Journal of Nanoparticle Research

, Volume 12, Issue 1, pp 75–82

Dispersion of multi-walled carbon nanotubes in biocompatible dispersants

Authors

  • J.-P. Piret
    • Laboratory of Biochemistry and Cellular Biology (URBC)University of Namur
  • S. Detriche
    • Laboratory of Chemistry and Electrochemistry of SurfacesUniversity of Namur
  • R. Vigneron
    • Laboratory of Analysis by Nuclear Reactions (LARN)University of Namur
  • S. Vankoningsloo
    • Laboratory of Biochemistry and Cellular Biology (URBC)University of Namur
  • S. Rolin
    • Laboratory of PharmacyUniversity of Namur
  • J. H. Mejia Mendoza
    • Laboratory of Chemistry and Electrochemistry of SurfacesUniversity of Namur
  • B. Masereel
    • Laboratory of PharmacyUniversity of Namur
  • S. Lucas
    • Laboratory of Analysis by Nuclear Reactions (LARN)University of Namur
  • J. Delhalle
    • Laboratory of Chemistry and Electrochemistry of SurfacesUniversity of Namur
  • F. Luizi
    • Nanocyl s.a.
  • C. Saout
    • Laboratory of Biochemistry and Cellular Biology (URBC)University of Namur
    • Laboratory of Biochemistry and Cellular Biology (URBC)University of Namur
Special focus: Safety of Nanoparticles

DOI: 10.1007/s11051-009-9697-8

Cite this article as:
Piret, J., Detriche, S., Vigneron, R. et al. J Nanopart Res (2010) 12: 75. doi:10.1007/s11051-009-9697-8

Abstract

Owing to their phenomenal electrical and mechanical properties, carbon nanotubes (CNT) have been an area of intense research since their discovery in 1991. Different applications for these nanoparticles have been proposed, among others, in electronics and optics but also in the medical field. In parallel, emerging studies have suggested potential toxic effects of CNT while others did not, generating some conflicting outcomes. These discrepancies could be, in part, due to different suspension approaches used and to the agglomeration state of CNT in solution. In this study, we described a standardized protocol to obtain stable CNT suspensions, using two biocompatible dispersants (Pluronic F108 and hydroxypropylcellulose) and to estimate the concentration of CNT in solution. CNT appear to be greatly individualized in these two dispersants with no detection of remaining bundles or agglomerates after sonication and centrifugation. Moreover, CNT remained perfectly dispersed when added to culture medium used for in vitro cell experiments. We also showed that Pluronic F108 is a better dispersant than hydroxypropylcellulose. In conclusion, we have developed a standardized protocol using biocompatible surfactants to obtain reproducible and stable multi-walled carbon nanotubes suspensions which can be used for in vitro or in vivo toxicological studies.

Keywords

Multi-walled CNTPluronic F108HPCSonicationCentrifugationHealth safetyNanomedicineEHS

Introduction

Due to their unique physical and chemical features, nanostructured materials have a wide range of industrial applications. This explains why nanotechnology has become one of the leading technologies over the past 10 years (Stix 2001). For example, the global market for carbon nanotubes (CNT) is predicted to grow up to between $1 billion and $2 billion by 2014 (Thayer 2007). Although the effects of nanoparticles on health and the environment are becoming more of a concern, studies on toxicology and the environmental impact of nanoparticles are still scarce (Colvin 2003).

One of the main issues to investigate the potential toxic effect of isolated CNT on human health using in vitro or in vivo models, is the poor solubility of CNT in water or biological fluids (Buford et al. 2007).

In order to obtain CNT solutions suitable for the in vivo or in vitro studies, there are currently two main ways available. First, CNT surface modifications using water-soluble chemical groups enhance CNT solubilization (Schipper et al. 2008; Wang et al. 2004; Zhao et al. 2008). However, the presence of chemical groups on CNT surface could modify the physico-chemical properties of CNT and render difficult to evaluate the potential toxic effect of the as-received manufactured CNT.

The second way to obtain well-dispersed CNT solutions consist in selection of dispersants for direct dispersion of CNT in water solutions (Jia et al. 2005). The use of dispersants allows dispersion of CNT while avoiding physico-chemical properties modification of the native CNT. However, the direct dispersion is especially difficult in order to achieve a well-dispersed, reproducible, stable CNT suspending solution which is mostly required and important in the biological studies/applications.

For example, there have been several published methodologies for the dispersion of CNT using organic solvents or polymers (Bahr et al. 2001; Holman and Lackner 2006; Thayer 2007). Although these dispersants are very effective at creating stable CNT dispersions, they are often uncompatible with biological experiments due to their intrinsic toxicity. Biocompatible surfactants, like Pluronic F108 (an amphiphilic copolymer composed by hydrophilic ethylene oxide and hydrophobic propylene oxide and used as drug delivery system) and Arabic gum have already been described to be helpful for CNT dispersion (Sayes et al. 2005; Wang et al. 2008). In addition, some suspension agents like hydroxypropylmethylcellulose (HPMC) or hydroxypropylcellulose (HPC), have also been used as dispersant for nanoparticles in different studies (Chen et al. 2006; Meng et al. 2007). However, in these different studies, the protocols for CNT dispersion preparation are not well described and most often only sketchy information about CNT dispersion state is provided (still presence of CNT bundles, only individual CNT in solution, etc.). In this study, we describe a standardized protocol to obtain stable multi-walled carbon nanotubes (MWCNT) suspensions using biocompatible dispersants and to estimate the concentration of MWCNT in solution.

Materials and methods

Carbon nanotubes

Nanocyl®-7000 series thin MWCNT were a kind gift of Nanocyl (Sambreville, Belgium) (Fig. 1). MWCNT are produced via the catalytic carbon vapor deposition process. The synthesis process uses ethylene gas with an alumina supporting cobalt and iron as catalysts.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9697-8/MediaObjects/11051_2009_9697_Fig1_HTML.jpg
Fig. 1

TEM and SEM images of the CNT for primary characterization. CNT morphology was investigated by TEM (a) or SEM (b)

The size and shape of MWCNT were characterized using transmission electron microscopy (TEM) and scanning electron microscopy (SEM). TEM images were taken using a Tecnai 10 apparatus (Philips) at an acceleration voltage of 80 kV. SEM images were taken using a JEOL 7500F TEM operating at an acceleration voltage of 15 kV. The JEOL 7500F microscope, allowing observation of nanoparticles with a resolution of 1.6 nm, is also equipped with an Energy Dispersive X-ray (EDX) detector for the analysis of the elementary composition (qualitative and quantitative) of nanoparticles. Dry powder of MWCNT was dispersed in ethanol and droplets were placed on TEM grids covered with non-porous or porous formvar before being dried. The surface area of MWCNT was studied by BET analysis.

The size and agglomeration state of MWCNT dispersed in biocompatible surfactants or cell medium were investigated by TEM using the Tecnai 10 apparatus (Philips).

Dispersion of MWCNT

Biocompatible dispersants Pluronic F108 and HPC were purchased from BASF (Ludwigshafen, Germany) and Fargon (Waregem, Belgium), respectively. A total of 1% w/v Pluronic F108 or HPC aqueous solutions were used to obtain homogeneous suspensions of MWCNT. MWCNT solutions were sonicated using a probe sonicator (VC 50T model, power 50 watts and frequency 20 KHz, Sonics & Materials Inc.) at amplitude of 20% for 90 min after 30 min agitation with a magnetic stirrer. MWCNT suspensions were centrifuged at 13000 rpm for 30 min unless stated otherwise.

Culture medium is Dulbecco’s minimal essential medium (DMEM) (Invitrogen) supplemented with 10% fetal bovine serum (FBS) and 5 mL of penicillin–streptomycin/500 mL medium (Biowhittaker Europe).

Estimation of MWCNT concentration

In order to estimate the concentration of MWCNT in suspension after centrifugation, a standard curve was established using non-centrifugated MWCNT-1% Pluronic F108 suspension obtained by sonication. The standard curve was obtained by plotting the absorbance for each standard value against the known concentration. Absorbance of the different standard points was obtained by wavelength scan between 700 and 600 nm using a UV–visible Uvikon spectrophotometer (BRS, Brussels, Belgium). Absorption values of 1% Pluronic F108 or 1% HPC were subtracted from those of MWCNT suspensions. The absorbance values measured between 700 and 600 nm were then used to obtain the average absorbance value for each standard value.

Results

MWCNT characterization

The MWCNT diameter, estimated by TEM analysis, and surface area (assayed by BET analysis) agree well with the average diameter and surface area data, respectively, specified by the manufacturer (Table 1; Fig. 1).
Table 1

CNT principal characteristics

Property

Data sheet

Analysis

 

Average diameter (nm)

9.5

12.7

 

Average length (μm)

1.5

ND

 

Surface area (m2/g)

250–300

324

 

Carbon purity (%)

90

86.2 ± 0.3

 

Metal oxide impurities (%)

10

O

8.8 ± 2.9

  

Al

4.4 ± 0.4

  

Fe

0.4 ± 1.5

  

Co

0.2 ± 1.9

Informations described by manufacturer’s data sheet are compared with those obtained during this study

ND not determined, O oxygen, Al aluminum, Fe iron, Co cobalt

Elementary analysis (using EDX detector) detected catalytic metals (aluminum, iron, cobalt) and oxygen. A Proton Induced X-ray Emission (PIXE) analysis confirmed the presence of aluminum, iron, and cobalt (data not showed).

Other metals may potentially be found depending on the production method and the sensitivity of the analytical method used (see for example Ge et al. 2008), but in any case, if present, they are below 100 ppm level.

MWCNT dispersion in biocompatible surfactants

As observed in Fig. 2, MWCNT form bundles and ropes in water. In order to allow dispersion of MWCNT in aqueous solution, two suspension agents, Pluronic F108 and HPC were used.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9697-8/MediaObjects/11051_2009_9697_Fig2_HTML.jpg
Fig. 2

Dispersion of CNT in water. CNT, at a concentration of 1 mg/mL, were dispersed in water by agitation (30 min) before analysing by TEM

The potential toxic effect of these two surfactants was tested by 3H-thymidine incorporation assay on different cell types like HepG2 cells, but they did not show any effect on cell viability (Pluronic F108: 91% ± 4 and HPC: 87% ± 3 versus control: 100% ± 11; non significative difference between dispersants and control; data analyzed by Student’s t tests). MWCNT, at 1 mg/mL (0.1%), were dispersed by sonication (90 min) in Pluronic F108 1% or HPC 1% solutions after 30 min of agitation. Agitation broke the biggest macroscopic agglomerates of MWCNT. As shown in Fig. 3, although both surfactants allow a good dispersion of MWCNT, still some MWCNT bundles and agglomerates were observed.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9697-8/MediaObjects/11051_2009_9697_Fig3_HTML.jpg
Fig. 3

Dispersion of CNT in two biocompatible suspension agents. CNT, at a concentration of 1 mg/mL, were dispersed in (a, c) Pluronic F108 1% or (b, d) HPC 1% by sonication (90 min) before analysing by TEM (c and d)

In order to eliminate these agglomerates, MWCNT suspension were first centrifuged for 30 min at 3500 rpm. Figure 4 shows that the MWCNT pellet obtained after centrifugation is bigger for the MWCNT suspension in HPC. This observation indicates that dispersion of MWCNT is better in Pluronic F108 solution than in HPC, i.e., there are more MWCNT bundles remaining in HPC solution after sonication.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9697-8/MediaObjects/11051_2009_9697_Fig4_HTML.jpg
Fig. 4

Pellets obtained after centrifugation of CNT suspensions. CNT were dispersed by sonication in (a) Pluronic F108 1% or (b) HPC 1% at 5 mg/mL before being centrifuged at 3500 rpm for 30 min

In order to improve the sedimentation of all MWCNT bundles, centrifugation speed was increased to 13000 rpm. The dispersion state of MWCNT in both solutions was then investigated by TEM (Fig. 5). Both suspensions were characterized by the total absence of agglomerates of MWCNT and only individualized MWCNT could be observed.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9697-8/MediaObjects/11051_2009_9697_Fig5_HTML.jpg
Fig. 5

Dispersion state of CNT in two biocompatible suspension agents after centrifugation. CNT, at a concentration of 1 mg/mL, were dispersed by sonication (90 min) in (a) Pluronic F108 1% or (b) HPC 1% and centrifuged 30 min at 13000 rpm before being analyzed by TEM

These data suggest that the protocol followed in this study allows a stable dispersion of MWCNT in biocompatible surfactants.

Estimation of the MWCNT concentration

One important parameter is the determination of MWCNT concentration after centrifugation. Because MWCNT are black body, they absorb steadily at all wavelengths of the visible spectrum. The absorption spectrum of a known concentration MWCNT suspension (1 mg/mL) was used to estimate the concentration of both MWCNT dispersed in Pluronic F108 and HPC solutions after centrifugation. As dispersion of MWCNT seems to be better and more homogeneous in Pluronic F108 solution after sonication (reduced proportion of MWCNT bundles, Fig. 4) and because of the similar dispersion state of MWCNT in both Pluronic F108 and HPC 1% solutions (after centrifugation) (Fig. 5), a Pluronic F108 solution containing MWCNT at a concentration of 1 mg/mL was used to set up a standard curve. As shown in Fig. 6, the establishment of this standard curve is reproducible allowing a good estimation of the MWCNT concentration after centrifugation in both dispersants.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9697-8/MediaObjects/11051_2009_9697_Fig6_HTML.gif
Fig. 6

Absorbance standard curve established from a suspension of CNT dispersed in Pluronic F108 1%

Dispersion of MWCNT in culture medium

As described above, Pluronic F108 and HPC allow a good dispersion of MWCNT leading to the individualization of MWCNT in solution. However, a crucial issue for the in vitro and in vivo cytotoxic studies is to determine whether these MWCNT suspensions could be diluted in culture medium, used for cells growing without altering MWCNT dispersion suspended in Pluronic F108 1% or HPC 1% solutions or water were diluted to 100 μg/mL in culture medium. While MWCNT dispersed in water form big agglomerates in culture medium, MWCNT suspended in Pluronic F108 1% or HPC 1% solutions appear well dispersed and stay as individual particles in culture medium without any re-agglomeration (Fig. 7).
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9697-8/MediaObjects/11051_2009_9697_Fig7_HTML.jpg
Fig. 7

Micrographs of CNT dispersion in culture medium. CNT dispersed in (a) Pluronic F108 1%, (b) HPC 1% or (c) water were diluted at 100 μg/mL in DMEM + 10% serum and analyzed by TEM

Discussion

MWCNT are becoming increasingly studied given their possible applications not only in electronics, optics, and mechanical materials but also for biological applications, such as imaging and drug delivery (Lacerda et al. 2006; Martin and Kohli 2003). Thus, it is imperative to assess the possible toxicity of these carbon-based nanostructures. Data already available in literature are controversial. Several toxicological studies have already demonstrated that CNT could exhibit in vitro and in vivo toxicity (Karlsson et al. 2008; Monteiro-Riviere et al. 2005; Muller et al. 2005; Shvedova et al. 2003, 2005), while other authors showed very low or no effect of CNT on cell viability (Davoren et al. 2007; Fiorito et al. 2006; Pulskamp et al. 2007). Different hypotheses could explain these discrepancies. First, MWCNT showed, for example, cytotoxic effects in human epidermal keratinocytes (Monteiro-Riviere et al. 2005) but not in NR8383 macrophage cell line and A549 alveolar cell line (Pulskamp et al. 2007), underlining the fact that cytotoxicity could depend on the cell type used. Another possible explanation may be differences in metal impurities content, which may be as high as 50% in a sample of CNT, depending on the synthesis process (Lam et al. 2006). Moreover, Ge et al. (2008), using neutron activation analysis and inductively coupled plasma mass spectrometry analysis, have recently shown that about 15 different impurity elements could be present in CNT that could influence the potential toxic effect of CNT. However, some studies have shown that effects of MWCNT on cell viability and DNA damage were not dependent on the soluble metals impurities released from MWCNT (Karlsson et al. 2008). Besides cell type and metal impurities, dispersion rate and agglomeration state of CNT in biological fluids and cell culture media could also affect toxicity. In fact, several toxicological studies have been done using agglomerated CNT (Davoren et al. 2007; Karlsson et al. 2008; Pulskamp et al. 2007; Rotoli et al. 2008). The effect of MWCNT on cell viability could be due, in part, to differences in dispersion state (Smart et al. 2006).

In this study, we described the setting of a standardized protocol using biocompatible dispersants to obtain reproducible and stable MWCNT suspensions which can be used for in vitro or in vivo toxicological studies. MWCNT appear to be greatly individualized in these two dispersants without detectable presence of any remaining bundles or agglomerates after sonication and centrifugation. These MWCNT suspensions will be useful to study the biological effect of individualized MWCNT and to compare their effect with previous toxicological data using non-perfectly dispersed MWCNT. Indeed, MWCNT remain well dispersed when added in cell culture medium. Moreover, the use of two different suspension agents will allow comparing the potential effect of dispersants on MWCNT toxicity.

The establishment of a reproducible standard curve using non-centrifugated MWCNT-Pluronic F108 1% solution allows a good estimation of the MWCNT concentration after centrifugation in both dispersants. Lastly, we also showed that the two dispersants do not have the same capability to fully disperse MWCNT, Pluronic F108 providing the best suspensions of isolated MWCNT into water and cell culture medium.

Another interesting point for future investigations is related to the interaction of CNT with the components of cell culture medium. This medium contains various molecules such as amino acids and sugars, as well as fetal calf serum rich in proteins and growth factors necessary for maintaining cells in culture. Although CNT dispersion state in solution could be a important parameter participating to the potential cytotoxicity of CNT, the association of CNT with molecules present in the cell environment could also be of interest when considering the toxicity of CNT in biological fluids.

In conclusion, the reproducible dispersion protocol using different biocompatible surfactants described in this study could be useful to set up a common MWCNT dispersion protocol in order to decrease the discrepancy in toxicological results depending on the variety of strategies regarding CNT dispersion into biological media.

Acknowledgments

This work is supported by the “Direction Générale des Technologies de la Recherche et de l’Energie” (DGTRE) of the Walloon Region of Belgium (Nanotoxico Project, RW/FUNDP research convention No 516252). O. Toussaint is a Research Associate of the Belgian FNRS.

Copyright information

© Springer Science+Business Media B.V. 2009