Article pubs.acs.org/Langmuir
Adsorption of a Carboxylic Acid-Functionalized Aminoxyl Radical onto SiO2 Hidenori Murata,† Martha Baskett,† Hiroyuki Nishide,‡ and Paul M. Lahti*,† †
Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003 United States Department of Applied Chemistry, Waseda University, Tokyo 169-855, Japan
‡
S Supporting Information *
ABSTRACT: Silicon wafers both without and with silicon(IV) oxide surface coverage were covered with benzene solutions of stable organic radical 3-(N-tert-butyl-N-aminoxyl)benzoic acid (mNBA). X-ray photoelectron spectroscopy supported the presence of the radical on both surface-cleaned (oxide-reduced) and oxide-covered surfaces. Optical waveguide spectroscopy showed that the radical retained its structure while adsorbed to the surface of the wafers, without noticeable decomposition. AFM and MFM imaging showed that the radical formed blocky particles with a change in rms roughness from 0.3 nm premodification to 1.7 nm postmodification on the surface-cleaned silicon. Similar experiments using oxide-coated silicon showed that the radical adsorbed to form much smoother layers, with a small change in rms roughness from 0.2 to 0.3 nm. Contact angle measurements of water on the premodified and postmodified samples showed a large, hydrophobic change in the silicon oxide surface but only a modest change in the surface-cleaned silicon surface. Samples of mNBA adsorbed onto silica gel showed strong electron-spin resonance signals from the aminoxyl spin, even years after production. The results demonstrate the prospects for treating and coating oxide-covered silicon wafers and silicon oxide-coated particles with a paramagnetically active organic substrate, without major chemical modification of the pretreatment surface; the resulting organic spin sites can be stable for years.
■
INTRODUCTION Recent interest in designed molecular magnetic materials has included efforts to deposit, adsorb, or attach them to surfaces in a manner that does not destroy or substantially alter their unpaired spins or other magnetic properties, as noted in a recent review by Mas-Torrent et al.1 Simple spin-probe-type aminoxyls such as TEMPO and close analogues adsorb onto metal, alumina, silica, and zeolite surfaces.2−10 Stable radicals with pendant long-chain thioalkyl groups can adsorb onto gold with more specificity.11 More recently, single-molecule magnets (SMMs) have also adsorbed onto a number of different types of surfaces.12−17 Persistent organic polymeric polyradicals can even be deposited onto surfaces and imaged with magnetic force microscopy.18,19 A few studies of self-assembled monolayers with radicals have also been reported,20,21 including a recent study by Crivillers et al.22 who self-assembled monolayers of polychlorotriphenylmethyl radicals onto aminefunctionalized SiO2. Work has also been reported describing the Langmuir−Blodgett23 or vapor deposition24−26 of nitronylnitroxides onto surfaces, with an eye to building up layers and testing interradical and packing magnetic behavior under these conditions. Quite recently, Casu and co-workers reported27 adsorbing 2-pyrenylnitronylnitroxide (PyrNN) onto rutile TiO2, finding that the PyrNN either retained or lost its radical spin integrity depending on the local substrate nature (with the initial deposition layer possibly serving to protect subsequent layers). © 2014 American Chemical Society
One important strategy for adsorbing magnetically active molecules takes advantage of hydrogen bonding interactions at the surface.28,29 In 2005, Baskett and Lahti reported that stable hydrogen-bonding radical 3-(N-tert-butyl-N-aminoxyl)benzoic acid (mNBA, Scheme 1) exhibits a strong affinity for silica Scheme 1. Structure of 3-(N-tert-Butyl-N-aminoxyl)benzoic Acid (mNBA)
gel,30 raising the possibility that it could form well-behaved layers on hydroxylated SiO2 surfaces. Given the ubiquity of SiO2 in many uses, especially as a surface layer on silicon, it seemed reasonable to establish the ability of mNBA to adhere selectively to SiO2 and to determine whether the adsorbed mNBA is persistent under these conditions. In this article, we report that mNBA is readily adsorbed onto oxide-coated silicon wafers from benzene solutions to give smooth layers of radicals that are persistent for years and are readily detectable by Received: January 8, 2014 Revised: February 28, 2014 Published: March 31, 2014 4026
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
■
RESULTS AND DISCUSSION mNBA has a strong affinity to adsorb onto silica and related silicon oxide surfaces, especially from benzene solutions. It adsorbs so strongly to the inside of a standard Pyrex roundbottomed flask that attempts to scrape out mNBA remnants with a metal spatula cracked the flask first; this problem was solved by rinsing the inside of glassware with dichlorodimethylsilane before adding NBA solution. This latter behavior is consistent with the need for free surface OH groups to enable the strong adsorption. The addition of 200 mesh (74 μm maximum size) silica gel particles to freshly made solutions of mNBA in benzene essentially removes the UV−vis band associated with the radical from the solution (reddish color for mNBA in solution, a broad solution absorption at 400−450 nm). The adsorbed radical imparts a yellow color to the silica that remains for months to years without apparent visual change. Although mNBA solutions lose electron spin resonance (ESR) activity within about a week of exposure to air at ambient temperature, silica gel modified by exposure to mNBA in benzene still shows readily seen ESR signals from the aminoxyl radical spins after years under ambient conditions. Figure 1 shows ESR spectra obtained at different temperatures
spectroscopic methods as well as by atomic and magnetic force microscopy.
■
Article
EXPERIMENTAL PROCEDURES
Materials. mNBA was synthesized by the procedure of Baskett and Lahti.30 Benzene was freshly distilled from sodium prior to use. Ultrapure water was refined using a Millipore Mill-Q water-purification system. Other reagent chemicals were commercially obtained and used as received. Silicon wafers were obtained from KST World Corp (10 mm × 10 mm × 525 μm with about 200 nm of silicon oxide film on the surface according to commercial specifications). The wafers were soaked in a 4:1(v/v) sulfuric acid/hydrogen peroxide solution at 70 °C for 10 min to remove organic contaminants and metal particles on the surface and then rinsed with ultrapure water. To reduce the oxide coating for some experiments, wafers were soaked in 1% HF(aq) solution for 1 min and rinsed with ultrapure water and then used at once. The surface-cleaned silicon showed 6% residual O 1s in X-ray photoelectron spectra compared to the integrated peak intensity of the untreated, oxidecoated wafer; this may have been present as a result of handling in air in a typical laboratory environment after cleaning. Radical Deposition on SiO2-Covered Silicon. An oxide-coated silicon wafer was immersed in a 1.0 mM solution of mNBA in dry benzene for 10 min at room temperature and then rinsed with dry benzene. The wafer was then dried under a flow of nitrogen. Radical Deposition on Surface-Cleaned Silicon. The same procedure was used as for radical deposition on SiO2, except that the silicon wafer was freshly cleaned with HF(aq) solution before immersion in the benzene solution of the radical. Characterization. Static water drop contact angles were measured on surface-cleaned and oxide-coated silicon wafers both before and after radical deposition (Supporting Information Figure S1). X-ray photoelectron spectra (XPS) were obtained within a week of sample preparation using a JEOL JPS-9010 TR instrument equipped with an Al Kα source, a monochromator, a concentric hemispherical analyzer, and a multichannel detector. A takeoff angle of 15° from the surface was used. Spectral regions for O 1s (525−540 eV binding energy), N 1s (395−405 eV binding energy), and C 1s (265−290 eV binding energy) were recorded for all samples (Supporting Information Figures S2−S4). Binding energies were calibrated31 at 284.8 eV for the main C 1s peak of adventitious carbon on both surface-cleaned and oxide-coated silicon wafers. XPS line-shape fitting by a sum of Gaussian and Lorentzian forms yielded individual peak contributions that are summarized in Table 2. (See also Supporting Information Figures S3 and S4.) Elemental N/C composition ratios were determined by integration under the XPS peak envelopes. Force microscopy experiments were performed using a Nanoscope IIIa Multimode AFM/MFM instrument (Digital Instruments, Inc.). A Digital Instruments single-crystal silicon cantilever for AFM was used for NCH; tip height, 10−15 μm; radius curvature at front, ca. 5 nm; length, 125 μm; width, 30 μm; thickness, 3−5 μm; and a spring constant between 20 and 100 N m−1. Resonance peaks in the frequency response of the cantilever were selected in the range between 280 and 320 kHz for the tapping-mode oscillation. Electron spin resonance (ESR) spectra were obtained using a Bruker Elexsys E500 spectrometer at X-band frequencies (9.4−9.6 GHz). Samples of mNBA adsorbed onto silica gel were placed in 4 mm o.d. quartz tubes and studied first within 24 h of adsorption and then studied after extended times of being stored under ambient conditions in a clear, colorless glass screw-top vial inside a laboratory bench drawer. Variable-temperature ESR spectra were obtained using a Bruker ER-4131 cold nitrogen gas flow unit. Surface absorption spectra of mNBA were studied using a System Instruments Inc. SIS-50 optical waveguide (OWG) spectrophotometer. A quartz plate of thickness 0.2 mm and refractive index 1.46 was used as an optical waveguide. The angle of incidence of the light was 79.2° from the normal of the surface.
