Direct Photopatterning and SEM Imaging of Molecular Monolayers on

The photochemical grafting of the alkenes to the surface of diamond films was also .... Gray represents regions of transparent fused quartz; black rep...
0 downloads 0 Views 354KB Size
Download PDF
Langmuir 2007, 23, 11623-11630

11623

Direct Photopatterning and SEM Imaging of Molecular Monolayers on Diamond Surfaces: Mechanistic Insights into UV-Initiated Molecular Grafting Xiaoyu Wang,† Paula E. Colavita,† Kevin M. Metz,† James E. Butler,‡ and Robert J. Hamers*,† Department of Chemistry, UniVersity of WisconsinsMadison, 1101 UniVersity AVenue, Madison, Wisconsin 53706, and Gas Dynamics DiVision, NaVal Research Laboratory, Washington, DC 20375 ReceiVed June 18, 2007. In Final Form: July 30, 2007 We have used X-ray photoelectron spectroscopy (XPS), infrared reflection-absorption spectroscopy, and fieldemission scanning electron microscopy (SEM) to investigate the formation of single- and two-component molecular patterns by direct photochemical grafting of alkenes onto hydrogen-terminated diamond surfaces using sub-band gap 254 nm ultraviolet light. Trifluoroacetamide-protected 1-aminodec-1-ene (TFAAD) and 1-dodecene were used as model systems for grafting. Illumination with sub-band gap light can induce several different kinds of excitations, including creation of mobile electrons and holes in the bulk and creation of radicals at the surface and in the adjacent fluid, which induce grafting of the alkenes to the surface. SEM images of patterned molecular layers on nanocrystalline diamond surfaces reveal sharp transitions between functionalized and nonfunctionalized regions consistent with diffraction-limited excitation. However, identical experiments on type IIb single-crystal diamond yield a significantly more extended transition region in the molecular pattern. These data imply that the spatial resolution of the direct molecular photopatterning is affected by diffusion of charge carriers in the bulk of the diamond samples. The molecular contrast between surfaces with different terminations is consistent with the expected trends in molecular electron affinity. These results provide new mechanistic insights into the direct patterning and imaging of molecular monolayers on surfaces.

1. Introduction Organic monolayers are widely used as a means to tailor the chemical and physical properties of the surfaces of materials. Advances in fields such as microelectronics and biotechnology have provided increased interest in developing methods for controlling the spatial distribution of molecular and/or biomolecular functional groups.1-8 While molecular patterning can be accomplished using a variety of “soft” lithography methods,2,9,10 the direct, single-step photopatterning of molecules can be a simple, yet elegant method for precisely controlling the locations of molecules on surfaces. Previous studies have shown that selfassembled monolayers on gold11,12 can be patterned using ultraviolet light to selectively remove molecules via oxidation.13-15 An even more direct approach is to use light to induce grafting, * Corresponding author. E-mail: [email protected]. † University of WisconsinsMadison. ‡ Naval Research Laboratory. (1) Kane, R. S.; Takayama, S.; Ostuni, E.; Ingber, D. E.; Whitesides, G. M. Biomaterials 1999, 20, 2363. (2) Shi, H. Q.; Tsai, W. B.; Garrison, M. D.; Ferrari, S.; Ratner, B. D. Nat. Mater. 1999, 398, 593. (3) Tarlov, M. J.; Burgess, D. R. F.; Gillen, G. J. Am. Chem. Soc. 1993, 115, 5305. (4) Ye, T.; McArthur, E. A.; Borguet, E. J. Phys. Chem. B 2005, 109, 9927. (5) Perring, M.; Dutta, S.; Arafat, S.; Mitchell, M.; Kenis, P. J. A.; Bowden, N. B. Langmuir 2005, 21, 10537. (6) Krupke, R.; Malik, S.; Weber, H. B.; Hampe, O.; Kappes, M. M.; von Lohneysen, H. Nano Lett. 2002, 2, 1161. (7) Mack, N. H.; Dong, R.; Nuzzo, R. G. J. Am. Chem. Soc. 2006, 128, 7871. (8) Laibinis, P. E.; Nuzzo, R. G.; Whitesides, G. M. J. Phys. Chem. 1992, 96, 5097. (9) Mrksich, M.; Whitesides, G. M. Trends Biotechnol. 1995, 6, 228. (10) Whitesides, G.; Ostuni, E.; Takayama, S.; Jiang, X. Y.; Ingber, D. E. Ann. ReV. Biomed. Eng. 2001, 3, 335. (11) Bain, C. D.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321. (12) Nuzzo, R. G.; Allara, D. L. J. Am. Chem. Soc. 1983, 105, 4481. (13) Zhang, Y.; Terrill, R. H.; Bohn, P. W. Chem. Mater. 1999, 11, 2191. (14) Huang, J. Y.; Dahlgren, D. A.; Hemminger, J. C. Langmuir 1994, 10, 626.

