Zinc Oxide Growth Morphology on Self-Assembled Monolayer

Apr 10, 2008 - Julia W. P. Hsu,*,† W. Miles Clift,‡ and Luke N. Brewer†. Sandia National Laboratories, Albuquerque, New Mexico 87185, and Sandia...
0 downloads 0 Views 2MB Size
Langmuir 2008, 24, 5375-5381

5375

Zinc Oxide Growth Morphology on Self-Assembled Monolayer Modified Silver Surfaces Julia W. P. Hsu,*,† W. Miles Clift,‡ and Luke N. Brewer† Sandia National Laboratories, Albuquerque, New Mexico 87185, and Sandia National Laboratories, LiVermore, CA ReceiVed December 14, 2007. ReVised Manuscript ReceiVed February 22, 2008 Using organic molecules to direct inorganic crystal growth has opened up new avenues for controlled synthesis on surfaces. Combined with soft lithography to form patterned templates, self-assembled monolayers (SAMs) have been shown to be a powerful approach for the assembly of inorganic nanostructures. In this work, we show that the surface free energy of SAM-modified silver, which depends on end groups and deposition method of SAMs, has a dramatic effect on the nucleation and growth of crystalline ZnO, a technologically important material, from supersaturated solutions. For SAMs with inert methyl end groups, ZnO nucleation is inhibited. For SAMs with chemically active (carboxylic or thiol) end groups, the ZnO morphology is found to be three-dimensional nanorods on low-surfaceenergy surfaces and two-dimensional thin films on high-energy surfaces.

I. Introduction In nature it is well known that organic macromolecules play a critical role in the formation of inorganic biominerals to create crystals with desired shape and/or strength. During the past decade, much research activity has focused on controlling inorganic crystal growth using small molecules that bear critical molecular recognition, with the goal to mimic biological processes to synthesize nonbiominerals or on artificial substrates in laboratories. In particular, growth of various materials has been demonstrated on metal substrates functionalized with selfassembled monolayers (SAMs).1–4 Surface-sensitive growth processes, such as heterogeneous nucleation from supersaturated solutions or atomic layer deposition (ALD), can be greatly affected by small changes in surface conditions. In addition, the orientation,5–8 nucleation density,9,10 and polytype11 of inorganic crystals have been shown to depend strongly on the structure and end groups of the SAMs. An additional advantage of using SAMs as organic templates is the ease of forming patterned templates via soft lithography techniques, namely microcontact printing (µCP). Most commonly, patterned SAMs with inert methyl end * To whom correspondence should be addressed. E-mail: jwhsu@ sandia.gov. Phone: 505-284-1173. Fax: 505-844-5470. † Sandia National Laboratories, Alburquerque, NM. ‡ Sandia National Laboratories, Livermore, CA. (1) Mann, S.; Heywood, B. R.; Rajam, S.; Birchall, J. D. Nature 1988, 334, 692–695. (2) Ku¨ther, J.; Bartz, M.; Seshadri, R.; Vaughan, G. B. M.; Tremel, W. J. Mater. Chem. 2001, 11, 503–506. (3) Kang, J. F.; Zaccaro, J.; Ulman, A.; Myerson, A. Langmuir 2000, 16, 3791–3796. (4) Onuma, K.; Oyane, A.; Kokubo, T.; Trboux, G.; Kanzaki, N.; Ito, A. J. Phys. Chem. B 2000, 104, 11950–11956. (5) Aizenberg, J.; Black, A. J.; Whitesides, G. M. Nature 1999, 398, 495. (6) (a) Han, Y.-J.; Aizenberg, J. Angew. Chem., Int. Ed. 2003, 42, 3668–3670. (b) Aizenberg, J.; Black, A. J.; Whitesides, J. Am. Chem. Soc. 1999, 121, 4500– 4509. (7) (a) Travaille, A. M.; Donners, J. J. M.; Gerritsen, J. W.; Sommerdijk, N. A. J. M.; Nolte, R. J. M.; van Kempen, H. AdV. Mater. 2002, 14, 492–495. (b) Travaille, A. M.; Kaptijn, L.; Verwer, P.; Hulsken, B.; Elemans, J. A. A. W.; Nolte, R. J. M.; van Kempen, H. J. Am. Chem. Soc. 2003, 125, 11571–11577. (8) Biemmi, E.; Scherb, C.; Bein, T. J. Am. Chem. Soc. 2007, 129, 8054–8055. (9) Toworfe, G. K.; Composto, R. J.; Shapiro, I. M.; Ducheyne, P. Biomaterials 2006, 27, 631–642. (10) Dube, A.; Sharma, M.; Ma, P. F.; Ercius, P. A.; Muller, D. A.; Engstrom, J. R. J. Phys. Chem. C 2007, 111, 11045–11058. (11) Ku¨ther, J.; Seshadri, R.; Knoll, W.; Tremel, W. J. Mater. Chem. 1998, 8, 641–650.

