Sulfate-Mediated End-to-End Assembly of Gold ... - ACS Publications

Jan 18, 2017 - Virginia Tech, Department of Civil and Environmental Engineering, ... Virginia Tech, Institute for Critical Technology and Applied Scie...
0 downloads 0 Views 3MB Size
Subscriber access provided by HACETTEPE UNIVERSITESI KUTUPHANESI

Article

Sulfate Mediated End-to-End Assembly of Gold Nanorods Seyyed Mohammad Hossein Abtahi, Nathan D. Burrows, Fred A. Idesis, Catherine J. Murphy, Navid B. Saleh, and Peter J. Vikesland Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.6b04114 • Publication Date (Web): 18 Jan 2017 Downloaded from http://pubs.acs.org on January 25, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

1

Sulfate Mediated End-to-End Assembly

2

of Gold Nanorods

3 4

S.M.H. Abtahi,1, 2, 3 Nathan D. Burrows,4 Fred A. Idesis,4 Catherine J. Murphy,4

5

Navid B. Saleh,5 and Peter J. Vikesland*1, 2, 3 1.

6

Blacksburg, VA, USA

7 8

2.

Virginia Tech, Institute for Critical Technology and Applied Science (ICTAS) Center for Sustainable Nanotechnology (VTSuN), Blacksburg, VA, USA

9 10

3.

Center for the Environmental Implications of Nanotechnology (CEINT), Duke University, Durham, NC, USA

11 4.

12

15

Department of Chemistry, University of Illinois at Urbana-Champaign, 600 S. Matthews Ave., Urbana, IL, 61801, USA

13 14

Virginia Tech, Department of Civil and Environmental Engineering,

5.

The University of Texas at Austin, Department of Civil, Architectural and Environmental Engineering, Austin, TX 78712, USA

16

ACS Paragon Plus Environment

1

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 33

17 18

Abstract

19

There is interest in the controlled aggregation of gold nanorods (GNRs) for the

20

production of extended nanoassemblies. Prior studies have relied upon chemical

21

modification of the GNR surface to achieve a desired final aggregate structure. Herein,

22

we illustrate that control of electrolyte composition can facilitate end-to-end assembly of

23

cetyltrimethylammonium bromide (CTAB) coated GNRs. By adjusting either the sulfate

24

anion concentration or the exposure time it is possible to connect GNRs in chain-like

25

assemblies. In contrast, end-to-end assembly was not observed in control experiments

26

using monovalent chloride salts. We attribute the end-to-end assembly to the localized

27

association of sulfate with exposed quaternary ammonium head groups of CTAB at the

28

nanorod tip. To quantify the assembly kinetics, visible-near infrared extinction spectra

29

were collected over a pre-determined time period and the colloidal behavior of the GNR

ACS Paragon Plus Environment

2

Page 3 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

30

suspensions was interpreted using plasmon band analysis. Transmission electron

31

microscopy and atomic force microscopy results support the conclusions reached via

32

plasmon band analysis and the colloidal behavior is consistent with Derjaguin-Landau-

33

Verwey-Overbeek (DLVO) theory.

34

Introduction

35

Gold and silver nanoparticles are of interest because of the unique electro-optical

36

properties (e.g., localized surface plasmon resonance [LSPR])1-5 that originate from the

37

collective behavior of their surface electrons. In the case of LSPR, visible-infrared light

38

interacts with a noble metal nanoparticle smaller than the incident radiation wavelength1,

39

6

40

conduction band electrons. The LSPR is a function of the local dielectric environment as

41

well as the size and shape of the nanoparticle.1,

42

extinction spectrum at the wavelength where the frequency of the incident light matches

43

the frequency at which the conduction electrons oscillate.1,

44

dependent on nanoparticle size and shape and shifts to longer wavelengths with an

45

increase in particle size.14-16 Anisotropic nanoparticles, such as gold nanorods (GNRs),

