remove unreacted DSP, then lysed in RIPA buffer (10) before immunoprecipitation.
38. Immunoprecipitation and immunoblotting of Flag and HA epitope–taggedb2ARs, HA epitope–tagged AT1A receptors, Flag epitope–taggedb-arrestins, and c-Src were performed with commercially available antisera as described (9, 31). Immunoprecipitation and immunoblotting of His6-tagged b-arrestin 1 were performed with a rabbit polyclonal antiserum raised against a GST–b-arrestin 1 COOH-terminus fusion protein (32). Immune complexes on nitrocel- lulose were visualized by enzyme-linked chemilumi- nescence and quantified by scanning laser densitom- etry. The stoichiometry ofb-arrestin 1 to c-Src in receptor immunoprecipitates was determined by normalizing immunoblot intensities to the signal ob- tained from known amounts of purified recombinant His6-tagged b-arrestin 1 and c-Src resolved along with the receptor immunoprecipitates.
39. Flag epitope–tagged truncation or deletion mutants of b-arrestin 1 were prepared by the polymerase chain reaction (PCR), incorporating an Eco RI site, minimal Kozak sequence (ACC), and initiator methionine codon into the 59 primer, and the Flag epitope sequence, stop codon, and an Xho I site into the 39 primer. PCR
products were subcloned into a peptide minigene ex- pression cassette as described (33). DNA sequences were confirmed by dideoxynucleotide sequencing.
40. Recombinant His6-taggedb-arrestin 1 and recombinant human c-Src or c-Src SH2 or SH3 domain GST fusion proteins were combined for 20 min at room tempera- ture in 20ml of 10 mM Pipes (pH 7.0) before samples were diluted with 0.5 ml RIPA buffer, and His6-tagged b-arrestin 1 was immunoprecipitated using rabbit poly- clonal anti-His6immunoglobulin G (IgG). After resolu- tion by SDS-PAGE, both the His6-taggedb-arrestin 1 and coprecipitated Src-GST fusion proteins were de- tected simultaneously by protein immunoblotting with rabbit polyclonal anti-GST–b-arrestin 1. Competition for binding between c-Src and GST–Src SH2 or GST–Src SH3 was performed with the GST-Src fusion protein in 20-fold excess.
41. 12CA5 epitope–taggedb2ARs (500 to 1000 fmol/mg whole-cell protein) expressed in HEK-293 cells in six-well plates were treated with or without isopro- terenol for 30 min in serum-free medium at 37°C.
Cell surface receptors were labeled with 12CA5 mAb, using fluorescein isothiocyanate–conjugated goat an- tibody to mouse IgG as secondary antibody. Receptor sequestration was quantified as loss of cell surface
receptors, as measured by flow cytometry (34). Re- ceptor expression was determined by saturation binding of [125I]pindolol.
42. Aliquots of whole-cell lysate from appropriately stimulated cells (30mg of protein per lane) were resolved by SDS-PAGE, and Erk 1 and Erk2 phosphor- ylation was detected by protein immunoblotting with rabbit polyclonal phospho-MAP kinase–specific IgG. Quantitation of Erk 1 and Erk2 phosphorylation was performed with a Storm Phosphorimager. After quantitation of Erk 1 and Erk2 phosphorylation, ni- trocellulose membranes were stripped of Ig and re- probed with rabbit polyclonal anti-Erk 2 IgG to con- firm equal loading of Erk protein.
43. R.J.L. and M.G.C. are investigators with the Howard Hughes Medical Institute. Supported in part by NIH grants DK02352 and DK55524 (L.M.L.), HL16037 (R.J.L.), NS19576 (M.G.C.), and Heart and Stroke Foundation of Ontario Grant NA3349 (S.S.G.F.). G.J.D.R. is supported by NIH Medical Scientist Training Program Grant T32GM-07171. We thank M. J. Eck for purified recom- binant c-Src and D. Addison and M. Holben for excellent secretarial assistance.
27 May 1998; accepted 17 November 1998
“Dip-Pen” Nanolithography
Richard D. Piner, Jin Zhu, Feng Xu, Seunghun Hong, Chad A. Mirkin*
A direct-write “dip-pen” nanolithography (DPN) has been developed to deliver collections of molecules in a positive printing mode. An atomic force micro- scope (AFM) tip is used to write alkanethiols with 30-nanometer linewidth resolution on a gold thin film in a manner analogous to that of a dip pen.
Molecules are delivered from the AFM tip to a solid substrate of interest via capillary transport, making DPN a potentially useful tool for creating and functionalizing nanoscale devices.
Lithographic methods are at the heart of mod- ern-day microfabrication, nanotechnology, and molecular electronics. These methods often rely on patterning of a resistive film, followed by a chemical etch of the substrate. Dip-pen technol- ogy, in which ink on a sharp object is transport- ed to a paper substrate via capillary forces, is approximately 4000 years old (1) and has been used extensively throughout history to transport molecules on macroscale dimensions. Here we report experiments that merge these two related but, with regard to scale and transport mecha- nism, disparate concepts to develop a new type of dip-pen nanolithography (DPN). DPN uses an atomic force microscope (AFM) tip as a
“nib,” a solid-state substrate (in this case, Au) as “paper,” and molecules with a chemical af- finity for the solid-state substrate as “ink.” Cap- illary transport of molecules from the AFM tip
to the solid substrate is used in DPN to directly “write” patterns consisting of a relatively small collection of molecules in submicrometer dimensions.
DPN is not the only lithographic method that allows one to directly transport molecules to substrates of interest in a positive printing mode. For example, microcontact printing,
which uses an elastomer stamp, can deposit patterns of thiol-functionalized molecules di- rectly onto Au substrates (2–6). This method is a parallel technique, allowing one to deposit an entire pattern or series of patterns on a substrate of interest in one step (2–19), which is an advantage over a serial technique such as DPN unless one is trying to selectively place different types of molecules at specific sites within a particular type of nanostructure. In this regard, DPN complements microcontact printing and many other existing methods of micro- and nanofabrication (2–19). Finally, there are a va- riety of negative printing techniques that rely on scanning probe instruments, electron beams, or molecular beams to pattern substrates, using self-assembled monolayers (SAMs) and other organic materials as resist layers (7–19), that is, to remove material for subsequent processing or adsorption steps. However, DPN can deliver relatively small amounts of a molecular sub- stance to a substrate in a nanolithographic fash- ion that does not rely on a resist, a stamp, complicated processing methods, or sophisti-
Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208, USA.
*To whom correspondence should be addressed. E- mail: [email protected]
Fig. 1. Schematic rep- resentation of DPN. A water meniscus forms between the AFM tip coated with ODT and the Au substrate. The size of the meniscus, which is controlled by relative humidity, af- fects the ODT trans- port rate, the effec- tive tip-substrate con- tact area, and DPN resolution.
RE S E A R C H AR T I C L E
www.sciencemag.org SCIENCE VOL 283 29 JANUARY 1999 661
R
E P O R T Scated noncommercial instrumentation.
The inspiration for this methodology came from the study of a problem that has plagued the AFM since its invention. The narrow gap capillary formed between the AFM tip and the sample when an experiment is conducted in air (20–23) condenses water from the ambient and substantially influences imaging experiments especially those attempting to achieve nanome- ter or even angstrom resolution (20–23). In our own work (23), we have shown that this is a dynamic problem and that water, depending on relative humidity and substrate wetting proper- ties, will either be transported from the sub- strate to the tip or vice versa. In the latter case, metastable nanometer-length–scale patterns could be formed from very thin layers of water deposited from the AFM tip (23). We now show that when the transported molecules can anchor themselves to the substrate through chemisorption, stable surface structures are formed, resulting in a new type of nanolithog- raphy, DPN (Fig. 1).
