Hybrid Nanostructures:  Filling multiwalled carbon nanotubes with quantum dots

Christopher M. Evans

Department of Chemistry

University of Rochester



Semiconducting quantum dots (QDs) and single wall carbon nanotubes (SWNTs) are being evaluated for a variety of optoelectronic applications. Each material has shown a rich chemistry for functionalization, and several recent reports have evaluated the direct coupling of these two materials. Previous approaches to chemical attachment of QDs to SWNTs relied on covalent methods, which require the breaking of carbon–carbon bonds and the reaction of the resulting carboxylic acid groups to give products via nucleophilic acyl substitution. Unfortunately, all covalent reaction schemes alter the SWNT structural and electronic properties by disrupting the sp2 hybridized conjugation, thereby reducing conduction efficiency. Coupling of QDs to SWNTs is expected to produce a composite material which facilitates selective wavelength absorption, charge transfer to SWNTs, and efficient electron transport. These properties are essential for additives in polymer photovoltaics utilizing CdSe QD–SWNT junctions. The potential energy level diagram for CdSe QD–SWNT complexes indicates that efficient photo-conversion is possible for properly engineered material junctions. A synthetic strategy which minimizes structural disruption at the QD–SWNT junctions is expected to advance nanomaterial development for polymeric solar cells. Therefore, I propose a methodology, whereby, no chemical bonds are broken, but instead, to fill the volume of the nanotube and rely on steric effects to maintain an electronic coupling between the two nanostructures.

For this report, multiwalled carbon nanotubes will be filled with PbSe QDs and processed such that the inner volume of the MWNT will be entirely filled with a single crystal of PbSe. It is the intent of this project to both develop novel nanoscale heterostructures as well as to develop growth templates for PbSe quantum wires.


Quantum Dot Growth

PbSe QDs were grown from literature preparations that allow for precise control over nanocrystal size. In brief, PbO was heated to 150°C with oleic acid and octadecene for one hour under a flowing N2 environment. Upon dissolution of the yellow solid the temperature was reduced to 120 °C and 8 mL of 1 M trioctylphosphine-selenide was injected through the septa fitting. After 5 minutes the temperature was quenched by immersion in an ice bath and upon reaching room temperature the reaction product was rinsed sequentially with MeOH/BuOH to induce flocculation. The QD product was then centrifuged to isolate a solid and redissolved in tetrachloroethylene.

MWNT Filling

MWNTs were purchased from pyrograf industries and are purified prior to filling. A concentrated volume of QDs in tetrachloroethylene is combinmed with several milligrams of MWNTs and sonicated for 30 minutes prior to centrifugation. Rinsing excess QDs was achieved by adding 10 mL of hexane to the MWNT pellet and vortexing briefly. The resulting solution was centrifuges further to isolate a washed MWNT pellet.

Quantum Wire growth

A filled MWNT sample was placed in a quartz boat and heated to the desired temperature in a tube furnace under 10 sccm flowing Ar gas. The furnace was isothermal at the desired temperature for one hour prior to cooling back to room temperature.


In attempting to fill the volume of a MWNT with QDs it is necessary to choose appropriately sized nanostructures to allow for a high packing density. Figure 1a is a transmission electron microscope (TEM) image of PbSe QDs and shows that their synthesis results in uniform spherical particles with a diameter distribution centered at 5.5 nm (Figure 2b). Analogously to a particle in a box concept, the size of the QD is proportional to the energy level of the 1st excited state (bandgap). The absorption/photoluminescense energies of the QD can be controlled by changing their size. For instance, 5.5 nm PbSe QDs bandgap is quantum confined from the bulk value of 0.25 eV (4960 nm) to 0.80 eV (1550 nm). This is a unique phenomena that occurs for nanoscale dimensions and highlights the potential applications of QDs.

Figure 1. a.) TEM of PbSe QDs b.) SEM of MWNTs and c.) TEM of native MWNTs. Residual iron catalyst can be seen as dark spots between layers.

Figure 2. a.) Size histogram for the inside diameter (red) and outside diameter (blue). The sizes for this sample are polydisperse and it was observed that the average I.D. was 48 ± 21 nm with an average of 82 ± 51 individual layers per MWNT. b.) histogram for the diameters of PbSe QDs shown in Figure 1a. The diameter distribution was observed to be 5.4 ± 0.9 nm.

MWNTs are also nanoscale and Figure 1b is a scanning electron microscopy (SEM) image encompassing several hundred nanotubes. It can be seen that these MWNTs are roughly 10 microns long but vary dramatically in diameter. A TEM image (Figure 1c) of several purified MWNTs allows for the observation of the concentric layers of cylindrical graphite that make up a MWNT. Also, we can see that the inner volume varies greatly between these few nanotubes, ranging between 20 and 130 nm. A more detailed particle analysis is shown in Figure 2a and results in an average outer diameter of 102 ± 44 nm and an inner diameter of 48 ± 21 nm. The inter-planar distance between graphitic MWNT layers is 3.3 Å and given our distribution of inner and outer diameters it can be calculated that on average each MWNT has 82 ± 51 individual layers. The large standard deviations highlight the size dispersity for these particular MWNTs. For future work it would be desirable to use MWNTs with controlled diameters.

