Interaction between manufactured gold nanoparticles and naturally occurring organic macromolecules

SC IE N CE OF TH E TOTA L E N V I RO N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / s c i t o t e n v

Interaction between manufactured gold nanoparticles and naturally occurring organic macromolecules Sara Diegoli a , Adriana L. Manciulea b , Shakiela Begum a , Ian P. Jones c , Jamie R. Lead b , Jon A. Preece a,⁎ a

School of Chemistry, the University of Birmingham, Edgbaston, Birmingham B15 2TT, UK Department of Geography, Earth and Environmental Sciences, the University of Birmingham, Birmingham B15 2TT, UK c Department of Metallurgy and Materials, the University of Birmingham, Birmingham B15 2TT, UK b

AR TIC LE I N FO

ABS TR ACT

Article history:

The increasing exploitation of nanomaterials into many consumer and other products is

Received 29 October 2007

raising concerns as these nanomaterials are likely to be released into the environment. Due

Received in revised form

to our lack of knowledge about the environmental chemistry, transport and ecotoxicology of

17 April 2008

nanomaterials, it is of paramount importance to study how natural aquatic colloids can

Accepted 19 April 2008

interact with manufactured gold nanoparticles as these interactions will determine their

Available online 4 June 2008

environmental fate and behaviour. In this context, our work aims to quantify the effect of

Keywords:

(IHSS) Suwannee River Humic Acid Standard (SRHA) – on citrate- and acrylate-stabilized

Humic substances

gold nanoparticles. The influence of SRHA on the stability of the gold colloids was studied as

Ecotoxicity

a function of pH by UV–visible absorption spectroscopy, dynamic light scattering (DLS) and

Gold nanoparticles

transmission electron microscopy (TEM). At high ionic strengths (0.1 M), extensive and rapid

UV–visible absorption spectroscopy

aggregation occurred, while more subtle effects were observed at lower ionic strength

TEM

values. Evidence was found that SRHA enhances particle stability at extreme pH values

naturally occurring riverine macromolecules – International Humic Substances Society

(ionic strength b 0.01 M) by substituting and/or over-coating the original stabilizer on the gold nanoparticle surface, thus affecting surface charge and chemistry. These findings have important implications for the fate and behaviour of nanoparticles in the environment and their ecotoxicity. © 2008 Elsevier B.V. All rights reserved.

1.

Introduction

Gold nanoparticles have attracted extensive attention in interdisciplinary research because of their potential applications spanning electronics (Lee et al., 2002; Khomutov et al., 2003), optics (Sharma and Gupta, 2005; Chau et al., 2006), catalysis (Sau et al., 2001; Hutchings, 2005) and drug delivery (Yang et al., 2005; Hong et al., 2006). The development of new cost-effective routes for the production of metallic particles has already led to the pilot-scale production of noble metal nanoparticles which are compatible with polymer processing. These composites have applications in many consumer

products such as clothing, footwear and plastic containers (NanoHorizons, 2007). Other examples of the widespread use of nanoparticles are titanium dioxide, which is present in sunscreens and cosmetics (Colvin, 2003) and carbon nanotubes which are employed in bicycles frames and tennis rackets (Kanellos, 2006). As nanotechnology is contributing to a variety of consumer products (Roco, 2005), the release of nanomaterials into the environment is raising concerns (Lead and Wilkinson, 2006). Due to our current lack of knowledge on the environmental chemistry, transport and ecotoxicology of nanomaterials, more research is necessary in this field (SCHENIR, 2007)

⁎ Corresponding author. Tel.: +44 121 414 3528; fax: +44 121 414 4403. E-mail address: [email protected] (J.A. Preece). 0048-9697/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2008.04.023

52

SC IE N CE OF TH E TOTA L E N V IR O N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

Since aggregation is one of the primary controls on both transport and toxicity of nanomaterials in the aquatic environment, it is of paramount importance to study how naturally occurring organic colloids can interact with manufactured nanoparticles and influence their aggregation (Buffle, 2006). The most important class of natural organic colloids is represented by humic substances. Despite the complexity and heterogeneity of humic matter, a great deal is known about its structure and chemistry (Dewit et al., 1993a,b; Baalousha et al., 2006), and well characterized reference materials are available (Lead et al., 2000). Previous reports have studied interactions of humic substances with multi walled carbon nanotubes (Hyung et al., 2007) and zerovalent iron (Giasuddin et al., 2007). Both the reports concluded that interaction with humic substances conferred extra stability from aggregation to the manufactured nanoparticles. In particular, Giasuddin et al. (2007) demonstrated that the absorption of humics on ion nanoparticles interferes with pollutants absorption. In addition, Dos Santos et al. (2005) have studied the effects of using humic substances to control the shape and size in the manufacture of gold nanoparticles. This report aims to investigate the effect of the interactions of nanoparticles with humic substances — IHSS Suwannee River Humic Acid Standard (SRHA). Of particular interest is the effect of humic substances on the stability and aggregation of gold nanoparticles. Variables such as pH and ageing has also been independently varied as a means of understanding aggregation and surface characteristics under environmentally relevant conditions and thus attempting a more realistic explanation of the likely fate and behaviour of the nanoparticles in natural waters.

2.

Methods

2.1.

Materials

Commercially available chemicals and solvents were purchased from Aldrich and Fisher Scientific. Suwannee River Humic Acid Standard (SRHA) was purchased for IHSS and was used as received. Ultra High Purity (UHP) water with a maximum resistivity of 18 MΩ cm− 1 was used throughout the experiments.

2.2.