Figure 1. ESR spectra of mNBA adsorbed onto silica gel from benzene at various temperatures; ν0 = 9.373 GHz. Spectra are offset for ease of viewing.
on a sample of mNBA absorbed from benzene onto silica gel. The line shape reversibly changes when cooling from room temperature to below 240 K but shows no other significant qualitative changes. The line-shape changes are consistent with changes in the conformational mobility of the aminoxyl unit on the surface. There was no ESR-based evidence, such as half-field transitions or additional features in the g ≈ 2 region, that would indicate the formation of triplet states from dimerized spins resulting from the close interaction between mNBA radicals. With a modest loss of spin intensity, the same ESR spectrum was found in the same set of samples after up to 3 years of storage under ambient conditions. 4027
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
interacting at the surface. If wide-coverage adsorption occurred on the surface-cleaned silicon, then the contact angle should be similar after treating the radical-treated, oxide-coated surface. Instead, the surface-cleaned silicon becomes just slightly more hydrophobic. X-ray photoelectron spectra (XPS) of both surface-cleaned and SiO2-coated silicon wafers were also obtained before and after modification by “dip-and-wash” treatment with benzene solutions of mNBA. Comparisons of pre- and post-treatment results are summarized in Figure 2 and Table 2. Both surfacecleaned and oxide-coated silicon wafers show adventitious carbon peaks; some spectra showed a small C 1s shoulder at about 282.5 eV that corresponds32 to Si−C. Both silicon wafers showed mNBA deposition after modification, especially on the basis of the formation of N 1s peaks for the mNBA-modified samples; there are no N 1s peaks before modification. There is also an increase in the C 1s peak for both surface-cleaned and oxide-coated silicon wafers after mNBA modification. Although the mNBA-modified, surface-cleaned sample shows a strong increase in the O 1s peak intensity (Supporting Information Figure S2), the oxide-coated silicon sample shows a decrease in the O 1s peak intensity after modification. This is presumably due to the effective coverage of the SiO2 surface with the radical, which has only three oxygen atoms among multiple carbons and a nitrogen atom. The main peak of the N 1s region is located at about 400 eV, in good agreement with reports for other aminoyxl and related N−O-containing systems.33 The oxide-coated silicon sample has higher-energy N 1s shoulders at about 401.5 and 402.6 eV, possibly as a result of aminoxyl moieties that are hydrogen bonded to the oxide surface34 as suggested in Scheme 1. No prominent shoulders are present in the mNBA-modified, surface-cleaned sample; this fits the expectation that there is little surface oxide to bind the aminoxyl group. The integrated ratios of the N 1s/C 1s peak region intensities for the surfacecleaned and oxide-coated mNBA-modified silicon wafers were N/C = 0.115 and 0.102, respectively: for mNBA, the theoretical elemental ratio is N/C = 0.091. The postmodified, oxide-coated silicon sample C 1s XPS peak shows multiple features that are not present in the adventitious carbon XPS of the premodified sample. The main carbon peak envelope at 285 eV corresponds to expectations for aryl C 1s overlapped with higher-energy aliphatic tert-butyl C 1s transitions (Table 2, Figure 2). A higher-energy peak at 286.5 eV probably arises from the two carbons bonded to the N−O group, and the 288.2 eV peak fits the expectations33 for the carboxylic acid carbon. Peak fitting (Supporting Information, Figures S3 and S4) for the C 1s region is in reasonably good agreement with this progression of overlapping peak contributions. Analogous C 1s XPS peaks with similar intensities are seen in the postmodified, surfacecleaned silicon sample after mNBA modification. Although the XPS results show mNBA deposition onto both surface-cleaned and oxide-coated silicon, it does not show whether the radical adsorbs relatively smoothly onto either surface or just forms accretions that adhere despite washing with nonpolar benzene. The water contact angle studies showed a clear difference between the two types of modified surfaces. Atomic force microscopy (AFM) and magnetic force microscopy (MFM) imaging of both types of silicon wafers was therefore carried out before and after mNBA treatment. Figure 3 shows 2-D and 3-D AFM images from the same surfacecleaned silicon wafer used for the XPS measurements given in Figure 2. Figure 4 shows the corresponding results for the
The use of relatively nonpolar benzene as a solvent for the adsorption process is important. mNBA solutions in diethyl ether also enable adsorption onto silica gel, giving the typical characteristic coloration of this aminoxyl radical on the gel particles and a clear aminoxyl ESR spectrum from the particles. However, the ether solution was not decolorized in the manner that benzene was, showing that the radical is partitioned between the solution and silica surface. Attempts to adsorb the radical onto silica gel from methanol solutions yield no coloration of the silica gel; the gel samples from these experiments also show no ESR signal. Presumably, as the solvent polarity increases, its affinity for the strongly hydrogenbonding carboxylic acid also increases, with commensurately decreased competition of adsorption on the silica surface. Thus, one can adsorb mNBA onto silica from benzene, separate and dry the silica, and then extract the majority of the radical with methanol away from the silica back into solution, judging by changes in the silica gel color and the loss of most of the aminoxyl ESR signal. The reversibility of the adsorption process from benzene rules out a strong covalent attachment of mNBA to silica, for example, by the mNBA carboxylic acid carbonyl to the surface OH groups. It seems likely that hydrogen bonds form between the OH groups on the surface of the silica and the carboxylic acid groups of the radicals. An additional interaction as shown in Scheme 2 between the moderately basic Scheme 2. Possible Interactions between mNBA and the Oxide-Coated Silicon Surface
aminoxyl oxygen and the silica OH groups might also occur because of the geometrical matching of the silica OH groups and the meta geometry between the aminoxyl unit and the carboxylic acid. To test the extent of modification of wafers soaked in benzene solutions of mNBA, water drop contact angles were measured before and after treatment (Table 1, Supporting Information Figure S1). They show only a small relative change for surface-cleaned silicon but a large change after mNBA treatment of the oxide-coated surface. The results show significantly hydrophobic modification of the oxide-coated surface, as expected if the radicals were extensively surface adsorbed to have their hydrophilic carboxylic acid groups Table 1. Static Water Drop Contact Angles for Pre- and Post-mNBA Modification of Surface-Cleaned and OxideCoated Silicon Wafers
a
sample
contact angle (θ/deg)a
Si modified Si SiO2 modified SiO2
73 80 35 66
See also Supporting Information Figure S1. 4028
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Figure 2. X-ray photoelectron (XPS) spectral traces for a surface-cleaned silicon wafer (left column) and an oxide-coated silicon wafer (right column) before (broken lines) and after (solid lines) modification by a benzene solution of mNBA. Each individual plot compares pairs of traces having the same ordinate scale for the same wafer, but all ordinate scales are different from one another.