Figure 3. a.) TEM of unwashed and b.) washed QD-MWNT filled sample. The unwashed sample shows outstanding coverage (interior and exterior) of the MWNT, however, after washing the packing density on the exterior has been reduced.

Figure 4. Treatments without first washing away excess QDs results in particle formation on the exterior of the MWNT.

For filling MWNTs, PbSe QDs were chosen because Lead is a large electron rich element and gives good contrast in a TEM. However, filling is by no means limited to this particular stoiciometry and could easily be extended to any number of QDs compositions. The un-washed filled MWNTs are shown in Figure 3a and are completely covered and filled with QDs. In addition, there is such an excess of QDs that numerous quantities can be observed on the formvar grid as well. It is important to remove any excess QDs by washing to preferentially remove any/all QDs at the surface. Upon heating the exterior QDs can also nucleate and will grow into tiny semiconductor tumors on the surface of the MWNT (Figure 4). A gentle wash with a strong QD solvent (hexane) preferentially removes QDs on the surface and only those on the inside of the MWNT will remain. However, for larger inner diameters the washing step greatly reduced the packing efficiency of QDs. It would be desirable to close the ends of the MWNT prior to washing to encapsulate the QDs inside the MWNT permanently.

Heating the QD-MWNT sample to elevated temperatures was expected to force nucleation of individual QDs and result in growth. Two temperatures were evaluated (400 and 600 °C) and both show startlingly different results. After a 400 °C treatment the QDs aggregated and grew into small cubes ranging from 8 to 40 nm in size. Considering that the bulk crystal structure of PbSe is rock salt it is not too surprising to see the QDs grow in cubes. However, they did not agglomerate together to fill the inner-volume of the MWNT. At an even higher temperature (600 °C), the QDs have completely coalesced filling the inner volume and creating a single crystal of PbSe. It appears that elevated temperatures will successfully produce quantum wires inside the MWNT. Furthermore, it shows that MWNTs can be utilized as a molecular template to grow specific diameters/lengths of wires.

Figure 5. a.) 400 and b.) 600 C heat treatments for washed QD-MWNT samples resulting in cubic and wire PbSe respectively.

Additional PbSe wires can be seen in Figure 6 in both bright (a,c) and dark (b,d) field TEM. The crystalline nature of the wires results in strong diffraction of the incident electron beam. Therefore, electrons incident on the PbSe wires will diffract differently than on the graphitic-like MWNT resulting in improved contrast. Dark-field imaging is very useful for these particular samples since the PbSe wires show up as extremely bright spots on the image differentiating themselves from the primarily carbon sample.

Figure 6. a.) c.) Bright-field b.) d.) dark-field TEM images for PbSe quantum wires inside MWNT. Dark-field imaging results in outstanding contrast between the graphitic MWNT and the crystalline PbSe.

Imaging processes such as TEM and SEM are sample limiting techniques that do not sample a representative section of an entire sample but merely interrogate a sliver of the total. Therefore, techniques that examine larger sections may give more representative information than simply imaging a few specific features of a much larger sample. In energy dispersive spectroscopy (EDS) the primary electron beam ionizes atoms within the interaction volume and via inelastic collisions with nuclei produces X-rays of characteristic energy for each atom in the periodic table. This technique can be used quantitatively investigate the elemental composition of a sample. From the observed intensities/area for each x-ray line a software program corrects the intensities for elemental factors such as size, absorption and fluorescence and results in atomic percentages. However, these technique is very sensitive to the surface reconstruction of the sample and is ideally suited for a flat and uniform sample. As seen in Figure 1b, our MWNT sample is far from ideal and the EDS spectra shown in Figure 7 is not absolute. However, the observed atomic ratio for Pb:Se was 1.65:1.0 consistent with a Pb terminated QD which will have an excess of Pb. Additionally, we observe iron and copper in the EDS data which can be accounted for by residual catalyst from synthesis (Figure 1c) and contributions from the formvar TEM grid respectively.

Figure 7. EDS data for the PbSe-MWNT sample. Copper and iron are from the formvar TEM grid and residual nanotube growth catalyst respectively.

In addition to TEM, we can also observe the formation of PbSe quantum wires by using scanning tunneling electron microscopy (STEM). In this particular technique, a SEM is outfitted with a special sample holder where the detector is placed below the sample. The detector is split into several zones and by careful choice of zones you can acquire either a bright-field or dark-field image. Figure 8 shows both bright (a) and dark-field imaging for a PbSe quantum wire in a MWNT. The resolution/contrast are comparable to the TEM, however, these images were produced in a much less expensive and easier to use instrument and also were acquired in a fraction of the time.

Figure 8. a.) Bright-field and b.) dark-field STEM images for a single PbSe quantum wire grown inside a MWNT.


PbSe quantum wires can be successfully produced within the inner diameter of a MWNT.  The inner diameter acts as a template that can be used to control the size of the resulting quantum wire.

Christopher M. Evans, April 30th, 2008

Please enter any comments, criticisms, questions, etc. below.

Your name:

Email address: