer Nanocomposites by Electron Tomography and Small-Angle X-ray Scattering Lawrence F. Drummy,* ,²,‡ Y. C. Wang, § Remco Schoenmakers, § Keith May, | Mike Jackson, O Hilmar Koerner, ²,# B. L. Farmer, ² Benji Mauryama, ² and Richard A. Vaia ² Materials and Manufacturing Directorate, Air Force Research Laboratory, AFRL/RXBP, 2941 Hobson Way, Wright-Patterson AFB, Ohio 45433, UES Inc., Dayton, Ohio 45432, FEI Company, Hillsboro, Oregon 97124, Triune Software, BeaVercreek, Ohio 45431, InnoVatiVe Management and Technology SerVices, Fairmount, West Virginia 26554, and UniVersal Technology Corporation, Dayton, Ohio 45432 ReceiVed October 5, 2007; ReVised Manuscript ReceiVed December 19, 2007 ABSTRACT: A basis for quantitative analysis of layered silicate- (nanoclay-) polymer nanocomposite morphology using two characterization methods, electron tomography and small-angle X-ray scattering (SAXS), is provided. For tilt greater than 15°, the contrast of a single montmorillonite layer experimentally decreases below the detectable limit of high-angle annular dark-field scanning transmission electron microscopy (HAADF- STEM). Calculations based on Z-contrast imaging of a 1 nm thick aluminosilicate layer predict this tilt angle (15°) should produce 17% contrast, consistent with a reasonable limit of HAADF-STEM detection for this system. This result implies that segmentation or thresholding of 2-dimensional Z-contrast projection images of randomly oriented, highly anisotropic nanoparticles, such as layered silicates in polymer nanocomposites, will be extremely inaccurate. For example, nearly 75% of the volume of montmorillonite layers in an epoxy matrix will not be identified in the segmentation, owing to their orientation alone. Using electron tomography, this number is reduced to below 15% and tomographic reconstruction reveals three-dimensional information. The corresponding 3D fast Fourier transformation (FFT) indicates that the image volume (10 -1 µm 3 ) does not contain sufficient distribution of local environments (interlayer correlation length ∼ 16.1 nm) to directly correspond to the global average as revealed by SAXS (scattering volume, 10 7 µm 3 ; interlayer correlation length ∼ 12.3 nm). Nevertheless, in contrast to SAXS, the tomographic reconstruction provides precise details of the distribution of morphological features, in addition to statistical averages over the sample volume. I. Introduction Polymer nanocomposites (PNCs) are of significant interest for a wide array of applications including sensors, 1 barrier materials 2 and high performance aerospace components. 3 PNCs can be defined as multiphase inorganic/organic hybrid materials in which one of the constituents has at least one dimension on the nanometer length scale (<100 nm). Many natural materials, such as bone, 4 conch shells, 5 or diatoms 6 also belong to this class of materials. Arguably, the most examined class of PNCs are those containing layered aluminosilicates (nanoclays). 7 They consist of pseudo-two-dimensional crystalline aluminosilicate layers 8 dispersed in a polymer matrix. Compared to traditional micro or millimeter scale composite materials, large improve- ments in several property areas can be engineered with the addition of a very small volume fraction of layered silicate owing to the small dimension (1 nm thickness) and high aspect ratio (>100) of the layers. 9 While these improvements have been impressive in certain select cases, in many instances PNC properties fail to meet such lofty expectations. There are several possible reasons for this, two of which are insufficient control of the interface between the nanoparticle and the matrix, and inability to reproducibly create and quantitatively verify uniform morphologies. For example, mediocre increases in mechanical performance may be tied to weak, ill-defined coupling at the interface between the polymer matrix and the nanoparticle. Furthermore, insuf- ficient quantification of the often complex hierarchical morphol- ogy present in PNCs have hindered detailed comparison of results between laboratories and with theoretical models, thus inhibiting the emergence of general structure-properties relation- ships for this class of materials. Characterization of nanocomposite morphology must over- come many significant challenges before these structure- properties relationships can mature. Methods for quantitative morphology characterization can be grouped into four catego- ries: real space (microscopy), reciprocal space (scattering), interfacial area (NMR, optical spectroscopy, dielectric spec- troscopy), and physical effects (rheology, mechanical properties, barrier properties). 10 The first (real space) provides a direct view of the morphology; however, care must be taken to avoid artifacts, to interpret images correctly, and to image a statistically significant amount of material. Reciprocal space scattering techniques are extremely powerful and they typically sample large amount of materials, although similar problems with proper data interpretation exist. Spectroscopy techniques such as dielectric and NMR can be sensitive to changes in structure and dynamics at the nanoparticle/matrix interface, and therefore can be used to probe the amount of interfacial area in nanocomposites. The use of physical effects such as mechanical ² Materials and Manufacturing Directorate, Air Force Research Labora- tory. ‡ UES Inc. § FEI Company. | Triune Software. O Innovative Management and Technology Services. # Universal Technology Corporation, Dayton, Ohio 45432 * Corresponding author. 2135 Macromolecules 2008, 41, 2135-2143 10.1021/ma702232f CCC: $40.75 © 2008 American Chemical Society Published on Web 02/23/2008