5.1 The influence of growth conditions on the growth of single-wall carbon nanotubes
Carbon nanotubes have attracted great attention in the past decade due to their interesting properties and potential applications.
The chemical, physical, and electronic properties of CNTs are dependent on their geometry and structure, which are determined by the preparation procedures. The influence of catalyst, the optimal amount of chemical components in the catalyst and the properties of produced SWNTs on silicon wafers were reported in our previous sections. It was found that the catalyst has a strong influence not only on the yield but also on the growth and on the properties of CNTs. In this part, we perform a detailed study of the various process factors (gases flows rates, growth temperature and growth time) influencing the growth of single walled carbon nanotubes based on optimal catalyst. The growth of the SWNTs relies on the
decomposition of methane (CH4) over patterned catalyst material on a silicon substrate covered with a silicon dioxide layer.
5.1.1 Influence of the deposition temperature
The synthesis temperature, which is the source of energy in the thermal CVD method, is of great importance in order to achieve the growth of individual single walled carbon nanotubes.
The influence of the synthesis temperature on the catalytic ability of the catalyst (the lower limit of the growth temperature based on the optimal catalyst), the yield and the type of carbonaceous material are studied by varying the temperature between 700 to 900°C. All of the other growth factors (optimal catalyst, gases rate and deposition time) were fixed
a b
c d
Figure 5.1 Scanning electron microscopy images of samples after CVD.
(a). 700°C. (d). 730°C. (g). 800°C.
(b). 710°C. (e). 740°C. (h). 850°C (c). 720°C. (f). 750°C. (i). 900°C.
e f
g h
i j
Fig 5.1 illustrates SEM images of representative samples. When the CVD growth is carried out at temperature < 710°C, no carbon nanotubes can be detected under SEM (as shown in fig 5.1 a and b). However from 720 to 750°C a low density of carbon nanotubes are grown (fig 5.1c to 5.1f). At above 800°C, especially at 850-900°C (fig 5.1g to 5.1i), a film of single walled carbon nanotubes is produced. The CNTs grown at 950°C are well obtained and uniform in size, but the density seems lower than the products at 850-900°C.
From fig. 5.1 one can observe that the reaction temperature has an evident influence on the deposited products. The density seems to decrease with decreasing of temperature. However, the SEM images demonstrate that the growth of carbon nanotubes with optimal catalyst occurs above 720°C.
The Raman spectra in fig 5.2 indicate that the films of CNTs, which are observed in fig 5.1, contain a large proportion of high quality single walled carbon nanotubes.
150 200 250 300
150 200 250 300
Frequency (cm-1)
T9LF700
Intensity (arb. units)
T9LF750 T9LF800 T9LF850 T9LF900 T9LF950
1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650
Frequency (cm-1) T9HF700
T9HF750
Intensity (arb. units)
T9HF800 T9HF850 T9HF900 T9HF950
Figure 5.2 Experimental Raman spectra of as-grown SWNT samples at various growth temperatures
In fig 5.2 all Raman spectra show RBM peaks in the range of 120 – 300 cm-1, except Raman spectra of sample at 700°C where no carbon nanotubes species are detected. This is an evidence for the growth of SWNTs at temperature in the range 750°C to 950°C and the spectra indicate that the SWNT diameters are approximately about 0.8 to 1.8nm (as show in table 5.1).
Table 5.1 Nanotube diameters calculated from RBM peaks of as-grown nanotubes SWNTs at various temperatures
Temperature (°C)
Diameter (calculated with
RBM peaks)
700 ~
750 0,76 – 1,65
800 0,76 – 1,58
850 0,76 – 1,58
900 0,77 – 1,7
950 0,8-1,4 nm
0 5 10 15 20
750 800 850 900 950
Temperature (°C)
IG/ID ratio
Figure 5.3 Peak intensities ratios from Raman spectra of samples grown at 750-950°C The G and D bands are observed at ~1590 and ~1350 cm−1, respectively. As already explained in previous sections, we have used the G/D ratio to follow the amount of defects introduced to the nanotubes versus temperature variation. The G/D ratios are found to be 8.7, 13.6, 16.5 and 17.4 for the nanotubes obtained from CVD process at 750°C, 800°C, 850°C, 900°C and 950°C, respectively (fig. 5.3). The increase in G/D ratio when we increase the
growth temperature from 750°C to 900°C indicates that the ratio of the carbon impurities or nanotubes with carbon nanotubes has been decreased.