remains comparable to the premodification roughness. Also, the MFM images for the mNBA-treated, surface-cleaned silicon show the magnetic detection of larger features in the AFM image, indicating discrete accretions of the spin-active radical on the surface rather than a coating of the overall surface in a smooth manner. By comparison, both AFM and MFM images for oxide-coated silicon after mNBA treatment are essentially featureless in Figures 4 and 5, consistent with the relatively smooth adsorptive formation of a layer of radical under these conditions. The following permissive evidence for the surface deposition of mNBA has been described to this point: (1) strong ESR aminoxyl radical signals are seen in silica gel modified by soaking in benzene solutions of mNBA, signals that are removed by soaking the silica gel in polar solvents; (2) XPS shows strong N 1s peaks on both surface-cleaned and oxidecoated silicon wafer surfaces after dip-and-wash modification with a benzene solution of mNBA whereas no N 1s peak is seen before modification and only mNBA contains nitrogen in the modification treatment reagents; (3) water contact angles are strongly altered on an oxide-coated silicon surface after modification by an mNBA solution, consistent with surface coverage to make it less hydrophilic; (4) MFM detects features seen in AFM tapping mode for mNBA-modified, surfacecleaned silicon, showing that the mNBA spin units survive as discrete radicals in these features. To support further the survival of mNBA discrete radicals on the oxide-coated silicon wafer, Figure 6 compares a solution mNBA UV−vis absorption spectrum to the wafer’s solid film optical waveguide (OWG) spectrum after mNBA modification. The two spectra correspond quite well, with both showing 300 nm strong absorption peaks, particularly the long-wavelength band at ∼450 nm that is characteristic of the phenyl-aminoxyl moiety. Notably, the mNBA samples adsorbed to the oxide-coated
Table 2. XPS Results for Surface-Cleaned Si and OxideCoated Si, Premodified and Postmodified by Treatment with mNBA sample cleaned Si, premodified cleaned Si, postmodified
oxide-coated Si, premodified
oxide-coated Si, postmodified
C 1s/eV 284.8a (70)b 286.5 (30) [283.2 (1)c] 285.3 (68) 287.0 (21) 288.7 (10) [282.7 (20)c] 284.8a (73) 287.0 (7) [282.5 (5)c] 285.0 (63) 286.7 (19) 288.4 (13)
N 1s/eV
O 1s/eV
nil
532.8
400.3 (86) 402.7 (14)
532.9
nil
532.9
399.9 (80) 401.5 (9) 402.6 (11)
532.9
a
Calibration peak, eV. b(Percent) contribution to peak envelope; see Supporting Information Figures S3 and S4. cSi−C (ref 32).
oxide-coated silicon wafer used for the XPS measurements in Figure 2. Figure 5 compares AFM to MFM images of the same regions for both surface-cleaned and oxide-coated silicon wafers after the deposition of mNBA. The AFM results show that the surface-cleaned silicon surface becomes very rough after radical modification, with numerous large surface features including crystallites 30−100 nm in diameter. By comparison, oxide-coated silicon shows a relatively smooth postmodification surface. Table 3 summarizes rms roughness measurements obtained by AFM for portions of the premodification and postmodification surfaces that are shown in Supporting Information Figures S4 and S5. The roughness of the surface-cleaned silicon wafer increases more than 6-fold after modification, but the oxide-coated surface 4029
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Figure 3. AFM images of a surface-cleaned silicon wafer before and after modification with mNBA. (a) 2-D, premodification, 20 nm height; (b) 2-D, postmodification, 20 nm height; (c) 3-D, premodification, 20 nm height; (d) 3-D, postmodification, 20 nm height. These profiles were obtained using tapping-mode AFM in air under ambient conditions, with a 1.0 Hz scan rate in all cases. The 2-D plot inset in the upper row shows z-height color coding.
Figure 4. AFM images of an oxide-coated silicon wafer before and after modification with mNBA. (a) 2-D, premodification, 10 nm lift; (b) 2-D, postmodification, 20 nm height; (c) 3-D, premodification, 20 nm height; (d) 3-D, postmodification, 20 nm height. The 2-D plot insets in the upper row show z-height color codings.
silicon surface still showed this characteristic aminoxyl OWG absorbance after more than 2 years under ambient conditions in air. This is consistent with the extended lifetime of mNBA radicals adsorbed onto silica gel, by comparison to the much shorter ESR lifetime for mNBA in toluene solution under ambient conditions (7−10 days).
The combination of XPS, force microscopy, and OWG spectral results shows that mNBA adsorbs onto SiO2 to form persistent films of the radical. The results support a qualitative picture of the adsorption shown in Scheme 3, where grainy crystallites of mNBA form on surface-cleaned silicon after the evaporation of benzene solvent. The surface coverage is not 4030
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Scheme 3. Deposition of mNBA on Surface-Cleaned versus Oxide-Coated Silicon
smooth, even adsorption on the surface of the oxide-coated silicon.
■
CONCLUSIONS Although 3-(N-tert-butyl-N-aminoxyl)benzoic acid is a structurally simple molecule, its strong affinity for SiO2 causes it to be adsorbed as relatively smooth layers on oxide-coated silicon surfaces, which is quite different from its behavior on surfacecleaned silicon. Notably, the adsorbed organic radicals are persistent for years on both particulate and wafer surfaces. This constitutes a simple method for coating an inorganic oxide, presumably having surface coverage with covalently bound OH groups, with an organic paramagnetic radical substrate and maintaining readily detectable paramagnetism of the radical spin sites. This sort of strategy is worth further investigation as part of efforts to achieve the technologically useful adsorption of persistently spin-bearing species as thin films on solid surfaces.
Figure 5. Comparison of force microscopy images of silicon wafers after modification with mNBA. (a) Surface-cleaned silicon, AFM tapping mode, 100 nm z range; (b) surface-cleaned silicon, MFM mode with 5 nm lift, 5° z-range phase; (c) oxide-coated silicon, AFM, 5 nm lift, 10 nm z range; (d) oxide-coated silicon, MFM mode with 5 nm lift, 2° z-range phase.
■
S Supporting Information *
Table 3. Roughness of Pre- and Post-mNBA Modification of Surface-Cleaned and Oxide-Coated Silicon Surfaces
a
sample
rms roughness (nm)
Si modified Si SiO2 modified SiO2
0.3 1.7 0.2 0.3
ASSOCIATED CONTENT
Static contact water droplet images; XPS peak fitting deconvolution for premodification and mNBA-modified silicon wafers (surface-cleaned and oxide-coated); AFM images of premodified and mNBA-modified silicon wafers (surfacecleaned and oxide-coated) showing regions used for surface roughness evaluation. This material is available free of charge via the Internet at http://pubs.acs.org.
a
■
See also Figures S5 and S6 in the Supporting Information.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS M.B. and P.M.L. acknowledge support from the U.S. National Science Foundation (CHE 0415716). REFERENCES
(1) Mas-Torrent, M.; Crivillers, N.; Rovira, C.; Veciana, J. Attaching Persistent Organic Free Radicals to Surfaces: How and Why. Chem. Rev. 2012, 112, 2506−2527. (2) Zhang, Z.; Berg, A.; Levanon, H.; Fessenden, R. W.; Meisel, D. On the Interactions of Free Radicals with Gold Nanoparticles. J. Am. Chem. Soc. 2003, 125, 7959−7963. (3) Gutjahr, M.; Boettcher, R.; Poeppl, A. Characterization of the Ditert-butyl Nitroxide: Li+ Adsorption Complex in LiY Zeolites by Oneand Two-Dimensional Electron Spin-Echo Envelope Modulation Spectroscopy. J. Phys. Chem. B 2002, 106, 1345−1349. (4) Gutjahr, M.; Boehlmann, W.; Boettcher, R.; Poeppl, A. Adsorption of Di-tert-butyl Nitroxide at Monovalent Cations in Zeolite Y as Studied by Electron Spin Resonance Spectroscopy. Stud. Surf. Sci. Catal. 2001, 135, 2189−2196.
Figure 6. Solution UV−vis spectra of MBNA (○, in chloroform; shown normalized at the 300 nm region and with 7.5 × abscissa expansion to emphasize the long wavelength region), OWG spectrum for an mNBA-treated quartz waveguide (black line), and OWG spectrum for an mNBA-treated oxide-coated silicon wafer (gray line).
smooth because mNBA has no particular affinity for silicon. By contrast, the affinity of mNBA for SiO2 leads to relatively 4031
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Monolayers of a Multifunctional Organic Radical. Angew. Chem., Int. Ed. 2007, 46, 2215−2219. (23) Le Moigne, J.; Gallani, J. L.; Wautelet, P.; Moroni, M.; Oswald, L.; Cruz, C.; Galerne, Y.; Arnault, J. C.; Duran, R.; Garrett, M. Nitronyl Nitroxide and Imino Nitroxide Mono- and Biradicals in Langmuir and Langmuir-Blodgett Films. Langmuir 1998, 14, 7485− 7492. (24) Caro, J.; Fraxedas, J.; Jürgens, O.; Santiso, J.; Rovira, C.; Veciana, J.; Figueras, A. The First Oriented Thin Films Based on a Nitronyl Nitroxide Radical. Adv. Mater. 1998, 10, 608−610. (25) Caro, J.; Fraxedas, J.; Figueras, A. Thickness-Dependent Spherulitic Growth Observed in Thin Films of the Molecular Organic Radical p-Nitrophenyl Nitronyl Nitroxide. J. Cryst. Growth 2000, 209, 146−158. (26) Molas, S.; Coulon, C.; Fraxedas, J. Magnetic Properties of Thin α-p-Nitrophenyl Nitronyl Nitroxide Films. CrystEngComm 2003, 5, 310−312. (27) Kakavandi, R.; Savu, S.-A.; Caneschi, A.; Chassé, T.; Casu, M. B. At the Interface between Organic Radicals and TiO2(110) Single Crystals: Electronic Structure and Paramagnetic Character. Chem. Commun. 2013, 49, 10103−10105. (28) Markaryan, G. L.; Lunina, E. V. Adsorption Complexes of Nitroxyl Radicals of the Imidazoline and Imidazolidine Type of the Surface of Silica Gel. Zh. Fiz. Khim. 1996, 70, 1670−1673. (29) Mingalyov, P. G.; Fadeev, A. Yu.; Staroverov, S. M.; Lisichkin, G. V.; Lunina, E. V. Nitroxide Radicals in Studies of the Fine Bonded Layer Structure of Modified Silicas. J. Chromatogr. 1993, 646, 267. (30) Baskett, M.; Lahti, P. M. Crystallography and Magnetism of 3N-tert-Butyl-N-aminoxylbenzoic Acid. Polyhedron 2005, 24, 2645− 2652. (31) (a) Barr, T. L.; Seal, S. Nature of the Use of Adventitious Carbon as a Binding Energy Standard. J. Vac. Sci. Technol. A 1995, 13, 1239−1246. (b) Busolo, F.; Franco, L.; Armelao, L.; Maggini, M. Dynamics of a Nitroxide Layer Grafted onto Porous Silicon. Langmuir 2010, 26, 1889−1893. (32) Binner, J.; Zhang, Y. J. Mater. Sci. Lett. 2001, 20, 123−126. (33) (a) Caro, J.; Fraxedas, J.; Jürgens, O.; Santiso, J.; Rovira, C.; Veciana, J.; Figueras, A. The First Oriented Thin Films Based on a Nitronyl Nitroxide Radical. Adv. Mater. 1998, 10, 608−610. (b) Geneste, F.; Moinet, C.; Ababou-Girard, S.; Solal, F. Covalent Attachment of TEMPO onto a Graphite Felt Electrode and Application in Electrocatalysis. New J. Chem. 2005, 29, 1520−1526. (c) Cougnon, C.; Boisard, S.; Cador, O.; Dias, M.; Levillain, E.; Breton, T. A Facile Route to Steady Redox-Modulated Nitroxide SpinLabeled Surfaces Based on Diazonium Chemistry. Chem. Commun. 2013, 49, 4555−4557. (34) Katter, U. J.; Hill, T.; Risse, T.; Schlienz, H.; Beckendorf, M.; Klüner, T.; Hamann, H.; Freund, H.-J. Adsorption of the Stable Radical Di-tert-butyl Nitroxide (DTBN) on an Epitaxially Grown Al2O3 Film. J. Phys. Chem. B 1997, 101, 552−560.
(5) Yu, X.; Somasundaran, P. Structure of Sodium Dodecyl Sulfate and Polyacrylic Acid Adsorption Layer Using Nitroxide Spin Labeled Alumina. Langmuir 2000, 16, 3506−3508. (6) Damian, G.; Miclaus, V.; Znamirovschi, V.; Cozar, O.; Dulamita, N.; Chis, V.; David, L. The Dynamics of Nitroxide Radicals in Chloroform and Deuterated Chloroform Solutions Adsorbed on Zeolites. Prog. Catalysis 1998, 7, 61−66. (7) Katter, U. J.; Risse, T.; Schlienz, H.; Beckendorf, M.; Kluner, T.; Hamann, H.; Freund, H.-J. ESR and TPD Investigations of the Adsorption of Di-tert-butyl Nitroxide on Au(111) and NiO(111). Evidence for Long-Range Interactions. J. Magn. Reson. 1997, 126, 242−247. (8) Katter, U. J.; Hill, T.; Risse, T.; Schlienz, H.; Beckendorf, M.; Kluener, T.; Hamann, H.; Freund, H.-J. Dynamics of the Stable Radical Di-tert-butyl Nitroxide on an Epitaxially Grown Al2O3 Film. J. Phys. Chem. B 1997, 101, 552−560. (9) Ottaviani, M. F.; Venturi, F. Physicochemical Study on the Adsorption Properties of Asbestos. 1. EPR Study on the Adsorption of Organic Radicals. J. Phys. Chem. 1996, 100, 265−273. (10) Malbrel, C. A.; Somasundaran, P.; Turro, N. J. Adsorption of Nitroxide Spin Probes at the Alumina/Cyclohexane Interface in the Presence of Aerosol OT. Langmuir 1989, 5, 490−494. (11) Sagara, T.; Midorikawa, T.; Shultz, D. A.; Zhao, Q. Electrochemical and Spectroelectrochemical Study of a Bis-arylgalvinol-substituted Alkyldisulfide Monolayer and Mixed Monolayers on Polycrystalline Gold. Langmuir 1998, 14, 3682−3690. (12) Mannini, M.; Bonacchi, D.; Zobbi, L.; Piras, F. M.; Speets, E. A.; Caneschi, A.; Cornia, A.; Magnani, A.; Ravoo, B. J.; Reinhoudt, D. N.; Sessoli, R.; Gatteschi, D. Advances in Single-Molecule Magnet Surface Patterning through Microcontact Printing. Nano Lett. 2005, 5, 1435− 1438. (13) Zobbi, L.; Mannini, M.; Pacchioni, M.; Chastanet, G.; Bonacchi, D.; Zanardi, C.; Biagi, R.; Del Pennino, U.; Gatteschi, D.; Cornia, A.; Sessoli, R. Isolated Single-Molecule Magnets on Native Gold. Chem. Commun. 2005, 1640−1642. (14) Kim, G.; Kim, T. Molecular Electronics in Molecular Magnets. J. Korean Phys. Soc. 2005, 46, 684−688. (15) Coronado, E.; Forment-Aliaga, A.; Romero, F. M.; Corradini, V.; Biagi, R.; De Renzi, V.; Gambardella, A.; Del Pennino, U. Isolated Mn12 Single-Molecule Magnets Grafted on Gold Surfaces via Electrostatic Interactions. Inorg. Chem. 2005, 44, 7693−7605. (16) Abdi, A. N.; Bucher, J. P.; Rabu, P.; Toulemonde, O.; Drillon, M.; Gerbier, Ph. Magnetic Properties of Bulk Mn12pivalates16 Single Molecule Magnets and Their Self Assembly on Functionalized Gold Surface. J. Appl. Phys. 2004, 95, 7345−7347. (17) Gómez-Segura, J.; Díez-Pérez, I.; Ishikawa, N.; Nakano, M.; Veciana, J.; Ruiz-Molina, D. 2-D Self-Assembly of the Bis(phthalocyaninato)terbium(III) Single-Molecule Magnet Studied by Scanning Tunnelling Microscopy. Chem. Commun. 2006, 2866−2868. (18) Nishide, H.; Ozawa, T.; Miyasaka, M.; Tsuchida, E. A Nanometer-Sized High-Spin Polyradical: Poly(4-phenoxyl-1,2-phenylenevinylene) Planarily Extended in a Non-Kekulé Fashion and Its Magnetic Force Microscopic Images. J. Am. Chem. Soc. 2001, 123, 5942−5946. (19) Miyasaka, M.; Saito, Y.; Nishide, H. Magnetic Force Microscopy Images of a Nanometer-Sized, Purely Organic High-Spin Polyradical. Adv. Funct. Mater. 2003, 13, 113−117. (20) Kashiwagi, Y.; Uchyama, K.; Kurashima, F.; Anazi, J.; Osa, T. Enantioselective Oxidation of Amines on a Gold Electrode Modified by a Self-Assembled Monolayer of a Chiral Nitroxyl Radical Compound. Anal. Sci. 1999, 15, 907−909. (21) Matsushita, M. M.; Ozaki, N.; Sugawara, T.; Nakamura, F.; Hara, M. Formation of Self-Assembled Monolayer of Phenylthiol Carrying Nitronyl Nitroxide on Gold Surface. Chem. Lett. 2002, 6, 596−597. (22) Crivillers, N.; Mas-Torrent, M.; Perruchas, S.; Roques, N.; Vidal-Gancedo, J.; Veciana, J.; Rovira, C.; Basabe-Desmonts, L.; Ravoo, B. J.; Crego-Calama, M.; Reinhoudt, D. N. Self-Assembled 4032
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Adsorption of a Carboxylic Acid-Functionalized Aminoxyl Radical onto SiO2 Hidenori Murata,† Martha Baskett,† Hiroyuki Nishide,‡ and Paul M. Lahti*,† †
Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003 United States Department of Applied Chemistry, Waseda University, Tokyo 169-855, Japan
‡
S Supporting Information *
ABSTRACT: Silicon wafers both without and with silicon(IV) oxide surface coverage were covered with benzene solutions of stable organic radical 3-(N-tert-butyl-N-aminoxyl)benzoic acid (mNBA). X-ray photoelectron spectroscopy supported the presence of the radical on both surface-cleaned (oxide-reduced) and oxide-covered surfaces. Optical waveguide spectroscopy showed that the radical retained its structure while adsorbed to the surface of the wafers, without noticeable decomposition. AFM and MFM imaging showed that the radical formed blocky particles with a change in rms roughness from 0.3 nm premodification to 1.7 nm postmodification on the surface-cleaned silicon. Similar experiments using oxide-coated silicon showed that the radical adsorbed to form much smoother layers, with a small change in rms roughness from 0.2 to 0.3 nm. Contact angle measurements of water on the premodified and postmodified samples showed a large, hydrophobic change in the silicon oxide surface but only a modest change in the surface-cleaned silicon surface. Samples of mNBA adsorbed onto silica gel showed strong electron-spin resonance signals from the aminoxyl spin, even years after production. The results demonstrate the prospects for treating and coating oxide-covered silicon wafers and silicon oxide-coated particles with a paramagnetically active organic substrate, without major chemical modification of the pretreatment surface; the resulting organic spin sites can be stable for years.