Sodium acrylate-stabilized gold nanoparticles

Nanoparticles were prepared as described previously (Hussain et al., 2003). Briefly, an aqueous solution of HAuCl4d 3H2O (1 mM, 50 mL) was heated under reflux for 5–10 min, and a warm (50–60 °C) aqueous solution of sodium acrylate (80 mM, 15 mL) was added quickly. Heating was continued for another 30 min until a deep-red solution was observed. The particle solution was centrifuged for 10 min at 3500 rpm and the supernatant collected.

2.3.

Citrate-stabilized gold nanoparticles

Nanoparticles were synthesized by the Frens method (Frens, 1973). Briefly, an aqueous solution of HAuCl4d 3H2O (0.25 mM, 100 mL) was heated under reflux for 5–10 min, and an aqueous

solution of sodium citrate (38.8 mM, 2 mL) was added. Heating was continued for another 10 min to ensure complete reduction of the gold salt. The colloidal gold solution was centrifuged for 10 min at 3500 rpm and the supernatant collected.

2.4.

Solutions

In order to investigate the stability of the gold nanoparticle dispersions as a function of dilution with water and dilution with an aqueous SRHA solution, two aliquots (100 mL) each of the citrate- and acrylate-stabilized gold nanoparticle dispersions (CN and AN respectively) were either diluted with water (W) (100 mL) or an aqueous solution (100 mL of a 10 mg/L) of SRHA, to afford 3 pairs of dispersions (i) CN and AN (undiluted), (ii) CN + W and AN + W (diluted), and (iii) CN + SRHA and CN + SRHA (diluted and containing humic acid).

2.5.

pH Measurements

pH analyses were performed with a IQ150 pH meter from IQ Scientific Instruments, Inc. To aliquots (4 mL) of the 6 dispersions (CN, AN, CN + W, AN + W, CN + SRHA and AN + SRHA) was added either HCl or NaOH to adjust the pH over the range 1.5 to 13 in increments of 0.5 pH units. In altering the pH, care has been taken not to alter significantly the concentration of the colloidal dispersions since this would affect UV–visible spectra. Thus, different NaOH and HCl solution of optimum concentration were used to achieve the desired pH vales mentioned above and the added volume of acid or base solution was kept ≤50 µl. The addition of acid and base to the gold nanoparticles changes the ionic strength of the dispersion. Therefore, we attempted to use NaCl as a background electrolyte. It was found that the gold nanoparticles were not stable in dispersion with a NaCl concentration equal or above 0.1 M and almost instantaneously aggregated. All experiments reported were thus performed at ionic strengths lower than 10 nM, except those at extreme pH values (pH 1.5–2.0 and pH 13.0).

2.6.

UV–visible absorption spectroscopy experiments

UV–visible absorption spectra were colleted using a HewlettPackard 8452A spectrometer. The UV–visible absorption spectra of all the “pH adjusted” dispersions were recorded immediately after addition of the acid or base (t = 0). The same spectra were also recorded after 24 h to allow for the humic substances to equilibrate in solution. For the experiments reported in Fig. 3a and b the wavelength of the maximum of the surface plasmon band and the associated error at each pH value were obtained as the average of two replicates and their standard deviation. To make sure the free SRHA in solution were not responsible for the shift of the surface plasmon band reported in Fig. 3a and b the spectra of SRHA were mathematically subtracted from the spectra of CN + SRHA and AN + SRHA at t = 0 and the results recalculated using the corrected spectra (data not shown). Exactly the same trends of the shift of the maximum of the surface plasmon band with pH change appeared. However, we chose to report the unsubtracted results as they represent the less processed

53

SC IE N CE OF TH E TOTA L E N V I RO N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

Table 1 – Characterization of the citrate- and acrylate-stabilized gold nanoparticle dispersions Dispersion

CN AN

pH a

5.8 ± 0.1 4.9 ± 0.1

Z potential (mV) a

Plasmon resonance band (nm)

− 43.7 ± 0.7 − 34.0 ± 1.3

519 530

Nanoparticle diameter (nm) TEM

b

15.8 ± 1.9 19.7 ± 3.3

Z average size c

PDI d

18.60 ± 0.53 21.64 ± 0.04

0.44 ± 0.08 0.19 ± 0.01

a

The mean value reported and the standard deviation has been obtained from 8 measurements on each of two independent replicates. The diameter reported and the associated error are the mean and associated standard deviation. c Diameter measured by dynamic light scattering. The mean value reported and the standard deviation has been obtained from 3 measurements on each of two independent replicates. d Polydispersity index (PDI) measured by dynamic light scattering. The mean value reported and the standard deviation has been obtained from 3 measurements on each of two independent replicates. b

data. For the experiments reported in Figs. 5 and 6, UV–visible absorption spectra of 4 mL of dispersions at the acid and basic aggregation pHs, were collected as a function of time. The first spectrum was collected immediately after the pH change and was followed by 19 spectra collected every 4 min and 20 spectra recorded every 10 min (however to simplify the graphs only one spectra every 8 min and one every 20 min are shown). Each of the remaining spectra to a maximum total time of 24 h was then collected every hour.

2.7.

taken at 25 °C and repeated eight times on each of two independent replicates. Before, and in between measurements, the flow through cell was washed with ultra high pure (UHP) water (10 ml).

2.9.

Transmission electron microscopy (TEM)

Micrographs of citrate- and acrylate-stabilized gold nanoparticles were collected on a Philips CM20 transmission electron

Dynamic light scattering (DLS)

Dynamic light scattering measurements have been performed at 20 °C on a Malvern High Performance Particle Sizer (HPPS 5001) using disposable low volume polystyrene cuvettes. The zeta average diameter and polydispersity index (PDI) were automatically provided by the instrument using cumulant analysis(Rasi et al., 2003; Maulucci et al., 2005). The zeta average diameter and PDI reported herein were obtained as the average of three measurements performed on each sample immediately after pH change.