46

exhibit two different plasmon oscillations: one along the short axis (the transverse

47

oscillation) and the other along the long axis (the longitudinal oscillation).5, 17-21 Each of

48

these oscillations exhibits its own specific LSPR band. The transverse LSPR band

49

typically occurs at a wavelength between 500-550 nm and shifts to longer wavelengths

50

with an increase in diameter. The longitudinal LSPR band is aspect ratio (AR:

51

length/diameter) sensitive, and occurs at longer wavelengths for GNRs with higher

52

AR.19, 22, 23 The extinction (absorbance + scattering) of these peaks is linearly related to

and resonant interactions result in the coherent oscillation of the nanoparticle's

2, 7-12

ACS Paragon Plus Environment

A peak is observed in the

6, 8, 13

Peak location is

3

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 33

53

the concentration of nanoparticles in suspension. Changes in the number concentration,

54

the AR, or the local dielectric environment can be detected by a change in extinction or

55

by the shift of these two plasmon peaks.14, 24

56

The electro-optical properties of GNRs can be significantly altered by nanoparticle

57

assembly. We use the term controlled assembly to refer either to end-to-end, end-to-

58

side, or side-by-side assembly.25,

59

plasmon preferentially and are of particular interest for sensing technologies that rely on

60

large electric fields.26-28 Extended end-to-end GNR assemblies have previously been

61

produced by manipulation of GNR surface chemistry,29-32 by alteration of the chemistry

62

of the solution in which the GNRs are suspended,13, 30, 33, 34 by applying polystyrene and

63

surface enhanced Raman spectroscopy (SERS) markers,35 or by attachment of pH-

64

responsive DNA36 to the GNR surface. Side-by-side assemblies of GNRs have been

65

prepared by cysteine conjugated GNRs in the presence of lead37 and in adiptic acid

66

treated GNR suspensions.38 Interchangeable end-to-end and side-by-side assemblies

67

of GNRs have been prepared by addition of tetrahydrofuran,39 dithiol polyethylene

68

glycol,40 bifunctional poly(ethylene glycol, PEG),26 or antibodies,25 onto GNR surfaces

69

where the final assembled configuration is controlled by the concentration of the organic

70

molecule. Each of these approaches requires extensive nanoparticle functionalization.

26

End-to-end assemblies couple the longitudinal

71

In this work, we illustrate a simple method in which sulfate ions are used to facilitate

72

end-to-end GNR assembly in aqueous suspension. The coordinating activity of sulfate

73

anions enables conversion of >80% of well-dispersed GNRs in a suspension into

74

extended structures (chains of dimers, trimers, tetramers, etc.). For these experiments

75

we used cetyltrimethylammonium bromide capped GNRs (CTAB-GNRs) because CTAB

ACS Paragon Plus Environment

4

Page 5 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

76

is the de facto GNR capping agent in GNR synthesis5, 18, 41-44. Visible-near infrared (Vis-

77

NIR) extinction spectra were collected over a pre-determined time period and the

78

colloidal behavior of the GNR suspensions was interpreted using plasmon band

79

analysis. Transmission electron microscopy (TEM) and atomic force microscopy (AFM)

80

were used to further characterize the samples. These results are self-consistent and

81

can be explained by Derjaguin-Landau-Verwey-Overbeek (DLVO) theory.

82

Materials and Methods

83

Preparation of CTAB-coated GNRs

84

GNRs were synthesized via the well-established seed-mediated surfactant-directed

85

method.45, 46 In brief, 3-4 nm spherical gold nanoparticle seeds were produced through

86

the reduction of chloroauric acid (HAuCl4•3H2O, 2.5×10-4 M in aqueous solution) in the

87

presence of CTAB ([CH3(CH2)15N(CH3)3]+Br-, 0.10 M) by addition of ice-cooled and

88

freshly prepared sodium borohydride (NaBH4, 0.010 M). A rapid color change to light

89

brown indicated seed formation. In the next step, GNRs of aspect ratio 4.4 were grown