groups are used to regulate crystal growth in selective areas in both solution growth5,12–14 and ALD.15,16 Assembly of complex three-dimensional nanostructures on surfaces that cannot be achieved by conventional lithography has also been demonstrated with this approach.17–19 However, little work has explored the use of SAMs to control the growth mode of inorganic crystals. Lee et al. showed that using mixed SAMs to reduce surface free energy of Si altered ALD growth of TiO2 from two-dimensional (2D) to three-dimensional (3D).20 ZnO is a technologically important material. It is a wideband-gap (3.37 eV) semiconductor with a large exciton binding energy (∼60 meV), making it an attractive candidate for UV/ blue solid-state lighting. Because of its wurtzite crystal structure, and hence lack of inversion symmetry in the direction, ZnO is piezoelectric and is used for actuation in microelectronics. By alloying with In and/or Sn, ZnO is a transparent conductor that has uses in display and solar cell technology. Furthermore, ZnO nanostructures are readily formed by various methods, including vapor-21–24 and solution-phase deposition techniques.25–27 Such nanostructures have an enhanced surface-tovolume ratio in addition to showing quantum-confined effects. (12) Rieke, P. C.; Tarasevich, B. J.; Wood, L. L.; Engelhard, M. H.; Baer, D. R.; Fryxell, G. E.; John, C. M.; Laken, D. A.; Jaehnig, M. C. Langmuir 1994, 10, 619–622. (13) Meldrum, G. C.; Flath, J.; Knoll, W. Thin Solid Films 1999, 348, 188– 195. (14) Chen, C.-C.; Lin, J.-J. AdV. Mater. 2001, 13, 136–139. (15) Seo, E. K.; Lee, J. W.; Sung-Suh, H. M.; Sung, M. M. Chem. Mater. 2004, 16, 1878–1883. (16) Jiang, X.; Chen, R.; Bent, S. F. Surf. Coating Technol. 2007, 201, 8799– 8807. (17) Hsu, J. W. P.; Tian, Z. R.; Simmons, N. C.; Matzke, C. M.; Voigt, J. A.; Liu, J. Nano Lett. 2005, 5, 83–86. (18) Sounart, T. L.; Liu, J.; Voigt, J. A.; Hsu, J. W. P.; Spoerke, E. D.; Tian, Z. R.; Jiang, Y. AdV. Func. Mater. 2006, 16, 335–344. (19) Morin, W. A.; Amos, F. F.; Jin, S. J. Am. Chem. Soc. 2007, 129, 13776– 13777. (20) Lee, J. P.; Jang, Y. J.; Sung, M. M. AdV. Func. Mater. 2003, 13, 873–876. (21) Pan, Z. W.; Dai, Z. R.; Wang, Z. L. Science 2001, 291, 1947–1949. (22) Park, W. I.; Kim, D. H.; Jung, S.-W.; Yi, G-C. Appl. Phys. Lett. 2002, 80, 4232–4234. (23) Baxter, J. B.; Aydil, E. S. J. Cryst. Growth 2005, 274, 407–411. (24) Yan, M.; Zhang, H. T.; Widjaja, E. J.; Chang, R. P. H. J. Appl. Phys. 2003, 94, 5240–5246. (25) Vayssieres, L.; Keis, K.; Lindquist, S.-E.; Hagfelt, A. J. Phys. Chem B 2001, 105, 3350–3352. (26) Tian, Z.; Voigt, J. A.; Liu, J.; McKenzie, B.; McDermott, M. J.; Rodriguez, M. A.; Konishi, H.; Xu, H. Nat. Mater. 2003, 2, 821–826.

10.1021/la703919w CCC: $40.75  2008 American Chemical Society Published on Web 04/10/2008

5376 Langmuir, Vol. 24, No. 10, 2008

Hsu et al.

Figure 1. Reflection optical microscope images of ZnO nanorods grown on µCP patterned Ag surfaces using (a) 1-hexadecanethiol and (b) 11mercaptoundecanoic acid as templates. The darker the image is, the denser the ZnO nanorods are. An example of type A and B regions on the -COOH patterns are indicated in (b). The bottom area in (b) is outside of the -COOH SAM pattern, with the horizontal line being the edge of the elastomer stamp used in µCP.

ZnO nanostructures are explored for sensing applications28,29 and as the electron transporter in polymer-oxide hybrid solar cells.30,31 Hence, the ability to place and control ZnO growth on surfaces is critically important. Previously, Turgeman et al. found the structure and end groups of silane SAMs affected the orientation, density, and size of ZnO nanorods grown on Si substrates.32,33 Here we report using alkanethiols with different end groups to control the nucleation and morphology of ZnO crystals grown from supersaturated solution on Ag surfaces. Our results can be understood in terms of changing surface energy by the SAMs, which depends not only on the chemistry of the end groups but also on the method of SAM formation.

