Carbon materials comprising carbon nanotubes and methods of making carbon nanotubes

Abstract

The present invention relates to carbon materials comprising carbon nanotubes, powders comprising carbon nanotubes and methods of making carbon nanotubes. In the methods of the present invention, the size and/or formation of floating catalyst particles is closely controlled. The resulting carbon nanotubes typically exhibit armchair chirality and typically have metallic properties. The carbon nanotubes produced by this method readily form bulk materials, which typically have a conductivity of at least 0.7?106 Sm1 in at least one direction. The invention has particular application to the manufacture of components such as electrical conductors. Suitable electrical conductors include wires (e.g. for electrical motors) and cables (e.g. for transmitting electrical power).

Claims

1. A method of producing carbon nanotubes, the method comprising: providing a plurality of floating catalyst particles, wherein at least 70% by number of the catalyst particles have a diameter less than or equal to 4.5 nm; and contacting the floating catalyst particles with a gas phase carbon source at a carbon nanotube formation temperature of at least 900? C. to produce carbon nanotubes, wherein the floating catalyst particles are provided by: initiating growth of the catalyst particles by thermal degradation of a catalyst source substance, the thermal degradation of the catalyst source substance beginning at a first onset temperature being 700? C. or less, and subsequently arresting the growth of the catalyst particles using an arresting agent the arresting agent being sulfur, the arresting agent being provided to the catalyst particles by thermal degradation of an arresting agent source substance comprising carbon disulphide (CDS), the thermal degradation of the arresting agent source substance beginning at a second onset temperature being 800? C. or less, wherein the second onset temperature is in the range of temperatures from 10? C. more than the first onset temperature to 350? C. more than the first onset temperature; and wherein the arresting agent source substance, the catalyst source substance and the carbon source pass through the reaction chamber in a gas phase and in a flow direction, wherein the temperature in the reaction chamber varies along the flow direction from the first onset temperature to the second onset temperature to the carbon nanotube formation temperature.

2. The method according to claim 1 wherein the first onset temperature is in the range from 300? C. to 700? C., and wherein the second onset temperature is in the range from 350? C. to 750? C.

3. The method according to claim 1, further comprising performing the contacting step substantially continuously for at least 10 minutes.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Preferred embodiments of the invention will now be described, with reference to the accompanying drawings in which:

(2) FIG. 1 shows a Kataura plot.

(3) FIG. 2 shows a schematic illustration of typical diffraction patterns obtained for (A) armchair carbon nanotubes, and (B) zigzag carbon nanotubes.

(4) FIG. 3 illustrates an embodiment of the method of the present invention, showing (A) a schematic illustration of the apparatus used, (B) a typical temperature gradient within the reaction chamber, and (C) a schematic flow chart illustration of the method.

(5) FIG. 4 shows SEM images of carbon materials produced in the examples.

(6) FIG. 5 shows typical Raman spectra of carbon materials produced in the examples.

(7) FIG. 6 shows (A) and (B) extracts from a typical Raman spectrum of a carbon material produced in the examples, and (C) an annotated Kataura plot.

(8) FIG. 7 shows TEM images of carbon materials produced in the examples.

(9) FIG. 8 shows the carbon nanotube diameter distribution of carbon materials produced in the examples, determined using TEM.

(10) FIG. 9 shows (A) a HREM image of catalyst particles withdrawn from the reactor in the examples, and (B) the diameter distribution of these particles.

(11) FIG. 10 shows the thermal degradation temperatures of reactants used in the examples, and relates them to the temperature profile in the reactor.

(12) FIG. 11 shows (A) an example electron diffraction pattern obtained from a bundle of carbon nanotubes having armchair chirality, and (B) a marked up version of the pattern of FIG. 11A with oval marks indicating the location of the diffraction spots.

(13) FIG. 12 shows (A) an example electron diffraction pattern obtained from a bundle of carbon nanotubes having armchair chirality, and (B) a marked up version of the pattern of FIG. 11A with oval marks indicating the location of the diffraction spots.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

(14) A preferred embodiment of the method of the present invention will now be described. A schematic flow chart illustration of the method is shown in FIG. 3C, and a schematic illustration of the apparatus used is shown in FIG. 3A.

(15) A gaseous mixture of carbon source (e.g. methane), catalyst source substance (e.g. ferrocene) and arresting agent source substance (e.g. carbon disulphide) is fed into a furnace, carried in a stream of gas (e.g. hydrogen and/or helium). The gas mixture flows through the furnace in a flow direction.

(16) The temperature increases along the flow direction, so that the mixture is first subjected to a first onset temperature, at which temperature the catalyst source substance degrades to initiate growth of catalyst particles. For example, iron atoms may be released, to form catalyst particles comprising iron. Further along the flow direction, the mixture is subjected to a second onset temperature, at which temperature the arresting agent source substance degrades. The arresting agent is thus released, and acts to arrest the growth of the catalyst nanoparticles. The mixture is then subjected to a carbon nanotube formation temperature, and carbon nanotubes are produced.

(17) As illustrated in FIG. 3A, the resulting carbon nanotubes may be densified by supplying a densification agent (e.g. acetone). The carbon nanotubes may be drawn into a fibre. A typical winding rate is from 10 m s.sup.?1 to 20 m s.sup.?1. It will be understood that much higher winding rates may be employed.

(18) A typical temperature gradient within the furnace is illustrated in FIG. 3B.

EXAMPLES

(19) Continuous production of carbon nanotube fibres (schematic as shown in FIG. 3A) was carried out in a vertical ceramic reactor (d=80 mm, l=2 m), with a temperature profile as shown in FIG. 3B. The feedstock contained a carbon precursor (methane) and vapours of a catalyst source substance (ferrocene) and a sulphur source substance, carried by helium. The feedstock was introduced in to the reactor through a steel injector tube (d=12 mm, l=90 mm).

(20) On thermolysis of the feedstock components in a reductive atmosphere of hydrogen followed by synthesis of nanotubes, a plume composed of entangled nanotubes was obtained which was continuously drawn at 20 m min.sup.?1 and densified with an acetone spray into a fibre. The fibre had a typical diameter of 10 ?m.

(21) The effect of two different sulphur precursors, thiophene and carbon disulphide (CDS), on the morphology of the carbon nanotubes constituting the fibres produced was investigated. The input concentrations of the various precursors were optimised experimentally to provide continuous spinning of the fibre. The elemental ratios used are presented in Table 1.

(22) TABLE-US-00001 TABLE 1 Input precursor concentrations and the elemental ratios Precursors mol min.sup.?1 Elemental ratio (?10.sup.?5) (?10.sup.?3) Carbon Catalyst Promoter Fe/S Fe/C Case 1 170 0.21 C.sub.4H.sub.4S 2.5 80 8 Case 2 170 0.21 CS.sub.2 18 6 8

(23) (The higher input concentration of sulphur where CDS is used, compared to where thiophene is used, reflects the fact that smaller catalyst particles are formed where CDS is used (see below); the surface area to volume ratio of these catalyst particles is higher. Additionally, the CDS becomes available at an earlier stage of catalyst particle growth (again, see below), at which stage there is a higher number density of forming catalyst particles.)

(24) The analyses of the fibre microstructure and the constituting nantotubes were carried out by electron microscopy (FEI Tecnai F20-G2 FEGTEM, JEOL 2000FX and JEOL 6340 FEG), Raman Spectroscopy using a Renishaw Ramanscope 1000 system (incident light of wavelength 633 nm and 514 nm; acquisition time=10 s; laser spot size=1 ?m). The mechanical properties of the fibres were investigated with tensile tests using Textechno Favimat, a dedicated fibre testing equipment which employs a load cell with a force and displacement measurement range of 0-2 N (resolution=0.0001 cN) and 0-100 mm (resolution=0.1 micrometer) respectively. Testing was carried out at a standard gauge length of 20 mm and a test-speed of 2 mm min.sup.?1 to acquire the specific strength and specific stiffness (expressed in N Tex.sup.?1, these values are numerically equivalent to GPa SG.sup.?1) of the fibres.

(25) Results

(26) Fibre Composition, Microstructure and Nanostructure

(27) SEM Analysis

(28) Typical SEM images of the condensed and the internal structure of the uncondensed fibres from both the CDS and thiophene runs are presented in FIG. 4. FIG. 4A shows a typical condensed fibre, FIGS. 4B and 4C shows the internal structure of the fibre prior to acetone densification, where CDS is used as the sulphur precursor (B), and thiophene is used as the sulphur precursor (C). The nanotubes shown are orientated in the fibre direction. It can also be seen that the CDS fibre shows minimal presence of extraneous materials (which are generally by-products of most CVD processes) in comparison to the thiophene fibre.

(29) Raman Spectroscopy

(30) Raman spectra were acquired on the fibre samples with the polarisation of the incident light parallel to the fibre axis. At least 10 spectra were collected along the length of the fibre per fibre sample. The fibre samples were 1 cm in length, and acquisitions were spaced at equal intervals along the length. At least five samples obtained from each sulphur precursor were examined. The typical spectra for fibres obtained using CDS and thiophene are presented in FIG. 5. FIG. 5A shows a typical spectrum for a fibre obtained using CDS, and FIG. 5B shows a typical spectrum for a fibre obtained using thiophene. FIG. 5C shows the M and iTOLA regions of a typical Raman spectrum for a fibre obtained using CDS, and FIG. 5D shows the IFM region of a typical Raman spectrum for a fibre obtained using CDS.

(31) The positions of the peaks are in the spectra of FIGS. 5A and 5B and the D/G ratios (indicative of fibre purity and crystallinity of the nanotubes) are presented in Table 2.

(32) TABLE-US-00002 TABLE 2 The list of positions of the major peaks in the Raman spectra Position cm.sup.?1 Fibre RBM D G G I.sub.D/I.sub.G CDS 194.5 ? 3.8 1320.9 ? 1589.7 ? 2627.2 ? 0.010 ? 1.1 0.3 1.9 0.003 Thiophene Absent 1331.3 ? 1583.8 ? 2656.4 ? 0.3 ? 1.5 1.8 2.5 0.04

(33) The low D/G ratios shown by the fibres obtained using CDS in comparison to those obtained using thiophene are in agreement with the SEM results. They suggest minimal presence of extraneous materials and low density of defects in the nanotubes produced using CDS. The distinctive intense low frequency ring breathing modes (RBMs) occurring in the spectra from the CDS fibres indicate the presence of single-walled carbon nanotubes. In addition, the upshifted G peak (to 1590 cm.sup.?1), the downshifted D peak (1320 cm.sup.?1), the presence of M (1750 cm.sup.?1), i-TOLA (1950 cm.sup.?1) and intermediate frequency vibration modes (IFM modes 600-1200 cm.sup.?1) confirm that the fibres obtained with CDS as the sulphur precursor are composed of mainly single-walled carbon nanotubes. All these vibrational features are completely absent in fibres obtained with thiophene as the sulphur precursor and the G band and D band position occurring at 1582 cm.sup.?1 and 1331 cm.sup.?1 are suggestive of the presence of carbon nanotubes with more than one wall.

(34) G Band and RBM Analysis of CDS Fibres

(35) Further analysis of the G band (FIG. 6A) reveals an internal structure and in addition to the G+ feature at 1590 cm.sup.?1 (Lorentzian fit), the G-band occurs as a broad feature at 1552 cm.sup.?1 fit with the Breit-Wigner-Fanoline shape which indicates the predominant presence of metallic nanotubes. The Fano line shape is given by:

(36) $l\left(?\right)=l_0\frac{{\left[1+\left(?-{?}_{\mathrm{BWF}}\right)/q?\right]}^2}{1+{.Math.\left(?-{?}_{\mathrm{BWF}}\right)/q?.Math.}^2}$
where l.sub.0, ?.sub.0, ? and q are intensity, normalised frequency, broadening parameter and line shape parameter respectively.

(37) (FIG. 6A shows the internal structure of the G band, with the Lorentzian G+ and the G? exhibiting the Fano lineshape (see the above equation) with fit parameters l.sub.0, ?.sub.0, ? and q=2256, 1556, 49.5 and ?0.20 respectively.)

(38) The position of the radial breathing modes (RBMs) can be utilised to obtain the diameters of the nanotubes, as described above. It was observed that all the RBM frequencies noticed in the CDS fibre occur around 200 cm.sup.?1 at the excitation wavelength of 633 nm (FIG. 6B) corresponding to the diameter range of 1.2?0.2 nm (d=239/?.sub.RBM). This can be mapped to the Kataura plot, which is a theoretical model that relates the diameter of the nanotubes to the optical transition energies. Nanotubes of the same diameter can be either metallic or semiconducting, and the difference in the behaviour is shown in the differences in their optical transition energies. From the Kataura plot (FIG. 1 and FIG. 6C) it can be inferred that nanotubes in the diameter range of 1.1-1.4 nm with optical transition energies in the range of 1.96?0.1 eV are metallic, while those with transition energies in the range of 2.41?0.1 eV are semiconducting (the energy range of 0.1 eV takes in to account any transition energy shifts caused due to environmental effects such as nanotube bundling).

(39) Only those tubes with optical transition energies that are in resonance with the excitation energy (in the case of Raman spectroscopy, the incident laser light) will yield an RBM. While intense RBMs could be obtained with incident light of 633 nm (E.sub.excitation=1.96 eV), no resonance, and hence no RBMs, was observed when an incident light of 514 nm (E.sub.excitation=2.41 eV) was used (FIGS. 4B and 4C). This further confirms that the single-walled carbon nanotubes that constitute the CDS fibre are metallic.

(40) (FIG. 6B shows the representative RBM region of a typical Raman spectrum for a fibre obtained using CDS. It has a peak at 195 cm.sup.?1 with ?.sub.excitation=633 nm and the absence of the RBM peak with ?.sub.excitation=514 nm. FIG. 6C shows the metallic and semiconducting window in the Kataura plot (non-filled circles=metallic nanotubes, filled circles=semiconducting nanotubes) are marked red and green respectively on the original colour version of this drawing for nanotubes in the diameter range of 1.1 to 1.4 nm in correlation to the excitation energies used to acquire the Raman spectra (green region=2.41?0.1 eV, 514 nm; red region=1.96?0.1 eV, 633 nm).)

(41) TEM and Electron Diffraction

(42) Analysis by transmission electron microscopy indicates that the bundles that constitute the fibres, from both CDS and thiophene, are typically in the diameter range of 30-60 nm (FIGS. 7A and 7B respectively). From HREM analysis, the CDS fibres are composed of SWCNTs and those obtained with thiophene as sulphur precursor are composed of collapsed DWCNTs (FIGS. 7C and 7D respectively), confirming the findings from Raman spectroscopic analysis.

(43) The diameters and diameter distribution of the nanotubes are presented in Table 3 and FIG. 8. It can be seen that the diameters obtained from TEM analysis of the CDS fibres are in close agreement with those obtained from Raman spectroscopy (bulk characterisation).

(44) TABLE-US-00003 TABLE 3 Average diameters of the single-walled carbon nanotubes obtained using CDS, and collapsed double-walled carbon nanotubes obtained using thiophene, from TEM and RBM. Diameter.sub.TEM Diameter.sub.RBM Fibre (nm) (nm) CDS: Metallic SWCNT 1.4 ? 0.3 1.2 ? 0.2 Thiophene: DWCNT 7.6 ? 2.3 N/A

(45) The diameter distributions, determined using TEM are presented in FIG. 8, which shows (A) the diameter distribution of carbon nanotubes obtained using CDS, and (B) the diameter distribution of carbon nanotubes obtained using thiophene.

(46) Electron diffraction was carried out on fibre bundles (e.g. those represented in FIGS. 7A and B). The electron pattern from fibres obtained using CDS showed a pattern of clear spots, positioned to indicate armchair (m,n) tubes, with a chiral angle of 30?. In correlation with the diameter measurements, this suggests that the tubes are (10,10) tubes. Armchair tubes are metallic and hence, these results are in agreement with the characterisation by Raman spectroscopy. The electron diffraction patterns from the fibres obtained from thiophene are composed of continuous rings corresponding to (10-10) and (11-20) reflections, which shows that the nanotubes have a continuous distribution of helicities (i.e. there is a mixture of different chiralities).

(47) An example electron diffraction pattern is shown in FIG. 11A, for a bundle of carbon nanotubes having armchair chirality. Only half of the hexagonal pattern of spots is shown, the remaining spots are obscured by a shade. The arrow indicates the principal axis of the carbon nanotubes. FIG. 11B shows the same image, which has been marked up to show the location of the diffraction spots. The position of the three visible diffraction spots of the hexagonal pattern is indicated with white oval marker points. A similar example electron diffraction pattern is shown in FIG. 12A. FIG. 12B shows the same image as FIG. 12A, marked up to highlight the location of the diffraction spots.

(48) Catalyst Particles

(49) The catalyst particles formed when CDS is used were examined. The particles were frozen and withdrawn from the zone of the reactor where they form (in the temperature range 400-600 C. A HREM image of the withdrawn catalyst particles is shown in FIG. 9A. The diameter distribution of these particles, determined using HREM is shown in FIG. 9B. This figure shows that the catalyst particles have a narrow size distribution.

(50) The average diameter values of the frozen catalyst particles is 2.5?0.8 nm and the ratio of the average diameter of the catalyst particles to that of the single-walled carbon nanotubes is about 1.8, which is in close agreement with that reported in the literature.

(51) Fibre Properties

(52) Mechanical and Electrical Properties

(53) The mechanical properties of the metallic single-walled carbon nanotube fibre (obtained using CDS) and the double-walled carbon nanotube fibre (obtained using thiophene) are presented in Table 4. The fibres composed of collapsed DWCNT fibres are expected to be superior mechanically, due to the large contact area between the nanotubes held by van der Waals forces within the bundles, which is evinced in the tensile strength and stiffness values.

(54) TABLE-US-00004 TABLE 4 Fibre characteristics and mechanical attributes properties of the metallic SWCNT and DWCNT fibres along with those of copper wire of electrical wiring grade. Fibre characteristics Mechanical properties Diameter Linear density Sp. Strength Sp. Stiffness Material (?m) (g Km.sup.?1) (GPa SG.sup.?1) (GPa SG.sup.?1) Metallic SWCNT 10-15 0.04 0.5 10 fibre DWCNT fibre 10-15 0.04 1 20 Copper AWG 10 1820 2.3 ? 10.sup.4 0.03 14

(55) Table 5 below illustrates typical properties of materials, including a non-optimised fibre within the scope of the present invention (final row).

(56) TABLE-US-00005 Volumetric Linear Specific Current Specific Specific Conductivity density density conductivity density Strength Stiffness Material S/m ? 10.sup.6 g/m.sup.3 ? 10.sup.6 g/km S/m/g/m.sup.3 (A/mm.sup.2) GPa/SG GPa/SG Copper 58 8.9 6.5 2-10 0.025 13 (electrolytic) Aluminium 38 2.7 14.1 4 0.026 26 Steel/Iron 10 7.9 1.3 0.038 27 Carbon fibre 0.06 1.8 0.03 1.96 128 T300 TORAY High 0.14 1.9 0.07 2.06 309 performance carbonfibre (M60J) TORAY CNT yarn 0.1-0.7 0.8-1.2 0.02-0.1 0.1-0.7 30 0.8-1.2 60-140 non (depending metallic on winding rate) CNT yarn 0.7-3 (so far) 0.8-1.2 0.02-0.1 0.7-3 80 0.8-1 60-140 metallic (depending on winding rate)

(57) In this table, specific conductivity parameter which takes into account both electrical conductivity and the density of a conductor. In Table 5 below, the values for specific conductivity were estimated, from experimentally obtained values for conductance, length and linear density, using the equation below:

(58) ${?}^{}=\frac{G.Math.L}{\mathrm{LD}}{10}^3$
wherein G is conductance (in Siemens), L is length (in metres), and LD is linear density (in tex, or g km.sup.?1), and ? is specific conductivity in:

(59) $\frac{S{m}^{-1}}{gc{m}^{-3}}$

(60) It will be understood that both the electrical conductivity and the density of a conductor are important in many engineering applications, for example in overhead power lines.

(61) In the above examples, without wishing to be bound by theory, it is believed that the thermal degradation of ferrocene in hydrogen atmosphere begins at about 673K to yield iron atoms (d=0.3 nm) which subsequently grow into nanoparticles (which act as catalysts for the nucleation and growth of nanotubes). These nanoparticles are believed to pass through the reactor in the flow direction, along the temperature profile. The thermal degradation of the sulphur precursor (CDS or thiophene) in the reaction feedstock leads to the interaction of the iron nanoparticles with sulphur. We call this sulphudisation. The addition of sulphur, a recognised promoter in carbon nanotube growth, allows the production of long carbon nanotubes (typically mm). It is believed that this can enhance the mechanical integrity of the mass of carbon nanotubes produced. This can facilitate the production of carbon materials, such as fibres and films, from the carbon nanotubes.

(62) The above examples probe the effect of different sulphur precursors, with varied thermal degradation behaviour. This alters when the sulphur becomes available to the growing iron nanoparticles. Sulphur is believed to act as an arresting agent, stopping or slowing the nanoparticle growth. This is believed to affect the structure of the carbon nanotubes formed.

(63) The thermal stability of CDS is lower than thiophene, especially in a reductive hydrogen atmosphere. The adjacent double bonds in CDS are expected to readily undergo hydrogenation followed by elimination of sulphur in the form of H.sub.2S. This compound readily sulphudises the iron nanoparticles. Thiophene on the other hand is resistive to hydrogenolysis owing to its stability as an aromatic compound. Where CDS is used, the temperature at which sulphur becomes available to the catalyst particles is lower than the temperature at which sulphur becomes available to the catalyst particles when thiophene is used.

(64) As shown in FIG. 10A, the thermal degradation temperatures of ferrocene and CDS are close to each other. Therefore, it is believed that the catalyst particles are sulphudised in the early stages of their growth. This is believed to result in catalyst particles wherein at least 70% by number of the catalyst particles have a diameter in the range from 0.5 nm to 4.5 nm. It is believed that the small catalyst particles tend to provide single-walled carbon nanotubes.

(65) In contrast, there is a much larger difference between the thermal degradation temperatures of thiophene and ferrocene, as shown in FIG. 10B. Therefore, the nanoparticles grow for much longer before they encounter sulphur obtained by degradation of thiophene. Before this, the iron particles grow to 8-10 nm in diameter. These larger nanoparticles tend to produce larger diameter carbon nanotubes, which tend to be double-walled and collapsed.

(66) The experiments were repeated with ethanol as the carbon source (ethanol decomposition yields a carbon supply at much lower temperatures (about 873K) than methane). In the case of carbon nanotube fibres obtained from ethanol, CDS lead to the formation of metallic single-walled carbon nanotubes, while thiophene yielded collapsed double-walled carbon nanotubes. In addition, the effect of the presence of helium (recently reported to play a role in the formation of metallic nanotubes; Reference 3) on the formation of metallic nanotubes was tested, by carrying out the ethanol runs in the absence and presence of helium. Both yielded identical results and the presence of helium did not seem to significantly affect the process.

(67) The preferred embodiments have been described by way of example only. Modifications to these embodiments, further embodiments and modifications thereof will be apparent to the skilled person and as such are within the scope of the present invention.

REFERENCES

(68) The content of each of the following references is incorporated herein in its entirety. 1. Koziol, K. et al; High-Performance Carbon Nanotube Fiber; Science 318, 1892 (2007) 2. Motta, M. S. et al; The Role of Sulphur in the Synthesis of Carbon Nanotubes by Chemical Vapour Deposition at High Temperatures; J. Nanosci. Nanotech. 8 1-8 (2008) 3. Harutyunyan et al; Preferential Growth of Single-Walled Carbon Nanotubes with Metallic Conductivity; Science 326, 116 (2009) 4. Carbon Nanotubes; Ed: Jorio, Dresselhaus and Dresselhaus Springer Verlag Heidelberg 2008 5. Kataura et al; Optical Properties of Single-Wall Carbon Nanotubes; Syn. Met. 103 2555-2558 (1999)