■
INTRODUCTION Recent interest in designed molecular magnetic materials has included efforts to deposit, adsorb, or attach them to surfaces in a manner that does not destroy or substantially alter their unpaired spins or other magnetic properties, as noted in a recent review by Mas-Torrent et al.1 Simple spin-probe-type aminoxyls such as TEMPO and close analogues adsorb onto metal, alumina, silica, and zeolite surfaces.2−10 Stable radicals with pendant long-chain thioalkyl groups can adsorb onto gold with more specificity.11 More recently, single-molecule magnets (SMMs) have also adsorbed onto a number of different types of surfaces.12−17 Persistent organic polymeric polyradicals can even be deposited onto surfaces and imaged with magnetic force microscopy.18,19 A few studies of self-assembled monolayers with radicals have also been reported,20,21 including a recent study by Crivillers et al.22 who self-assembled monolayers of polychlorotriphenylmethyl radicals onto aminefunctionalized SiO2. Work has also been reported describing the Langmuir−Blodgett23 or vapor deposition24−26 of nitronylnitroxides onto surfaces, with an eye to building up layers and testing interradical and packing magnetic behavior under these conditions. Quite recently, Casu and co-workers reported27 adsorbing 2-pyrenylnitronylnitroxide (PyrNN) onto rutile TiO2, finding that the PyrNN either retained or lost its radical spin integrity depending on the local substrate nature (with the initial deposition layer possibly serving to protect subsequent layers). © 2014 American Chemical Society
One important strategy for adsorbing magnetically active molecules takes advantage of hydrogen bonding interactions at the surface.28,29 In 2005, Baskett and Lahti reported that stable hydrogen-bonding radical 3-(N-tert-butyl-N-aminoxyl)benzoic acid (mNBA, Scheme 1) exhibits a strong affinity for silica Scheme 1. Structure of 3-(N-tert-Butyl-N-aminoxyl)benzoic Acid (mNBA)
gel,30 raising the possibility that it could form well-behaved layers on hydroxylated SiO2 surfaces. Given the ubiquity of SiO2 in many uses, especially as a surface layer on silicon, it seemed reasonable to establish the ability of mNBA to adhere selectively to SiO2 and to determine whether the adsorbed mNBA is persistent under these conditions. In this article, we report that mNBA is readily adsorbed onto oxide-coated silicon wafers from benzene solutions to give smooth layers of radicals that are persistent for years and are readily detectable by Received: January 8, 2014 Revised: February 28, 2014 Published: March 31, 2014 4026
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
■
RESULTS AND DISCUSSION mNBA has a strong affinity to adsorb onto silica and related silicon oxide surfaces, especially from benzene solutions. It adsorbs so strongly to the inside of a standard Pyrex roundbottomed flask that attempts to scrape out mNBA remnants with a metal spatula cracked the flask first; this problem was solved by rinsing the inside of glassware with dichlorodimethylsilane before adding NBA solution. This latter behavior is consistent with the need for free surface OH groups to enable the strong adsorption. The addition of 200 mesh (74 μm maximum size) silica gel particles to freshly made solutions of mNBA in benzene essentially removes the UV−vis band associated with the radical from the solution (reddish color for mNBA in solution, a broad solution absorption at 400−450 nm). The adsorbed radical imparts a yellow color to the silica that remains for months to years without apparent visual change. Although mNBA solutions lose electron spin resonance (ESR) activity within about a week of exposure to air at ambient temperature, silica gel modified by exposure to mNBA in benzene still shows readily seen ESR signals from the aminoxyl radical spins after years under ambient conditions. Figure 1 shows ESR spectra obtained at different temperatures
spectroscopic methods as well as by atomic and magnetic force microscopy.
■
Article
EXPERIMENTAL PROCEDURES
Materials. mNBA was synthesized by the procedure of Baskett and Lahti.30 Benzene was freshly distilled from sodium prior to use. Ultrapure water was refined using a Millipore Mill-Q water-purification system. Other reagent chemicals were commercially obtained and used as received. Silicon wafers were obtained from KST World Corp (10 mm × 10 mm × 525 μm with about 200 nm of silicon oxide film on the surface according to commercial specifications). The wafers were soaked in a 4:1(v/v) sulfuric acid/hydrogen peroxide solution at 70 °C for 10 min to remove organic contaminants and metal particles on the surface and then rinsed with ultrapure water. To reduce the oxide coating for some experiments, wafers were soaked in 1% HF(aq) solution for 1 min and rinsed with ultrapure water and then used at once. The surface-cleaned silicon showed 6% residual O 1s in X-ray photoelectron spectra compared to the integrated peak intensity of the untreated, oxidecoated wafer; this may have been present as a result of handling in air in a typical laboratory environment after cleaning. Radical Deposition on SiO2-Covered Silicon. An oxide-coated silicon wafer was immersed in a 1.0 mM solution of mNBA in dry benzene for 10 min at room temperature and then rinsed with dry benzene. The wafer was then dried under a flow of nitrogen. Radical Deposition on Surface-Cleaned Silicon. The same procedure was used as for radical deposition on SiO2, except that the silicon wafer was freshly cleaned with HF(aq) solution before immersion in the benzene solution of the radical. Characterization. Static water drop contact angles were measured on surface-cleaned and oxide-coated silicon wafers both before and after radical deposition (Supporting Information Figure S1). X-ray photoelectron spectra (XPS) were obtained within a week of sample preparation using a JEOL JPS-9010 TR instrument equipped with an Al Kα source, a monochromator, a concentric hemispherical analyzer, and a multichannel detector. A takeoff angle of 15° from the surface was used. Spectral regions for O 1s (525−540 eV binding energy), N 1s (395−405 eV binding energy), and C 1s (265−290 eV binding energy) were recorded for all samples (Supporting Information Figures S2−S4). Binding energies were calibrated31 at 284.8 eV for the main C 1s peak of adventitious carbon on both surface-cleaned and oxide-coated silicon wafers. XPS line-shape fitting by a sum of Gaussian and Lorentzian forms yielded individual peak contributions that are summarized in Table 2. (See also Supporting Information Figures S3 and S4.) Elemental N/C composition ratios were determined by integration under the XPS peak envelopes. Force microscopy experiments were performed using a Nanoscope IIIa Multimode AFM/MFM instrument (Digital Instruments, Inc.). A Digital Instruments single-crystal silicon cantilever for AFM was used for NCH; tip height, 10−15 μm; radius curvature at front, ca. 5 nm; length, 125 μm; width, 30 μm; thickness, 3−5 μm; and a spring constant between 20 and 100 N m−1. Resonance peaks in the frequency response of the cantilever were selected in the range between 280 and 320 kHz for the tapping-mode oscillation. Electron spin resonance (ESR) spectra were obtained using a Bruker Elexsys E500 spectrometer at X-band frequencies (9.4−9.6 GHz). Samples of mNBA adsorbed onto silica gel were placed in 4 mm o.d. quartz tubes and studied first within 24 h of adsorption and then studied after extended times of being stored under ambient conditions in a clear, colorless glass screw-top vial inside a laboratory bench drawer. Variable-temperature ESR spectra were obtained using a Bruker ER-4131 cold nitrogen gas flow unit. Surface absorption spectra of mNBA were studied using a System Instruments Inc. SIS-50 optical waveguide (OWG) spectrophotometer. A quartz plate of thickness 0.2 mm and refractive index 1.46 was used as an optical waveguide. The angle of incidence of the light was 79.2° from the normal of the surface.