II. Experimental Section Ag films (500 Å) were deposited on (001) Si wafers by electron beam evaporation at 15Å/s. The masters for making polydimethalsiloxane (PDMS) (Dow Corning, Sylgard 184) stamps used in µCP were SU8-5 (Microchem Corp.) photoresist patterns on Si substrates. PDMS was cured over the masters to form stamps with complementary relief structures. Flat (unpatterned) PDMS stamps were formed by casting over a Si wafer. SAM solutions (∼5 mM) were (27) Peterson, R. B.; Fields, C. L.; Gregg, B. A. Langmuir 2004, 20, 5114– 5118. (28) Wan, Q.; Li, Q. H.; Chen, Y. J.; Wang, T. H.; He, X. L.; Li, J. P.; Lin, C. L. Appl. Phys. Lett. 2004, 84, 3654–3656. (29) Wang, H. T.; Kang, B. S.; Ren, F.; Tien, L. C.; Sadik, P. W.; Norton, D. P.; Pearton, S. J.; Lin, J. Appl. Phys. Lett. 2005, 86, 243503. (30) Olson, D. C.; Piris, J.; Collins, R. T.; Shaheen, S. E.; Ginley, D. S. Thin Solid Films 2006, 496, 26–29. (31) Law, M.; Greene, L. E.; Johnson, J. C.; Saykally, R.; Yang, P. Nat. Mater. 2005, 4, 455–459. (32) Turgeman, R.; Gershevitz, O.; Deutsch, M.; Ocko, B. M.; Gedanken, A.; Sukenik, C. N. Chem. Mater. 2005, 17, 5048–5056. (33) Turgeman, R.; Gershevitz, O.; Palchik, O.; Deutsch, M.; Ocko, B. M.; Gedanken, A.; Sukenik, C. N. Cryst. Growth Des. 2004, 4, 169–175.

Figure 2. SEM images of µCP -COOH SAM-templated samples after 2 h of ZnO growth. The lighter regions contain ZnO nanorods. (a) A low-magnification image showing that type B regions predominantly occur at the corners and edges of the stamped patterns. (b) A closeup image near a type A-B boundary on a -COOH pattern. The ZnO nanorods are clearly present in the type A region and absent in the type B region. The inset in (b) shows a higher magnification image of nanorods in a type A region. (c) An example of patterned growth of ZnO nanorods on bare Ag regions that were surrounded by -COOH templates with type B behavior, i.e., lack of ZnO nanorods.

formed by dissolving 1-hexadecanethiol (Fluka, >95%), 11mercaptoundecanoic acid (Aldrich, 95%), and 1,8-octanedithiol (Aldrich, 97+%) in ethanol (Aaper Alcohol and Chemical Co., 200 proof). All chemicals were used as received. To coat the PDMS stamps with SAM molecules, a drop of SAM solution was deposited on the stamp and quickly spun dry on a spin-coater (Laurell Technologies Corp. Model WS-400B). The spinning speed was typically 4000 rpm. To chemically pattern the Ag surface, the stamp inked with SAM was placed on the Ag for 10-15 s, resulting in SAM molecules on the portions of PDMS stamp that came into contact with the surface being transferred and assembled onto Ag. The patterns used in this experiment had squares and rectangles of bare Ag surfaces surrounded by SAM-covered Ag regions. For comparison, SAMs were also formed from traditional solution deposition from 5 mM ethanol solutions for 24 h. In the case of 11-mercaptoundecanoic acid, we also made solution-deposited samples following the procedures given by Wang et al. by adding

Zinc Oxide Growth Morphology on SilVer Surfaces

Langmuir, Vol. 24, No. 10, 2008 5377

Figure 4. Auger spectra averaged over 10 µm × 10 µm areas after 2-h growth (curves displaced vertically for clarity) taken on bare Ag (black), µCP -CH3 SAM regions (red), and type A (blue) and type B (green) of µCP -COOH SAM regions.

Figure 3. Auger maps of (a) Zn, (b) O, (c) S, and (d) C on -CH3 (left) and -COOH (right) SAM-templated samples after 30 min of growth time. The molecules are stamped on the regions outside of the rectangular areas shown in the figures. The rectangular regions are bare Ag. The lighter areas in the images indicate regions with higher Auger signals of that element.

2% of trifluoroacetic acid in the ethanol (the TFA method),34 which resulted in a more uniform monolayer without bilayer regions. SAMs formed by µCP and solution deposition were characterized by contact-angle (Kru¨ss DSA1) and single-wavelength ellipsometry (DRE EL X-01R). For these large area measurements, a flat PDMS was used to simulate the SAMs formed by µCP. All sample preparation procedures (inking, printing, and rinsing) were kept the same as preparing a SAM patterned growth template. These chemically modified Ag surfaces were immersed in aqueous solutions (MilliQ, >17.8 MΩ) containing 0.02 M zinc nitrate (Zn(NO3)2 · 6H2O) (Fluka) and 0.02 M hexamethylenetetramine (C6H12N4) (Fisher) at 60 °C for 5 min to 3 h. Under our growth condition, the predominant growth species of ZnO is Zn2+ ions.35 After growth, the samples were rinsed in deionized water thoroughly and air-dried at room temperature. Two SAM patterns using molecules with different end groups were made on a given sample for simultaneous growth of ZnO nanorods to minimize variations from different growth runs when comparing the effects of different molecules. Auger electron spectra (AES) were collected using a Physical Electronics Model 680 with accelerating beam voltage of 5 kV and beam current of 20 nA. The spectra were averaged over 10 µm × 10 µm areas. Sputter Auger profiling was carried out using a 500 eV, 50 nA Ar+ ion beam at a 45° angle of incidence, rastered over a 500 µm × 500 µm area. All depth scales for the Auger depth profiles were normalized to the sputter rate for SiO2. No attempt was made to calibrate the different sputtering rates of the substrate or interfacial regions. A Zeiss (Supra 55 VP-FEG) field-emission source (34) Wang, H.; Chen, S.; Li, L.; Jiang, S. Langmuir 2005, 21, 2633–2636. (35) Mesmer, Jr., R. E. In The Hydrolysis of Cations; Krieger Publishing Company: Malabar, FL, 1986; Chapter 13, pp 287-294.

Figure 5. Auger maps of µCP -CH3 (left) and -COOH (right) SAM patterns after being immersed in the growth solution for 5 min at room temperature: (a) Zn and (b) S. The molecules are stamped on the outside of the rectangular/square regions. The rectangles/squares are bare Ag surfaces.

scanning electron microscope (SEM) was used for imaging and energy dispersive X-ray (EDX) mapping. Transmission electron microscopy (TEM) specimens were prepared using an FEI DB235 Ga+ ion focused ion beam operating at 30 keV. The lift-out sections were placed onto a carbon film held by a copper support grid. The structural and compositional data at the nanoscale were collected using an FEI Tecnai TF30, 300 keV TEM/STEM with an EDAX EDX detector. The EDX data were collected in spectrum image mode, with a 2048 channel spectrum collected for each pixel in the map. The compositional data were statistically analyzed using the Sandia AXISA software package.36

III. Results III.A. ZnO Growth on Different Molecular Templates. Figure 1a shows an example of directed ZnO nanorod growth on a hexadecanethiol template. For -CH3-terminated alkanethiols, we reproducibly found ZnO nanorod growth on bare Ag regions (squares or rectangles) only and no nanorod growth on the SAM-covered areas. The selectivity was high, with occasional isolated nanorods found on the SAM regions, which probably nucleate from defects in the molecular monolayer. Furthermore, Auger spectra taken on -CH3-terminated SAM regions (outside of the rectangles in Figure 1a) showed Zn signals at the noise level (Table 1 and Figure 4) even after 2 h of growth time. Thus, methyl groups very effectively inhibit the attachment of Zn2+ growth species on the surface. This is expected on the basis of (36) Kotula, P. G.; Keenan, M. R.; Michael, J. R. Microsc. Microanal. 2003, 9, 1–17.

5378 Langmuir, Vol. 24, No. 10, 2008

Hsu et al. Table 1. Atomic Percentages (from Auger Spectra) of Zn, O, S, and C on Various µCP SAM Surfaces after 2 h of ZnO Growth bare Ag -CH3 -COOH type A -COOH type B -SH

Figure 6. Time dependence of Zn signals (atomic percentage) taken from Auger spectra on bare Ag (black), -CH3 SAM regions (red), and type A (blue) and type B (green) for 10 µm × 10 µm areas. The error bars represent the standard deviation found in different areas.

