Article pubs.acs.org/Langmuir
Atomic Force Microscopy Characterization and Lithography of CuLigated Mercaptoalkanoic Acid “Molecular Ruler” Multilayers Chad I. Drexler, Kevin B. Moore, III, Corey P. Causey,* and Thomas J. Mullen* Department of Chemistry, University of North Florida, Jacksonville, Florida 32224, United States ABSTRACT: Hybrid chemical patterning strategies that combine the sophistication of lithography with the intrinsic precision of molecular self-assembly are of broad interest for applications including nanoelectronics and bioactive surfaces. This approach is exemplified by the molecular-ruler process where the sequential deposition of mercaptoalkanoic acid molecules and coordinated metal ions is integrated with conventional lithographic techniques to fabricate registered, nanometer-scale spacings. Herein, we illustrate the capabilities of atomic force microscopy characterization and lithography to investigate the morphology, quality, and local thickness of Culigated mercaptohexadecanoic acid multilayers on Au{111} substrates. These multilayers are a key component utilized in the molecular-ruler process. The rich and varied topographic features of each layer are investigated via contact-mode atomic force microscopy. Using nanoshaving, an atomic force microscopy lithographic strategy that reveals the underlying Au{111} substrate via tip-induced desorption of a molecular film, the local thicknesses of these multilayers are ascertained; these thicknesses are consistent with the anticipated heights for Cu-ligated mercaptohexadecanoic acid multilayers as well as previous ensemble surface analytical measurements. By regulating the force set point utilized during nanoshaving, the upper layer of a Cu-ligated mercaptohexadecanoic acid bilayer is removed, revealing the carboxyl moiety of the lower mercaptohexadecanoic acid layer. This selective nanoshaving demonstrates a simple and practical means to generate three-dimensional multilayers and to reveal buried chemical functionalities within metal-ligated multilayers.
■
INTRODUCTION The study of self-assembled monolayers (SAMs) of alkanethiolate molecules adsorbed on noble metal surfaces has garnered tremendous interest in recent years.1 Among the reasons for the interest in these systems are the rich surface features and wide variety surface functionalities that can be imparted by this molecular architecture.2,3 Numerous chemical patterning strategies have been developed to arrange and to register alkanethiolate SAMs across metal substrates. These techniques have been employed for applications including molecular resists for lithography and templates for directing the assembly and growth of metals, polymers, and biomolecules.4−13 Of these strategies, those that combine conventional top-down lithographic techniques to define the large-scale, registered features with molecular self-assembly to control the physical and chemical interactions of the small features show significant promise.14−17 Such hybrid lithographic strategies are of particular interest because they couple a key aspect afforded by conventional lithography, the ability to create complex architectures over large areas, to the precision and resolution of molecular self-assembly. The coupling of conventional patterning methods with molecular self-assembly is exemplified by the molecular-ruler process.18−29 Employing this methodology, nanogaps between registered noble metal surface features are generated across a Si substrate using conventional lithographic techniques, such as © 2014 American Chemical Society
photolithography or electron-beam lithography. The multilayer utilized in the molecular-ruler process is formed by depositing sequential alternating layers of mercaptoalkanoic acid molecules and cupric (Cu2+) ions across the sample. The formation of the Cu-ligated mercaptoalkanoic acid multilayers only occurs on the noble metal surface and not on the underlying Si substrate. The overall thickness of the multilayer is governed by the number of treatments with Cu2+ ions and mercaptoalkanoic acid molecules. Once the desired thickness of the Cu-ligated mercaptoalkanoic acid multilayer is achieved, a second metal is deposited across the entire substrate, including the exposed Si regions and atop the multilayer. Using chemical lift-off, the multilayer and the second metal atop of the multilayer are removed, while the metal deposited on the Si substrate remains. This process creates a tailored nanogap ranging from 4 to 100 nm between two metal structures; the size of the nanogap is defined by the thickness of the Cu-ligated mercaptoalkanoic acid multilayer. Therefore, the molecular-ruler process provides a general and widely applicable methodology to fabricate registered, nanometer-scale features that have potential applications in areas such as nanoelectronics and bioactive surfaces/sensors. Received: April 28, 2014 Revised: May 28, 2014 Published: June 4, 2014 7447
dx.doi.org/10.1021/la501645w | Langmuir 2014, 30, 7447−7455
Langmuir
Article
information outside of film thicknesses (elipsometry) and surface energy (contact angle goniometry) can be ascertained for multilayers beyond the third layer.30−32 Local AFM imaging has only been utilized to study topography, surface coverages, and domain structures for the upper layer of Cu-ligated mercaptoalkanoic acid bilayers.32,33 Given the limitations of current surface analytical techniques, it is imperative that additional methods and strategies be developed in order to gain further insights into the assembly and structures of Cu-ligated mercaptoalkanoic acid multilayers. In this work, we report the characterization of mercaptohexadecanoic (MHDA) acid multilayers, composed of up to four layers MHDA molecules ligated with Cu+ ions, and highlight the varied and rich local features of each layer. These multilayers are a key component utilized in the molecularruler process. Utilizing nanoshaving, a form of atomic force lithography that revels the underlying Au{111} substrate via tip-induced desorption of a chemical film, the local thicknesses of these multilayers are investigated and verified. Additionally, the upper layer of a Cu-ligated MHDA bilayer is removed by regulating the force set point utilized during nanoshaving. This selective nanoshaving reveals the underlying chemical functionalities and demonstrates a possible means to generate threedimensional multilayers with varied surface chemistries.