exemplified by studies showing that light of an appropriate wavelength can induce grafting of alkenes onto porous silicon16 and planar silicon surfaces,5,17 allowing patterns to be formed by controlling the photoinitiation process. Recent studies have shown that alkenes will graft to hydrogenterminated surfaces of polycrystalline- and single-crystal diamond when illuminated with 254 nm ultraviolet (UV) light,18-24 yielding functional molecular and biomolecular layers exhibiting extremely good chemical and thermal stability.19,20 The use of UV light to initiate grafting onto diamond,18-24 silicon,5,16,17 and other semiconductors25 raises important questions about the mechanism of functionalization and how the physical and chemical processes involved may affect the attainable resolution in direct molecular photopatterning. We recently showed that grafting of alkenes to H-terminated surfaces of nanocrystalline and single-crystal diamond is initiated by photoemission of electrons into the liquid alkenes (“internal (15) Brewer, N. J.; Rawsterne, R. E.; Kothari, S.; Leggett, G. J. J. Am. Chem. Soc. 2001, 123, 4089. (16) Stewart, M. P.; Buriak, J. M. Angew. Chem., Int. Ed. 1998, 37, 3257. (17) Cicero, R. L.; Linford, M. R.; Chidsey, C. E. D. Langmuir 2000, 16, 5688. (18) Strother, T.; Knickerbocker, T.; Russell, J. N., Jr.; Butler, J. E.; Smith, L. M.; Hamers, R. J. Langmuir 2002, 18, 968. (19) Yang, W. S.; Auciello, O.; Butler, J. E.; Cai, W.; Carlisle, J. A.; Gerbi, J.; Gruen, D. M.; Knickerbocker, T.; Lasseter, T. L.; Russell, J. N., Jr.; Smith, L. M.; Hamers, R. J. Nat. Mater. 2002, 1, 253. (20) Lu, M. C.; Knickerbocker, T.; Cai, W.; Yang, W. S.; Hamers, R. J.; Smith, L. M. Biopolymers 2004, 73, 606. (21) Nichols, B. M.; Metz, K. M.; Tse, K. Y.; Butler, J. E.; Russell, J. N.; Jr.; Hamers, R. J. J. Phys. Chem. B 2006, 110, 16535. (22) Nichols, B. M.; Butler, J. E.; Russell, J. N., Jr.; Hamers, R. J. J. Phys. Chem. B 2005, 109, 20938. (23) Lasseter, T. L.; Clare, B. H.; Abbott, N. L.; Hamers, R. J. J. Am. Chem. Soc. 2004, 126, 10220. (24) Yang, W.; Auciello, O.; Butler, J. E.; Cai, W.; Carlisle, J. A.; Gerbi, J. E.; Gruen, D. M.; Knickerbocker, T.; Lasseter, T. L.; Russell, J. N., Jr.; Smith, L. M.; Hamers, R. J. Mat. Res. Soc. Symp. Proc. 2003, 737, F4.4.1. (25) Kim, H.; Colavita, P. E.; Metz, K. M.; Nichols, B. M.; Sun, B.; Uhlrich, J.; Wang, X. Y.; Kuech, T. F.; Hamers, R. J. Langmuir 2006, 22, 8121.