0 10 20 30 40 50 60
700 750 800 850 900 950
Reaction temperature (°C)
Yield of carbon (%)
Figure 5.4 Yield of as-grown carbon product at different growth temperature (mg) Fig 5.4 shows the influence of reaction temperature on the carbon yield under constant reaction time for 10 min and CH4/H2 flow rate ratio of 1000/250 sccm. These results were calculated from eq. 6 (chapter 3). The carbon product obviously increases with the increase of the growth temperature, and reaches a maximum at 900°C, and then declines over 900°C.
The bar chart demonstrates that the synthesis process at 700°C provides ~7 %wt. yield.
However, combining with SEM and Raman spectra, we deduce that all carbon products are amorphous carbon.
As we know, methane decomposition is a moderately endothermic cracking reaction.
CH4 ↔ C + 2H2 ∆H = 74.5 kJmol-1
The equilibrium constant (Ke) of this reaction can be calculated by eq:
RT G
CH H C
e e
P P K a
°
−∆
=
=
4 2
2
(Eq. 5.1)
where:
Ke : the equilibrium constant for the reaction
H2
P : the partial pressure of hydrogen in reactor
CH4
P : the partial pressure of methane in reactor aC : carbon potential (aC=Ke/K)
K : calculated from the measured H2 and CH4 concentrations R : the universal gas constant 1.98585 J/(cal K)
T : the temperature (K).
∆G° : the free enthalpy of H2-CH4 system
An increase in temperature results in an exponential increase of the equilibrium constant. A high temperature promotes the decomposition of CH4 and consequently increases the concentration of carbon atoms, so the yield rises with increasing temperature.
When the temperature is too high, the formation rate of carbon over the catalyst surface might exceed the growth rate of CNTs. This can result in a dissolution rate of carbon radicals that is higher than the rates of diffusion and precipitation. In that case carbon atoms will accumulate on the surface of the catalyst to form a carbon shell. The catalyst particles then lose their catalytic activity and the growth of CNTs stops. Moreover, a chemical reaction may take place between the carbon and the active catalyst, leading to the formation of metal carbide, which decreases the catalytic ability of catalyst nanoparticles. Therefore, with increasing temperature (900–950°C), the amount of produced nanotubes becomes smaller.
5.1.2 Influence of the hydrogen flow rate
The second important factor we have found is the effect of hydrogen content. Figure 5.5 shows typical results with different hydrogen concentration in the feed gas. All prepared substrates have same density of catalyst nanoparticles before SWNT growth. The methane flow is kept constant at 1000 sccm while the H2 flow is varied. It is clearly shown that the concentration of hydrogen obviously affects SWNTs growth.
Figure 5.5 SEM images of as-grown CNTs with various H2 flow rates
(a). 0 sccm (c). 100 sccm (e). 200sccm (b). 50 sccm (d). 150 sccm (f). 250 sccm
a b
c d
e f
There is a low density of CNTs when 1000sccm methane without any hydrogen is used. It can be considered that CNTs grow thanks to catalyst even with low or no hydrogen gas content. High densities of CNTs are formed when hydrogen addition increases. We find that the same nanotubes densities are obtained for H2 concentrations from 50 to 250sccm.
To confirm the carbon nanotubes structures seen by SEM, Raman spectra have been recorded on these samples
150 200 250 300
150 200 250 300
Frequency (cm-1) T90H2LF
T950H2LF T9100H2LF Intensity (arb. units) T9150H2LF T9200H2LF T9250H2LF
1300 1400 1500 1600
1300 1400 1500 1600
D peak G_
Frequency (cm-1) T90H2HF
G_
G+ G+ T950H2HF
D peak
G_ G+
Intensity (arb. units)
T9100H2HF D peak
G_ G+ T9150H2HF
D peak GG__
T9200H2HF
G_
G+ T9250H2HF
G+ D peak
Figure 5.6 Raman spectra of as-grown SWNT samples with various hydrogen flow rates Fig 5.6 shows the Raman spectra from nanotubes grown at varying CH4: H2 ratio. The spectra are composed as before of two characteristic zones. The first zone consists in two peaks in the high frequency region 1300-1600cm-1. Five spectra are displayed, which were taken with the various hydrogen flows. We first carried out Lorentzian and/or Breit-Wigner- Fano fit of G-peak. The results are shown in fig 5.6 (green lines). The G-band exhibits two main features G+ and G-. The G+ peak always appears at around 1590 cm-1 and the G- peak at around 1560 cm-1. From the fitting function for G- peak and the peak width, we also confirm that our CNTs products contain semiconducting and metallic SWNTs.
The second zone, in the low frequency range of 120-300 cm-1, consists of the radial breathing modes (RBM) of SWNTs whose frequencies depend on the tube diameters.
So, the Raman spectra of our SWNTs have been characterized. The high IG/ID ratio (table 5.2) confirms that we obtain the high quality carbon nanotubes products.
Table 5.2 Comparison of IG/ID ratios and diameters of the CNTs samples with various hydrogen flow rates
Hydrogen flow rate (sccm)
G :D Diameters of CNTs (nm)
0 9,86156 0.8-1.72
50 12,02587 0.78-1.79
100 14,55124 0.84-1.72
150 14,75676 0.84-1.82
200 16,783085 0.72-1.84
250 15,18971 0.8-1.72
0 10 20 30 40 50 60
0 50 100 150 200 250
Hydrogen flow (sccm)
Yield of carbon (%)
Figure 5.7 Yield of as-grown carbon product with various hydrogen flow rates
Fig 5.7 displays the effect of hydrogen flow rate on carbon yield under constant reaction time, growth temperature and CH4 flow rate. The figure shows that the carbon products yield increases with increasing H2 flow rate, but fall off as the flow rate reaches 250 sccm.
To understand the effect of hydrogen, we have to focus on the chemical reactions that happen in our system. It is known that methane decomposition is a moderately endothermic cracking reaction. There are two kind of chemical reactions in methane chemical vapor deposition:
CH4 C* + 2H2 by high temperature (5.1)
CH4 C** + 2H2 by catalyst decomposition (5.2)
where: C* represent amorphous carbon and C** denote de carbon atoms which are adsorbed on the catalyst
When the carbon atoms in the catalyst are saturated, nanotubes will grow from the bulk.
Moreover, in the presence of H2 and the carbon containing species co-produced during CNT growth over catalyst surface, the reversible reaction occurs:
H2 + CHx CH4 (5.3)
This reaction removes carbon species on the catalyst surface, which favours self-cleaning of the catalyst surface by inhibiting the decomposition of the “encapsulating” carbon, and slow down deactivation of catalyst.
So, hydrogen plays an important role in our system. First, it acts as a carrying gas.
Hydrogen is carried into the reactor by the hydrogen flow, and its rate influences the partial pressure of hydrocarbon. Secondly, hydrogen affects the catalytic ability of catalyst nanoparticles. However, a high content of hydrogen reduces considerably methane conversion during the process because hydrogen is one of the reaction products. The optimal ratio found is 200 sccm of hydrogen diluted in 1000 sccm of methane.
5.1.3 Influence of the growth time
The yield of carbon nanotubes has a close relationship with the injection time of the hydrocarbon source. The growth of carbon nanotubes is a rapid process. Here even at the shortest reaction times nanotubes were observed by SEM (fig 5.8a).
Figure 5.8 Scanning electron microscopy images of as-grown CNTs with influence of time
(a). 1 minutes (c). 3 minutes (e). 20 minutes (b). 2 minutes (d). 10 minutes (f). 30minutes
a b
c d
e f
Fig 5.8 shows the SEM images of SWNTs film grown for (a) 1, (b) 2, (c) 3, (d) 10, (e) 20, and (f) 30min, respectively. For all growth times, the process factors were fixed. The SEM images in figures reveal that high densities network CNTs are formed when increasing the growth time from 2 to 30 min. No significant relation between the properties of CNTs products (the length, the diameter and the kind) and reaction time was found in the SEM images.
150 200 250 300
150 200 250 300
Wavelengtn (cm-1) T91MLF
T92MLF
Intensity (arb. units)
T93MLF T910MLF T920MLF T930MLF
1200 1300 1400 1500 1600
1200 1300 1400 1500 1600
Wavelength (cm-1) T91MHF
T92MHF T93MHF
Intensity (arb. units)
T910MHF T920MHF T930MHF
Figure 5.9 Raman spectra of as-grown SWNT samples under time influence Raman spectra of produced-SWNTs are shown in fig 5.9 The G-peaks around 1590 cm-1 are very strong. Also some RBM peaks in the low frequency range and the weak D-peak are observed.