The quality of the Cu-ligated mercaptoalkanoic acid multilayers used in the molecular-ruler process is an important characteristic in generating reproducible and precise nanometer-scale surface structures. Given the versatility of the molecular-ruler process for generating complex and registered nanoscale features over large areas, and given the general versatility of chemical multilayers, significant efforts have been made to better understand the assembly process and molecular structure of Cu-ligated mercaptoalkanoic acid multilayers. Initial elipsometry studies of Cu-ligated mercaptoalkanoic acid multilayers by Ulman and co-workers showed a linear relationship between the number of layers within the multilayer film and overall film thickness.30 However, few details related to the interlayer binding chemistry and molecular packing within the various layers of the Cu-ligated mercaptoalkanoic acid multilayers were described. Bard and co-workers demonstrated that the Cu2+ ions remain in their 2+ oxidation state upon exposure to and ligation with the carboxyl moieties of a mercaptoalkanoic acid monolayer using both X-ray photoelectron spectroscopy (XPS) and fluorescence spectroscopy.31 However, they observed that the Cu2+ ions are reduced to Cu+ (cuprous) ions upon exposure and chemisorption of subsequent mercaptoalkanoic acid molecules. A more recent study by Allara and co-workers demonstrated a 1:1 stoichiometry between the upper layer mercaptoalkanoic acid and Cu+ ions in a Cu-ligated mercaptoalkanoic acid bilayer resulting in a ∼50% coverage of the upper mercaptoalkanoic acid layer relative to the lower mercaptoalkanoic acid layer.32 When analyzed by atomic force microscopy (AFM) imaging and infrared spectroscopy (IRS) studies, the upper mercaptoalkanoic acid layer was distributed as discrete islands with conformational ordering indicating chain aggregation. Building upon these investigations, Wälti and co-workers examined the assembly, quality, and morphology of Cu-ligated mercaptoalkanoic acid bilayers while systematically varying the number of methylene groups of the mercaptoalkanoic acid molecules.33 When the upper mercaptoalkanoic acid layers were assembled with relatively short mercaptoalkanoic acid molecules (n = 6), incomplete upper layers composed of disordered molecules, with surface coverages of ∼6%, were observed across the underlying monolayer. In contrast, the upper layers of Cu-ligated mercaptoalkanoic acid bilayers tended to aggregate into discrete islands with surface coverages approaching 50% when the upper mercaptoalkanoic acid layers were assembled with relatively long mercaptoalkanoic acid molecules (n = 11 and n = 16). The island formation and increased surface coverage of these multilayers are governed solely by the molecules that comprise the upper layer of the bilayers and result from the gained energy associated with maximizing the intermolecular interactions of longer mercaptoalkanoic acid molecules. Recently, Wälti and co-workers observed structural reorganizations of Cu-ligated mercaptoalkanoic acid multilayers upon ligation of Cu2+ ion as well as during chemisorption of additionally mercaptoalkanoic acid molecules, indicating that these multilayers are dynamic and complex chemical systems.34 While the combination of ensemble surface analytical techniques (e.g., elipsometry, XPS, IRS, etc.), local AFM imaging, and theoretical simulations can provide valuable insights into the assembly process and structures of Cu-ligated mercaptoalkanoic acid multilayers used in the molecular-ruler process, these methodologies are not without substantial limitations. For instance, XPS and IRS are typically limited to the first two to three layers of the multilayer, and little
■
MATERIALS AND METHODS
Materials and Reagents. Mercaptohexadecanoic acid (90%), copper(II) perchlorate hexahydrate (Cu(ClO4)2·6H2O, 98%), octadecyltrichlorosilane (OTS, >90%), and acetic acid (>99%) were purchased from Sigma-Aldrich (St. Louis, MO). Chloroform (HPLC grade) was purchased from J.T. Baker (Center Valley, PA), and hexadecane (>98%) was purchased from Fisher Chemical (Pittsburgh, PA). Absolute ethanol was purchased from Pharmco-Aaper (Brookfield, CT). All reagents were used as received. 18 MΩ water was generated using a Milli-Q system (Q-GARD 2, Millipore, Billerica, MA). Preparation of MHDA SAMs. All MHDA SAMs were fabricated on commercially available Au{111} substrates evaporated onto freshly cleaved mica substrates (Agilent Technologies, Tempe, AZ). The Au{111} substrates were annealed using a hydrogen flame for 30 s prior to immersion into base-bath-cleaned glass V-vials filled with 0.01 mM ethanolic solutions of MHDA with 10 vol % acetic acid for 24 h. Acetic acid suppressed the dimerization and agglomeration of the MHDA molecules by competing for hydrogen-bonding interactions.35−37 Subsequently, the Au{111} substrates with the MHDA SAMs were rinsed with absolute ethanol and dried under a stream of N2. All MHDA SAMs were imaged or utilized to generate multilayers immediately after preparation. Preparation of Cu-Ligated MHDA Multilayers. Cu-ligated MHDA multilayers were produced by a sequential-alternating-layer assembly strategy outlined in Figure 1. Freshly prepared MHDA SAMs on Au{111} substrates were immersed into 5 mM Cu(ClO4)2·6H2O ethanolic solutions for 3 min. Upon removal from the Cu(ClO4)2· 6H2O ethanolic solutions, the Au{111} substrates were rinsed with absolute ethanol and dried under a stream of N2. Subsequently, the Au{111} substrates were immersed into 1 mM MHDA ethanolic solutions for 1 h. Upon removal from the MHDA ethanolic solutions, the Au{111} substrates were rinsed with absolute ethanol and dried under a stream of N2. This sequence of immersion into Cu(ClO4)2· 6H2O and MHDA solutions was repeated until the number of desired layers was achieved. It is important to note that only the initial MHDA monolayer in a Cu-ligated MHDA multilayer was assembled from a 0.01 mM ethanolic solution of MHDA with 10 vol % acetic acid; the subsequent layers of the multilayers were assembled from 0.01 mM ethanolic solutions of MHDA without acetic acid. We observed that a 0.01 mM ethanolic solution of MHDA with 10 vol % acetic acid 7448
dx.doi.org/10.1021/la501645w | Langmuir 2014, 30, 7447−7455
Langmuir
Article
Collider”), which is an open-source software freely available on the Internet and supported by the Czech Metrology Institute. Nanoshaving. Nanoshaving is an in situ AFM-based lithographic technique developed by Liu and co-workers to expose underlying Au regions.43,44 Chemical patterns are generated by moving an AFM tip in predetermined track at a high force set point (40−100 nN) to induce desorption within a chemical film. Figure 2 outlines the use of
Figure 1. Schematic of the assembly of the Cu-ligated MHDA multilayers. (A) A MHDA SAM is fabricated on a Au{111} substrate via immersion into a 0.01 mM ethanolic solution of MHDA with 10 vol % acetic acid for 24 h, rinsed with absolute ethanol, and dried with N2. (B) Au{111} substrate with the MHDA SAM is subsequently immersed into a 5 mM Cu(ClO4)2·6H2O solution for 3 min, rinsed with absolute ethanol, and dried with N2. The Cu2+ ions chelate with carboxyl moieties of the MHDA molecules. (C) Au{111} substrate is then immersed into a 1 mM ethanolic solution of MHDA for 1 h. The Cu2+ ions are reduced to Cu+ upon chemisorption of the MHDA molecules.31,32 This sequential-alternating-layer assembly strategy is repeated until the number desired layers is achieved.
Figure 2. Schematic of the selective nanoshaving Cu-ligated MHDA bilayers. Using a force set point of less than 1 nN (small red arrow) under absolute ethanol, a region of a Cu-ligated MHDA bilayer on a Au{111} substrate is selected for nanoshaving. To expose the underlying Au{111} substrate and to remove the Cu-ligated MHDA bilayer, a high force set point (40−100 nN, wide green arrow) is applied during nanoshaving. To expose the lower MHDA layer of the Cu-ligated MHDA bilayer and reveal the buried carboxyl moiety, a low force set point (10−30 nN, narrow blue arrow) is applied during nanoshaving. Once patterning is complete, a force set point of less than 1 nN (small red arrow) is utilized to image resulting features.
disrupted the assembly of the Cu-ligated MHDA multilayers and typically resulted in a MHDA monolayer. Atomic Force Microscopy. Contact-mode AFM images were acquired using an Agilent 5420 scanning probe microscope with sharpened Si3N4 cantilevers (DNP-S, Bruker AFM Probes, Santa Barbara, CA) with nominal force constants of 0.35 N/m. Force constants for individual cantilevers were measured via the thermal noise method and ranged from 0.25 to 0.50 N/m.38 The Si3N4 cantilevers were functionalized with an OTS monolayer to prevent adsorption of MHDA molecules during imaging and nanoshaving.39−42 Prior to the OTS monolayer formation, the Si3N4 cantilevers were cleaned using an UV ozone cleaner (Novascan, PSDP-UVT, Ames, IA) for 45 min to remove surface contamination. The UVozone-cleaned Si3N4 cantilevers were then immersed into a 1 mM OTS solution of a 7:3 mixture of hexadecane and chloroform for 1 h and subsequently rinsed with chloroform and dried under a stream of N2. Imaging force set points below 1 nN were utilized to minimize damage to and disruption of the MHDA monolayers and Cu-ligated MHDA multilayers, and scan rates were set to 1 Hz to maximize topographic tracking. All AFM images were acquired at 256 points per line under absolute ethanol to minimize surface contamination and to promote nanoshaving. Image processing and analysis of the AFM images were performed using Gwyddion (version 2.36, “Casual
nanoshaving to remove the various layers of a Cu-ligated MHDA bilayer, selectively. It is important to note that this strategy is not limited to Cu-ligated MHDA bilayers and has been applied to Culigated MHDA multilayers of varying thicknesses. Initially, a Au{111} substrate with a Cu-ligated MHDA bilayer was imaged at a force set point of less than 1 nN under absolute ethanol to characterize the substrate and to locate an appropriate region for patterning.35,45 Using a high shaving force set point (40−100 nN depending on the sharpness of the AFM tip), the underlying Au{111} substrate was exposed due to the tip-induced desorption of the Cu-ligated MHDA bilayer. Using a low shaving force set point (10−30 nN depending on the sharpness of the AFM tip), the Cu+ ions and upper MHDA layer were removed, exposing the lower MHDA monolayer. Subsequently, the patterned region was imaged at a force set point of less than 1 nN to characterize the resulting features. The force set points for imaging and nanoshaving were quantified by using force−distance curves before each image and nanoshaving step. The high and low force set points for nanoshaving were determined by systematically increasing 7449
dx.doi.org/10.1021/la501645w | Langmuir 2014, 30, 7447−7455
Langmuir
Article
the force set point until selective desorption was observed.35,46 All nanoshaved regions were generated at 512 points per line and under absolute ethanol to minimize surface contamination and to promote nanoshaving. Determining the Apparent Heights and RMS Roughnesses of MHDA Monolayers and Cu-Ligated MHDA Multilayers. The apparent heights of the preexisting MHDA monolayers and Cu-ligated MHDA multilayers were determined from at least three cursor profiles across the preexisting MHDA monolayer or Cu-ligated MHDA multilayer and underlying Au{111} substrate exposed during nanoshaving for two AFM images of each chemical film. To account for incomplete multilayer formation, these measurements were calculated from the top of the preexisting monolayer/multilayer to the nanoshaved region on the same Au{111} terrace. To enable comparison between the various nanoshaved regions, the RMS roughnesses of the preexisting MHDA monolayer and Cu-ligated MHDA multilayers as well as underlying Au{111} substrate exposed during nanoshaving were calculated using a 50 nm × 50 nm square on the same Au{111} terrace in at least three locations for two AFM images of each preexisting chemical film and subsequent nanoshaved region.
■
RESULTS AND DISCUSSION AFM Characterization of MHDA Monolayers and CuLigated MHDA Multilayers. The morphology and structure of MHDA monolayers and Cu-ligated MHDA multilayers fabricated across Au{111} substrates were investigated using contact-mode AFM. Figure 3A shows a representative AFM topographic image of a 1 μm × 1 μm region of a MHDA monolayer fabricated from a 0.01 mM ethanolic solution with 10 vol % acetic acid, and Figure 3E shows a cursor profile across the monolayer as indicated by the black line in Figure 3A. The surface morphology of the MHDA SAM is consistent with a highly ordered SAM as evidenced by a RMS roughness of 0.11 ± 0.01 nm and the observation of Au terraces.33,35,37,47 Isolated and protruding features with apparent heights of 1.89 ± 0.13 nm are observed across the substrate surface. These features are attributed to the presence of dimerized MHDA molecules by hydrogen bonding of solute MHDA molecules to the carboxylate moiety of the MHDA SAM.32,35,37 Figure 3B shows a representative AFM topographic image of a 1 μm × 1 μm region of a Cu-ligated MHDA bilayer, and Figure 3E shows a cursor profile across the bilayer as indicated by the green line in Figure 3B. Protruding domains are observed across the surface; these features are consistent with previous AFM topographic images of Cu-ligated MHDA bilayers.32,33 This surface morphology results in an increase in the RMS roughness to 0.97 ± 0.11 nm. The apparent height of these domains is 2.20 ± 0.15 nm, which is characteristic of an incomplete second layer of MHDA. In addition to this incomplete second layer, isolated and more protruding features are observed across the surface ranging in apparent heights from 2 to 6 nm. These features are attributed to the dimerization and agglomeration of MHDA molecules. The absence of acetic acid during the multilayer assembly helps account for the appearance of the more protruding dimerized and agglomerated features on the Cu-ligated MHDA bilayer as well as subsequent Cu-ligated multilayers when compared to the MHDA monolayers. Figures 3C and 3D show representative AFM topographic images of 1 μm × 1 μm regions of a Cu-ligated MHDA trilayer and tetralayer, respectively. Figure 3E show cursor profiles across the Cu-ligated MHDA trilayer and tetralayer as indicated by the red and blue lines, respectively. The morphologies of these multilayers appear to be more complete as indicated by
Figure 3. Comparison of the morphology and local structure of a MHDA monolayer and Cu-ligated MHDA multilayers. Representative AFM topographic images of (A) a MHDA monolayer, (B) a Culigated MHDA bilayer, (C) a Cu-ligated MHDA trilayer, and (D) a Cu-ligated MHDA tetralayer fabricated on Au{111} substrates. (E) Corresponding cursor profiles across the substrates as indicated in the AFM images. All AFM images were acquired in contact mode with force set points of less than 1 nN and scan rates of 1 Hz.
fewer isolated domains compared to the Cu-ligated MHDA bilayers and result in a relative decrease in the RMS roughnesses to 0.48 ± 0.07 and 0.74 ± 0.08 nm for the Culigated MHDA trilayer and tetralayer, respectively. These more complete morphologies may be a result of the increased intermolecular interactions associated with the increase in the number of MHDA molecules within the Cu-ligated MHDA trilayers and tetralayers, which is consistent with the increased surface coverages observed by Wälti and co-workers for Culigated MHDA bilayers composed of long mercaptoalkanoic molecules.33 Protruding features, with apparent heights ranging from 2 to 6 nm, are also observed on these multilayers and are attributed to the dimerized and aggregated MHDA molecules as previously discussed (vide supra). Depressed features, which are attributed to incomplete multilayer formation, are observed for both the Cu-ligated MHDA trilayer and tetralayer. For the Cu-ligated MHDA trilayer, the apparent height for the depressed defects is 2.09 ± 0.10 nm, corresponding to a single-layer defect. For the Cu-ligated MHDA tetralayer, the apparent height for the depressed defects is 4.07 ± 0.15 nm, corresponding to a two-layer defect. This observation suggests that the multilayer film continues to grow across a large 7450
dx.doi.org/10.1021/la501645w | Langmuir 2014, 30, 7447−7455
Langmuir
Article
percentage (>90%) of the substrate, but there is a small portion (