Although we have investigated the transport process with several molecules, we focus on the transfer of 1-octadecanethiol (ODT) to Au sur- faces, a system that has been studied extensive- ly (24–28). This moderately air-stable mole- cule, when immobilized on Au, can be easily differentiated from unmodified Au by means of lateral force microscopy (LFM). Our studies suggest that when an AFM tip coated with ODT is brought into contact with a sample surface, the ODT flows from the tip to the sample by capillary action, a process much like that of a dip pen (Fig. 1). We have studied this process using a conventional AFM (29) on thin film substrates that were prepared by thermally evaporating 300 Å of polycrystalline Au onto mica at room temperature. For these experi- ments, a silicon nitride tip (Park Scientific, Mi- crolever A) was coated with ODT by dipping of the cantilever into a saturated solution of ODT in acetonitrile for 1 min. The cantilever was blown dry with compressed difluoroethane be- fore being used. A simple demonstration of this process involves raster scanning of a tip that was prepared in this manner across a section of a Au substrate measuring 1mm by 1 mm (Fig.
2A). An LFM image of this section within a larger scan area (3mm by 3 mm) shows two areas of differing contrast. The interior dark area, or region of lower lateral force, is a de- posited monolayer of ODT, and the exterior lighter area is bare Au (22). Formation of high- quality SAMs occurs when the deposition pro- cess is carried out on Au(111)/mica, which was prepared by annealing our Au thin film sub- strates at 300oC for 3 hours (28). In this case, it was possible to obtain a lattice-resolved image of an ODT SAM (Fig. 2B). The hexagonal lattice parameter of 5.06 0.2 Å compares well with reported values for SAMs of ODT on Au(111) (28) and shows that ODT, rather than some other adsorbates (water or acetonitrile), Fig. 2. (A) Lateral force
image of a square of ODT measuring 1mm by 1 mm, deposited onto a Au substrate by DPN. This pattern was generated by scanning the 1-mm2 area at a scan rate of 1 Hz for a period of 10 min at a relative humidity of 39%. Then the scan size was increased to 3 mm, and the scan rate was increased to 4 Hz while the image was recorded. The faster scan rate prevents ODT transport. (B) Lattice-resolved, later- al force image of an ODT SAM deposited onto Au(111)/mica by DPN. The image has been filtered with a fast Fourier transform (FFT ), and the FFT of the raw data is shown in the lower right in- sert. The monolayer was generated by scanning a 1000 Å
square area of the Au(111)/mica five times at a rate of 9 Hz at 39% relative humidity. (C) Lateral force image of a 30-nm-wide line (3mm long) deposited onto Au/mica by DPN. The line was generated by scanning the tip in a vertical line repeatedly for 5 min at a scan rate of 1 Hz. (D) Lateral force image of a 100-nm line deposited on Au by DPN. The method of depositing this line is analogous to that used to generate the image in (C), but the writing time was 1.5 min. In all images, darker regions correspond to areas of relatively lower friction.
Fig. 3. (A) Lateral force image of an Au substrate after an AFM tip, which was coated with ODT, had been in contact with the substrate for 2, 4, and 16 min (left to right); the relative hu- midity was held con- stant at 45%, and the image was recorded at a scan rate of 4 Hz. (B) Lateral force image of dots of 16-mercapto- hexadecanoic acid on a Au substrate. To gen- erate the dots, an AFM tip coated with 16- mercaptohexadecanoic acid was held on the Au substrate for 10, 20, and 40 s (left to right).
The relative humidity was 35%. The images show that the trans- port properties of 16- mercaptohexadecanoic acid and of ODT differ substantially. (C) Later- al force image of an ar- ray of dots generated
by DPN. Each dot was generated by holding an ODT-coated tip in contact with the surface for
;20 s. Writing and recording conditions were the same as in (A). (D) Lateral force image of a molecule-based grid. Each line is 100 nm in width and 2mm in length and required 1.5 min to write.
R E P O R T S
29 JANUARY 1999 VOL 283 SCIENCE www.sciencemag.org 662
are transported from the tip to the substrate.
Although the experiments performed on Au(111)/mica provide important information about the chemical identity of the transported species in these experiments, Au(111)/mica is a poor substrate for DPN. The deep valleys around the small Au(111) facets make it diffi- cult to draw long (micrometer) contiguous lines with nanometer widths.