10.1021/la701803g CCC: $37.00 © 2007 American Chemical Society Published on Web 10/16/2007

11624 Langmuir, Vol. 23, No. 23, 2007

photoemission”),21,22 initiating a cascade of events that ultimately grafts the alkenes to the surface. However, many questions remain about the mechanism of the grafting, the origin of sub-band gap photoemission from diamond,26,27 and how excitations in the diamond and in the adjacent fluid may impact the spatial distribution of molecules during direct molecular photopatterning. Understanding low-energy electronic excitations is also crucial for understanding the origin of the contrast in SEM images of molecular layers only ∼1 nm thick.28-32 Here, we report investigations of the direct photochemical patterning of molecular monolayers onto H-terminated nanocrystalline diamond surfaces using direct masking techniques and the imaging of these layers using scanning electron microscopy. Using SEM to directly image the layers, we demonstrate that the transition between functionalized and nonfunctionalized surface regions is limited by optical diffraction on nanocrystalline diamond but is significantly more diffuse on single-crystal diamond. Computational results suggest that the observed contrast between different surface functional groups is correlated with the molecular electron affinities. These studies provide important insights into the mechanisms of the UV-initiated grafting of molecular layers to diamond and the implications of this mechanism for controlling the fabrication of molecular patterns on small length scales. 2. Methods 2.1. Sample Preparation. Nanocrystalline diamond (NCD) films (p-type, B-doped, 0.49 µm thick) were grown on silicon (111) wafers at the Naval Research Laboratory using previously described methods.33 These samples have grains with an average diameter of ∼150 nm at the surface. A type IIb single-crystal sample cleaved along the (111) direction was used in some experiments. All data reported here used the nanocrystalline diamond samples unless otherwise stated. All diamond samples were terminated with hydrogen by exposing to a 13.56 MHz radio frequency hydrogen plasma while heating to ∼650 °C for 10 min; the samples were cooled in the plasma for 20 min and then further cooled in pure H2 for >30 min to ensure full H-termination.22,34 2.2. Sample Functionalization and Patterning. Two types of photomask were used to photopattern molecular monolayers onto diamond samples: (a) a large-scale mask consisting of a Mo sheet with 1 mm diameter holes and (b) a chromium mask having a variety of features with dimensions on the micron scale, fabricated using electron-beam lithography to pattern a photoresist on a UVtransmitting fused quartz substrate, and then depositing 70 nm Cr on the surface via standard lift-off procedures. Figure 1 shows the reaction scheme and the experimental setup of the photochemical modification of diamond. Experiments reported here were conducted with two molecules: trifluoroacetamideprotected 1-aminodec-1-ene (“TFAAD”) and 1-dodecene, as depicted in Figure 1a. In the photopatterning process, depicted in Figure 1b, a small droplet of an organic alkene (such as the TFAAD or 1-dodecene molecules depicted here) was placed onto the H(26) Bandis, C.; Pate, B. B. Phys. ReV. B 1995, 52, 12056. (27) Takeuchi, D.; Ri, S. G.; Kato, H.; Nebel, C. E.; Yamasaki, S. Diam. Relat. Mater. 2006, 15, 698. (28) Hickman, J. J.; Bhatia, S. K.; Quong, J. N.; Schoen, P.; Stenger, D. A.; Pike, C. J.; Cotman, C. W. J. Vac. Sci. Technol. A 1994, 12, 607. (29) Lopez, G. P.; Biebuyck, H. A.; Harter, R.; Kumar, A.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 10774. (30) Lopez, G. P.; Biebuyck, H. A.; Whitesides, G. M. Langmuir 1993, 9, 1513. (31) Rezek, B.; Sauerer, C.; Garrido, J. A.; Nebel, C. E.; Stutzmann, M.; Snidero, E.; Bergonzo, P. Appl. Phys. Lett. 2003, 82, 3336. (32) Saito, N.; Wu, Y.; Hayashi, K.; Sugimura, H.; Takai, O. J. Phys. Chem. B 2003, 107, 664. (33) Philip, J.; Hess, P.; Feygelson, T.; Butler, J. E.; Chattopadhyay, S.; Chen, K. H.; Chen, L. C. J. Appl. Phys. 2003, 93, 2164. (34) Thoms, B. D.; Russell, J. N., Jr.; Pehrsson, P. E.; Butler, J. E. J. Chem. Phys. 1994, 100, 8425.

Wang et al.

Figure 1. (a) Structures of TFAAD and 1-dodecene; (b) Reaction scheme for photochemical grafting of alkenes to H-terminated diamond. terminated diamond sample and the appropriate photomask was placed directly onto the wet sample; this traps a thin layer of reactant molecules between the diamond sample and the mask. The liquid layer thickness was measured to be 12-15 µm in thickness using side-view optical microscopy. The sample and the mask were placed inside a nitrogen-purged reaction chamber with a quartz window and the sample was illuminated with a low-pressure mercury lamp (254 nm, ∼1 mW/cm2) with a slow flow of dry N2 gas. Reactions were typically allowed to proceed between 13∼15 h, a time which we previously have shown is sufficient to achieve a self-terminating monolayer coverage with TFAAD.21,22 In some experiments, a second alkene, 1-dodecene (95%, Aldrich) was used in the functionalization and patterning of diamond surfaces. After the photochemical reaction, the diamond samples were sonicated in chloroform twice (5 min for each time), followed by methanol (5 min) and finally dried under N2 before analysis. 2.3. Characterization. The samples were characterized using X-ray photoelectron spectroscopy (XPS) in an ultrahigh vacuum system equipped with a load-lock for sample introduction, a monochromatized Al KR source (1486.6 eV), and a hemispherical analyzer with a multichannel detector. The photochemical grafting of the alkenes to the surface of diamond films was also verified by infrared reflection-absorption spectroscopy (IRRAS). Spectra were collected using a Bruker Vertex 70 FTIR spectrometer using a variable angle reflectance accessory (VeeMaxII, Pike) equipped with a wire grid polarizer and were referenced against a H-terminated diamond film. Spectra shown here were obtained using 65° angle of incidence (from the surface normal) for TFAAD on diamond and 70° for dodecene on diamond, using p-polarized light. Photopatterned diamond samples were imaged using a Leo Supra55 VP scanning electron microscope (SEM). Data shown here were obtained using 2 kV incident electron energy. Images were obtained using two different detectors: (1) a conventional, off-axis EverhartThornley secondary electron detector biased at +380 V and (2) the standard in-lens detector that is located above the final Gemini lens in the column of the SEM. The in-lens detector has a high efficiency for detection of secondary electrons produced by the incident beam (SE1 electrons) and low sensitivity for detection of back-scattered electrons and secondary electrons produced by the backscattered electrons (SE2 electrons) or other scattering processes (SE3 electrons).18,19,35-37 Because the in-lens detector provided higher contrast in all cases, images presented here used the in-lens detector unless otherwise stated. (35) Cazaux, J. J. Microsc. 2004, 214, 341. (36) Seiler, H. J. Appl. Phys. 1983, 54, R1. (37) Tse, K.-Y.; Nichols, B. M.; WYang, W. S.; Butler, J. E.; Russell, J. N., Jr.; Hamers, R. J. J. Phys. Chem. 2005, 109, 8523.

UV-Initiated Molecular Grafting

Langmuir, Vol. 23, No. 23, 2007 11625

Figure 3. IRRAS spectra of (a) TFAAD-terminated nanocrystalline diamond and (b) C-H region for 1-dodecene terminated nanocrystalline diamond.

Figure 2. XPS spectra of patterned TFAAD/H-terminated diamond surfaces. (a) XPS spectrum of sample covered by mask (b) XPS spectrum of region exposed to UV light, grafting TFAAD to the surface. The spectra are scaled so the C(1s) peaks are of equal height.

