Tuning the number density of nanoparticles by

1 downloads 0 Views 371KB Size Report
Dec 15, 2006 - Department of Chemical and Biomolecular Engineering, North Carolina .... solutions with varying pH but same ionic strength (∼10 mM). ... (1 μm × 1 μm) by manual counting. ... pdf and [18]. ... (This figure is in colour only in the electronic version) ..... [12] Atkins P W 1993 The Elements of Physical Chemistry.
INSTITUTE OF PHYSICS PUBLISHING

NANOTECHNOLOGY

Nanotechnology 18 (2007) 025301 (6pp)

doi:10.1088/0957-4484/18/2/025301

Tuning the number density of nanoparticles by multivariant tailoring of attachment points on flat substrates Rajendra R Bhat1 and Jan Genzer Department of Chemical and Biomolecular Engineering, North Carolina State University, Raleigh, NC 27695-7905, USA E-mail: Jan [email protected]

Received 27 September 2006, in final form 26 October 2006 Published 15 December 2006 Online at stacks.iop.org/Nano/18/025301 Abstract We report on the organization of nanoparticles on a flat surface when there is strong yet tunable interaction between the particles and the surface. Specifically, we tailor the number density of citrate-stabilized gold nanoparticles on flat substrates by varying the concentration of the grafted amino groups on the surfaces and their degree of ionization. While the former effect is accomplished by decorating silica-based substrates with a molecular gradient of (3-aminopropyl)triethoxysilane (APTES), the latter effect is achieved by varying the degree of ionization of the −NH2 groups in APTES by varying the pH of the gold sol. We show that the measurement of particle number density on an APTES concentration gradient substrate at different pH values provides a simple, non-spectroscopic means to deduce the relative molecular concentration profile of APTES on the substrate.

one can use structured surfaces comprising chemically welldefined regions, which may govern the assembly of nanosized objects [3]. The chemical patterns on such surfaces can either comprise patterns of various shapes and dimensions having tunable wettabilty separated by sharp boundaries [4] or can be prepared such that the affinity to the nanosized adsorbates changes gradually as a function of the position on the substrate [5, 6]. In our previous paper [7] we demonstrated that a number density gradient of adsorbed nanoparticles can be prepared by first forming a one-dimensional molecular gradient of amino groups (−NH2 ) on the substrate by vapour deposition [8] of amine-terminated (3-aminopropyl)triethoxy silane (APTES) molecules, followed by attachment of gold nanoparticles to −NH2 functional groups by immersing the substrate in a colloidal gold solution. When carrying out the experiments at moderately low pH (≈6), the nanoparticles bind to the substrate due to strong electrostatic interactions acting between the negatively charged citrate groups present on the surface of gold nanoparticles and positively charged −NH+ 3 groups on the substrate. We demonstrated that by varying the density of the −NH+ 3 groups within the gradient, a particle number density (PND) gradient of gold nanoparticles was formed. PND

1. Introduction The emerging fields of nanoscience and nanotechnology are leading to unprecedented understanding of the structure and function of the fundamental building blocks of all physical entities. These developments are likely to change the way almost everything—from vaccines to computers to automobile tyres to objects not yet imagined—is designed and made. An important aspect of the aforementioned activities involves understanding how to design and build materials from nanosized objects by controlling their assembly in space. Over the past few years, many ingenious technologies have been conceived and explored that facilitate the organization of nanosized objects in two and three dimensions [1]. To this end, nanoparticle assembly in 3D patterns can be accomplished by utilizing chemically modulated matrices, such as multiphase polymer blends or melts of amphiphilic block copolymers, the latter forming periodic structures on different length and time scales [2]. Assembly of nanoobjects in 2D structures requires adopting different strategies, however. Specifically, 1 Present address: Becton Dickinson Technologies, Research Triangle Park,

NC 27709, USA.