the chemical inertness of the methyl end groups and agrees with previously published reports. In contrast, when -COOH-terminated alkanethiol SAMs were used as templates, two types of behavior (labeled A and B in Figure 1b) were observed. Most of the SAM regions had dense ZnO nanorods (type A behavior, Figure 2b, inset), albeit with slightly lower density than in bare Ag regions. However, near the edges of the microcontact printed patterns, as shown in Figures 1b and 2a, we often found regions that did not show any ZnO nanorods on the SAM regions (type B behavior). Figure 2a shows a low-magnification SEM image that clearly depicts type B regions mostly occurring at the edges of the stamped patterns. Figure 2b shows a high-magnification SEM image near a type A-B boundary on a -COOH pattern, depicting the density contrast of ZnO nanorods found in the type A and B regions, with the inset showing that the growth in the type A regions are 3D nanorods. The selectivity of ZnO nanorods in the type B areas using -COOH-terminated SAMs can be much higher than using -CH3-terminated alkanethiols. In the case of long contact time and using an old SAM solution, large areas of type B regions surrounding bare Ag regions (with ZnO nanorods) have been observed, as if the -COOH templates also inhibiting ZnO growth similar to -CH3 templates (Figure 2c).17 The apparent absence of ZnO nanorods on -COOH-terminated surfaces is surprising because, under the growth conditions (pH ≈ 6.7), the carboxylic groups are deprotonated and the -COO- should attract the positively charged Zn2+ species to the surface, favoring ZnO nucleation and growth.32 III.B. Spatial Distribution of Elements. To gain information on the distribution of individual elements on the surface, we employed Auger mapping using a Zn signal at 62 eV, a C signal at 272 eV, an O signal at 510 eV, and a S signal at 151 eV. Figure 3 shows the Auger maps for (a) Zn, (b) O, (c) S, and (d) C on µCP -CH3 (left) and -COOH (right) templates after 30 min of ZnO growth. For the -CH3 template (Figure 3, left column), Zn and O are higher in the bare Ag regions, indicating ZnO nucleation on Ag, with the higher C and S signals outside of the square clearly indicating the locations of SAMs. On the -COOH template (Figure 3, right column), Zn and O signals are higher outside of the square, i.e., on top of the -COOH SAM regions. The contrast in the S map between the bare Ag and the SAM areas is not obvious, except near the edges of the patterns, because ZnO has covered the SAM after 30 min of growth. These observations hold true for both type A and B regions on the -COOH templates. The results in Figure 3 clearly show that the distributions of Zn and O depend strongly on the end groups of the SAMs used as the templates. Figure 4 shows Auger spectra taken on bare Ag, a -CH3 SAM, and a type A and a type B -COOH SAM region after 2 h

Zn

O

S

C

14.6 ( 2.9 0.9 ( 0.3 18.5 ( 2.9 8.2 ( 1.3 4.4 ( 0.2

15.4 ( 2.1 0.8 ( 0.2 21.6 ( 3.7 9.1 ( 1.3 5.4 ( 0.3

0.9 ( 0.2 2.4 ( 0.1 1.5 ( 0.4 3.3 ( 0.2 5.2 ( 0.1

35.5 ( 2.7 62.3 ( 0.5 32.6 ( 3.2 48.3 ( 1.8 37.3 ( 0.6

of growth. Table 1 displays the corresponding quantitative elemental analyses on Zn, O, S, and C obtained from several Auger spectra. It is clear that enhanced Zn and O signals were found in all but the -CH3 SAM regions. The S signal in the bare Ag regions was in the noise, while definitive S signals were found on SAM regions for all molecules, indicating the presence of the molecular layer. The uptake of Zn at -COOH SAM regions occurred immediately after the sample was inserted into the growth solution. Auger images and spectra taken after immersing a sample in the growth solution for 5 min at room temperature showed a clear enhancement of Zn and O signals in the -COOH SAM regions. Figure 5 shows the (a) Zn and (b) S maps of a -CH3 (left column) and a -COOH (right column) pattern. At such a short reaction time, the S signal was clearly visible in both patterns, indicating the presence of the thiol binding groups, and hence the SAMs, in both cases. The strong correlation between the Zn and S maps on the -COOH pattern indicates that Zn is preferentially deposited on the -COOH region. The Zn signal is barely above the background on the bare Ag region on the -CH3 pattern for such a short growth time. This confirms that the carboxylic end groups attract Zn growth species, even more than Ag, probably due to the favorable interaction between -COO- on the surface and Zn2+ ions in the growth solution. Figure 6 shows the time dependence of Zn signals on the different surfaces. While the type B regions showed an uptake of Zn faster than type A, the Zn signal in the type B regions ceased to increase after ∼45 min. After 1 h of growth, when nanorods could be clearly observed, the Zn signal in the type B regions became substantially smaller than in type A areas. The higher ZnO density found in the bare Ag regions surrounded by type B regions in Figure 1b was the result of enhanced local supersaturation arising from no additional consumption of Zn species by the type B surfaces. In summary, we found that while type B regions of -COOH-covered surfaces contained no ZnO nanorods, similar to -CH3 surfaces, Zn and O were present on the type B regions, unlike the -CH3 surfaces. III.C. Characterization of Thin ZnO Films. High-resolution Auger and EDX spectrum imaging revealed that the Zn signal was uniformly distributed on the type B -COOH-terminated SAM regions. The increase in Zn signal was always accompanied by an increase in O signal, suggesting the presence of Zn-O compounds. A logical conclusion is that, on the type B regions of -COOH templates, ZnO takes on a 2D thin film, rather than a 3D nanorod, morphology. Auger profiling showed that the Zn signal dropped to zero within 15 Å, while the Ag signal was about 100% at this depth, indicating that the thickness of this ZnO layer was ∼1 nm. (Figure 7) This 2D film thickness is very small compared to the length of 3D nanorods, which is typically 500 nm. In addition to Auger, a Zn signal was also clearly found at the Ag surfaces in EDX spectra taken on the cross-sectional TEM samples (Region 1 in Figure 8) and on planar surfaces using scanning electron microscopy. A Zn signal was not found in the bulk of Ag substrate (Region 2 in Figure 8) or on -CH3 surfaces. However, under high-resolution TEM, we did not consistently find a layer of ZnO on the SAM surfaces, most