90

from an aqueous growth solution consisting of 0.10 M CTAB, 5.0×10-4 M HAuCl4,

91

8.0×10-5 M AgNO3, and 5.5×10-4 M ascorbic acid, which became colorless upon acid

92

addition. Once the gold seeds were added to the growth solution, it changed color to

93

dark brown over 1-2 h indicating GNR formation. Spheres, cubes, and other shaped

94

byproducts

95

centrifugation (25 min at 8000 rcf), repeated 5 times with pellet resuspension in 800 µM

96

CTAB, yielding CTAB-GNRs. ICP-MS measurements indicated that the GNR stock had

97

a 41.0±0.3 mg/L atomic gold concentration. Accordingly, a concentration of 2.1×1014

(minor

component)

and

chemical

byproducts

ACS Paragon Plus Environment

were

removed

by

5

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 33

98

GNRs/L was calculated based on the density of gold (19.3 g/cm3)47 and the TEM

99

determined average GNR size of 28.8 ± 0.1 nm length and 6.6 ± 0.1 nm diameter.

100

End-to-end linkage of CTAB-GNRs in CaSO4 and MgSO4

101

Suspensions of CTAB-GNRs (1.1×1014 GNRs/L) were incubated over a range of

102

CaSO4 and MgSO4 concentrations (1-5 mM) for 24 h. To initiate an experiment, CTAB-

103

GNRs were washed by centrifugation (20 min at 11000 rcf) and 90% of the supernatant

104

was replaced with DI water, followed by bath sonication for 1 min. The resuspended

105

GNRs were readily dispersed and were colloidally stable. This step removed excess

106

CTAB from the GNR suspension. Following a second centrifugation step (20 min at

107

11000 rcf), 90% of the supernatant was replaced with a given CaSO4 or MgSO4 solution

108

followed by 1 min of bath sonication. The GNR suspensions were then transferred to

109

Vis-NIR cuvettes that were used for colloidal stability monitoring over 24 h. During the

110

first hour, 10 min sampling intervals were used. Subsequently, one-hour intervals were

111

used for the remaining 23 h. Vis-NIR extinction spectra were collected between 400-

112

1200 nm. TEM samples were prepared after 3 and 10 h of exposure to salt solutions. To

113

quantitatively capture the temporal changes in the extinction of the GNR suspensions,

114

individual peaks within the Vis-NIR spectrum were fit with Lorentzian distributions48,

115

using Grams/AI software (Thermo-Fisher, Waltham, MA).

49

116

TEM samples were prepared by submerging 200 mesh carbon film coated copper

117

grids in a given GNR suspension and then fast dried by putting the grid on top of an

118

80 °C hot plate for 10 s. An AFM sample was prepared following 3 h GNR exposure to

119

salt solutions. Negatively charged AFM metal specimen discs were submerged

ACS Paragon Plus Environment

6

Page 7 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

120

vertically in a petri dish containing 5 mL GNR suspension sample for half an hour

121

followed by DI water rinsing prior to AFM measurement.

122

Instrumentation

123

An Agilent Cary 5000 spectrometer (Santa Clara, CA) and disposable polystyrene

124

cuvettes (10 mm pathlength) were used for Vis-NIR extinction spectroscopy. TEM

125

images were taken using a Philips (Thermo-Fisher, Waltham, MA) EM420 conventional

126

transmission electron microscope. A Malvern (Worcestershire, UK) Zetasizer Nano-ZS

127

was used for electrophoretic mobility measurements. A Bruker (Billerica, MA)

128

Nanoscope IIIa AFM and Ted Pella (Redding, CA) 15 mm metal specimen discs were

129

used for atomic force microscopy. Sonication was performed using a Fisher Scientific

130

FS20H bath sonicator. DI water (18.2 MΩ) was produced by a Thermo Scientific

131

Barnstead nanopure system.

132

Results and Discussion

133

GNR characterization

134

The GNRs were characterized via TEM, Vis-NIR spectroscopy, electrophoretic

135

mobility, and ICP-MS measurements. The GNRs (inset to Figure 1; Figure S1) have an

136

average length of 28.8 ± 0.1 nm and an average diameter of 6.6 ± 0.1 nm (the noted

137

errors reflect 95% confidence interval based on n = 2104 measurements obtained using

138

ImageJ). These relative dimensions result in an average AR of 4.4 ± 0.1. These GNRs

139

exhibit a transverse extinction band at 518 nm and a longitudinal band at 756 nm, which

140

are the expected wavelengths for GNRs of these dimensions.45 The electrophoretic

141

mobility of the GNRs in DI water was determined to be +3.58 µm•cm/V•s. Applying

ACS Paragon Plus Environment

7

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 33

142

Henry’s equation while considering Smoluchowski equation for Henry’s constant

143

resulted in calculated ζ-potential of +45.6 mV. We note that for the GNRs and the

144

range of ionic strengths tested in this study,

145

the nanoparticle effective diameter in nm) varies from 4.09 to 9.13. Because

146

the Smoluchowski equation is an acceptable approximation of ζ-potential (details and

147

example calculations are presented in SI; pages 2-3).50,

148

consistent with the literature and illustrate that the rods have a positive effective surface

149

charge.44, 51

150 151 152 153

Figure 1. Vis-NIR extinction spectrum for the synthesized CTAB-GNRs after centrifugal purification and washing. Inset: Transmission electron micrograph of CTABGNRs.

ૂa (ૂ is 1/Debye length in 1/nm and “a” is

51, 52

ૂa>>1,

These values are

154

ACS Paragon Plus Environment

8

Page 9 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

155

Langmuir

End-to-end assembly of CTAB-coated gold nanorods in CaSO4 and MgSO4

156

The in situ development of GNR assemblies was probed via Vis-NIR extinction

157

spectroscopy. It has been shown previously that during side-to-side and end-to-side

158

assemblies, the extinction of the initial transverse LSPR band increases and the band

159

slightly red shifts to higher wavelengths; while the extinction of the longitudinal band

160

decreases and blue shifts to lower wavelengths, thus resulting in the convergence of the

161

two bands.26 The calculated change in extinction intensity versus the incident light

162

wavelength and the corresponding transverse and longitudinal bands of these two

163

assemblies are presented in Table S3 and Figure S6. These calculations were

164

developed based on our experiment parameters, while employing extended Mie theory

165

modeling. In end-to-end assembly the bands behave quite differently. For aggregates

166

with this configuration, neither the transverse nor the longitudinal band shifts; however,

167

the extinction of the longitudinal band significantly decreases as the total average length

168

of the GNR chain increases. Under these conditions, a third plasmon band appears at

169

wavelengths larger than that of the original longitudinal band (Figure S5). Because of

170

extended plasmonic coupling this new plasmon band red shifts to longer and longer

171

wavelengths as the GNR chain length increases.22, 26, 53-55

172

To evaluate the aggregation behavior of the GNRs, CTAB-GNR suspensions were

173

exposed to 1, 2, and 5 mM CaSO4 or MgSO4. We observed salt concentration

174

dependent changes in the visual characteristics of the suspensions. In the presence of

175

DI water the suspension is brownish in color, while at higher salt concentrations the

176

color darkens. At a salt concentration of 5 mM the colloidal stability of the suspension

177

was lost within 24 h (Figure S2). As illustrated in representative plots for 2 mM CaSO4

ACS Paragon Plus Environment

9

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 33

178

and 2 mM MgSO4 (Figure 2) there are readily observable declines in the intensities of

179

both the transverse and longitudinal bands and additional bands appear at longer

180

wavelengths. Such behavior is consistent with the previously described end-to-end

181

aggregation process.

182

concentrations.

