Method for the Production of Carbon Nanotube Structures

Abstract

The present invention relates to a method for the production of carbon nanotube structures.

Claims

1. A method for the production of carbon nanotube structures comprising: (a) introducing a flow of metal catalyst or metal catalyst precursor into a temperature-controlled flow-through reactor; (b) exposing the flow of metal catalyst or metal catalyst precursor to a first temperature zone sufficient to generate particulate metal catalyst, wherein the first temperature zone includes a region of peak particle concentration; (c) releasing an axial or radial flow of a source of carbon into the temperature-controlled flow-through reactor at a release point; (d) exposing the particulate metal catalyst and source of carbon to a second temperature zone downstream from the first temperature zone, wherein the second temperature zone is sufficient to produce carbon nanotube structures; (e) exposing the particulate metal catalyst and source of carbon to a third temperature zone downstream from the second temperature zone, wherein the third temperature zone is sufficient to evaporate the particulate metal catalyst; (f) exposing the particulate metal catalyst and source of carbon to a fourth temperature zone downstream from the third temperature zone, wherein the fourth temperature zone is sufficient to re-nucleate the particulate metal catalyst and to produce carbon nanotube structures; and (g) discharging the carbon nanotube structures from a discharge outlet of the temperature-controlled flow-through reactor, wherein either the release point is substantially between the beginning of the first temperature zone and the end of the second temperature zone or the method further comprises: (f′) releasing an axial or radial flow of a source of sulphur into the temperature-controlled flow-through reactor at a release location, wherein the release location is at or near to the fourth temperature zone.

2. The-method as claimed in claim 1 wherein the release point is substantially between the beginning of the first temperature zone and the end of the second temperature zone and the method further comprises: (f′) releasing an axial or radial flow of a source of sulphur into the temperature-controlled flow-through reactor at a release location, wherein the release location is at or near to the fourth temperature zone.

3. The method as claimed in claim 1 wherein the release point is substantially coincident with the region of peak particle concentration.

4. The method as claimed in claim 1 wherein the release location is upstream from and near to the fourth temperature zone.

5. The method as claimed in claim 1 wherein the temperature profile in the temperature-controlled flow-through reactor is substantially parabolic.

6. The method as claimed in claim 1 wherein in step (c), the source of carbon is released in an axial counterflow.

7. The method as claimed in claim 1 wherein in step (c) the source of carbon is released radially.

8. The method as claimed in claim 1 wherein the source of carbon is methane optionally in the presence of an optionally substituted and/or optionally hydroxylated aromatic or aliphatic, acyclic or cyclic hydrocarbon which is optionally interrupted by one or more heteroatoms.

9. The method as claimed in claim 1 wherein the metal catalyst is iron.

10. The method as claimed in claim 1 wherein the metal catalyst precursor is a sulphur-containing organometallic.

11. The method as claimed in claim 10 wherein the metal catalyst precursor is ferrocene and a sulphur-containing ferrocenyl derivative.

12. The method as claimed in claim 1 wherein the metal catalyst or metal catalyst precursor is introduced in step (a) together with a sulphur-containing additive.

13. The method as claimed in claim 12 wherein the metal catalyst precursor is ferrocene optionally together with a sulphur-containing additive being thiophene or carbon disulphide.

14. The method as claimed in claim 1 wherein the source of sulphur is thiophene or carbon disulphide.

15. The method as claimed in claim 1 further comprising: measuring the particle size distribution of the particulate metal catalyst in the first temperature zone.

16. The method as claimed in claim 1 further comprising: measuring the particle size distribution of the particulate metal catalyst in the fourth temperature zone.

Description

[0069] The present invention will now be described in a non-limitative sense with reference to Examples and the accompanying Figures in which:

[0070] FIG. 1 illustrates the temperature gradients between the wall and centreline of a tubular reactor;

[0071] FIGS. 2a-c illustrate FT-IR analysis of the exhaust gases at different temperature set points to show the decomposition behaviour of thiophene and ferrocene alone and together;

[0072] FIG. 3 illustrates TSI-Scanning Mobility Particle Sizer 3080 (SMPS) measurements of particle size distribution in the tubular reactor;

[0073] FIG. 4 illustrates the total mass of catalyst nanoparticles along the tubular reactor;

[0074] FIG. 5 illustrates scanning electron microscopy (SEM) images showing different morphologies of the CNT fibre along the axis of the tubular reactor;

[0075] FIG. 6 illustrates the temperature profile at a set point of 1200° C., a particle size distribution measured at that set point along the tubular reactor, modelling results of particle growth and ferrocene decomposition, a reaction snapshot and a resistance measurement;

[0076] FIG. 7 illustrates schematically the experimental set-up used in Example 1;

[0077] FIG. 8 illustrates particle size distribution, decomposition and deposition along the tubular reactor at a set point of 1200° C.;

[0078] FIG. 9 illustrates different reactants in the tubular reactor at 1200° C., namely a) ferrocene, thiophene and methane (5 minute run); b) thiophene; and c) methane;

[0079] FIG. 10 illustrates a tubular reactor operated (A) conventionally and (B-C) in accordance with embodiments of the method of the invention;

[0080] FIGS. 11a-c illustrate results from the embodiment of the method of the invention compared with the conventional method;

[0081] FIG. 12 illustrates schematically an embodiment of the method of the invention; and

[0082] FIGS. 13a-b illustrate various multi-diameter discharge outlets of the tubular reactor.

EXAMPLE 1

Experimental

[0083] Experiments were carried out in a horizontal tubular reactor in ambient conditions and at variable temperatures. A schematic of the experimental setup is shown in FIG. 7. A sampling system was used to extract catalyst particles from axial locations and information was taken from velocity and thermal profiles and from axial analysis of carbon species synthesized from the particles.

[0084] A bulk flow of 0.5 slpm hydrogen (purity grade hydrogen N5.0 (BOC)) carrying ferrocene (.sup.˜2 wt %) and thiophene (.sup.˜0.3 wt %) entered the tubular reactor (40 mm ID and 700 mm length) through a showerhead injector. This ensured a uniform and laminar inflow with typical Re.sup.˜25<<2300 at the inlet. Flow rates were controlled by mass flow controllers (Alicat). The injector face was placed 70 mm from the inlet of the tubular reactor which was below 400° C. to avoid decomposition of ferrocene in the injector.

[0085] Particle measurements were carried out in situ by means of a TSI-Scanning Mobility Particle Sizer 3080 (SMPS) system including a TSI-Ultrafine Condensation Particle Counter 3776 (UCPC) and TSI-Differential Mobility Analyzers 3081 and 3085 (DMA) in an alumina (basis 99% Al.sub.2O.sub.3) tubular reactor. Samples along the centreline were taken through a 1.9 mm ID alumina probe at flow rates inside the probe of 0.300-0.450 slpm and were assisted by an ejector system including a 16/1000 inch orifice. Typically this was implemented at the end of the probe by a 1:50 dilution with pure and filtered ambient temperature nitrogen. The data presented was corrected for dilution, diffusion and thermophoretic losses in the probe. Exhaust gases were vented to atmosphere through a low pressure drop HEPA filter.

[0086] FT-IR was used to analyze gases along the axis of the tubular reactor and in the exhaust. The sampler for the FTIR was located in the hottest part of the tubular reactor and the set temperature was sequentially raised and the effect on the IR spectrum was recorded. This allowed the detection of the onset of (for example) ferrocene or thiophene decomposition and of the temperature at which complete decomposition occurred.

[0087] In a further configuration, methane was introduced at 0.06 slpm to serve as a source of carbon for carbon nanotube growth. Optical access was gained by using a quartz tubular reactor at 1200° C. and performing a rapid shut down of all reactant gases while allowing the experiment to cool down in an argon (99.998% (BOC)) atmosphere. The morphology of the carbon nanotube web that formed was then investigated by means of scanning electron microscopy (SEM).

Results and Discussion

Flow Dynamics

[0088] The gas velocity inside the tubular reactor and along its axis varies according to the ideal gas law. This means that the centreline velocity increases with increasing temperature. Conservation of mass, momentum and energy are taken into account in a 2D axisymmetric reactor model. Since the concentration of ferrocene and thiophene in the hydrogen bulk flow is very low, it can be ignored and its dynamics are decoupled from fluid dynamics. In a simulation of the flow dynamics within the tubular reactor, a uniform inflow, a temperature profile as shown in FIG. 1 at the outside wall of the tubular reactor and atmospheric pressure are assumed. Because of a no slip condition at the wall, the gradient of velocity increases along the tubular reactor until the temperature decreases beyond the mid-point which results in greater shear within the flow. The contribution of CNTs to viscous forces are omitted but may play an important role in the velocity profile. No radiative heat transfer between CNTs and the wall is taken into account.

[0089] A parabola shaped temperature profile is present in the tubular reactor (see FIG. 6). The actual temperature in the gas flow almost follows this profile. The radial temperature gradient illustrated in FIG. 1 at different locations within the tubular reactor represents the thermal diffusivity of energy from the walls into the tubular reactor. The gradient in temperature at the centreline is dT/dr|.sub.r=0>0 until x≈480 mm and becomes negative for x>480 mm. Unlike the velocity gradient, the gradient in temperature is at a minimum near the hottest region of the tubular reactor and at a maximum near the entry and exit of the tubular reactor. Given the temperature gradients between wall and centreline, thermophoretic forces drive particles away from the wall when dT/dr>0 for x<480 mm and towards the wall when dT/dr<0 for x>480 mm.

Decomposition of Ferrocene and Thiophene

[0090] FT-IR analyses of the exhaust gases at different temperature set points showed the decomposition behavior of thiophene and ferrocene alone and together (see FIGS. 2a-c).

[0091] For thiophene entering the tubular reactor alone, a CH.sub.4 mode and a C-triple bond mode appeared in the IR spectra as the temperature increased. This indicated thiophene decomposition. In addition, the appearance of a CH bending mode indicated that thiophene was disappearing as the temperature increased. Decomposition started at a set point of .sup.˜750° C. and was complete at a set point of .sup.˜1050° C.

[0092] For ferrocene entering the tubular reactor alone, the appearance of a C-triple bond mode indicated ferrocene decomposition at a set point of .sup.˜550° C. which reached its maximum at .sup.˜750° C.

[0093] For thiophene and ferrocene entering the tubular reactor together, the disappearance of the CH bending mode was independent of the presence of ferrocene. At .sup.˜750° C., a C-triple bond mode was already present indicating that ferrocene had already been decomposed. The intensity of the C-triple bond and CH.sub.4 mode increased as temperature increased which indicated that thiophene was decomposing.

[0094] Deposits of iron and soot on the tubular reactor wall in the respective temperature zones were observed (see FIG. 9). The deposition and therefore decomposition of reactants are mutually independent and can be detected for each species individually. Alongside the thermal experiments, it is clear that ferrocene and thiophene are a source of carbon for CNT nucleation (see FIGS. 5 and 6).

Particle Size Distributions

[0095] Axial measurement of particle size distributions along the centreline of the tubular reactor in the absence of methane showed almost instantaneous nucleation of catalyst nanoparticles (see FIG. 3). The first location of detection was in good agreement with the location of ferrocene and thiophene decomposition. The bulk nucleation and detection of nanoparticles occurred only after thiophene decomposed. Downstream measurements showed an almost immediate disappearance of particles at .sup.˜1200° C.

[0096] SMPS measurements of particle size distribution showed an onset of formation at 200 mm which corresponds to a tubular reactor wall temperature of .sup.˜960° C. Nucleated particles appeared with a particle size of about 4 nm. It is likely that the sampling method did not retroactively correct for all losses of small particles. Moving from 200 mm to 300 mm, the particle concentration and diameter grew. The peak of particle concentration was 7×10.sup.8 #/cm.sup.−3 at .sup.˜1100° C. for the 1150° C. tubular reactor and 10×10.sup.8 #/cm.sup.−3 at .sup.˜1100° C. for the 1300° C. tubular reactor. After the peak particle concentration, the measured particle diameters decreased as the temperature increased towards the hottest zone indicating potential particle-to-gas conversion ie evaporation. This trend was strongly temperature dependent and led to almost complete particle evaporation at a set point of 1300° C. For lower set points, the concentration of hydrogen-carried catalyst nanoparticles remained high. The maximum particle concentration occurred at a set point of 1300° C. and the lowest was measured at 1150° C. Regardless of the set points, the highest temperatures corresponded to the lowest particle concentrations.

[0097] As the temperature fell downstream from the maximum, re-nucleation occurred which resulted in an increase in particle concentration but to a concentration less than the concentration which was observed upstream. At a set point of 1250° C., the total mass was at a maximum which indicates that at higher set point temperatures, evaporation is competing with decomposition, nucleation and coagulation (see FIG. 4).

CNT Formation and Particle Size Distributions

[0098] Samples of the CNT fibre formed in experiments with methane as a source of carbon were taken along the axis of the tubular reactor. SEM images revealed different morphologies of the CNT fibre along the axis (see FIG. 5). The fibre extended through the tubular reactor and attached to the wall where ferrocene decomposed and thiophene started to decompose at .sup.˜170 mm which corresponds to a temperature of .sup.˜700° C. At the beginning of the tubular reactor were found CNT bundles (diameter 15-35 nm) and individual CNTs mainly oriented along the gas flow. Some catalyst nanoparticles were attached to these small CNT bundles close to the entrance. The first CNTs that started to grow in the zone of ferrocene decomposition showed only some attached nanoparticles. The carbon for CNT formation in this region is likely to be coming from the thermal breakdown of ferrocene and catalytic decomposition of methane. As the temperature increased, there was more ferrocene and thiophene break down and the initially grown CNTs act as a surface for heterogeneous nucleation of nanoparticles. CNT growth and nanoparticle nucleation occur in parallel. Bundling and branching of CNTs due to Van der Waals forces were observed. Many nanoparticles were found in between CNTs and bundle-junctions. Nanoparticles attached to CNTs and CNT bundles might act as a catalyst and growing point for further CNT growth.

[0099] The fibre further downstream the axis of the tubular reactor showed almost no attached catalyst nanoparticles. The nanoparticles disappeared as the temperature increased indicating particle evaporation. Some nanoparticles attached to initially grown CNTs potentially catalyze further CNT growth. As the temperature dropped downstream from the maximum temperature, re-nucleation occurred resulting in an increase in impurities in the web whereas the fibre in and just before the hottest zone appears to be nearly free of impurities and catalyst nanoparticles. The bulk CNT web formed near the exit had impurities and catalyst nanoparticle clusters were present. Since carbon is mainly available upstream from ferrocene and thiophene decomposition, carbon from methane only becomes available as the temperature rises and when catalyst nanoparticles start to evaporate. Once re-nucleation of nanoparticles occurs as the temperature drops, there is observed a bulk formation of CNTs from decomposed methane present alongside re-nucleated catalyst nanoparticles.

Overview of Results

[0100] Experiments with just a single reactant (ferrocene, thiophene or methane) reveal that a first wall deposit is related to ferrocene decomposition and a second one to thiophene decomposition. The simulated ferrocene decomposition along the centreline of the tubular reactor according to

[00001]$\frac{\mathrm{DC}}{\mathrm{Dt}}=D_c.Math.{\nabla }^2.Math.C-k_c.Math.{\mathrm{Ce}}^{-E_a/\left(R_u.Math.T\right)}$

is in good agreement with experimental observations and matches the location at which iron can be detected as a deposit at the tubular reactor wall.

[0101] A simulation including surface growth and coagulation of iron based nanoparticles along the centreline shows the evaporation of nanoparticles. A first particle size distribution measured at a temperature set point of 1200° C. and .sup.˜1100° C. (240 mm) is where the total mass of airborne nanoparticles is at a maximum. A second particle size distribution measured at .sup.˜700° C. (540 mm) is where re-nucleation and surface growth occurs as the temperature drops. The first and second particle size distribution are used as an input for the simulation. At these locations a saturation ratio of S=actual monomer concentration/saturation concentration of monomers=1 is assumed.

[0102] A snapshot of the process including methane as a carbon precursor shows the continuous CNT fibre reaching through almost the entire tubular reactor. Further downstream where the temperature drops after the hottest zone, the fibre is attached to the bulk CNT web. In the middle of the tubular reactor, the fibre is partly attached to the wall. Resistance measurements between the tubular reactor wall and the exit were carried out after the fibre which reached through the tubular reactor was taken out (see FIG. 6). Starting near the hottest zone, the wall becomes conductive indicating a CNT coating at the wall. Microscopy analysis of material gained from the wall confirmed these results. In this region, methane thermally decomposes and most of the carbon becomes available. At the same time, particle measurements showed that the catalyst nanoparticles evaporate and are no longer available to catalyze CNT growth.

Conclusions

[0103] Particle size distributions along the axis of the tubular reactor show a distinctive, temperature dependent appearance and disappearance of catalyst nanoparticles. Four specific regions of CNT web production showing different structural features that directly correlate to the nanoparticle behavior have been confirmed using SEM imaging. At the beginning of the reaction zone, homogeneous nanoparticle nucleation in a carbon-lean environment stimulates the growth of CNT material with a low impurity profile. This is followed by a region characterised by a higher impurity concentration where there is a contribution from heterogeneous nucleation of catalyst particles on the existing CNT structures and agglomeration of nanoparticles. The lowest impurity concentration is seen in the hottest zone of the tubular reactor where non-encapsulated nanoparticles evaporate. Towards the exit, the decrease in temperature profile stimulates the re-nucleation of iron-based nanoparticles from a saturated vapour in a carbon rich environment which leads to a rapid increase in the growth of carbon nanostructures which is dominated by undesirable impurities. To eliminate the possibility that re-nucleation is an aberration caused by the closed system, an experiment was carried out in an open-ended system and showed similar results.

[0104] As shown by IR data and the location of deposits on the quartz tubular reactor walls, ferrocene and thiophene decompose independently within narrow temperature dependent zones indicating that the onset of decomposition for each is thermally rather than catalytically driven.

EXAMPLE 2

[0105] (A) FIG. 10A illustrates a tubular reactor 1 operated conventionally to introduce hydrogen, ferrocene and thiophene via injector 12 at the downstream end. Methane and hydrogen were introduced at the upstream end. CNT growth was determined by locating the position at which a sample probe was blocked. The onset of CNT nucleation (the fibre nucleation point) occurred downstream from a region of peak particle concentration (5×10.sup.8) and from an incubation region over which the particle count fell.

[0106] (B) FIG. 10B illustrates a tubular reactor 1 operated according to an embodiment of the method of the invention. In this embodiment, a counterflow of hot methane and hydrogen was released into the region of peak particle concentration using a lance 11 inserted into the tubular reactor 1. A flow rate of 19 to 45 cm.sup.3/min was sufficient for the growth of CNTs to occur in the region of peak particle concentration. A flow rate of 45 to 55 cm.sup.3/min moved the growth of CNTs into a low temperature zone and closer to the injector 12.

[0107] (C) FIG. 10C illustrates a tubular reactor operated according to an embodiment of the method of the invention. In this embodiment, a counterflow of hot methane and hydrogen was released at the fibre nucleation point.

[0108] The experimental setup was as follows: [0109] Quartz tube 50 mm OD, 46 mm ID, 1200° C. set point, Carbolite STF 15/180 [0110] Showerhead injector 85 mm into tube [0111] Ferrocene 75 C, 40 sccm in H.sub.2 [0112] Thiophene 0.1 C, 10 sccm in H.sub.2 [0113] CH.sub.4 (40 sccm) supply line entering through centre port on exit flange, through ¼″ OD alumina line whose tip is positioned 20 cm from tube entrance [0114] CH.sub.4 diluted 1:4 with H.sub.2 (160 sccm H.sub.2+40 sccm CH.sub.4) with the dilution happening at the entrance to the ceramic tube to limit CH.sub.4 breakdown in the counterflow alumina line [0115] Showerhead: 340 sccm H.sub.2+40 sccm H.sub.2/Fe+10 sccm H.sub.2/thiophene [0116] Total flow through tubular reactor: 590 sccm.

[0117] The experimental protocol was as follows: [0118] 11:30—Flush with H.sub.2 (340 sccm entrance, 160 sccm counterflow) [0119] 11:33—Start Fe (40 sccm) and thiophene (10 sccm) through showerhead (pause for .sup.˜1 min) [0120] 11:40—Start CH.sub.4 40 sccm counterflow [0121] 11:47:34—CNT web formed in a third zone and reached the exit [0122] 11:50:00—Switched off all reagents, 10 sccm Ar through injector and counterflow line, heat off. Ferrocene heat off. Once tubular reactor had reached .sup.˜80° C., the entrance flange was removed and the tubular reactor was extracted with the counterflow line still in place.

[0123] FIG. 11a illustrates the results from this embodiment of the method of the invention compared with results from the conventional method (see FIG. 11b-c). In FIG. 11a, the counterflow of methane and hydrogen encouraged the growth of CNT material in a short time span (5 minutes) over which the conventional method produces no fibre. The downstream depositions in the method of the invention (FIG. 11a) are less dense than those of the conventional method (FIGS. 11b-c).

EXAMPLE 3

[0124] According to an embodiment of the method of the invention illustrated schematically in FIG. 12, a 10 sccm counterflow of thiophene was released into the region of the tubular reactor (R) in which re-nucleation occurred. The effect on the particle distribution and concentration was studied. Methane was subsequently introduced to see how the counterflow of thiophene affects fibre generation compared with the conventional arrangement.

[0125] The experimental setup was as follows: [0126] Alumina tube reactor 50 mm OD, 40 mm ID [0127] Set point 1200° C. [0128] Insulation: same as that used with previous quartz runs to keep temperature profile as similar as possible [0129] Showerhead injector [0130] Particle sampling probe: 3 mm OD alumina (1.9 mm ID) through the centre port of the exit flange, 16/1000 critical orifice [0131] Counterflow thiophene: alumina tube inserted through exit port so that the tip was 25 cm from the exit of the tubular reactor just upstream from region R.

[0132] Thiophene appears to condition re-nucleating catalyst particles for CNT growth. After five minutes flow of methane, a visual inspection revealed that a thick web-cap had formed and fibre appeared down the length of the tubular reactor. The thread looked thicker than that formed with no counterflow of thiophene and particulates were visible along the thread.

EXAMPLE 4

[0133] An experiment was carried out to determine whether narrowing the diameter of the tubular reactor towards the discharge outlet led to a change in flow velocity to assist the retrieval of the CNT material as a sock.

[0134] In the necked arrangement shown in FIG. 13a, an experiment in which different lengths extended beyond the tubular reactor showed that a CNT sock could be generated in a 28 mm diameter tube with lower reagent feed rates to permit the sock to be withdrawn upstream of the third zone.

[0135] In the trumpet arrangement shown in FIG. 13b, an experiment in which different lengths extended beyond the tubular reactor showed that a CNT sock could be generated in a 50 mm diameter tube to permit the sock to be withdrawn upstream of the third zone.