Although the nonannealed Au substrates are relatively rough (the root mean square rough- ness is 2 nm), we could deposit 30-nm lines with DPN; this distance is the average Au grain diameter of our thin film substrates and repre- sents the resolution limit of DPN on this type of substrate (Fig. 2C). The 30-nm molecule-based line prepared on this type of substrate is discon- tinuous and follows the grain edges of the Au.
Smoother and more contiguous lines can be drawn by increasing the line width to 100 nm (Fig. 2D) or presumably by using a smoother Au substrate. The width of the line depends on tip scan speed and the rate of transport of the alkanethiol from the tip to the substrate (relative humidity can change the transport rate). Faster scan speeds and a smaller number of traces give narrower lines.
We also used DPN to prepare molecular dot features to demonstrate the diffusion properties of the “ink” (Fig. 3, A and B). The ODT-coated tip was brought into contact (set point5 1 nN) with the Au substrate for a set period of time.
For example, ODT dots 0.66mm, 0.88 mm, and 1.6mm in diameter were generated by holding the tip in contact with the surface for 2, 4, and 16 min, respectively (Fig. 3A, left to right). The uniform appearance of the dots probably re- flects an even flow of ODT in all directions from the tip to the surface. Opposite contrast images were obtained by depositing dots of an alkanethiol derivative, 16-mercaptohexade- canoic acid, in an analogous fashion (Fig. 3B).
This not only provides additional evidence that the molecules are being transported from the tip to the surface but also demonstrates the molec- ular generality of DPN.
We could generate arrays and grids in addi- tion to individual lines and dots. An array of 25 ODT dots 0.46mm in diameter, spaced 0.54 mm apart (Fig. 3C), was generated by holding an ODT-coated tip in contact with the surface (1 nN) for 20 s at 45% relative humidity with- out lateral movement to form each dot. A grid consisting of eight intersecting lines 2mm in length and 100 nm wide (Fig. 3D) was gener- ated by sweeping the ODT-coated tip on a Au surface at a 4-mm/s scan speed with a 1-nN force for 1.5 min to form each line.
The resolution of DPN depends on several parameters, and its ultimate resolution is not yet clear. First, the grain size of the substrate affects DPN resolution much as the texture of paper controls the resolution of conventional writing.
Second, chemisorption and self assembly can be used to limit the diffusion of the molecules
after deposition. The ODT patterns are stable whereas water forms metastable patterns (23).
Third, the tip-substrate contact time and thus the scan speed influence DPN resolution.
Fourth, relative humidity seems to affect the resolution of the lithographic process by controlling the rate of ODT transport from the tip to the substrate. The size of the water meniscus that bridges the tip and substrate depends on relative humidity (23). For example, the 30-nm-wide line (Fig. 1C) required 5 min to generate in a 34% relative humidity environment, where- as the 100-nm line (Fig. 1D) required 1.5 min to generate in a 42% relative humidity environment.
DPN is a simple but powerful method for transporting molecules from AFM tips to substrates at resolutions comparable to those achieved with much more expensive and so- phisticated competitive lithographic methods, such as electron-beam lithography. It should be especially useful for the detailed function- alization of nanoscale devices prepared by more conventional lithographic methods (30, 31).
References and Notes
1. A. C. Ewing, The Fountain Pen: A Collector’s Compan- ion (Running Press, Philadelphia, PA, 1997).
2. Y. Xia and G. M. Whitesides, Angew. Chem. Int. Ed.
Engl. 37, 550 (1998).
3. E. Kim, Y. Xia, G. M. Whitesides, Nature 376, 581 (1995).
4. Y. Xia et al., Science 273, 347 (1996).
5. L. Yan, X.-M. Zhao, G. M. Whitesides, J. Am. Chem.
Soc. 120, 6179 (1998).