Qualitatively similar results were observed at the other salt

183

To quantify the declines in peak intensity we plotted the change in extinction for

184

both the transverse and longitudinal bands for each salt concentration. As shown in

185

Figure 2C-F, a sharp drop in the longitudinal peak intensity was observed within the first

186

hour of exposure of the GNRs to all concentrations of CaSO4 and MgSO4, accompanied

187

by a slower decrease in intensity over the next 23 h. In contrast, the transverse band

188

intensity did not change as noticeably. To interpret the temporal changes in the

189

longitudinal and transverse bands we fitted our collected data using a simple second

190

order aggregation model56, 57 (details are presented in the SI; page 4):

191

ௗ[௥௢ௗ] ௗ௧

= −݇[‫]݀݋ݎ‬ଶ

(1)

192

Where [rod] is the number concentration of individual, fully dispersed rods and k is the

193

observed second order rate constant. Other more complex aggregation models may be

194

more appropriate for this system due to its complexity; however, our purpose in using

195

this model is to capture the differential behavior of the two extinction bands. Number

196

concentration-based rate constants and associated 95% confidence intervals for single

197

CTAB-GNRs in the presence of CaSO4 and MgSO4 are presented in Table 1. We note

198

that for 5 mM CaSO4 the second order model did not fit the collected data as well as it

199

did for other experimental conditions. We attribute this behavior to the more rapid

200

aggregation observed with CaSO4 and the greater degree of colloidal instability.

ACS Paragon Plus Environment

10

Page 11 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

201 202 203 204 205 206 207 208

Langmuir

Figure 2. Vis-NIR extinction spectra of 1.1×1014 GNRs/L exposed to A) 2 mM CaSO4 and B) 2 mM MgSO4 from zero hours (red spectra) to 24 hours (blue spectra). Changes in the average normalized transverse band and longitudinal band extinction maxima of CTAB-GNRs (1.1×1014 GNRs/L) in 1-5 mM CaSO4 and MgSO4 solutions as a function of exposure time. (C) Transverse band-CaSO4 (D) Longitudinal band-CaSO4 (E) Transverse band-MgSO4 (F) Longitudinal bandMgSO4.

209

ACS Paragon Plus Environment

11

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

210 211 212 213

Table

1.

Second

order

number

Page 12 of 33

concentration

rate

constants

L ) for single CTAB-GNRs in the presence of different 1.07 × 10 ( goldnanoro ds ). S concentrations of CaSO4 and MgSO4. Both the changes in the transverse band and the longitudinal band are provided. Errors are reported at the ±95% confidence intervals.

(

14

Salt Concentration (mM) 1 2 5

Transverse band CaSO4 MgSO4 0.0006 ± 0.0001 0.0007 ± 0.0001 0.0065 ± 0.0000 0.0032 ± 0.0001 0.22 ± 0.049 0.012 ± 0.0006

Longitudinal band CaSO4 MgSO4 0.025 ± 0.0020 0.019 ± 0.0021 0.05 ± 0.0051 0.028 ± 0.0026 0.49 ± 0.072 0.051 ± 0.0053

214 215

An increase in ionic strength significantly increased the rate at which the GNRs

216

assembled. In all cases, the rate of change of the longitudinal band was greater than

217

that of the transverse band (Table 1). Over the 24 h period that the GNRs were exposed

218

to 1 mM and 2 mM of CaSO4 and MgSO4 there was minimal change in the transverse

219

band intensity, while simultaneously there was a 40-60% decline in the longitudinal

220

band intensity. If the assemblies were colloidally unstable and were precipitating out of

221

suspension and leaving only single GNRs behind then the rate constants for the change

222

in the transverse and longitudinal bands of single GNRs should be equal (Table S1 and

223

Figure S3). However, the much slower change in the transverse band relative to the

224

longitudinal band supports the conclusion that the GNRs are assembling into extended

225

structures. In end-to-end nanoparticle assembly the transverse band is not expected to

226

shift, while the longitudinal band will decrease in intensity and red-shift.

