Letter pubs.acs.org/NanoLett
Electron Microscopy Reveals Structure and Morphology of One Molecule Thin Organic Films Virginia Altoe,† Florent Martin,‡,§ Allard Katan,‡ Miquel Salmeron,*,†,‡,§ and Shaul Aloni† †
The Molecular Foundry and ‡Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States § Materials Science and Engineering Department, University of California at Berkeley, Berkeley, California 94720, United States S Supporting Information *
ABSTRACT: Transmission electron microscopy was used to determine the structure of molecular films of self-assembled monolayers of pentathiophene derivatives supported on various electron transparent substrates. Despite the extreme beam sensitivity of the monolayers, structural crystallographic maps were obtained that revealed the nanoscale structure of the film. The image resolution is determined by the minimum beam diameter that the radiation hardness of the monolayer can support, which in our case is about 90 nm for a beam current of 5 × 106 e−/s. Electron diffraction patterns were collected while scanning a parallel electron beam over the film. These maps contain uncompromised information of the size, symmetry and orientation of the unit cell, orientation and structure of the domains, degree of crystallinity, and their variation on the micrometer scale, which are crucial to understand the electrical transport properties of the organic films. This information allowed us to track small changes in the unit cell size driven by the chemical modification of the support film. KEYWORDS: Organic monolayer, electron diffraction, scanning transmission electron microscopy, low-dose, polythiopene, Langmuir−Blodgett
O
resolution imaging techniques, such as transmission electron microscopy, use high energy electrons that cause irreversible damage.7 Typically, the crystallographic order is destroyed with electron doses as low as 1e−/Å2 for amino acids, 10 e−/Å2 for paraffin, and up to 18 000 e−/Å2 for thin chlorinated Cu phthalocyanine crystals.8 Despite this, electron microscopy is an essential tool that can provide unique information on organic samples. Organic crystals were first studied by electron diffraction in the 1930s9 and the first high-resolution transmission electron microscopy (HRTEM) images of radiation hard phtalocyanine crystals were obtained in 1970.10,11 In the early 1980s, Downing12 pointed out the advantages of spot-scanned techniques to reduce sample motion for increased image stability and resolution. Since then, great improvement in cryogenic techniques facilitated determination of the structure of two-dimensional (2D) and three-dimensional (3D) organic crystals by tomography13 and related diffraction techniques.14 For inorganic materials that can sustain high doses, a small parallel beam has been used for extracting crystallographic information from individual nanometer size crystals.15 Automated crystal orientation and phase mapping in the TEM was used to bridge the characterization gap between the nanometer structural determination and the microstructure.16,17 However, the study of monolayers of organic molecules at room
rganic electronic devices are modern technologies based on the charge transport properties of molecular assemblies.1,2 They have received considerable attention due to their promise of building nanometer size electronic devices made of low cost materials synthesized and assembled by wet chemistry methods. The electronic properties of molecular assemblies, such as charge transport, depend crucially on their internal structure,3 which is difficult to determine. And yet, it is only through a molecular level understanding of this structure and its relationship with the material’s electronic properties that advances in this important field will emerge. Moreover, the ability to follow structural variations on the microscopic scale will allow us to understand the effect of structural imperfections on transport properties, which is essential for performance optimization of organic electronics devices.4 Here we use transmission electron microscopy (TEM) to study organic monolayer films supported on electron transparent substrates that reveal their internal crystallographic structure and its relation to the microstructure of the film. This crucial information could not be obtained by any other technique. Determining the structure of functional organic materials hinges on the availability of techniques capable to map local structure with molecular level resolution. Scanning probe techniques fulfill some of these requirements but do not provide details of the internal structure of the molecules. In addition, they are impractical for high-resolution large area mapping.5 X-ray scattering techniques on the other hand provide detailed crystallographic information but their spatial resolution is limited to the micrometer scale.6 Powerful high© 2012 American Chemical Society
Received: October 26, 2011 Revised: January 26, 2012 Published: February 16, 2012 1295
dx.doi.org/10.1021/nl203776n | Nano Lett. 2012, 12, 1295−1299
Nano Letters
Letter
Figure 1. Models of the pentathiophene buteric acid molecule (5TBA), two of its derivatives with alkyl chains (D5TBA and DH5TBA), trans-4stilbene TSB, and stearic acid used in this study. The table lists the critical dose for damage at 120 kV of the monolayers (ML) and bilayers (BL) on ultrathin amorphous carbon (a-C), the 15 nm thin silicon nitride membrane (SiN) and suspended over the holes in lacy carbon support (suspended). The critical dose is measured from the decay time of the diffraction patterns.
noise ratio. A detailed description of the TEM methodology can be found in the Supporting Information, including a discussion on technical and instrumental factors limiting the technique. We used the method described above to map the structure of molecularly thin films of oligothiophenes (molecular models shown in Figure 1), a well-studied and promising family of organic semiconductors.22,23 The films were self-assembled at the air−water interface of a Langmuir−Blodgett trough and transferred to electron transparent substrates for TEM analysis (see details in the Supporting Information). Many of the films used here were also imaged by atomic force microscopy (AFM) prior to TEM analysis to determine their thickness and morphology. Figure 2A shows a STEM image of a 5TBA monolayer on a hydrophilic 15 nm thick SiN membrane. The film consists of islands 0.5 to 2 μm in diameter as well as some discontinuous regions. Very similar morphological features were observed in images obtained by AFM (see Supporting Information). The spatially resolved diffraction patterns show that the islands are made up of one or two crystalline domains that differ by their lattice orientation. These islands are 2.3 ± 0.3 nm high (Figure 2C), consistent with the height of a fully extended molecule. The lack of the (0,1) and (1,0) reflections (Figure 2B) indicates that their symmetry belongs to the p2gg space group, which is consistent with a herringbone arrangement of two molecules per unit cell (Figure 2D). The measured lattice parameters are a = 5.35 ± 0.12 Å and b = 7.46 ± 0.15 Å. A simulated diffraction pattern obtained from this model (Figure 2D) shows excellent agreement with the experimental results. In many cases, reflections up to the (1,4) order could be resolved, corresponding to an information limit of ∼1.8 Å. The orientation of the [0,1] lattice direction in the domains is shown by yellow arrows in Figure 2A. The domain boundaries are sharp with complete transition from one direction to the other often within a single 100 nm pixel. Both nuclei and areas with the discontinuous phase exhibit similar diffraction patterns consisting of distinct diffraction spots along specific radius forming semicontinuous diffraction rings. This indicates that the analyzed pixel area is composed by numerous randomly oriented crystal domains. The interplanar distances corresponding to all diffraction spots match the same distances found in the ordered domains of the island indicating that these regions have the same molecular arrangements in the islands but lack long-range order. This morphology is characteristic of 2D
temperature poses new challenges due to the weak contrast of low Z elements and their high sensitivity to beam damage. To overcome these problems, we combined the advantages of electron diffraction based techniques16−18 with low-dose,19,20 scanned parallel beam,15 and automated diffraction analysis.16,20 This novel methodology allows for mapping the crystallographic information over micrometer scale domains with resolution defined by sample sensitivity to the beam. The diffraction patterns were acquired from small areas by scanning a parallel electron beam of small diameter. The diameter of the beam was defined by the beam sensitivity of the material, effectively spreading the required electron dose to allow detection of the diffraction pattern on the time scale of the experiment (about 1 s). We used the high-angle annular dark field (HAADF) technique, where images are formed by collecting electrons scattered to high angles (∼50 to 200 mrad) where the scattering is predominately incoherent.21 Scanning transmission electron microscopy (STEM) HAADF signals and diffraction patterns were recorded at each location followed by converged probe imaging to obtain high-resolution HAADF images of the region of interest. Crystallographic information is extracted from individual diffraction patterns16,20 and then combined with the sample’s morphology. We could perform this entire procedure without exceeding the measured critical damage dose8 for the material. The measured critical dose for the samples reported here, illustrated in Figure 1, was between 6 and 65 e−/Å2, depending on substrate and sample preparation. The achievable resolution of the