2.8.

Zeta potentiometry

Zeta potential measurements were performed using a Malvern Instruments Zetasizer 1000Hs, operating with a variable power (5–50 mW) He–Ne laser at 633 nm. Measurements were

Fig. 1 – Comparison between the UV–visible absorption spectra of SRHA, CN and AN dispersions.

Fig. 2 – Absorbance of the maximum of the plasmon resonance band as a function of pH and time (t = 0 and t = 24 h). (a) CN, CN + W and CN + SRHA dispersions and (b) AN, AN + W and AN + SRHA dispersions.

54

SC IE N CE OF TH E TOTA L E N V IR O N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

diately. Thus, the data reported below relates to ionic strength values always at least one order of magnitude lower than 0.1 M.

3.1.

Characterization of the gold nanoparticle dispersions

Citrate-stabilized gold nanoparticle (CN) and acrylate-stabilized gold nanoparticles (AN) dispersions were prepared as reported previously (Diegoli et al., 2006) and their chemical and physical data are summarized in Table 1. Initial characterization of the gold nanoparticles by UV–visible absorption spectroscopy revealed the absorption maximum (λmax) of the surface plasmon band at ~519 nm and ~530 nm, for CN and AN dispersions, respectively (Fig. 1). Citrate- and acrylate-stabilized gold nanoparticles are negatively charged as determined by zeta potential measurements and are between 15 and 22 nm in diameter as measured by DLS and TEM (Table 1).

3.2.

Characterization of SRHA solution

SRHA solution was prepared as detailed in the experimental section and was characterized by UV–visible absorption

Fig. 3 – Wavelength of the maximum of the plasmon resonance band as a function of pH and time (t = 0 and t = 24 h). (a) CN, CN + W and CN + SRHA dispersions and (b) AN, AN + W and AN + SRHA dispersions. The data reported and the associated error at each pH value were obtained as the average of two replicates and their standard deviation.

microscope operating at 200 kV. The micrographs used for the determination of the nanoparticle size were collected on a JEOL 1200EX transmission electron microscope operating at 120 kV. In both cases samples were prepared by slow evaporation of one drop of the nanoparticle dispersion on a carbon coated copper grid.

3.

Results

We report here studies (UV–visible absorption spectroscopy, dynamic light scattering (DLS) and transmission electron microscopy (TEM)) on citrate- and acrylate-stabilized gold nanoparticle dispersions as a function of pH, ageing and concentration in the absence and presence of IHSS Suwannee River Humic Acid Standard (SRHA) (Powell and Fenton, 1996; Balnois et al., 1999; Haiber et al., 2001; Her et al., 2002; Redwood et al., 2005), in order to investigate the influence of naturally occurring organic macromolecules on gold nanoparticle stability. Under high ionic strength conditions (0.1 M), extensive aggregation and sedimentation occurred almost imme-

Fig. 4 – Zeta average size determined by dynamic light scattering of (a) CN and CN + SRHA dispersions (b) AN and AN + SRHA dispersions at different pHs.

SC IE N CE OF TH E TOTA L E N V I RO N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

spectroscopy. The only feature in the UV–visible absorption spectra of the SRHA solution is a broad absorbance lifting off at ~450 nm and increasing towards the UV region of the spectra (Fig. 1), due to the aromatic moieties present in the humic structure. SRHA are known to be around 1–3 nm in size and aggregation is minimal over a wide range of environmentally relevant pH values (Lead et al., 2000; Lead et al., 2003; Lead and Wilkinson, 2006).

55

3.3. UV–visible absorption spectroscopy study of the stability of gold nanoparticles as a function of concentration, pH and the presence of SRHA Plots of the maximum absorbance of the transverse plasmon resonance band (λmax) of the three sets of both types of gold nanoparticle dispersions as a function of pH and time (t = 0 and t = 24 h) are shown in Fig. 2a (CN, CN + W, CN + SRHA) and Fig. 2b

Fig. 5 – Comparison between UV–visible absorption spectra recorded as a function of time and TEM micrographs of citrate-stabilized gold nanoparticle dispersions. (a) CN pH 2.0, (b) CN + SRHA pH 1.5, (c) CN pH 12.0 and (d) CN + SRHA pH 12.5 dispersions. Scale bar: 200 nm.

56

SC IE N CE OF TH E TOTA L E N V IR O N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

(AN, AN + W, AN + SRHA). The shift in the λmax of the plasmon resonance band for CN + SRHA and AN + SRHA dispersions is reported as a function of pH and time (t = 0 and t = 24 h) in Fig. 3a and b. For comparison it should be noted that the dispersions which do not contain the humic acid (CN + W and AN + W) show no significant shift (±1 nm) in the wavelength of

the plasmon resonance over the pH ranges in which the particles are stable (data not shown). Thus, CN + W and AN + W dispersions absorb at ~519 and ~525, respectively, independent of pH. Whereas CN + SRHA dispersions at pH 2.0 have a λmax centred around 525 nm, which blue shifts by about 6 nm to reach 519 nm at pH 7.5, and remains constant to pH 12.0

Fig. 6 – Comparison between UV–visible absorption spectra recorded as a function of time and TEM micrographs of acrylate-stabilized gold nanoparticle dispersions. (a) AN pH 2.0, (b) AN + SRHA pH 1.5, (c) AN pH 12.5 and (d) AN + SRHA pH 12.5 dispersions. Scale bar: 200 nm.

SC IE N CE OF TH E TOTA L E N V I RO N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

(Fig. 2a). AN + SRHA dispersions also show a blue shift of their λmax of about 4 nm (from ~529 to 525 nm) when the pH is changed from pH 2.0 to pH 11.0 (Fig. 3b). It can also be observed that both behaviours are largely independent of ageing over 24 h.