Figure 1. ESR spectra of mNBA adsorbed onto silica gel from benzene at various temperatures; ν0 = 9.373 GHz. Spectra are offset for ease of viewing.
on a sample of mNBA absorbed from benzene onto silica gel. The line shape reversibly changes when cooling from room temperature to below 240 K but shows no other significant qualitative changes. The line-shape changes are consistent with changes in the conformational mobility of the aminoxyl unit on the surface. There was no ESR-based evidence, such as half-field transitions or additional features in the g ≈ 2 region, that would indicate the formation of triplet states from dimerized spins resulting from the close interaction between mNBA radicals. With a modest loss of spin intensity, the same ESR spectrum was found in the same set of samples after up to 3 years of storage under ambient conditions. 4027
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
interacting at the surface. If wide-coverage adsorption occurred on the surface-cleaned silicon, then the contact angle should be similar after treating the radical-treated, oxide-coated surface. Instead, the surface-cleaned silicon becomes just slightly more hydrophobic. X-ray photoelectron spectra (XPS) of both surface-cleaned and SiO2-coated silicon wafers were also obtained before and after modification by “dip-and-wash” treatment with benzene solutions of mNBA. Comparisons of pre- and post-treatment results are summarized in Figure 2 and Table 2. Both surfacecleaned and oxide-coated silicon wafers show adventitious carbon peaks; some spectra showed a small C 1s shoulder at about 282.5 eV that corresponds32 to Si−C. Both silicon wafers showed mNBA deposition after modification, especially on the basis of the formation of N 1s peaks for the mNBA-modified samples; there are no N 1s peaks before modification. There is also an increase in the C 1s peak for both surface-cleaned and oxide-coated silicon wafers after mNBA modification. Although the mNBA-modified, surface-cleaned sample shows a strong increase in the O 1s peak intensity (Supporting Information Figure S2), the oxide-coated silicon sample shows a decrease in the O 1s peak intensity after modification. This is presumably due to the effective coverage of the SiO2 surface with the radical, which has only three oxygen atoms among multiple carbons and a nitrogen atom. The main peak of the N 1s region is located at about 400 eV, in good agreement with reports for other aminoyxl and related N−O-containing systems.33 The oxide-coated silicon sample has higher-energy N 1s shoulders at about 401.5 and 402.6 eV, possibly as a result of aminoxyl moieties that are hydrogen bonded to the oxide surface34 as suggested in Scheme 1. No prominent shoulders are present in the mNBA-modified, surface-cleaned sample; this fits the expectation that there is little surface oxide to bind the aminoxyl group. The integrated ratios of the N 1s/C 1s peak region intensities for the surfacecleaned and oxide-coated mNBA-modified silicon wafers were N/C = 0.115 and 0.102, respectively: for mNBA, the theoretical elemental ratio is N/C = 0.091. The postmodified, oxide-coated silicon sample C 1s XPS peak shows multiple features that are not present in the adventitious carbon XPS of the premodified sample. The main carbon peak envelope at 285 eV corresponds to expectations for aryl C 1s overlapped with higher-energy aliphatic tert-butyl C 1s transitions (Table 2, Figure 2). A higher-energy peak at 286.5 eV probably arises from the two carbons bonded to the N−O group, and the 288.2 eV peak fits the expectations33 for the carboxylic acid carbon. Peak fitting (Supporting Information, Figures S3 and S4) for the C 1s region is in reasonably good agreement with this progression of overlapping peak contributions. Analogous C 1s XPS peaks with similar intensities are seen in the postmodified, surfacecleaned silicon sample after mNBA modification. Although the XPS results show mNBA deposition onto both surface-cleaned and oxide-coated silicon, it does not show whether the radical adsorbs relatively smoothly onto either surface or just forms accretions that adhere despite washing with nonpolar benzene. The water contact angle studies showed a clear difference between the two types of modified surfaces. Atomic force microscopy (AFM) and magnetic force microscopy (MFM) imaging of both types of silicon wafers was therefore carried out before and after mNBA treatment. Figure 3 shows 2-D and 3-D AFM images from the same surfacecleaned silicon wafer used for the XPS measurements given in Figure 2. Figure 4 shows the corresponding results for the
The use of relatively nonpolar benzene as a solvent for the adsorption process is important. mNBA solutions in diethyl ether also enable adsorption onto silica gel, giving the typical characteristic coloration of this aminoxyl radical on the gel particles and a clear aminoxyl ESR spectrum from the particles. However, the ether solution was not decolorized in the manner that benzene was, showing that the radical is partitioned between the solution and silica surface. Attempts to adsorb the radical onto silica gel from methanol solutions yield no coloration of the silica gel; the gel samples from these experiments also show no ESR signal. Presumably, as the solvent polarity increases, its affinity for the strongly hydrogenbonding carboxylic acid also increases, with commensurately decreased competition of adsorption on the silica surface. Thus, one can adsorb mNBA onto silica from benzene, separate and dry the silica, and then extract the majority of the radical with methanol away from the silica back into solution, judging by changes in the silica gel color and the loss of most of the aminoxyl ESR signal. The reversibility of the adsorption process from benzene rules out a strong covalent attachment of mNBA to silica, for example, by the mNBA carboxylic acid carbonyl to the surface OH groups. It seems likely that hydrogen bonds form between the OH groups on the surface of the silica and the carboxylic acid groups of the radicals. An additional interaction as shown in Scheme 2 between the moderately basic Scheme 2. Possible Interactions between mNBA and the Oxide-Coated Silicon Surface
aminoxyl oxygen and the silica OH groups might also occur because of the geometrical matching of the silica OH groups and the meta geometry between the aminoxyl unit and the carboxylic acid. To test the extent of modification of wafers soaked in benzene solutions of mNBA, water drop contact angles were measured before and after treatment (Table 1, Supporting Information Figure S1). They show only a small relative change for surface-cleaned silicon but a large change after mNBA treatment of the oxide-coated surface. The results show significantly hydrophobic modification of the oxide-coated surface, as expected if the radicals were extensively surface adsorbed to have their hydrophilic carboxylic acid groups Table 1. Static Water Drop Contact Angles for Pre- and Post-mNBA Modification of Surface-Cleaned and OxideCoated Silicon Wafers
a
sample
contact angle (θ/deg)a
Si modified Si SiO2 modified SiO2
73 80 35 66
See also Supporting Information Figure S1. 4028
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Figure 2. X-ray photoelectron (XPS) spectral traces for a surface-cleaned silicon wafer (left column) and an oxide-coated silicon wafer (right column) before (broken lines) and after (solid lines) modification by a benzene solution of mNBA. Each individual plot compares pairs of traces having the same ordinate scale for the same wafer, but all ordinate scales are different from one another.