227

Following exposure of the GNRs to CaSO4 or MgSO4 a new shoulder with an

228

extinction maximum at a wavelength greater than the longitudinal band appears. The

229

intensity of this shoulder increases and red-shifts to longer wavelengths over time and

230

at increased salt concentrations.

231

assembly of GNRs to produce new structures with elongated AR (i.e., end-to-end

Formation of this shoulder is consistent with the

ACS Paragon Plus Environment

12

Page 13 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

232

assemblies). Past studies have illustrated that chain-like assemblies of GNRs display

233

the formation of an extinction band at longer wavelengths compared to individual

234

rods.25, 26, 30, 34 The peak wavelength is strongly related to the number of linked rods and

235

shifts to longer wavelengths as the chains become increasingly extended over time.

236

TEM and AFM images consistent with formation of end-to-end assemblies are

237

shown in Figure 3. Both the TEM and AFM images illustrate the formation of end-to-end

238

linkages. The AFM images unlike the TEM images were collected in situ without sample

239

drying. Observation of an extended end-to-end assembly in the AFM image supports

240

the argument that the presence of these assemblies in the TEM images is not

241

necessarily a result of drying mediated artifacts. TEM images taken at different times

242

and salt concentrations suggest that the end-to-end chains of GNRs grow longer over

243

time and with an increase in ionic strength (Figure 3B-C, Figure S4).

ACS Paragon Plus Environment

13

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 33

A

B

C

50 nm 244 245 246 247

Figure 3. A) Representative AFM image of CTAB-GNRs after 3 h exposure to 2 mM CaSO4, B) TEM image of CTAB-GNRs after 3 h exposure to 2 mM CaSO4, C) TEM image of CTAB-GNRs after 10 h exposure to 2 mM CaSO4.

248

ACS Paragon Plus Environment

14

Page 15 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

249

Langmuir

Extinction spectra peak fitting Prior studies have established that there is a linear relationship between the peak

250 251

wavelength location and AR.22,

58

252

assemblies of different length can be estimated by application of extended Mie theory

253

(see SI for details).22, 58-61 Knowing the physical and chemical properties of the GNRs

254

and the solution chemistry employed herein, we were able to calculate the expected

255

location of the extinction band for each end-to-end configuration (see SI for details of

256

calculations). Our GNRs have an AR of 4.4, therefore the calculated λmax values for

257

dimer (AR=8.8), trimer (AR=13.2), and tetramer (AR=17.6) end-to-end linked AuNRs

258

are 860, 965, and 1070 nm, respectively (Figure 4A). These values are qualitatively

259

similar to those theoretically calculated for end-to-end assemblies of GNRs of similar

260

AR.

Theoretical peak wavelengths for end-to-end GNR

ACS Paragon Plus Environment

15

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 33

261 262 263 264 265 266

Figure 4. A) Plasmon band wavelengths of different GNR end-to-end assemblies as calculated by extended Mie theory. B) Exemplary figure illustrating quality of fit for experimental vs. fitted Vis-NIR spectrum for GNRs exposed to 5 mM MgSO4 after 1 hr. The magnitudes of individual extinction bands are illustrated.

267

ACS Paragon Plus Environment

16

Page 17 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

268

Based upon these wavelengths we used a non-linear least squares iterative62

269

procedure (within GRAMS/AI) to fit the collected Vis-NIR spectra for each experiment as

270

the summation of five separate Lorentzian bands that correspond to the transverse

271

band and the longitudinal band for single, dimer, trimer, and tetramer GNRs. The

272

Lorentzian distribution can be explained as a Fourier-transform of exponentially

273

decaying oscillations63. We used extended Mie theory modeling to calculate the change

274

in extinction intensity versus the incident light wavelength and to locate the plasmon

275

bands of different end-to-end assemblies as has been done previously22 (refer to SI

276

page 7 for calculation details). As a result, Lorentzian distributions were used for curve