3.4. UV–visible absorption spectroscopy study of the SRHA solution as a function of pH UV–visible absorption spectra of the SRHA solution as a function of pH were also collected at t = 0 and t = 24 h. The absorption of the SRHA increases in intensity and shifts towards longer wavelengths with increasing pH. This trend is independent from aging of the solution over 24 h.

3.5. DLS Study of gold nanoparticle stability as a function of pH and the presence of SRHA DLS studies were performed on CN, CN + SRHA, AN and AN + SRHA dispersions at pH 1.5–2.5 and pH 12.0–13.0 (Fig. 4). The pH values investigated are those at which significant changes in the stability of the colloidal dispersions were observed by UV–visible absorption spectroscopy. The size of CN, CN + SRHA (Fig. 4a) and AN, AN + SRHA (Fig. 4b) particles were measured immediately after pH adjustment (t = 0) and reveal increases in particle size at the same pH at which the surface plasmon band began to reduce in intensity (Fig. 2).

3.6. Time dependent UV–visible absorption spectroscopy study of the aggregation behaviour UV–visible absorption spectroscopy as a function of time in conjunction with TEM observations has also been used to investigate the influence of SRHA on the citrate- and acrylatestabilized gold nanoparticles aggregation, as reported in Figs. 5a-d and 6a-d, respectively. These time dependent UV– visible spectra were recorded only at pH values at which the nanoparticle dispersion were observed to undergo aggregation as suggested by (i) reduction in intensity of the surface plasmon band in the UV–visible spectra at t = 0 (Fig. 2), and (ii) increases in the particle diameters by DLS (Fig. 4). TEM micrographs of the gold nanoparticles are reported in Figs. 5a-d and 6a-d. The samples were prepared by drop casting each dispersion onto a TEM grid immediately after pH adjustment. The grids were then left to dry in air as detailed in the experimental section. Two types of behaviour can be observed; fusion of the gold cores (Figs. 5a and 6a) and nonfusion of the gold cores (Figs. 5b-c, 6b and d).

4.

Discussion

It is well known that the optical spectrum of colloidal gold suspensions is dominated by a strong absorption in the visible region called the surface plasmon resonance band, which is due to the collective dipole oscillations of the free electrons in the conduction band of the gold (Mie, 1908). The position, intensity and shape of the plasmon resonance band is influenced by many factors such as particle size (Logunov et al., 1997) and shape (Wiesner and Wokaun, 1989; Jian et al., 2004),

57

dispersity and degree of aggregation (Weisbecker et al., 1996; Schmitt et al., 1999; Shipway et al., 2000; Norman et al., 2002; Liao et al., 2003; Wang et al., 2005; Sendroiu et al., 2006). Other contributions come from the dielectric constant of the medium in which the particles are embedded (Mulvaney, 1996) and the presence of stabilizing agents absorbed on the nanoparticle surface (Henglein, 1993). When gold nanoparticles deviate from spherical symmetry (i.e. synthesis of nanorods or formation of chain-like aggregates) a degeneration of the plasmon resonance absorption into two bands is observed (Norman et al., 2002; Rechberger et al., 2003). The plasmon resonance band at shorter wavelengths is called the transverse plasmon resonance band, whereas the absorption at longer wavelengths is called the longitudinal plasmon resonance band. Moreover, the appearance of the longitudinal plasmon resonance band in gold nanoparticle dispersions has also been attributed to the coupling of the plasmon bands of gold nanoparticles coming in close proximity during aggregation. Thus, UV–visible absorption spectroscopy has been used to probe the stability, the surface chemistry and the aggregation behaviour of citrate- and acrylate-stabilized gold nanoparticles as a function of pH and ageing. Moreover, DLS and TEM studies have corroborated the results obtained by UV– visible absorption spectroscopy and helped generate a conceptual model of the effect the SRHA has on the gold nanoparticle stability. In Table 1 a difference of 11 nm in the λmax of the plasmon resonance band for the CN and AN colloidal dispersions can be noted. This difference is due to the variation of ~20% between CN and AN diameter measured by TEM (Table 1). As a consequence, the surface plasmon band of AN dispersions is red shifted, consistent with the larger particle size. Also of note in Table 1 is that the diameters measured in dispersion by DLS are slightly larger than the diameters measured by TEM. This difference can be rationalized by (i) taking into account that the organic stabilizer is transparent to electrons, and therefore does not contribute to the diameter obtained by TEM, and (ii) the DLS method is sensitive to the double layer surrounding the gold nanoparticles in dispersion and is expected to overestimate the particles diameter. Fig. 1 shows a comparison between the UV–visible absorption spectra of (i) SRHA, (ii) CN and (iii) AN dispersions. It is evident that the absorbance of SRHA in the region of the plasmon resonance band of CN and AN dispersions is negligible. Moreover, a red shift of the UV absorption of the SRHA was observed as the pH of the solution was increased. However, the absorbance of this shift does not significantly overlap with the plasmon resonance of the gold nanoparticles (λmax = 519 nm for CN and 530 nm for AN). The increase in absorbance of the SRHA by 0.02 units from pH 1.5 to pH 13.0, corresponds to less than 3% of the absorbance of CN + SRHA and AN + SRHA at these wavelengths. Thus, the presence of ‘free’ SRHA in the CN + SRHA and AN + SRHA dispersions does not significantly affect the intensity of the surface plasmon band compared with the gold nanoparticle dispersions that do not contain SRHA (CN, AN, CN + W, AN + W). The UV–visible absorption spectroscopy results reported in Fig. 3a-b show that the λmax of CN + SRHA and AN + SRHA dispersions undergo blue shifts as the pH of the dispersions is increased from 2.0 to 12.0. In contrast, no significant shift has