remains comparable to the premodification roughness. Also, the MFM images for the mNBA-treated, surface-cleaned silicon show the magnetic detection of larger features in the AFM image, indicating discrete accretions of the spin-active radical on the surface rather than a coating of the overall surface in a smooth manner. By comparison, both AFM and MFM images for oxide-coated silicon after mNBA treatment are essentially featureless in Figures 4 and 5, consistent with the relatively smooth adsorptive formation of a layer of radical under these conditions. The following permissive evidence for the surface deposition of mNBA has been described to this point: (1) strong ESR aminoxyl radical signals are seen in silica gel modified by soaking in benzene solutions of mNBA, signals that are removed by soaking the silica gel in polar solvents; (2) XPS shows strong N 1s peaks on both surface-cleaned and oxidecoated silicon wafer surfaces after dip-and-wash modification with a benzene solution of mNBA whereas no N 1s peak is seen before modification and only mNBA contains nitrogen in the modification treatment reagents; (3) water contact angles are strongly altered on an oxide-coated silicon surface after modification by an mNBA solution, consistent with surface coverage to make it less hydrophilic; (4) MFM detects features seen in AFM tapping mode for mNBA-modified, surfacecleaned silicon, showing that the mNBA spin units survive as discrete radicals in these features. To support further the survival of mNBA discrete radicals on the oxide-coated silicon wafer, Figure 6 compares a solution mNBA UV−vis absorption spectrum to the wafer’s solid film optical waveguide (OWG) spectrum after mNBA modification. The two spectra correspond quite well, with both showing 300 nm strong absorption peaks, particularly the long-wavelength band at ∼450 nm that is characteristic of the phenyl-aminoxyl moiety. Notably, the mNBA samples adsorbed to the oxide-coated
Table 2. XPS Results for Surface-Cleaned Si and OxideCoated Si, Premodified and Postmodified by Treatment with mNBA sample cleaned Si, premodified cleaned Si, postmodified
oxide-coated Si, premodified
oxide-coated Si, postmodified
C 1s/eV 284.8a (70)b 286.5 (30) [283.2 (1)c] 285.3 (68) 287.0 (21) 288.7 (10) [282.7 (20)c] 284.8a (73) 287.0 (7) [282.5 (5)c] 285.0 (63) 286.7 (19) 288.4 (13)
N 1s/eV
O 1s/eV
nil
532.8
400.3 (86) 402.7 (14)
532.9
nil
532.9
399.9 (80) 401.5 (9) 402.6 (11)
532.9
a
Calibration peak, eV. b(Percent) contribution to peak envelope; see Supporting Information Figures S3 and S4. cSi−C (ref 32).
oxide-coated silicon wafer used for the XPS measurements in Figure 2. Figure 5 compares AFM to MFM images of the same regions for both surface-cleaned and oxide-coated silicon wafers after the deposition of mNBA. The AFM results show that the surface-cleaned silicon surface becomes very rough after radical modification, with numerous large surface features including crystallites 30−100 nm in diameter. By comparison, oxide-coated silicon shows a relatively smooth postmodification surface. Table 3 summarizes rms roughness measurements obtained by AFM for portions of the premodification and postmodification surfaces that are shown in Supporting Information Figures S4 and S5. The roughness of the surface-cleaned silicon wafer increases more than 6-fold after modification, but the oxide-coated surface 4029
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Figure 3. AFM images of a surface-cleaned silicon wafer before and after modification with mNBA. (a) 2-D, premodification, 20 nm height; (b) 2-D, postmodification, 20 nm height; (c) 3-D, premodification, 20 nm height; (d) 3-D, postmodification, 20 nm height. These profiles were obtained using tapping-mode AFM in air under ambient conditions, with a 1.0 Hz scan rate in all cases. The 2-D plot inset in the upper row shows z-height color coding.
Figure 4. AFM images of an oxide-coated silicon wafer before and after modification with mNBA. (a) 2-D, premodification, 10 nm lift; (b) 2-D, postmodification, 20 nm height; (c) 3-D, premodification, 20 nm height; (d) 3-D, postmodification, 20 nm height. The 2-D plot insets in the upper row show z-height color codings.
silicon surface still showed this characteristic aminoxyl OWG absorbance after more than 2 years under ambient conditions in air. This is consistent with the extended lifetime of mNBA radicals adsorbed onto silica gel, by comparison to the much shorter ESR lifetime for mNBA in toluene solution under ambient conditions (7−10 days).
The combination of XPS, force microscopy, and OWG spectral results shows that mNBA adsorbs onto SiO2 to form persistent films of the radical. The results support a qualitative picture of the adsorption shown in Scheme 3, where grainy crystallites of mNBA form on surface-cleaned silicon after the evaporation of benzene solvent. The surface coverage is not 4030
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Scheme 3. Deposition of mNBA on Surface-Cleaned versus Oxide-Coated Silicon
smooth, even adsorption on the surface of the oxide-coated silicon.
■
CONCLUSIONS Although 3-(N-tert-butyl-N-aminoxyl)benzoic acid is a structurally simple molecule, its strong affinity for SiO2 causes it to be adsorbed as relatively smooth layers on oxide-coated silicon surfaces, which is quite different from its behavior on surfacecleaned silicon. Notably, the adsorbed organic radicals are persistent for years on both particulate and wafer surfaces. This constitutes a simple method for coating an inorganic oxide, presumably having surface coverage with covalently bound OH groups, with an organic paramagnetic radical substrate and maintaining readily detectable paramagnetism of the radical spin sites. This sort of strategy is worth further investigation as part of efforts to achieve the technologically useful adsorption of persistently spin-bearing species as thin films on solid surfaces.
Figure 5. Comparison of force microscopy images of silicon wafers after modification with mNBA. (a) Surface-cleaned silicon, AFM tapping mode, 100 nm z range; (b) surface-cleaned silicon, MFM mode with 5 nm lift, 5° z-range phase; (c) oxide-coated silicon, AFM, 5 nm lift, 10 nm z range; (d) oxide-coated silicon, MFM mode with 5 nm lift, 2° z-range phase.
■
S Supporting Information *
Table 3. Roughness of Pre- and Post-mNBA Modification of Surface-Cleaned and Oxide-Coated Silicon Surfaces
a
sample
rms roughness (nm)
Si modified Si SiO2 modified SiO2
0.3 1.7 0.2 0.3
ASSOCIATED CONTENT
Static contact water droplet images; XPS peak fitting deconvolution for premodification and mNBA-modified silicon wafers (surface-cleaned and oxide-coated); AFM images of premodified and mNBA-modified silicon wafers (surfacecleaned and oxide-coated) showing regions used for surface roughness evaluation. This material is available free of charge via the Internet at http://pubs.acs.org.
a
■
See also Figures S5 and S6 in the Supporting Information.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS M.B. and P.M.L. acknowledge support from the U.S. National Science Foundation (CHE 0415716). REFERENCES
(1) Mas-Torrent, M.; Crivillers, N.; Rovira, C.; Veciana, J. Attaching Persistent Organic Free Radicals to Surfaces: How and Why. Chem. Rev. 2012, 112, 2506−2527. (2) Zhang, Z.; Berg, A.; Levanon, H.; Fessenden, R. W.; Meisel, D. On the Interactions of Free Radicals with Gold Nanoparticles. J. Am. Chem. Soc. 2003, 125, 7959−7963. (3) Gutjahr, M.; Boettcher, R.; Poeppl, A. Characterization of the Ditert-butyl Nitroxide: Li+ Adsorption Complex in LiY Zeolites by Oneand Two-Dimensional Electron Spin-Echo Envelope Modulation Spectroscopy. J. Phys. Chem. B 2002, 106, 1345−1349. (4) Gutjahr, M.; Boehlmann, W.; Boettcher, R.; Poeppl, A. Adsorption of Di-tert-butyl Nitroxide at Monovalent Cations in Zeolite Y as Studied by Electron Spin Resonance Spectroscopy. Stud. Surf. Sci. Catal. 2001, 135, 2189−2196.
Figure 6. Solution UV−vis spectra of MBNA (○, in chloroform; shown normalized at the 300 nm region and with 7.5 × abscissa expansion to emphasize the long wavelength region), OWG spectrum for an mNBA-treated quartz waveguide (black line), and OWG spectrum for an mNBA-treated oxide-coated silicon wafer (gray line).