6. A. Kumar, H. A. Biebuyck, N. L. Abbott, G. M. White- sides, ibid. 114, 9188 (1992).
7. L. A. Bottomley, Anal. Chem. 70, 425R (1998).
8. R. M. Nyffenegger and R. M. Penner, Chem. Rev. 97, 1195 (1997).
9. K. K. Berggren et al., Science 269, 1255 (1995).
10. J. A. M. Sondag-Huethorst, H. R. J. van Helleputte, L. G. J. Fokkink, Appl. Phys. Lett. 64, 285 (1994).
11. J. K. Schoer and R. M. Crooks, Langmuir 13, 2323 (1997).
12. S. Xu and G. Liu, ibid., p. 127.
13. F. K. Perkins et al., Appl. Phys. Lett. 68, 550 (1996).
14. D. W. Carr et al., J. Vac. Sci. Technol. A 15, 1446 (1997).
15. M. J. Lercel, H. G. Craighead, A. N. Parikh, K. Seshadri, D. L. Allara, Appl. Phys. Lett. 68, 1504 (1996).
16. H. Sugimura and N. Nakagiri, J. Vac. Sci. Technol. A 14, 1223 (1996).
17. T. Komeda, K. Namba, Y. Nishioka, ibid. 16, 1680 (1998).
18. H. U. Muller, C. David, B. Volkel, M. Grunze, J. Vac. Sci.
Technol. B 13, 2846 (1995).
19. Y. Kim and C. M. Lieber, Science 257, 375 (1992).
20. L. Xu, A. Lio, J. Hu, D. F. Ogletree, M. Salmeron, J.
Phys. Chem. B 102, 540 (1998).
21. M. Binggeli and C. M. Mate, Appl. Phys. Lett 65, 415 (1994).
22. M. Fujihira et al., Chem. Lett. (1996), p. 499.
23. R. D. Piner and C. A. Mirkin, Langmuir 13, 6864 (1997).
24. C. D. Bain and G. M. Whitesides, Angew. Chem. Int.
Ed. Engl. 28, 506 (1989).
25. A. Ulman, An Introduction to Ultrathin Organic Films:
From Langmuir-Blodgett to Self-Assembly (Academic Press, Boston, MA, 1991).
26. L. H. Dubois and R. G. Nuzzo, Annu. Rev. Phys. Chem.
43, 437 (1992).
27. A. R. Bishop and R. G. Nuzzo, Curr. Opin. Coll. Interf.
Sci. 1, 127 (1996).
28. C. A. Alves, E. L. Smith, M. D. Porter, J. Am. Chem. Soc.
114, 1222 (1992).
29. A Park Scientific Model CP instrument was used to perform all experiments. The scanner was enclosed in a glass isolation chamber, and the relative humidity was measured with a hygrometer. All humidity mea- surements have an absolute error of65%.
30. M. A. Reed, C. Zhou, C. J. Muller, T. P. Burgin, J. M.
Tour, Science 278, 252 (1997).
31. D. L. Feldheim and C. D. Keating, Chem. Soc. Rev. 27, 1 (1998).
32. C.A.M. acknowledges the Air Force Office of Scientific Research and the NSF-funded Northwestern Univer- sity Materials Research Center for support of this work.
24 September 1998; accepted 22 December 1998
Visible Quantum Cutting in LiGdF 4 :Eu 31 Through
Downconversion
Rene´ T. Wegh, Harry Donker, Koenraad D. Oskam, Andries Meijerink*
For mercury-free fluorescent lamps and plasma display panels, alternative luminescent materials are required for the efficient conversion of vacuum ultraviolet radiation to visible light. Quantum cutting involving the emission of two visible photons for each vacuum ultraviolet photon absorbed is demon- strated in Eu31-doped LiGdF4 with the concept of downconversion. Upon excitation of Gd31with a high-energy photon, two visible photons can be emitted by Eu31through an efficient two-step energy transfer from Gd31to Eu31, with a quantum efficiency that approaches 200 percent.
Luminescent materials with lanthanides are found in fluorescent tubes, color televisions, x-ray photography, lasers, infrared (IR) to visible light upconversion materials, and fi- ber amplifiers (1–3). Such applications rely
on the luminescence properties of lanthanide ions (sharp lines and high efficiency). In flu- orescent lamps, phosphors on the inside wall of the glass tube convert the ultraviolet (UV) radiation (mainly with a wavelengthl of 254 R E P O R T S
www.sciencemag.org SCIENCE VOL 283 29 JANUARY 1999 663