0957-4484/07/025301+06$30.00

1

© 2007 IOP Publishing Ltd Printed in the UK

Nanotechnology 18 (2007) 025301

R R Bhat and J Genzer

order to generate a large number of hydroxyl groups, which are required for coupling silane molecules. A mixture of APTES and paraffin oil (PO) (1:1 w/w) was prepared and kept in a small rectangular container. This container was placed near the shorter edge of the UVO-treated silicon wafer and the whole system was enclosed in a Petri dish at ambient conditions. After a predetermined period of time the wafer was taken out of the container and washed with ethanol to remove physisorbed silane molecules. The APTES concentration gradient was characterized by measuring contact angle at various positions along the substrate. The substrate was subsequently immersed in an aqueous gold nanoparticle solution for 24 h, following which it was washed thoroughly with DI water and dried with nitrogen. In order to determine the number density of adsorbed gold nanoparticles, atomic force microscopy (AFM, Nanoscope III, Digital Instruments) tapping mode scans were taken at various positions along the longer side of the substrate For each x position, three scans were (x direction). taken along the transverse direction ( y direction). The particle densities were determined from the AFM micrographs (1 μm × 1 μm) by manual counting. An average of three transverse measurements was taken for each x position. To determine the concentration profile of grafted APTES silane, we used near-edge x-ray absorption fine structure (NEXAFS) spectroscopy [17]. The NEXAFS experiments were carried out on the NIST/Dow materials characterization end-station at the National Synchrotron Light Source at Brookhaven National Laboratory (NSLS BNL)2 .

profiles of the gold nanoparticles on the substrate deduced from atomic force microscopy experiments were shown to coincide with the concentration of APTES on the surface, measured by combinatorial near-edge x-ray absorption fine structure (NEXAFS) spectroscopy [9]. As the interaction between particles and APTES molecules is thought to be electrostatic in nature, we expect pH and ionic strength of the gold colloid solution to substantially influence the nanoparticle organization on monolayer surfaces [10]. In this paper we study systematically the effect of pH of the gold sol on the PND on the surface. We demonstrate that the PND profiles obtained at various pH deposition conditions can be scaled onto a single master curve and this scaling provides an alternative means of obtaining information about the concentration profile of APTES on the surface.

2. Materials and procedures 2.1. Preparation of gold nanoparticle solution Hydrogen tetrachloroaurate (III) trihydrate (HAuCl4 ·3H2 O), trisodium citrate dihydrate and (3-aminopropyl)triethoxysilane (APTES) were obtained from Aldrich and used as received. Concentrated HNO3 and HCl were purchased from Fisher Chemicals. Deionized (DI) water (resistivity >16 M cm) was produced using the Millipore water purification system. Silicon substrates with a ≈2 nm thick layer of native SiOx were purchased from Virginia Semiconductors. All glassware was washed with freshly prepared aqua regia solution (3:1 HCl:HNO3 ) and rinsed thoroughly with DI water. An aqueous solution of gold nanoparticles was prepared by citrate reduction of HAuCl4 following the method of Frens [15, 16]. The diameter of the particles was determined to be 16.8 ± 1.8 nm by transmission electron microscopy. To study the effect of pH on nanoparticle binding, deposition of particles was carried out using gold sol having different pH values. Since ionic strength also affects particle interaction with APTES SAM, we prepared gold nanoparticle solutions with varying pH but same ionic strength (∼10 mM). For this, we first centrifuged 17 nm gold sol at 8000 rpm for 40 min. After centrifugation, the colourless supernatant was carefully decanted and the remaining colloid increased its concentration by a factor of 7. This concentrated solution was used as the stock colloid in the present study. Aliquots of the stock colloid were redispersed in citrate buffered solutions maintained at different pH values but same ionic strength (∼10 mM). The concentration and ionic strength of the colloids were intentionally kept identical to facilitate unambiguous study of the pH effects. Ultraviolet–visible light (UV–vis) absorbance intensity at 520 nm, where the surface plasmon excitation band of nanosized gold particles appears, was kept at the same level for each buffered colloid so that all the colloids will have a nearly identical particle concentration (about 1.03 × 1012 particles ml−1 ).

3. Results and discussion In figure 1(a), we plot the particle number density (PND) of gold nanoparticles attached to an APTES self-assembled monolayer (SAM) having a homogeneous concentration of the APTES molecules upon carrying out particle deposition from gold solutions having different pH. The PND of surfacebound particles decreases upon increasing the solution pH. A dramatic reduction in the number of particles bound to the APTES SAM is observed in the pH range of 7–10, with the adsorption profile resembling a ‘reverse S-curve’. We explain this behaviour by considering the ionization behaviour of APTES on the substrate surface. The degree of ionization of primary amine groups in APTES at a given solution pH is described by their surface p K 1/2 , i.e. the pH at which half of the surface groups become ionized. Previous experiments using chemical force microscopy and contact angle measurements revealed that p K 1/2 of amino groups in APTES is ≈7.6 [11]. With this value of p K 1/2 , a simple calculation based on the Henderson–Hesselbach (H–H) equation [12]:

f NH+3 = [1 + 10pH−p K 1/2 ]−1 ,

(1)

where f NH+3 is the fraction of the −NH+ 3 groups in the APTES SAM, indicates that the rapid change in ionization state of

2.2. Creation and confirmation of PND gradient using molecular gradient template

2 For detailed information about the NIST/Dow Soft X-ray Materials Characterization Facility at NSLS BNL, see the November 1996 newsletter of NSLS at: www.nsls.bnl.gov/newsroom/publications/newsletters/1996/96-nov. pdf and [18].

The silicon wafer was cut into rectangular pieces (4.5 cm × 1.2 cm) that were exposed to ultraviolet/ozone (UVO) treatment for 30 min (Jelight Company, Inc., model 42) in 2

R R Bhat and J Genzer

amino groups in APTES occurs in the pH range of 6–9 (cf figure 1(b)). The dissociation behaviour of APTES is also ‘reverse S-shaped’, with −NH2 groups fully protonated below pH = 6 and totally neutral above pH = 9.5. Thus, decrease in PND is due to the decrease in the number of positively ionized amino groups as pH is increased. However, the inflection point of the particle ‘S-curve’ (≈9.4) shown in figure 1(a) is higher than the reported pK1/2 of APTES (≈7.6). The higher value of the inflection point sensed via particle binding is due to the large disparity between the sizes of gold nanoparticles and −NH+ 3 groups. If the particles were of the same dimension as the −NH+ 3 moieties, the experimentally measured p K 1/2 profile would correspond to the ‘true’ p K 1/2 profile of the −NH+ 3 groups. With increasing particle size, the position of the inflection point in the p K 1/2 profile shifts towards higher pH as a smaller number of surface-bound −NH+ 3 groups is needed to achieve 50% coverage of the surface by nanoparticles. The same arguments can be used to explain why the particle ‘S-curve’ (cf figure 1(a)) is spread over a narrower range of pH compared to the ionization profile of surface-grafted APTES (cf figure 1(b)). In order to fully understand the influence of pH on the assembly of particles on the APTES surface, we carried out particle deposition on APTES gradient surfaces at various pH. The procedure leading to the formation of such gradients is shown pictorially in figure 2 and described in detail in the Materials and procedures section. In figure 3 we plot PND profiles obtained at different pH values. The profiles exhibit the effect of two parameters simultaneously: (1) grafting

-2

Particle number density ( μm )

Nanotechnology 18 (2007) 025301

500 400 300 200 100

a) 0

+

Fraction of NH3 groups

1.0 0.8 0.6 0.4 pK1/2,NH

2

0.2 0.0

b) 5

6

7

8

9

10

11

solution pH

Figure 1. (a) Particle number density (PND) as a function of pH measured by exposing substrates covered with (3-aminopropyl)triethoxy silane (APTES) self-assembled monolayer (SAM) to gold nanoparticle (particle diameter ≈17 nm) aqueous solutions of various pH. The line is meant to guide the eye. (b) Fraction of −NH+ 3 groups as a function of solution pH calculated using the Henderson–Hesselbach equation.

Figure 2. Schematic illustrating the formation of gradients comprising number density (PND) of gold particles on solid substrates. First, a molecular gradient of (3-aminopropyl)triethoxy silane (APTES) is formed on a silica-covered substrate by utilizing the vapour deposition method described in the text. Such an APTES molecular gradient is immersed into a gold sol of a given pH. Negatively charged citrate groups present on the surfaces of the gold nanoparticles anchored the nanoparticles to positively charged surface-bound APTES molecules. (This figure is in colour only in the electronic version)

3

Nanotechnology 18 (2007) 025301

R R Bhat and J Genzer 500 solution pH 6.57 7.16 9.08

-2

400

Particle number density (μm )

-2

Particle number density ( μm )

500

300

200

100

0

0

5

10

15

20

400

300

200

100

0

25

Position on substrate (mm)

Position on substrate (mm) x=5 x=7 x=10 x=12 x=13 x=15 x=17 x=18 x=19 x=20

6

7

8

9

10

solution pH

Figure 3. Particle number density (PND) of gold nanoparticles measured at various positions along the (3-aminopropyl)triethoxy silane (APTES) gradient substrate. The figure depicts data collected from three different specimens, each of them prepared by exposing the substrate containing the APTES molecular gradient to a gold sol having a different pH, equal to: ( ) 6.67, ( ) 7.16, and () 9.08. The lines are meant to guide the eye.

Figure 4. Particle number density (PND) as a function of solution pH at different positions along the gradient substrate, x (= the distance in mm from the edge of the substrate that was closest to the APTES diffusion source). The data was taken from figure 3. The lines are meant to guide the eye.



distance from the edge of the sample that was closest to the APTES diffusing source. In contrast, at pH = 6.57, the same PND is observed at x = 18 mm. In order to achieve the same PND, the increase in the fraction of the charged groups on the surface, f NH+3 (due to the decrease in pH), has to be compensated by decreasing the overall grafting density of the surface-bound −NH2 groups, σNH2 . Hence a simple relationship should exist between the PND, f NH+3 , and σNH2 :

density of amino groups on the surface and (2) charged state of those groups. For a given pH, the PND decreases along the gradient because the density of amino groups with which gold particles bind diminishes as one moves along the gradient (in the direction of increasing values on the abscissa in figure 3). At pH = 6.57, amino groups exist predominantly as −NH+ 3 moieties; hence, they promote the attachment of negatively charged gold nanoparticles by strong electrostatic interaction. At pH = 7.16, a value that is closer to p K 1/2 , amino groups are less protonated than those at pH = 6.57, leading to the decrease in overall attractive electrostatic interactions between surface and particles. At pH = 9.08, an even smaller number of −NH2 groups is protonated, thus causing drastic reduction in electrostatic interactions between particles and the surface, manifested by a significantly smaller nanoparticle density on the surface. The ionization state of gold nanoparticles will have only modest influence on the particle attachment in the pH range studied because gold particles prepared by the citrate reduction method have an isoelectric point of about 2 [13]. Electrophoretic measurements on citrate-stabilized gold sol [13] as well as H–H calculation on citrate molecules3 indicate that there is little change in the charged state of particles in the pH range explored. Thus, our observations confirm the electrostatic nature of interaction between gold colloids and APTES monolayer. The second effect seen in the data in figure 3 is the interplay between the grafting density of the surface bound −NH2 groups and the fraction of those groups that are ionized. In order to illustrate this point, consider, for instance, a constant PND equal to 200 particles μm−2 . At pH = 9.08, PND equal to 200 particles μm−2 is detected at x = 8 mm, where x is the position along the gradient measured as a

PND ≈ f NH+3 · σNH2 .

(2)

In order to further comprehend the coupling between PND, f NH+3 , and σNH2 , we make one simple assumption. Namely, in the region of pH 6–9.5, where f NH+3 changes most rapidly, we can assume that f NH+3 decreases linearly with increasing pH (cf figure 1). We stress that this assumption is valid only within a narrow range of pH, corresponding to moderate values of f NH+3 . At very small and/or very large fNH+3 , equation (2) would change. Invoking the aforementioned reciprocal relationship between f NH+3 and σNH2 allows us to collapse the PND data collected at different pH and different positions along the gradient on a master curve. For this, we re-plot the data in figure 3 with PND at various positions on the APTES gradient substrate as a function of the pH of gold sol (cf figure 4). The fact that the PND versus pH exhibits approximately equal slopes (with the exception of the positions on the substrates that are close to the foot of the gradient, x > 18 mm) supports the aforementioned assumption that PND ∼ f NH+3 . We then horizontally move the data points corresponding to all the positions, except x = 5 mm, so as to follow the particle density trend observed at x = 5 mm (i.e. the position with the highest measured particle density). The data collapse onto a master curve, which is depicted in figure 5. By shifting the data points in figure 4 along the pH axis so as to match the trends observed at x = 5 mm, we artificially decrease f NH+3 in order to bring σNH2 of all the positions on the substrate to the value present at x = 5 mm, while at the same time keeping the PND unchanged (the ordinate in

3 In calculating the percentage of charged groups on the surface of gold nanoparticles, we assumed that the charge on gold particles is mainly due to the adsorbed citrate anions. Previous studies (see [14]) have shown that citrate anions are strongly adsorbed on gold surfaces.

4

Nanotechnology 18 (2007) 025301

R R Bhat and J Genzer

400

300

200

0

11

1 10

ΔpH

-2

Position on substrate (mm) x=5 x=7 x=10 x=12 x=13 x=15 x=17 x=18 x=19 x=20

2 9 3

100

Δ pH

4

8

NEXAFS

0

5

6

7

8

9

10

11

12

13

2

14

4

6

8

10

12

14

16

18

20

NEXAFS PEY from N-H bond (a.u.)

Particle number density ( μm )

500

22

Position on substrate (mm)

solution pH

Figure 6. pH shift (left ordinate, ) and concentration of N–H bonds obtained from near-edge x-ray absorption fine structure ) as a function of the position along spectroscopy (right ordinate,  the (3-aminopropyl)triethoxy silane (APTES) molecular gradient on silica substrate. The pH values were obtained by horizontally shifting particle number density (PND) data shown in figure 4 in order to collapse them onto a single ‘master curve’ shown in figure 5.

Figure 5. Particle number density (PND) as a function of solution pH. The points were obtained by horizontally shifting the data depicted in figure 4 by a certain pH value (cf figure 6). The line is meant to guide the eye.

figure 5 remains the same as in figure 4). Upon increasing x on the substrate, the grafting density of APTES, and hence σNH2 , decreases. However, as we are projecting the σNH2 at all points on the substrate to be as high as that existing at x = 5 mm, the fraction of charged amino groups, which are ultimately responsible for particle binding, must decrease in order to compensate for the increased grafting density. This is manifested in the form of an increase in pH of the solution. The extent of pH shift (pH), which represents the increase in pH (or decrease in fNH+3 ) required to maintain the same PND, would then represent a measure of the variation of σNH2 on the substrate. In figure 6 we plot pH as a function of the position on the substrate. In the same plot we present the APTES concentration profile determined via NEXAFS. The agreement between the two profiles is not a coincidence. The pH versus position profile represents a relationship between pH of gold sol and grafting density of APTES on the substrate. As mentioned earlier, in order to maintain a given PND, one can vary either the overall grafting density of APTES molecules or the fraction of APTES molecules that are charged. If grafting density of APTES increases then pH of nanoparticles sol has to increase (i.e. f NH+3 decreases) in order to maintain a given PND value. It is also worth noticing that the inflection point of the master curve is pH ≈ 9, which agrees well with that obtained from particle number density values measured on a homogeneous APTES SAM (cf figure 1(a)).

molecules, the particle number density decreases upon increasing the pH of gold sol, thus revealing the concentration of charged amino molecules as the main parameter dictating the extent of particle attachment. The coupling between grafting density of APTES molecules on the surface and their charged state was illustrated by forming number density gradients of nanoparticles at different pH values. The pH of the nanoparticle sol required to achieve a given particle density was found to be inversely related to the grafting density of anchoring molecules on the substrate. Simultaneous variation of grafting density of APTES molecules in the molecular gradient and pH of gold sol enabled us to indirectly derive the concentration profile of APTES molecules in the gradient.

Acknowledgments This research was funded by the National Science Foundation, Grant Nos. CTS-0403535 and EEC-0403903. NEXAFS experiments were carried out at the National Synchrotron Light Source, Brookhaven National Laboratory, which is supported by the US Department of Energy, Division of Materials Sciences and Division of Chemical Sciences. We thank Professor Gregory N Parsons (NCSU) for allowing us to use his AFM. We also thank Dr Daniel A Fischer (NIST/BNL) and Dr Kirill Efimenko (NCSU) for their assistance during the course of the NEXAFS experiments.

4. Conclusion We elucidated the principles governing the two-dimensional organization of nanoparticles on self-assembled monolayer templates that exhibit strong yet tunable interaction between the particles and the substrate. By employing a concentration gradient template of surface-grafted APTES molecules capable of strongly binding to gold nanoparticles, we demonstrated a direct relationship between the coverage of the surface by particles and concentration of APTES molecules on the surface. For a given concentration of grafted APTES

References [1] Shipway A N, Katz E and Willner I 2000 ChemPhysChem 1 18 [2] Krishnamoorty S, Hinderling C and Heinzelmann H 2006 Mater. Today 9 40 [3] Shenhar R, Norsten T B and Rotello V M 2005 Adv. Mater. 17 657 [4] Ciebien J F, Clay R T, Sohn B H and Cohen R E 1998 New J. Chem. 685

5

Nanotechnology 18 (2007) 025301

R R Bhat and J Genzer

[5] Bhat R R, Genzer J, Chaney B N, Sugg H W and Liebmann-Vinson A 2003 Nanotechnology 14 1145 [6] Bhat R R, Tomlinson M R, Wu T and Genzer J 2006 Adv. Polym. Sci. 198 51 [7] Bhat R R, Fischer D A and Genzer J 2002 Langmuir 18 5640 [8] Chaudhury M K and Whitesides G M 1992 Science 256 1539 [9] Genzer J, Fischer D A and Efimenko K 2003 Appl. Phys. Lett. 82 266 [10] Zhu T, Fu X Y, Mu T, Wang J and Liu Z F 1999 Langmuir 15 5197 [11] Zhang H, He H-X, Wang J, Mu T and Liu Z F 1998 Appl. Phys. A 66 S269 [12] Atkins P W 1993 The Elements of Physical Chemistry (New York: Freeman)

[13] Thompson D W and Collins I R 1992 J. Colloid Interface Sci. 152 197 [14] Chow M K and Zukoski C F 1994 J. Colloid Interface Sci. 165 97 Biggs S, Mulvaney P, Zukoski C F and Grieser F 1994 J. Am. Chem. Soc. 116 9150 Thompson D W and Collins I R 1992 J. Colloid Interface Sci. 152 197 [15] Frens G 1973 Nat. Phys. Sci. 241 20 [16] Turkevich J, Stevenson P C and Hiller J 1951 Discuss. Faraday Soc. 11 55 [17] St¨ohr J 1992 NEXAFS Spectroscopy (Berlin: Springer) [18] Genzer J et al 2000 Langmuir 16 1993 Genzer J et al 2002 Macromolecules 33 1882

6