smooth because mNBA has no particular affinity for silicon. By contrast, the affinity of mNBA for SiO2 leads to relatively 4031
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
Langmuir
Article
Monolayers of a Multifunctional Organic Radical. Angew. Chem., Int. Ed. 2007, 46, 2215−2219. (23) Le Moigne, J.; Gallani, J. L.; Wautelet, P.; Moroni, M.; Oswald, L.; Cruz, C.; Galerne, Y.; Arnault, J. C.; Duran, R.; Garrett, M. Nitronyl Nitroxide and Imino Nitroxide Mono- and Biradicals in Langmuir and Langmuir-Blodgett Films. Langmuir 1998, 14, 7485− 7492. (24) Caro, J.; Fraxedas, J.; Jürgens, O.; Santiso, J.; Rovira, C.; Veciana, J.; Figueras, A. The First Oriented Thin Films Based on a Nitronyl Nitroxide Radical. Adv. Mater. 1998, 10, 608−610. (25) Caro, J.; Fraxedas, J.; Figueras, A. Thickness-Dependent Spherulitic Growth Observed in Thin Films of the Molecular Organic Radical p-Nitrophenyl Nitronyl Nitroxide. J. Cryst. Growth 2000, 209, 146−158. (26) Molas, S.; Coulon, C.; Fraxedas, J. Magnetic Properties of Thin α-p-Nitrophenyl Nitronyl Nitroxide Films. CrystEngComm 2003, 5, 310−312. (27) Kakavandi, R.; Savu, S.-A.; Caneschi, A.; Chassé, T.; Casu, M. B. At the Interface between Organic Radicals and TiO2(110) Single Crystals: Electronic Structure and Paramagnetic Character. Chem. Commun. 2013, 49, 10103−10105. (28) Markaryan, G. L.; Lunina, E. V. Adsorption Complexes of Nitroxyl Radicals of the Imidazoline and Imidazolidine Type of the Surface of Silica Gel. Zh. Fiz. Khim. 1996, 70, 1670−1673. (29) Mingalyov, P. G.; Fadeev, A. Yu.; Staroverov, S. M.; Lisichkin, G. V.; Lunina, E. V. Nitroxide Radicals in Studies of the Fine Bonded Layer Structure of Modified Silicas. J. Chromatogr. 1993, 646, 267. (30) Baskett, M.; Lahti, P. M. Crystallography and Magnetism of 3N-tert-Butyl-N-aminoxylbenzoic Acid. Polyhedron 2005, 24, 2645− 2652. (31) (a) Barr, T. L.; Seal, S. Nature of the Use of Adventitious Carbon as a Binding Energy Standard. J. Vac. Sci. Technol. A 1995, 13, 1239−1246. (b) Busolo, F.; Franco, L.; Armelao, L.; Maggini, M. Dynamics of a Nitroxide Layer Grafted onto Porous Silicon. Langmuir 2010, 26, 1889−1893. (32) Binner, J.; Zhang, Y. J. Mater. Sci. Lett. 2001, 20, 123−126. (33) (a) Caro, J.; Fraxedas, J.; Jürgens, O.; Santiso, J.; Rovira, C.; Veciana, J.; Figueras, A. The First Oriented Thin Films Based on a Nitronyl Nitroxide Radical. Adv. Mater. 1998, 10, 608−610. (b) Geneste, F.; Moinet, C.; Ababou-Girard, S.; Solal, F. Covalent Attachment of TEMPO onto a Graphite Felt Electrode and Application in Electrocatalysis. New J. Chem. 2005, 29, 1520−1526. (c) Cougnon, C.; Boisard, S.; Cador, O.; Dias, M.; Levillain, E.; Breton, T. A Facile Route to Steady Redox-Modulated Nitroxide SpinLabeled Surfaces Based on Diazonium Chemistry. Chem. Commun. 2013, 49, 4555−4557. (34) Katter, U. J.; Hill, T.; Risse, T.; Schlienz, H.; Beckendorf, M.; Klüner, T.; Hamann, H.; Freund, H.-J. Adsorption of the Stable Radical Di-tert-butyl Nitroxide (DTBN) on an Epitaxially Grown Al2O3 Film. J. Phys. Chem. B 1997, 101, 552−560.
(5) Yu, X.; Somasundaran, P. Structure of Sodium Dodecyl Sulfate and Polyacrylic Acid Adsorption Layer Using Nitroxide Spin Labeled Alumina. Langmuir 2000, 16, 3506−3508. (6) Damian, G.; Miclaus, V.; Znamirovschi, V.; Cozar, O.; Dulamita, N.; Chis, V.; David, L. The Dynamics of Nitroxide Radicals in Chloroform and Deuterated Chloroform Solutions Adsorbed on Zeolites. Prog. Catalysis 1998, 7, 61−66. (7) Katter, U. J.; Risse, T.; Schlienz, H.; Beckendorf, M.; Kluner, T.; Hamann, H.; Freund, H.-J. ESR and TPD Investigations of the Adsorption of Di-tert-butyl Nitroxide on Au(111) and NiO(111). Evidence for Long-Range Interactions. J. Magn. Reson. 1997, 126, 242−247. (8) Katter, U. J.; Hill, T.; Risse, T.; Schlienz, H.; Beckendorf, M.; Kluener, T.; Hamann, H.; Freund, H.-J. Dynamics of the Stable Radical Di-tert-butyl Nitroxide on an Epitaxially Grown Al2O3 Film. J. Phys. Chem. B 1997, 101, 552−560. (9) Ottaviani, M. F.; Venturi, F. Physicochemical Study on the Adsorption Properties of Asbestos. 1. EPR Study on the Adsorption of Organic Radicals. J. Phys. Chem. 1996, 100, 265−273. (10) Malbrel, C. A.; Somasundaran, P.; Turro, N. J. Adsorption of Nitroxide Spin Probes at the Alumina/Cyclohexane Interface in the Presence of Aerosol OT. Langmuir 1989, 5, 490−494. (11) Sagara, T.; Midorikawa, T.; Shultz, D. A.; Zhao, Q. Electrochemical and Spectroelectrochemical Study of a Bis-arylgalvinol-substituted Alkyldisulfide Monolayer and Mixed Monolayers on Polycrystalline Gold. Langmuir 1998, 14, 3682−3690. (12) Mannini, M.; Bonacchi, D.; Zobbi, L.; Piras, F. M.; Speets, E. A.; Caneschi, A.; Cornia, A.; Magnani, A.; Ravoo, B. J.; Reinhoudt, D. N.; Sessoli, R.; Gatteschi, D. Advances in Single-Molecule Magnet Surface Patterning through Microcontact Printing. Nano Lett. 2005, 5, 1435− 1438. (13) Zobbi, L.; Mannini, M.; Pacchioni, M.; Chastanet, G.; Bonacchi, D.; Zanardi, C.; Biagi, R.; Del Pennino, U.; Gatteschi, D.; Cornia, A.; Sessoli, R. Isolated Single-Molecule Magnets on Native Gold. Chem. Commun. 2005, 1640−1642. (14) Kim, G.; Kim, T. Molecular Electronics in Molecular Magnets. J. Korean Phys. Soc. 2005, 46, 684−688. (15) Coronado, E.; Forment-Aliaga, A.; Romero, F. M.; Corradini, V.; Biagi, R.; De Renzi, V.; Gambardella, A.; Del Pennino, U. Isolated Mn12 Single-Molecule Magnets Grafted on Gold Surfaces via Electrostatic Interactions. Inorg. Chem. 2005, 44, 7693−7605. (16) Abdi, A. N.; Bucher, J. P.; Rabu, P.; Toulemonde, O.; Drillon, M.; Gerbier, Ph. Magnetic Properties of Bulk Mn12pivalates16 Single Molecule Magnets and Their Self Assembly on Functionalized Gold Surface. J. Appl. Phys. 2004, 95, 7345−7347. (17) Gómez-Segura, J.; Díez-Pérez, I.; Ishikawa, N.; Nakano, M.; Veciana, J.; Ruiz-Molina, D. 2-D Self-Assembly of the Bis(phthalocyaninato)terbium(III) Single-Molecule Magnet Studied by Scanning Tunnelling Microscopy. Chem. Commun. 2006, 2866−2868. (18) Nishide, H.; Ozawa, T.; Miyasaka, M.; Tsuchida, E. A Nanometer-Sized High-Spin Polyradical: Poly(4-phenoxyl-1,2-phenylenevinylene) Planarily Extended in a Non-Kekulé Fashion and Its Magnetic Force Microscopic Images. J. Am. Chem. Soc. 2001, 123, 5942−5946. (19) Miyasaka, M.; Saito, Y.; Nishide, H. Magnetic Force Microscopy Images of a Nanometer-Sized, Purely Organic High-Spin Polyradical. Adv. Funct. Mater. 2003, 13, 113−117. (20) Kashiwagi, Y.; Uchyama, K.; Kurashima, F.; Anazi, J.; Osa, T. Enantioselective Oxidation of Amines on a Gold Electrode Modified by a Self-Assembled Monolayer of a Chiral Nitroxyl Radical Compound. Anal. Sci. 1999, 15, 907−909. (21) Matsushita, M. M.; Ozaki, N.; Sugawara, T.; Nakamura, F.; Hara, M. Formation of Self-Assembled Monolayer of Phenylthiol Carrying Nitronyl Nitroxide on Gold Surface. Chem. Lett. 2002, 6, 596−597. (22) Crivillers, N.; Mas-Torrent, M.; Perruchas, S.; Roques, N.; Vidal-Gancedo, J.; Veciana, J.; Rovira, C.; Basabe-Desmonts, L.; Ravoo, B. J.; Crego-Calama, M.; Reinhoudt, D. N. Self-Assembled 4032
dx.doi.org/10.1021/la5000952 | Langmuir 2014, 30, 4026−4032