Zinc Oxide Growth Morphology on SilVer Surfaces

Figure 7. Auger profiling of a -COOH type B region. Zn signals (red solid circles) use the left and Ag signals (blue open squares) use the right axis.

Langmuir, Vol. 24, No. 10, 2008 5379

Figure 9. Cross-sectional TEM images of ZnO films on -COOHmodified Ag surfaces (a) type A region and (b) type B region. Table 2. Contact Angle (deg) Measured with Water and Methylene Iodine and Calculated Surface Free Energy (ergs/cm2) of Various SAM-Modified Ag Surfaces

bare Ag µCP -CH3 solution -CH3 µCP -COOH solution -COOH TFA solution -COOH µCP -SH solution -SH

Figure 8. (a) Cross-sectional TEM image of a -SH SAM templated sample. (b) EDX spectrum taken from Ag surface (blue box in a) showing clear Zn signal (blue curve) and that taken from deep inside the Ag film (red box in a) showing no Zn signal (red curve). The two curves are offset vertically for clarity.

likely because of the extremely small thickness of the ZnO layer. The thickness of the ZnO film is consistent with leveling off of the Zn Auger signal on the type B regions at long growth time (Figure 6). In some samples, the ZnO films were thicker. In this case, we can clearly identify the films as ZnO from the fringe spacings in the high-resolution TEM images (Figure 9). These films showed the ZnO lattice spacing (5.20 Å) perpendicular to the substrate, indicating c-axis-oriented films, the same orientation as the ZnO nanorods.37 In type A regions, we observe direct growth of ZnO nanorods on Ag surfaces with no additional materials, e.g., a ZnO or AgOx film, at the interface.38 There were also no detectable ZnO films in between the nanorods.

IV. Discussion To understand the two types of growth behaviors we observed on the µCP -COOH SAM surfaces, we examined the properties (37) Scrymgeour, D. A.; Sounart, T. L.; Simmons, N. C.; Hsu, J. W. P. J. Appl. Phys. 2007, 101, 014316. (38) Floro, J. A.; Michael, J. R.; Brewer L. N.; Hsu, J. W. P. Unpublished work.

water

methylene iodine

surface energy

101.8 ( 3.0 105.6 ( 0.4 106.2 ( 0.7 94.5 ( 1.8 61.8 ( 3.1 32.7 ( 0.5 84.8 ( 1.0 58.6 ( 0.6

61.6 ( 0.5 69.2 ( 0.8 70.8 ( 0.9 51.0 ( 4.1 38.7 ( 0.9 35.5 ( 0.2 40.2 ( 4.8 20.4 ( 0.2

27.7 ( 0.3 23.4 ( 0.6 22.4 ( 0.9 33.7 ( 2.3 47.6 ( 1.9 64.2 ( 0.4 39.8 ( 3.0 53.8 ( 0.4

of the SAMs made by µCP, as well as by conventional solution deposition, using Auger spectra mapping, contact angle, and ellipsometry measurements. Since the latter two SAM characterization techniques are not conducive to probing micrometerscale variation, we produced µCP samples using large (∼1 cm2) flat stamps. The surface free energy of each SAM-modified silver substrate was determined from the contact angles of water and methylene iodine, based on the procedures outlined by Owens and Wendt.39 For SAMs with methyl end groups, the wetting properties of the monolayers were independent of SAM formation methods (Table 2). However, µCP produced thinner SAMs (1.47 vs 2.14 nm),40 suggesting the density of molecules on the surface was lower for SAMs made by µCP. This is to be expected; Love et al. reported that a contact time of 1 h was required to produce an alkanethiol SAM with coverage comparable to solutiondeposited monolayers.41 However, for producing micropatterns in µCP, a short contact time is necessary because the diffusion of molecules during a long contact time causes smearing of the patterns. For both µCP and solution-deposited -CH3 SAM surfaces, very sparse ZnO nanorods were found. These surfaces also showed lower surface energies compared to bare Ag. We believe that the reason for inhibition of ZnO growth on the -CH3 surface is due to the chemical inertness of the methyl group. The isolated and rare ZnO nanorods probably nucleated on Ag from defect sites of the SAM. Solution-deposited 11-mercaptoundecanoic acid SAMs produced hydrophilic surfaces, as expected. In particular, the TFA method resulted in the most hydrophilic surfaces with SAM thicknesses ∼10% less than the conventional ethanolonly method (1.52 vs 1.69 nm). The S signals from Auger spectra on the -COOH SAMs made by these two solution (39) Owens, D. K.; Wendt, R. C. J. Appl. Polym. Sci. 1969, 13, 1741–1747. (40) The thicknesses of solution-deposited 1-hexadecanethiol agreed with previously reported value on Ag in Laibinis, P. E.; Whitesides, G. M.; Allara, D. L.; Tao, Y.-T.; Parikh, A. N.; Nuzzo, R. G. J. Am. Chem. Soc. 1991, 113, 7152–7167. (41) Love, J. C.; Wolfe, D. B.; Haasch, R.; Chabinyc, M. L.; Paul, K. E.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 2003, 125, 2597–2609.