58

SC IE N CE OF TH E TOTA L E N V IR O N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

been observed in the position of the λmax of CN and AN dispersions in the absence of SRHA (data not shown). A rationalization of the observed blue shift can be gleaned from considering Mie theory (Mie, 1908), which predicts that the plasmon resonance can shift due to changes in the dielectric constant on the nanoparticle surface. With this phenomenon in mind, it is generally accepted that humic substances change their size (i.e. aggregate and disaggregate) in solution depending on pH and ionic strength (Sutton and Sposito, 2005). When the concentration of protons and cations in solution is reduced, the total charge (potential) on the humic acid surface increases, whereas at low pH values or high ionic strength conditions the total charge on the humic is neutralized or screened. This change in the total charge of the humic molecules will influence the humic substances conformation and aggregation status (Baalousha et al., 2006). If the SRHA is absorbed on the gold surface, these changes of the humic structure may well influence the dielectric constant on the surface of the nanoparticles, and thus their optical spectra. These observations suggest that the SRHA is present on the gold nanoparticle surface and has (i) substituted and/or (ii) over-coated the citrate and acrylate anions on the surface of the gold colloids. We favour a substitution mechanism considering that it has been demonstrated in an interesting study by Dos Santos et al. (2005) that humic substances (fulvic acids) can be used both as a reducing and stabilizing agent for the synthesis of gold nanoparticles. Moreover, it has to be taken into account that displacement of the citrate and acrylate anions by the polyanionic SRHA would be favoured both on thermodynamic (cooperativity of binding) and entropic grounds (release of small anions upon substitution of a polyanion). From the spectra collected as a function of pH for CN, CN + W, CN + SRHA, AN, AN + W and AN + SRHA at t = 0 and t = 24 h the intensity and wavelength of the maximum of the transverse plasmon resonance band (520 nm) were extracted and plotted in Fig. 2. Fig. 2a-b clearly indicates that dilution of both citrate- and acrylate-stabilized gold nanoparticle dispersions has a deleterious effect on the colloidal stability. However, when SRHA is added the stability of the gold nanoparticles is much enhanced over a wider pH range despite the dilution effect. In fact, for citrate-stabilized gold nanoparticle dispersions aggregation occurs at pH 2.5 (CN), 3.0 (CN + W) and 1.5 (CN + SRHA) and for acrylate-stabilized gold nanoparticle dispersions at pH 2.0 (AN), 3.5 (AN + W) and 1.5 (AN + SRHA), respectively, i.e. the addition of SRHA has enhanced the stability of the colloidal dispersions by 1.5 pH units (comparing CN + W and CN + SRHA) and 2.0 pH units (comparing AN + W and AN + SRHA), respectively. The stability of citrate- and acrylate-stabilized gold nanoparticles was also studied by DLS at t = 0 as a function of pH in the absence and presence of SRHA. DLS measurements were collected immediately after pH change (t = 0). Thus, the onset of aggregation occurs at pH 2.5 for CN dispersions, where aggregates with sizes N200 nm were observed. Whereas for CN + SRHA it is not until the pH is lowered to 1.5 that the onset of aggregation is observed revealing particle aggregates with sizes N400 nm. CN dispersions at pH 12.0 and CN + SRHA dispersions at pH 12.0 and 12.5 show comparable sizes lying between 25 and 35 nm (Fig. 4a), as one might expect for

anionically stabilized colloids, i.e. at basic pHs the negative charge on the anionic stabilizer is maintained. AN dispersions at pH 2.0 exhibit particle size of ~30 nm, but are precipitating rapidly, whereas for AN + SRHA dispersions aggregation occurs at pH 1.5 (particle size of 60 nm). AN and AN + SRHA at pH 12.0 and pH 12.5 all show comparable sizes between 20 nm and 30 nm (Fig. 4b), again as might be expected for negatively charged stabilized colloids under basic conditions. It should be noted that the aggregation pHs obtained by UV (pH 2.5 for CN, pH 2.0 for AN, pH 1.5 for CN + SRHA and AN + SRHA) are confirmed by these DLS experiments (Fig. 4a and b). The influence of SRHA on the ageing and aggregation behaviour of citrate- and acrylate-stabilized gold nanoparticles has been examined by UV–visible absorption spectroscopy and TEM (Figs. 5a-d and 6a-d), at the pHs at which the gold nanoparticles were first observed to aggregate as suggested by (i) reduction in intensity of the surface plasmon band in the UV–visible spectra at t = 0 (Fig. 2), and (ii) increases in the particle diameters by DLS (Fig. 4). The UV–visible spectra of CN dispersion at pH 2.5 (Fig. 5a) and AN dispersion at pH 2.0 (Fig. 6a) show that three different regions (A, B and C) can be distinguished in the time dependent UV–visible spectra. Region A shows loss in absorbance and a red shift of the transverse plasmon resonance band (~520 nm) with elapsed time. In region B the longitudinal plasmon resonance band (~700 nm) appears, and progressively red shifts, until in region C absorbance is lost. The decrease in intensity of the transverse plasmon band (Region A) and the appearance of the longitudinal plasmon band (Region B) would imply close selfassociation (Sendroiu et al., 2006) of the gold cores, followed by precipitation (Region C). These changes in the UV–visible spectra of CN and AN dispersions are in stark contrast to the results obtained for CN + SRHA and AN + SRHA dispersions at pH 1.5 (Figs. 5b and 6b), where the presence of SRHA suppresses the longitudinal band. At basic pHs there are some significant differences in the aggregation behaviour of CN and AN dispersions (Figs. 5c and 6c). It can be observed in Fig. 5c that the transverse surface plasmon band of CN dispersions at pH 12.0 does not reduce significantly, and there is no appearance of the longitudinal band, i.e. the CN particle dispersion appear to aggregate very slowly in basic conditions. In contrast, the UV–visible spectra of AN dispersions at pH 12.5 (Fig. 6c) displays a behaviour similar to those observed for acidic CN and AN dispersions, where aggregation is accompanied by the appearance of the longitudinal plasmon band. Clearly the citrate anions are a more effective stabilizer than the acrylate moieties at basic pHs. As might be expected, the aggregation behaviour of CN + SRHA particles at pH 12.5 (Fig. 5d) is very similar to the one observed for CN particles at pH 12.0 (Fig. 5c), i.e. slow aggregation and absence of the longitudinal plasmon band. However, comparison of the spectra reported in Fig. 6c (AN pH 12.5) and Fig. 6d (AN + SRHA pH 12.5), shows that the presence of SRHA results in marked differences in the aggregation behaviour of acrylate gold nanoparticles at basic pHs. Although the surface plasmon band of AN + SRHA dispersion at pH 12.5 reduces in intensity there is no commensurate rise of a longitudinal band, suggesting no coalescence of the gold cores. Interestingly, in the majority of cases where the longitudinal plasmon band (Region B) is present in the UV–visible