3. Results 3.1. Surface Functionalization with Alkenes. Although the ability to photochemically functionalize H-terminated diamond with a number of different alkenes has been demonstrated previously,19,22 we used XPS to verify the ability to prepare well-defined, chemically distinct regions by controlling the illumination of different parts of a single sample. A hydrogenterminated diamond sample was covered with TFAAD; half the sample was covered with an opaque sheet and the sample was then illuminated with 254 nm light. Figure 2 shows XPS survey spectra of both the covered (Figure 2a) and exposed (Figure 2b) regions of the sample after 13 h of reaction. High-resolution XPS spectra of the individual core levels are included in the Supporting Information. The survey spectrum of the sample that was not exposed to UV light shows a strong carbon peak with a small amount of oxygen but no significant intensity from fluorine. In contrast, the spectrum of the sample that was illuminated shows a significant F(1s) peak. Higher-resolution spectra (Supporting Information) of the region that was illuminated show that the F(1s), O(1s), and N(1s) peak areas are in a ratio of AF(1s):AO(1s):AN(1s) ) 3.1:1.0:0.88, which is consistent with the F:O:N ratio of 3:1:1 expected from the molecular formula of the parent alkene. The XPS analysis demonstrates that grafting of TFAAD to H-terminated diamond leaves the protecting group intact and confirms the ability to selectively functionalize specific surface regions by controlling the local intensity of UV light impinging on the sample. Infrared reflection-absorption spectroscopy (IRRAS) was also used to confirm successful grafting of TFAAD and of 1-dodecene. Figure 3a shows the IRRAS spectrum of a diamond surface after grafting of TFAAD. The C-H peaks at 2860 and 2939 cm-1, the peak structure in the C-H region, the carbonyl peak near ∼1705 cm-1, and the C-F stretching peaks (1160-1225 cm-1) are nearly identical to those reported for TFAAD grafted onto amorphous carbon surfaces.38 Grafting of 1-dodecene is difficult to characterize by XPS since the C(1s) peak cannot be distinguished from that of the underlying diamond, but IRRAS confirms successful grafting of this molecule. Figure 3b shows the C-H stretching region of a diamond surface after UV-initiated functionalization with (38) Sun, B.; Colavita, P. E.; Kim, H.; Lockett, M.; Marcus, M. S.; Smith, L. M.; Hamers, R. J. Langmuir 2006, 22, 9598.

Figure 4. SEM images of (a), (b) photopatterned 1 mm TFAAD spot on hydrogen-terminated diamond. (c) photopatterned 1 mm 1-dodecene spot on hydrogen-terminated diamond. (d) photopatterned 1 mm TFAAD followed by 1-dodecene grafted on the rest diamond surfaces. (e) photopatterned 1 mm 1-dodecene followed by TFAAD grafted on the rest diamond surfaces. Image (a) was taken with the Everhart-Thornley detector. Images b-e were taken with the inlens detector. Scale bar: 200 µm.

1-dodecene. The grafting procedure leads to the appearance of two near peaks at 2922 and 2850 cm-1 that arise from the asymmetric and symmetric CH2 stretching vibrations, respectively, of the alkyl chain.39 These data confirm that 1-dodecene grafts to H-terminated diamond surfaces. 3.2. Direct Molecular Photopatterning with SEM Imaging. To test the formation of 1- and 2-component molecular patterns on H-terminated diamond surfaces, we used scanning electron microscopy. SEM was used because the different electron yields of the different surface terminations provide a way to achieve chemical discrimination with sub-micron spatial resolution even on relatively rough surfaces.7,21,30 Initial experiments were conducted on samples patterned on millimeter length scales because this enables low resolution to be used in the SEM, reducing the electron dosage and minimizing possible effects of electron-induced damage to the layers. Figure 4 shows SEM images of organic layers prepared using the large (1 mm aperture) mask. In these images, the intensity and contrast adjustments of the SEM were optimized for each individual sample to show the best image quality for presentation. Figure 4a shows an SEM image of a sample functionalized with TFAAD in a 1 mm diameter region in the center, surrounded by (unreacted) H-terminated diamond, obtained using a conventional, off-axis EverhartThornley secondary electron detector biased at +380 V. The TFAAD-functionalized area appears darker than the H-terminated diamond surface. Figure 4b is a SEM image for the same sample obtained using the in-lens detector. With this detector, the (39) Ara, M.; Yamada, R.; Tada, H. Thin Solid Films 2006, 499, 8.

11626 Langmuir, Vol. 23, No. 23, 2007

TFAAD-functionalized spot again appears darker than the surrounding H-terminated surface. However, the image contrast between the two different regions is much larger when using the in-lens detector than when using the Everhart-Thornley detector. Similar behavior has been reported in previous studies of molecular layers on silicon6 and was observed for all of our samples on diamond. Consequently, all remaining SEM images in this paper will be those obtained using the in-lens detector. Figure 4c shows the SEM image of a 1 mm diameter dodecenefunctionalized region surrounded by H-terminated diamond. The dodecene-functionalized region appears darker than the surrounding H-terminated surface. These results demonstrate that grafting of molecules onto the H-terminated diamond samples alters the electron yield and allows the functionalized and non-functionalized (H-terminated) regions to be identified, even though the ∼1 nm thickness of the molecular layers is small compared to the ∼20 nm roughness of the nanocrystalline samples. We therefore investigated if it was possible to sequentially graft two different molecules, TFAAD and dodecene, onto the same diamond sample, while maintaining a sharp boundary between the layers. To test this, a H-terminated sample was covered with TFAAD and illuminated only in a 1 mm diameter circle using a photomask; the sample was then rinsed and covered with 1-dodecene and the entire sample was illuminated with UV light. Ideally, this procedure should produce a central TFAAD-functionalized region that is surrounded by a dodecene-functionalized region. Figure 4d shows an SEM image of the resulting sample. The central (TFAAD-functionalized) region appears darker than the surrounding (dodecene-covered) area. We also performed the complementary experiment, functionalizing the central region with 1-dodecene, rinsing, covering the entire sample with TFAAD, and then illuminating a second time. As expected, the resulting SEM (Figure 4e) image shows a brighter circular dodecene functionalized region surrounded by darker TFAAD regions. In both of these cases, the transition between two functionalized regions is clearly visible and the contrast is consistent with TFAAD functionalized regions appearing darker. There results demonstrate that two different organic monolayers can be patterned onto a diamond surface by sequential grafting and that the pattern can be imaged using SEM. 3.3. Spatial Resolution of Direct Molecular Photopatterning. The 4.9 eV energy of the incident photons is sufficient to induce photoemission of electrons from the diamond into the reactant fluids21,22 but is less than the 5.5 eV band gap of diamond. Sub-band gap photoemission from H-terminated diamond into vacuum is well-known, but its origin is not well understood.27,40 When in contact with reactant liquids, “internal photoemission” of electron into the liquid may create a number of possible excited species, including excess electrons and holes in surface states or bulk states, as well as radicals in the reactant liquid. Since many of these excited species can undergo diffusion, there is a connection between the mechanism of the functionalization process and the spatial resolution that can be achieved in patterning of the molecular layers. To test the spatial resolution of the patterning process on small length scales, we used a quartz mask with stripes of progressively smaller size (80-1 µm). Figure 5a shows the mask structure, with rectangular stripes covered by chromium and the remaining areas transparent. Figure 5b shows the resulting SEM image. As expected, the regions that were exposed to UV light give rise to lower secondary electron emission and therefore appear darker than the H-terminated regions. Figure 5c shows the profile of the (40) Cui, J. B.; Ristein, J.; Ley, L. Phys. ReV. B 1999, 60, 16135.

Wang et al.

Figure 5. Photochemical patterning of TFAAD onto H-terminated diamond surface. (a) Top view of the original mask with larger feature used for patterning. Gray represents regions of transparent fused quartz; black represents regions of chromium. (b) SEM image of the patterned surface. Lighter colors represent regions of higher secondary electron yields. (c) Profile showing linear variation in secondary electron yield along the line indicated in (b); note the abrupt transition, 1 for stripes wider than 2 µm. Thus, for all of the patterns investigated here, we are in the Fresnel diffraction regime. Using a simple single-slit model,49 the light exposure pattern can then be estimated for various widths of the stripe patterns. (49) Hecht, E. Optics, 4th ed.; Addison-Wesley: San Francisco, CA, 2002.

UV-Initiated Molecular Grafting

Figure 8 shows the predicted profile of light intensity across the pattern for stripe widths of 15 (Figure 8a) and 2 µm (Figure 8b). There are some common features shared by both figures. The light intensity near the center is close to the value of the incident light (I0) but is significantly lower (I ) 0.262I0 for the 15 µm stripe and I ) 0.375I0 for 2 µm) at the edges of the stripes. For the 15 µm stripe, the light intensity is high (∼I0) for most of the stripe and has an abrupt cutoff at the stripe edges. However, for the 2 µm width stripe, there is no region where the light intensity is uniform; instead it varies continually from I0 at the center of the stripe to 0.375I0 at edges. These predictions agree qualitatively with the experimental SEM images, which show that the TFAAD-functionalized and H-terminated regions are clearly distinct when using the 15 µm pattern, but the contrast is slightly less sharp, with a more continuous gradation from H-terminated to TFAAD-terminated regions when using the smaller 2 µm pattern. The importance of these optical calculations is to demonstrate that the resolution observed on nanocrystalline diamond can be explained simply via the diffraction of light due to the ∼15 µm separation between mask and sample due to the liquid film. Thus, even though the UV excitation induces a direct photoemission process that creates radicals and other species in the liquid phase and positive charges in the nanocrystalline diamond grains, the diffusion of these species does not impact the ability to produce molecular patterns on 1 µm length scales. In contrast, the much broader transition on single-crystal diamond implies that diffusion of charge carriers in the bulk is responsible for the poorer spatial resolution that is achievable. To understand the broad transition width observed on singlecrystal diamond, we note that while UV-initiated photoemission of electrons into the alkenes initiates the functionalization reaction, maintaining overall charge neutrality also requires the involvement of positive charges, such as abstraction of hydrogen as H+ from the surface, as integral parts of the functionalization reaction. Previous studies have suggested that the sub-band gap photoemission from H-terminated diamond involves excitation from mid-gap defect states to the conduction band,46,48 while others have suggested that it occurs by excitation of electrons from the valence band to C-H antibonding surface states that lie within the diamond band gap but are still above the vacuum level.47 The significant spreading observed on H-terminated diamond and not nanocrystalline diamond strongly implies that there must be some significant migration of excitation, away from where it is created. This is compatible with two possible scenarios. (1) If the photoelectron emission occurs via a single-step process mediated by C-H antibonding states of the surface atoms,47 then the profile of electron emission would be expected to track the diffraction-limited intensity distribution of the incident light, but the holes produced in the valence band could undergo diffusion after the photoelectron is emitted.26,45 (2) If the photoelectron emission is attributed to excitation from mid-gap defect states to the conduction band,46,48 then these excess conduction-band electrons could diffuse and broaden the spatial region from which electrons are emitted, while the holes would be significantly less mobile. While our present results do not allow us to definitively distinguish between these two possibilities, they clearly show the importance of bulk properties on what is otherwise an interfacial chemical reaction 4.2. Molecular Contrast in SEM Imaging of Monolayers on Diamond. Our SEM measurements show that each type of molecule can be distinctly imaged on the diamond surface with SEM, and that molecular layers decrease the secondary electron yield compared with the H-terminated surface. Previous studies

Langmuir, Vol. 23, No. 23, 2007 11629

of self-assembled monolayers on gold found that thin molecular layers increased the secondary electron yield compared to the clean gold surface.7 We attribute the decrease in electron emission upon molecular grafting onto H-terminated diamond to the fact that the H-terminated diamond surface is an unusually facile electron emitter due to its negative electron affinity (NEA).26,42-45 The NEA arises because the surface H atoms are slightly positively charged, creating a surface dipole that eliminates the barrier to electron emission from the conduction band into vacuum. Since the NEA is a consequence of the C-H surface dipoles, it is perhaps surprising that replacing the C-H dipoles of the H-terminated diamond surface with the C-H bonds of a molecular layer such as dodecene (which yields a dodecane-like monolayer) reduces the local electron emission. No significant interface dipole is expected at the diamond-dodecene interface, and the Hterminated diamond surface and the dodecene-modified surface are both terminated in C-H bonds. However, there are clear differences in atomic structure and degree of order. Because the molecular layers of dodecene and TFAAD are thin compared to the inelastic mean free path of the incident 2 kV electrons, the vast majority of the electrons that are emitted and collected during SEM imaging are created by inelastic scattering of electrons within the diamond substrate, rather than from within the molecular layers. One possibility is that the net decrease in electron emission after functionalization with dodecene arises because the random orientations of the C-H bonds of the alkyl terminal groups provide a smaller dipole perpendicular to the surface than those at the H-terminated diamond surface. However, we believe a more likely explanation is that as electrons in the diamond impinge on the diamond-molecule interfaces, they are more effectively scattered than at the diamond-H interface because of the disordered nature of the molecular layers at the diamondmolecule interface. Previous experimental and theoretical studies of self-assembled monolayers on gold have shown reduced electron-transmission rates through disordered layers compared with ordered layers, which has been ascribed in part to enhanced scattering of the electrons passing through the disordered molecular layer.50-53 We believe these electron scattering effects are the primary reason why the grafted layers of dodecene have lower secondary electron yield than the starting H-terminated diamond surface. The even lower electron emission from the TFAAD-modified surface is attributed largely to the electronic properties of trifluoroacetamide group. Previous studies by Saito, et al. on silicon showed that the contrast variations between different molecules correlated closely with the ordering of the lowest unoccupied molecular orbitals (LUMOs), with higher-energy LUMOs giving rise to increased secondary electron yields.32 On this basis, they proposed a two-step process in which electrons were captured in the LUMO orbitals of the molecules before being emitted into vacuum, with higher-lying LUMO orbitals giving rise to higher secondary electron yields.32 A dependence on packing density was also observed, with denser layers producing higher electron yields. To test whether such a correlation was also valid for the monolayers investigated in the present study, we performed density functional calculations of the adiabatic electron affinities of saturated TFAAD (i.e., TFAAD with two additional H atoms on the alkene group) and 1-dodecane; these are similar to the (50) Evans, D.; Wampler, R. J. Phys. Chem. B 1999, 103, 4666. (51) Haran, A.; Waldeck, D. H.; Naaman, R.; Moons, E.; Cahen, D. Science 1994, 263, 948. (52) Kadyshevitch, A.; Ananthavel, S. P.; Naaman, R. J. Chem. Phys. 1997, 107, 1288. (53) Kadyshevitch, A.; Naaman, R. Thin Solid Films 1996, 288, 139.

11630 Langmuir, Vol. 23, No. 23, 2007

chemical state of TFAAD and 1-dodecene after grafting to the diamond surface. These calculations used Becke’s three-parameter hybrid exchange functional54 with the Lee, Yang, and Parr’s correlation functional55 (B3LYP). We used the Dunning/Huzinaga D95 full double-ζ basis set56 and performed geometry optimization and vibrational energy calculations with it. These calculations show that 1-dodecane and saturated TFAAD have very similar ionization potentials (8.82 and 8.99 eV, respectively), whereas their electron affinities are very different: the electron acceptor level lies 0.05 eV below the vacuum level for TFAAD, and 4.19 eV above the vacuum level for 1-dodecane. The contrast trend we observe, with the molecule with the higher-energy LUMO orbital (1-dodecene) giving rise to more intense secondary electron emission, is consistent with that reported by Saito, et al.32 However, other factors such as the molecular packing density may also affect the contrast. Determining the molecular density of organic molecules on diamond is difficult since both are predominantly carbon and has not been attempted here. Further experiments will be required to fully understanding the factors controlling the image contrast, but the experiments reported here demonstrate that SEM can be used to image hydrocarbon monolayers even when there is little or no “elemental” contrast between the substrate and the organic molecules.

5. Conclusion Our results demonstrate that nanocrystalline diamond surfaces can be photopatterned with organic alkenes down to length scales of 1 µm using a simple contact mask, and that this resolution is limited by diffraction effects associated with the use of a simple contact mask immersed directly in the reactant alkenes. The relatively high spatial resolution obtained demonstrates that while reactive species are formed in the liquid alkenes and in the (54) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (55) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (56) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry; Schaefer, H. F., III, Ed.; Plenum: New York, 1976; Vol. 3, p 1.

Wang et al.

nanocrystalline diamond by UV excitation, the diffusion of these species is limited in extent and therefore does not degrade the spatial resolution of the molecular layers. In contrast, the much longer transition region observed on single-crystal diamond samples demonstrates that diffusion of electrons or holes in the diamond bulk also plays an important role in the molecular patterning process. Our work also demonstrates that the grafted molecular monolayers can be directly imaged using SEM due to spatial variations in the emission of low-energy secondary electrons and that the contrast between molecules correlates with the expected differences in electron affinity of the molecules. The observation of significant contrast from simple molecules such as 1-dodecene suggests that factors such as interfacial electron scattering maybe important for understanding electron transmission through molecular systems. The ability to sequentially pattern two distinct molecules (here, TFAAD and 1-dodecene), while retaining a sharp boundary between them implies that it should be possible to create even more complex patterns of different chemical functional groups by sequential grafting. Overall, the ability to prepare well-defined molecular patterns and arrays on micron length scales via simple masking methods provides a simple way to take advantage of the very high stability of diamond to prepare complex molecular structures with excellent chemical and thermal stability. Acknowledgment. This material is based upon work supported by the National Science Foundation under Grant No. CHE0613010 and DMR-0425880. J.E.B. would like to acknowledge the support of the Office of Naval Research/Naval Research Laboratory. Supporting Information Available: High-resolution XPS spectra of molecular layers. This information is available free of charge via the Internet at http://pubs.acs.org. LA701803G