5380 Langmuir, Vol. 24, No. 10, 2008

deposition methods were comparable. It is well known that alkanethiol with carboxylic end groups tend to form bilayers due to hydrogen bonding of -COOH groups.34,42 Our results are consistent with previous report of TFA minimizing the formation of bilayers.34 In contrast, stamping 11-mercaptoundecanoic acid using µCP produced moderate hydrophobic surfaces. Ellipsometry data showed that the thickness ratio for the 11-mercaptoundecanoic acid (1.03 nm µCP vs 1.52 nm TFA) was comparable to that for the 1-hexadecanethiol, indicating similar SAM coverage for both µCP -CH3 and -COOH SAMs. The S Auger signal taken on the µCP -COOH SAMs (2%) was about half of that obtained from solution-deposited -COOH SAMs (4%) but comparable to S signal on µCP -CH3 SAMs (see next paragraph). However, the surface energy of the µCP -COOH SAMs was substantially smaller (33.3 ergs/cm2 µCP vs 64.2 ergs/cm2 TFA). Combined with ellipsometry and Auger results, this suggests that the µCP -COOH SAMs are significantly more disordered than µCP -CH3 SAMs, with -COOH end groups buried in the SAM. The growth of ZnO on the µCP -COOH blanket SAMs was found to be 3D nanorods. On the solution-deposited -COOH SAMs, many fewer rods were observed but strong Zn and O Auger signals were detected everywhere, suggesting formation of 2D ZnO thin films on the entire surface. To understand the difference between type A and B regions on µCP -COOH templates, we rely on Auger mapping and our knowledge of large area SAMs. The S signal in the µCP -CH3 region was found to be ∼2% over many samples and did not depend on growth time. Assuming that the S signal is proportional to the amount of molecules on the surface, we will take the 2% S signal to indicate a monolayer of alkanethiol on surfaces produced by µCP. The S signal of the µCP -COOH SAM in type A region was also ∼2%, indicating the molecular density of the two SAMs produced by µCP was similar. This is also consistent with the proportional thicknesses of the two SAMs. As seen in Table 1, the S signal in type B regions was twice of that in the type A regions and substantially higher than the 2% monolayer signal, suggesting the presence of a molecular bilayer rather than monolayer.43 The factor-of-two difference between type B and A regions was observed on all samples independent of growth time so that ZnO coverage could not be the explanation for the lower S intensities observed on type A regions. This is consistent with the observation that these regions were often found near the edges of the stamp (Figures 1b and 2a), where the molecular concentration was expected to be higher due to nonuniform drying in the spin-coating process. While we believe that bilayer formation is the explanation based on knowledge of alkanethiol with -COOH end groups and the Auger and ellipsometry results, we cannot absolutely rule out the possibility of a higher density of 11-mercaptoundecanoic acid molecules in type B regions. The difference in the growth mode of ZnO in these two different regions is not due to local concentration or pH variations because we always observed type B regions near the edges/corners of the patterns (Figure 2a) regardless of where the patterns were related to the solution depth, sample size, or vial size. Since the density of nanorods on the type A regions was comparable to that on bare Ag while ellipsometry and Auger data suggested a good SAM coverage, nanorod growth on Ag from SAM defect sites was not the primary reason. Hence, -COOH SAM supports ZnO growth. (42) Arnold, R.; Azzam, W.; Terfort, A.; Wo¨ll, C. Langmuir 2002, 18, 3980– 3992. (43) The S signal in the type A -COOH SAM regions is smaller than that of -CH3 SAM regions because Auger is a highly surface sensitive technique and the -COOH SAM are covered by ZnO.

Hsu et al.

When bilayers of 11-mercaptoundecanoic acid formed, the end groups that were exposed to the growth solution were -SH rather than -COO-. To test that -SH end groups change the ZnO morphology, we grew ZnO on octanedithiol templates. The S signal for the µCP -SH SAMs was ∼5%, consistent with a SAM of the same molecular density as µCP hexadecanethiol but with two sulfurs. Surface energy of the µCP octanedithiol SAM was 40 ergs/cm2, higher than the µCP -COOH SAMs. Similarly to the type B regions of the -COOH-terminated SAMs, no ZnO nanorods but clear Zn signals were observed in the octanedithiol patterns (Table 1), consistent with the presence of ZnO thin films. One explanation of 2D growth morphology found on -SH surfaces might be the higher bond energy for Zn-S (205 kJ/mol) compared to that for Zn-O (159 kJ/mol).44 Hence, planar formation of the nucleation layer on the thiol end groups of SAMs could be favored due to strong Zn-S interaction that effectively stabilizes the ZnO {0001} surfaces. However, this would not explain the 2D growth mode found on solutiondeposited -COOH SAMs. Instead, a more general explanation is that lower-energy surfaces (bare Ag, type A in µCP -COOH SAMs) favor a 3D growth mode with high densities of ZnO nanorods while 2D ZnO thin films are the preferred growth morphology on high-energy surfaces (solution-deposited -COOH SAMs, type B in µCP -COOH SAMs, and -SH SAMs). Our results are in agreement with ALD deposition of TiO2 on mixed SAMs with 3D growth observed for surface with free energies below 40 ergs/cm2.20 Hence, surface free energy can also be applied to understand growth morphology from supersaturated solutions. Given the different growth conditions, the similar results observed in supersaturated solution growth and ALD indicate that heteronucleation on surfaces is indeed dominated by the surface properties. Previously, we reported that -COOH SAMs formed an effective template to direct ZnO rod growth only in the bare Ag regions (see Figure 2c). A hypothesis of HMT-H+ complexes competing with Zn2+ ions for binding to the -COO- groups was proposed.17 Given what we learned in this study, we now believe that the true explanation for -COOH SAMs as a seemingly efficient blocker to ZnO nanorod growth is the formation of a ZnO thin film on the -COOH surfaces, i.e., type B regions surrounding bare Ag areas as shown in Figure 1b and 2c. The experiments in ref 17 were done with longer contact times and older thiol solutions, more likely leading to the formation of a -COOH bilayer. The observation that presoaking in Zn(NO3)2 solution prior to growth led to ZnO nanorod growth on -COOH regions, i.e., pattern reversal, was probably due to the removal of weakly bound molecules and producing -COO- groups that are readily available for Zn2+ adsorption.45 Similarly, Morin et al. reported that -COOH groups created using UV light on polycarbonate and polyester surfaces blocked ZnO nanorod nucleation. The authors searched but did not find surface-bound HMT. The paper did not report on the possible presence or absence of a thin ZnO film through microscopy techniques.19 While the possibility of forming a -COOH bilayer or -SH end groups on the polymer surfaces does not exist, the -COOH-containing polymer surfaces could still have high enough surface energy (which was not reported in ref 19) to favor 2D ZnO film growth. The lack of 3D ZnO nanorods on the 2D ZnO films suggests that inhomogeneities play an important role in heteronucleation of nanocrystals beyond the binding of growth species on surfaces suggested previously.5,14,32 (44) Handbook of Chemistry and Physics, 84th ed.; Lide, D. R., Ed.; CRC Press: Boca Rotan, FL, 2003; Section 9, pp 52-64. (45) Jun, Y.; Zhu, X.-Y. J. Am. Chem. Soc. 2004, 126, 13224–13225.

Zinc Oxide Growth Morphology on SilVer Surfaces

V. Conclusions We find that nucleation and growth morphlogy of ZnO on surfaces are governed by the surface free energy of the substrate, which can be altered by SAMs with different end groups or formed by different methods. Alkanethiol SAMs with methyl end groups form a chemically inert surface, resulting in inhibition of nucleation and growth of ZnO. On the basis of chemical activity, surfaces terminated with carboxylic end groups are expected to support nucleation of ZnO. This is indeed observed, but the growth morphology of ZnO on the carboxylic surfaces is more complex. Low-energy surfaces (type A regions of µCP SAMs) favor 3D growth (ZnO nanorods), while 2D ZnO thin films are the common morphology on high-energy surfaces (solutionformed unpatterned SAMs and type B regions of µCP SAMs). Alkanedithiol SAMs with thiol end groups, regardless of

Langmuir, Vol. 24, No. 10, 2008 5381

deposition methods, produce fairly high energy surfaces that favor 2D growth of ZnO thin films. Our results show that surface free energy has a strong influence on growth mode in supersaturated solution growth, similarly to ALD growth reported previously.20 Acknowledgment. We thank N. Simmons and B. McKenzie for the SEM work, M. Rye for TEM specimen preparation, and J. Voigt, B. Bunker, and K. Leung for helpful discussion. This work was supported by the Sandia LDRD program and the DOE Office of Basic Energy Sciences. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under Contract No. DE-AC04-94AL85000. LA703919W