277

fitting. In this exercise, the center of each peak is fixed while the amplitude (height) and

278

width are unknown parameters. Using this constrained fitting approach, we are able to

279

fit and reproduce the collected Vis-NIR spectra quite readily (Figure 4B). By fitting each

280

of the spectra collected over the course of an experiment we can probe the relative

281

numbers of single, dimer, trimer, and tetramer assemblies present at any time. Figure 5

282

illustrates the temporal changes in each Lorentzian distribution band for the CaSO4

283

experiments, while the results for the MgSO4 experiments are shown in Figure S7.

ACS Paragon Plus Environment

17

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 33

284 285 286 287 288 289

Figure 5: Extinction intensities of different Lorentzian distribution band end-to-end assemblies of GNRs exposed to three different CaSO4 concentrations A) Transverse bands of single rod B) Longitudinal bands of single rod C) Longitudinal bands of dimer D) Longitudinal bands of trimer E) Longitudinal bands of tetramer

290

With an increase in salt concentration increasing numbers of GNRs participate in

291

end-to-end assembly and there is a concomitant decrease in the extinction of the

ACS Paragon Plus Environment

18

Page 19 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

292

longitudinal band that represents single GNRs (Figure 5B). At 1 and 2 mM CaSO4, the

293

extinction of the dimer band at 860 nm increases rapidly and then plateaus, while at 5

294

mM it increases and then decreases with time (Figure 5C). At 5 mM CaSO4, the dimer,

295

trimer, and tetramer extinction intensities rapidly increase but then decrease over

296

extended time. These results collectively show that at high salt concentrations the end-

297

to-end GNRs assemblies that are formed initially, grow further to be colloidally unstable,

298

and finally undergo precipitation. This precipitation trend is consistent with visual

299

observation of large dark pellets at the bottom of the sample cuvette (Figure S2).

300

The validity of our curve fitting technique was studied by analyzing 36 TEM images

301

taken from one sample (GNRs exposed to 2 mM CaSO4) at three different time periods

302

(3, 10, and 24 h). A total of 270 GNRs were counted via TEM image analysis. Figure 6

303

shows the change in percentage of these assemblies as a function of time. As

304

expected, the length of the end-to-end assemblies generally increase over time. As an

305

example, after 3 h almost 50% of GNRs are in the dimer form and no tetramers were

306

found; but after 24 h 50% of GNRs are in the tetramer form while only 12% single GNRs

307

were observed. Because of the potential for drying mediated artifacts and the potential

308

effects of the TEM substrate on aggregate formation the results presented in Figure 6

309

cannot be directly compared to the plasmon band analyses; however, the general trend

310

is the same with longer and longer assemblies forming with extended time.

ACS Paragon Plus Environment

19

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 33

311 312 313 314

Figure 6. Percentage of GNRs forming different end-to-end assemblies. Analysis of 36 TEM images taken from GNRs exposed to 2 mM CaSO4 at three different time periods. Representative TEM images are shown Figure S4.

315 316

Anion identity and valence dictate assembly

317

Past studies have suggested that CTAB forms a dense 3.5 nm thick vesicle-like

318

bilayer on the GNR side facets, while it may not fully coat the ends.44, 64, 65 The presence

319

of the positively charged vesicle-like bilayer around the side facets protects the GNRs

320

from undergoing side-by-side aggregation. To investigate the chemistry underlying the

321

observed end-to-end assembly we exposed CTAB-GNRs to CaCl2, MgCl2, NaCl, and

322

Na2SO4 salts at the same ionic strength used in the CaSO4 and MgSO4 experiments.

ACS Paragon Plus Environment

20

Page 21 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

323

The collected Vis-NIR extinction spectra over a period of 24 h are shown in Figures S8

324

and S9. The decline in the transverse and longitudinal absorption bands is shown in

325

Figures S10 and S11.

326

No change in the extinction of the transverse band, a