SC IE N CE OF TH E TOTA L E N V I RO N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

spectra, i.e. when SRHA is absent, the TEM images reveal significant fusion of the gold cores (Figs. 5a and 6a). Conversely when the longitudinal band is not present in the UV– visible spectra, i.e. when SRHA is present, the TEM images reveal that the gold cores are not fused (Figs. 5b-d and 6b).

59

Clearly the SRHA is having a significant effect on the structure of the aggregates (i) when they are in solution as observed in the UV–visible and DLS experiments and (ii) when they are deposited on a substrate (as observed in the TEM micrographs).

Fig. 7 – Schematic representation of the aggregation process for (a) CN and (b) CN + SRHA dispersions. (a) Citrate-stabilized gold nanoparticles are electrostatically stabilized by the negative citrate anions. When the pH of the colloidal dispersion is decreased protonation of the citrate anions occurs. Without their negative shell the particles are not prevented from coming into close proximity and aggregate. This causes the appearance of the longitudinal plasmon band. (b) When the SRHA is added to the citrate-stabilized gold nanoparticle dispersion, partial substitution of the small citrate anions by the polyanionic SRHA is likely to occur. When the pH of the solution is decreased protonation of the phenolic and carboxylic functionality of the SRHA occurs and the particles loose their electrostatic stabilization. However, aggregation is retarded by the steric stabilization offered by the SRHA. Moreover, when aggregates form in solution the SRHA prevents the gold nanoparticles from coming in close proximity and the longitudinal plasmon band is absent from the UV–visible spectra of aggregating dispersion containing SRHA.

60

SC IE N CE OF TH E TOTA L E N V IR O N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

We propose the following conceptual model to account for the increased stabilization of the gold colloids when SRHA is present. When the pH of the CN and AN dispersions is lowered, protonation of the anionic stabilizers occurs and the gold nanoparticles are no longer electrostatically stabilized (Fig. 7a, Step 1). This change in the surface charge allows the particles to come in close proximity during aggregation (Fig. 7a, Step 2) resulting in the reduction of the absorbance of the transverse plasmon band at 519 nm and 530 nm for CN and AN, respectively (Region A) and the commensurate increase of the longitudinal plasmon resonance band (Region B) (Figs. 5a and 6a), followed by precipitation (Fig. 7a, Step 3). We base this hypothesis on detailed work of Sendroiu et al. (2006) who have shown that the loss of the transverse plasmon band and commensurate appearance of the longitudinal plasmon band are dependent on the interparticle spacing (Blatchford et al., 1982; Weisbecker et al., 1996). This hypothesis is also corroborated by the corresponding TEM images in Figs. 5a and 6a showing that a high proportion of fused aggregates are observed. However, when SRHA is present we propose that the small citrate and acrylate anions are substituted (and/or overcoated) by the humic acid (Fig. 7b, Step 1) as supported by the blue shift in the transverse plasmon band (Fig. 3). Upon acidification the SRHA anionic moieties become protonated in a similar fashion to the citrate and acrylate anions, and no longer provide electrostatic stabilization. However, due to its macromolecular nature the SRHA has a larger foot print than the small molecular citric and acrylic acids compared with citrate and acrylate moieties (Stumm, 1992). Hence the neutralised SRHA now provides steric stabilization to the gold colloids (Fig. 7b, Step 2) and the longitudinal band is suppressed (Fig. 5b and c) as the gold nanoparticles do not come into close association and therefore cannot fuse (Fig. 7b, Step 3 compare with Fig. 7a, Step 2).

5.

Conclusions

Study of the interaction of citrate- and acrylate-stabilized gold nanoparticles with naturally occurring humic acid has provided an insight into the stability of these gold colloids in natural waters. Evidence has been found that SRHA are able to substitute and/or over-coat the citrate and acrylate anions on the gold nanoparticles surface. In particular, citrate- and acrylate-stabilized gold nanoparticles have shown increased stability against pH induced aggregation as a consequence of interactions with SRHA, which increases gold stability. This finding is of great interest from an environmental point of view, since the aggregation status of colloids influences their residence time in surface waters (Lead and Wilkinson, 2006) and their transport in ground water (Kretzschmar and Schafer, 2005). The results show that the increased stability associated with organic colloids is eliminated by high ionic strengths because of charge screening. Thus, it will be expected that in high ionic strength waters (hard waters and estuarine systems), aggregation will be rapid and extensive. In such cases, benthic (sediment dwelling) organism will be the biota most affected by nanoparticle pollution. In other soft freshwaters (at low ionic strength), increased stability will be likely

and our findings are of particular interest for acid affected environments (Benison et al., 1998; Lessmann et al., 2000; Koschorreck et al., 2003; Praharaj and Fortin, 2004; Sendroiu et al., 2006) and eutrophic systems (Scott et al., 2005) in which the water pH can be lower than pH 3.0 and higher than pH 10.0. Nevertheless, the general mechanism of surface sorption and modification and enhanced stability is likely to be a general mechanism in all aquatic and terrestrial systems and will need to he accounted for in ecotoxicological studies. Furthermore, one might well expect that even in normal environmental pH conditions such particles will have enhanced stability given the proposed model, and therefore, these results need to be taken into account when addressing the ecotoxicity of these materials.

Acknowledgements This work was financially sponsored by EU grants (Micro-Nano HPRN-CT-2000-00028 for studentship to SD, MEST-CT-2004504356 to ALM and Nano3D NMP4-CT-2005-014006), the EPSRC for a DTA to SB and the NERC (NE/D004942/1).

REFERENCES Baalousha M, Motelica-Heino M, Le Coustumer P. Conformation and size of humic substances: effects of major cation concentration and type, pH, salinity, and residence time. Colloids Surf A Physicochem Eng Asp 2006;272:48–55. Balnois E, Wilkinson KJ, Lead JR, Buffle J. Atomic force microscopy of humic substances: effects of pH and ionic strength. Environ Sci Technol 1999;33:3911–7. Benison KC, Goldstein RH, Wopenka B, Burruss RC, Pasteris JD. Extremely acid Permian lakes and ground waters in North America. Nature 1998;392:911–4. Blatchford CG, Campbell JR, Creighton JA. Plasma resonance enhanced Raman-scattering by adsorbates on gold colloids — the effects of aggregation. Surf Sci 1982;120:435–55. Buffle J. The key role of environmental colloids/nanoparticles for the sustainability of life. Environ Chem 2006;3:155–8. Chau LK, Lin YF, Cheng SF, Lin TJ. Fiber-optic chemical and biochemical probes based on localized surface plasmon resonance. Sens Actuators B Chem 2006;113:100–5. Colvin VL. The potential environmental impact of engineered nanomaterials. Nature Biotechnology 2003;21:1166–70. Dewit JCM, Vanriemsdijk WH, Koopal LK. Proton binding to humic substances .2. Chemical heterogeneity and adsorption models. Environ Sci Technol 1993a;27:2015–22. Dewit JCM, Vanriemsdijk WH, Koopal LK. Proton binding to humic substances .1. Electrostatic effects. Environ Sci Technol 1993b;27:2005–14. Diegoli S, Mendes PM, Baguley ER, Leigh SJ, Iqbal P, Garcia Diaz YR, et al. pH-dependent gold nanoparticle self-organization on functionalized Si/SiO2 surfaces. J Exp Nanosci 2006;1:333–53. dos Santos DS, Alvarez-Puebla RA, Oliveira ON, Aroca RF. Controlling the size and shape of gold nanoparticles in fulvic acid colloidal solutions and their optical characterization using SERS. J Mater Chem 2005;15:3045–9. Frens G. Controlled nucleation for regulation of particle-size in monodisperse gold suspensions. Nat Phys Sci 1973;241:20–2. Giasuddin ABM, Kanel SR, Choi H. Adsorption of humic acid onto nanoscale zerovalent iron and its effect on arsenic removal. Environ Sci Technol 2007;41:2022–7.

SC IE N CE OF TH E TOTA L E N V I RO N ME N T 4 02 ( 20 0 8 ) 5 1–6 1

Haiber S, Herzog H, Burba P, Gosciniak B, Lambert J. Two-dimensional NMR studies of size fractionated Suwannee river fulvic and humic acid reference. Environ Sci Technol 2001;35:4289–94. Henglein A. Physicochemical properties of small metal particles in solution — microelectrode reactions, chemisorption, composite metal particles, and the atom-to-metal transition. J Phys Chem 1993;97:5457–71. Her N, Amy G, Foss D, Cho JW. Variations of molecular weight estimation by HP-size exclusion chromatography with UVA versus online DOC detection. Environ Sci Technol 2002;36:3393–9. Hong R, Han G, Fernandez JM, Kim BJ, Forbes NS, Rotello VM. Glutathione-mediated delivery and release using monolayer protected nanoparticle carriers. J Am Chem Soc 2006;128:1078–9. Hussain I, Brust M, Papworth AJ, Cooper AI. Preparation of acrylate-stabilized gold and silver hydrosols and gold-polymer composite films. Langmuir 2003;19:4831–5. Hutchings GJ. Catalysis by gold. Catal Today 2005;100:55–61. Hyung H, Fortner JD, Hughes JB, Kim JH. Natural organic matter stabilizes carbon nanotubes in the aqueous phase. Environ Sci Technol 2007;41:179–84. Jian Z, Zhao JW, Wang YC. Influence of surface charge density on the plasmon resonance modes in gold nanoellipsoid. Phys B Condens Matter 2004;353:331–5. Kanellos, M. Carbon nanotubes enter Tour de France, CNET News. com, 07/07/2006. Khomutov GB, Kislov VV, Antipina MN, Gainutdinov RV, Gubin SP, Obydenov AY, et al. Interfacial nanofabrication strategies in development of new functional nanomaterials and planar supramolecular nanostructures for nanoelectronics and nanotechnology. Microelectron Eng 2003;69:373–83. Koschorreck M, Wendt-Potthoff K, Geller W. Microbial sulfate reduction at low pH in sediments of an acidic lake in Argentina. Environ Sci Technol 2003;37:1159–62. Kretzschmar R, Schafer T. Metal retention and transport on colloidal particles in the environment. Elements 2005;1:205–10. Lead JR, Wilkinson KJ. Aquatic colloids and nanoparticles: current knowledge and future trends. Environ Chem 2006;3:159–71. Lead JR, Wilkinson KJ, Starchev K, Canonica S, Buffle J. Determination of diffusion coefficients of humic substances by fluorescence correlation spectroscopy: Role of solution conditions. Environ Sci Technol 2000;34:1365–9. Lead JR, Starchev K, Wilkinson KJ. Diffusion coefficients of humic substances in agarose gel and in water. Environ Sci Technol 2003;37:482–7. Lee C, Kang Y, Lee K, Kim SR, Won DJ, Noh JS, et al. Molecular wires and gold nanoparticles as molewares for the molecular scale electronics. Curr Appl Phys 2002;2:39–45. Lessmann D, Fyson A, Nixdorf B. Phytoplankton of the extremely acidic mining lakes of Lusatia (Germany) with pH b= 3. Hydrobiologia 2000;433:123–8. Liao JH, Zhang Y, Yu W, Xu LN, Ge CW, Liu JH, et al. Linear aggregation of gold nanoparticles in ethanol. Colloids Surf A Physicochem Eng Asp 2003;223:177–83. Logunov SL, Ahmadi TS, ElSayed MA, Khoury JT, Whetten RL. Electron dynamics of passivated gold nanocrystals probed by subpicosecond transient absorption spectroscopy. J Phys Chem B 1997;101:3713–9. Maulucci G, De Spirito M, Arcovito G, Boffi F, Castellano AC, Briganti G. Particle size distribution in DMPC vesicles solutions undergoing different sonication times. Biophys J 2005;88:3545–50. Mie G. Articles on the optical characteristics of turbid tubes, especially colloidal metal solutions. Ann Phys 1908;25:377–445.

61

Mulvaney P. Surface plasmon spectroscopy of nanosized metal particles. Langmuir 1996;12:788–800. NanoHorizons. Polymer-Compatible Noble Metal Nanoparticles; 2007. http://www.nanohorizons.com/prodparticles.shtml, 27/04/. Norman TJ, Grant CD, Magana D, Zhang JZ, Liu J, Cao DL, et al. Near infrared optical absorption of gold nanoparticle aggregates. J Phys Chem B 2002;106:7005–12. Powell HKJ, Fenton E. Size fractionation of humic substances: effect on protonation and metal binding properties. Anal Chim Acta 1996;334:27–38. Praharaj T, Fortin D. Indicators of microbial sulfate reduction in acidic sulfide-rich mine tailings. Geomicrobiol J 2004;21:457–67. Rasi S, Mavelli F, Luisi PL. Cooperative micelle binding and matrix effect in oleate vesicle formation. J Phys Chem B 2003;107:14068–76. Rechberger W, Hohenau A, Leitner A, Krenn JR, Lamprecht B, Aussenegg FR. Optical properties of two interacting gold nanoparticles. Opt Commun 2003;220:137–41. Redwood PS, Lead JR, Harrison RM, Jones IP, Stoll S. Characterization of humic substances by environmental scanning electron microscopy. Environ Sci Technol 2005;39:1962–6. Roco MC. International perspective on government nanotechnology funding in 2005. J Nanopart Res 2005;7:707–12. Sau TK, Pal A, Pal T. Size regime dependent catalysis by gold nanoparticles for the reduction of eosin. J Phys Chem B 2001;105:9266–72. SCHENIR. The appropriateness of the risk assessment methodology in accordance with the Technical Guidance Documents for new and existing substances for assessing the risk of nanomaterials; 2007. 29 March. Schmitt J, Machtle P, Eck D, Mohwald H, Helm CA. Preparation and optical properties of colloidal gold monolayers. Langmuir 1999;15:3256–66. Scott DM, Lucas MC, Wilson RW. The effect of high pH on ion balance, nitrogen excretion and behaviour in freshwater fish from an eutrophic lake: a laboratory and field study. Aquat Toxicol 2005;73:31–43. Sendroiu IE, Mertens SFL, Schiffrin DJ. Plasmon interactions between gold nanoparticles in aqueous solution with controlled spatial separation. Phys Chem Chem Phys 2006;8:1430–6. Sharma AK, Gupta BD. Fiber optic sensor based on surface plasmon resonance with nanoparticle films. Photonics Nanostruct Fundam Appl 2005;3:30–7. Shipway AN, Lahav M, Gabai R, Willner I. Investigations into the electrostatically induced aggregation of Au nanoparticles. Langmuir 2000;16:8789–95. Stumm W. Chemistry of the Solid-Water Interface: Processes at the Mineral-Water and Particle-Water Interface in Natural Systems. John Wiley & Sons Inc; 1992. Sutton R, Sposito G. Molecular structure in soil humic substances: the new view. Environ Sci Technol 2005;39:9009–15. Wang ST, Yan JC, Chen L. Formation of gold nanoparticles and self-assembly into dimer and trimer aggregates. Mater Lett 2005;59:1383–6. Weisbecker CS, Merritt MV, Whitesides GM. Molecular self-assembly of aliphatic thiols on gold colloids. Langmuir 1996;12:3763–72. Wiesner J, Wokaun A. Anisometric gold colloids — preparation, characterization, and optical-properties. Chem Phys Lett 1989;157:569–75. Yang PH, Sun XS, Chiu JF, Sun HZ, He QY. Transferrin-mediated gold nanoparticle cellular uptake. Bioconjug Chem 2005;16:494–6.