structural map is a compromise between sample radiation hardness, detector sensitivity and noise, and data acquisition rate. Diffraction yields useful information at lower doses than real-space images as the elastically scattered electrons are concentrated into few diffraction spots, thus providing a better signal-to-noise ratio. In our system, using the smallest condenser aperture limits the total beam current to approximately 5 × 106 e−/s. Typical exposure was about 1 s per pixel but due to hardware limitations the diffraction pattern was acquired for only 0.1 s. To keep the dose below the critical value, the electron beam was spread to a 90 nm diameter, limiting the full pitch resolution of the structural mapping. Acquisition of the highresolution HAADF image requires a similar dose, leading to a total dose of less than 20 e−/Å2 for each data set. In summary, the structural map resolution is defined by the scanned beam diameter while the crystallographic information is limited by the microscope information limit and the diffraction signal-to1296
dx.doi.org/10.1021/nl203776n | Nano Lett. 2012, 12, 1295−1299
Nano Letters
Letter
nucleation, growth, and coalescence during solvent evaporation.24 The array of diffraction patterns contains a wealth of information about the morphology, structure, and “quality” of the 2D crystal. Figure 3 shows false color maps of the lattice orientation obtained from analysis of the data shown in Figure 2. We find that the unit cell size of the 5TBA monolayer lattice is uniform within the instrumental precision of ±0.1 Å. A map of the intensities of the diffraction peaks, shown in Figure 3C, demonstrates that the total intensity of diffracted peaks is uniform over the monolayer island except in a small region to the right of the center, below the domain boundary, where the intensity is lower. This intensity difference is more pronounced at the higher order {1,2} and {1,3} reflections (Figure 3C). Because this region is neither at the edge of an island nor a domain boundary we conclude that it contains a large number of defects that lower the total diffraction intensity. In organic electronics and photovoltaics, regions of reduced crystallinity such as the ones detected here are prime candidates for performance-reducing centers. The ability to detect their presence and distribution is a key capability for determining the underlying cause of a material’s electronic properties.4 One interesting finding made possible by our method is that the morphology and structure of the 5TBA monolayer is dependent on substrate and sample preparation. On hydrophobic substrates, such as untreated carbon or SiN films, the Langmuir−Blodgett deposition leads to the formation of bilayers, while similar substrates rendered hydrophilic by a 20 s exposure to argon/oxygen plasma are coated by monolayers. Under all these conditions, the crystallographic structure of the films retains the p2gg group symmetry with small but reproducible changes in the lattice size and sample morphology. When the 5TBA monolayer is prepared on a thin carbon film support, the lattice elongates by about 2% in each direction. Another unexpected finding is that a bilayer of 5TBA on SiN has 18% longer lattice parameters than the monolayer. AFM height analysis of the bilayer shows that it is only 1.4 times as high as the monolayer. These observations suggest that the molecules assume a 45° tilted configuration in the bilayer, while they stand up straight in the monolayer. A layer of 5TBA suspended in vacuum over the holes of the TEM grid support shows a lattice constant intermediate between that of the
Figure 2. (A) Scanning transmission electron microscopy image of a 5TBA monolayer island deposited on a SiN membrane. Scale bar: 500 nm. The yellow arrows indicate the orientation of the (0,1) crystallographic direction at each location. Green circles mark polycrystalline areas. (B) Diffraction patterns from the area marked by the blue rectangle in A. Scale bar: 4 nm−1 . (C) AFM cross-section across a gap between two 5TBA islands. (D) Simulated kinematic diffraction pattern based on the crystalline structure shown in E. (E) Top view (upper) and side view (lower) of 3D rendering of the proposed crystalline structure of the molecular film.
Figure 3. Analysis of the crystallographic information from a 5TBA monolayer. (A) STEM dark-field (HAADF) image, scale bar: 500 nm. (B) Color-coded orientation of the domains representing the angle of the [0,1] lattice vector measured counterclockwise from the horizontal line (see Figure 2). (C) Discreet dark-field images obtained by mapping the total diffracted beam intensities scattered along the {1,1},{0,2},{1,2} and {1,3} reflections. The dark radial line from the center of the island in the images corresponds to the domain boundary between the two main domains. (D) Graphic representation of the changes in lattice parameters along [1,0] (x-axis) and [0,1] (y-axis) of the self-assembled thin films studied here. Notice the variations of the lattice dimensions as a function of the environment (legend). 1297
dx.doi.org/10.1021/nl203776n | Nano Lett. 2012, 12, 1295−1299
Nano Letters
Letter
Figure 4. Dark-field (HAADF) STEM images showing the morphology of 5TBA (A), D5TBA, which shows elongated domains emanating from the center and producing the star-like appearance of the film (B), DH5TBA (C), and TSB (D) supported on TEM grids. (E) Correlation between the orientation of the long axis of the radial D5TBA monolayer domains and the [0,1] lattice direction. Scale bar: 500 nm.
systems, the resolution of the method is limited to ∼90 nm. However a fast and direct control of the beam position combined with the use of fast and ultrasensitive detectors will eliminate unnecessary sample exposure time and improve the signal-to-noise ratio and image acquisition speed. These modifications will allow for the use of smaller beams with a higher electron flux, resulting in a better than 10 nm resolution. The methodology is not limited to the particular monolayer/ substrate combination and we expect it to have widespread applications in materials research.
supported bilayer and monolayer. These results show the flexibility of molecularly thin organic films and their ability to adapt their structural parameters to optimize their interaction with the substrate or environment. AFM height analysis of the suspended layer however was not possible so that it is not clear whether this structure is a stretched monolayer or a compressed bilayer. Figure 3D summarizes the observed dependence of lattice parameter on substrate properties. Other molecular films structurally similar to oligothiophene were studied, such as DH5TBA, which has hexyl side chains on the third and fifth thiophene cycles (Figure 1). DH5TBA forms large sheets that do not generate distinct spot patterns indicating that DH5TBA layers are amorphous or highly beam sensitive with critical doses below 0.5 e−/Å2. In contrast, D5TBA forms micrometer size islands composed of elongated crystalline domains that grow radially from the center to the periphery of the island. D5TBA is more sensitive to damage than 5TBA, its nonalkylated relative (Figure 4A−D). An important result from the analysis of the data is that the [0,1] lattice vector of the D5TBA unit cell is always parallel to the long dimension of the domains (Figure 4E). This information was instrumental in understanding the anisotropic electron transport properties found recently when using conductive AFM measurements.25 In summary, using a novel and simple method we could study the crystalline structure of highly beam sensitive materials. Our results demonstrate the ability of our TEM methods to produce detailed information on the molecular scale structure of organic mono- and bilayers and to map the distribution of structural defects and domains in the monolayer organic samples. We also demonstrated a previously unknown flexibility of organic film structure, where the unit cell dimensions can vary as a function of its interaction with the supporting substrate. At present, for highly radiation sensitive
■
ASSOCIATED CONTENT
S Supporting Information *
Additonal information and figures. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: (M.S.)
[email protected]; (S.A.)
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was performed at the Molecular Foundry at LBNL and was supported by the Office of Science, Office of Basic Energy Sciences of the DOE under contract no. DE-AC0205CH11231.
■
REFERENCES
(1) Green, J. E.; Choi, J. W.; Boukai, A.; Bunimovich, Y.; Halperin, E. J-.; Deionno, E.; Luo, Y.; Sheriff, B. A.; Xu, K.; Shin, Y. S.; Tseng, H. −R.; Stoddart, J. F.; Heath, J. H. Nature 2007, 445, 414−417.
1298
dx.doi.org/10.1021/nl203776n | Nano Lett. 2012, 12, 1295−1299
Nano Letters
Letter
(2) Klauk, H.; Zschieschang, U.; Pflaum, J.; Halik, M. Nature 2007, 445, 745−748. (3) Smith, R. K.; Lewis, P. A.; Weiss, P. S. Prog. Surf. Sci. 2004, 75, 1− 68. (4) Karl, N. Synth. Met. 2003, 133, 649−657. (5) Cygan, M. T.; Dunbar, T. D.; Arnold, J. J.; Bumm, L. A.; Shedlock, N. F.; Burgin, T. P.; Jones, L. II; Allara, D. L.; Tour, J. M.; Weiss, P. S. J. Am. Chem. Soc. 1998, 120, 2721−2732. (6) Yuan, Q.; Mannsfeld, S. C. B.; Ming, L. T.; Toney, M. F.; Lunig, J.; Bao, Z. J. Am. Chem. Soc. 2008, 130, 3502−3508. (7) Massover, W. H. J. Synch. Rad. 2007, 14, 116−127. (8) Reimer, L.; Kohl, H. Transmission Electron Microscopy, 5th ed.; Springer-Verlag: New York, 2008; Chapter 11. (9) Veinstein, B.; Zviyagin, B.; Avilov, A. In Electron Diffraction Techniques; J. M..Cowley, Ed.; Oxford University Press: New York, 1992; Chapter 6 and references therein. (10) Uyeda, N.; Kobayashi, T.; Ishizuka, K.; Fujiyoshi, Y. Nature 1980, 285, 95−97 and references therein.. (11) Zemlin, F.; Reuber, E.; Beckmann, E.; Zeitler, E.; Dorset, D. L. Science 1985, 229, 461−462. (12) Downing, K. H. Science 1991, 251, 53−59. (13) Cheng, Y.; Walz, T. Annu. Rev. Biochem. 2009, 78, 723−742. (14) Kulb, U.; Gorelic, T. E.; Mugnaioli, E.; Stewart, A. Polym. Rev. 2010, 50, 385−409 and references within.. (15) Ganesh, K. J.; Kawasaki, M.; Zhou, J. P.; Ferreira, P. J. Microsc. Microanal. 2010, 16, 614−621 and references within.. (16) Rauch., E. F.; Veron, M.; Portilo, J.; Bultreys, D.; Maniette, Y.; Nicolopoulus, S. Microsc. Anal. 2008, 93, S5−S8. (17) Rauch., E. F.; Portilo, J.; Nicolopoulus, S.; Bultreys, D.; Rouvimov, S.; Moeck, P. Z. Kristallogr. 2010, 225, 103−109. (18) Alloyeau, D.; Ricolleou, C.; Oikawa, T.; Langlois, C.; Le Bouar, Y.; Loiseau, A. Ultramicroscopy 2008, 108, 656−662. (19) Buban, J. P.; Ramasse, Q.; Gipson, B.; Browning, N. D.; Stalhberg, H. J. Electron Microsc. 2010, 59, 103−112. (20) Sadler, K.; Brown, A.; Brydson, R.; Bleloch, A. Ultramicroscopy 2010, 110, 1324−1331. (21) Williams, D. B.; Carter, C. B. Transmission Electron Microscopy; Springer Science+Business Media Inc.: New York, 1996; Vol. 3. (22) Mishra, A.; Ma, C. −Q.; Bäuerle, P. Chem. Rev. 2009, 140, 1141−1276. (23) Murphy, A. R.; Fréchet, J. M. J. Chem. Rev. 2007, 107, 1066− 1096. (24) Doudevski, I.; Hayes, W. A.; Woodward, J. T.; Schwartz, D. K. Colloids Surf., A 2000, 174, 233−243. (25) Hendriksen, B. L. M.; Martin, F.; Qi, Y.; Maudin, C.; Vukmirovic, N.; Ren, J. R.; Wormeester, H.; Katan, A. J.; Altoe, V.; Aloni, S.; Frechet, J. M.; Wang, L.-W.; Salmeron, M. Nano Lett. 2011, 11, 4107−4112.
1299
dx.doi.org/10.1021/nl203776n | Nano Lett. 2012, 12, 1295−1299