Table.5.3 shows the ratio of G-peak and D-peak. In the initial stage of growth, especially when the growth time is less than 20 min, the purity of the tube product increases with the growth time. When the growth time is further increased, the IG/ID ratio decreases slightly.
This phenomenon may imply that the initial growth of SWNTs is kinetically controlled, as the catalytic sites are active and fresh for the formation of tubes. After growing for some time, the growth process might change to be thermodynamically dominated, which means that in this
period, a SWNT would modify its structure to a more perfect and stable state. In this situation, however, with the continuous injection of carbon and the probable poisoning of the catalytic particles, other forms of carbon can be deposited, reducing the final purity of the SWNT products. This becomes even worse at longer growth times. In this regard, there is a balance between the growth of SWNTs (density) and sacrifice of the purity of the product.
Considering both quality and quantity, we chose 20 min as the optimum growth time, which ensures the growth of SWNTs with a relatively small amount of amorphous carbon and gives a comparatively high productivity.
Table 5.3 Comparison of IG/ID ratios and diameters of the CNTs samples with various growth times
Growth times G/D RBM peaks Diameter (nm)
1 minutes 13,9188 128, 153, 172, 191, 206, 228, 268, 295 0,76-1,75 2 minutes 13,30086 132, 165, 173, 182, 186, 200, 223, 264, 288 0,78-1,69 3 minutes 16,93837 140, 159, 172, 192, 196, 209, 261, 279, 294 0,77-1,61 10 minutes
17,36966
131, 147, 168, 178, 184, 203, 228, 247, 268, 291
0,77-1,71 20 minutes 18,58866 135, 151, 166, 176, 196, 215, 244, 258, 284 0,79-1,66 30 minutes 16,071 140, 148, 163, 182, 202, 226, 244, 279 0,8-1,61
Fig 5.10 shows the relationship between the reaction time and the yield of carbon products under the constant reaction temperature of 900°C and CH4/H2 flow rate ratio of 1000/250 sccm, indicating that carbon yield increases with extended reaction time. Although the absolute yield of CNTs increases with increasing reaction time, the relative yield normalized by the total carbon amount from the vaporized hydrocarbon appears to reach an optimum value when the reaction time is about 20 min. When the reaction times is longer than 20 min, because the reaction zone is filled with the CNTs, the catalyst particles are instantly encapsulated by a large amount of carbon atoms decomposed from the hydrocarbon source under the high temperature, resulting in the fact this catalyst particles lose their activity which inhibits the growth of CNTs.
0 5 10 15 20 25 30 20
25 30 35 40 45 50 55 60
Yield of carbon product (%)
Growth time (min) 1m
2m 3m
10m
20m
30m
Figure 5.10 Yield of as-grown carbon product with various growth time
It can be observed that, during the initial stage of the growth, the yield of carbon deposits clearly increased with time up to 30 min. The carbon yield was around 50% with 10 min of growth time, while after 30 min, the yield only reached 56%, implying that the growth rate had slowed. This suggests that most of the growth had taken place in the first 30 min. The further increase in carbon yield during long growth times might be due to the deposition of amorphous carbon on the support, since in our experiments we found that CH4, could also readily deposit carbon on support without the presence of a catalyst.
In general, the longer the injection time the more hydrocarbon molecules pass over the catalyst particles and therefore the higher the carbon yield of the deposits. Along with the formation of nanotubes, increasing the reaction time leads to an enhanced appearance of amorphous soot covering the graphitic nanotructures up to high degrees. Then the aspect of the samples appears to be highly agglutinated. Therefore, increasing the reaction does not necessarily lead to a maximum of nanotubes production.
5.2 Synthesis of single wall carbon nanotubes in the CENTURA reactor
Our thermal CVD growth experiments demonstrate that we are able to achieve the growth of high-quality individual and bundles of single-walled carbon nanotubes networks on silicon substrates. The optimum catalyst and process conditions to growth SWNTs have been studied systematically presented above. In this section, we present the transfer of our process from the
small CVD tubular reactor (EST) to the industrial CENTURA tool for the deposition on 200 mm wafer.
EASYTUBE TUBULAR REACTOR 1000sccm CH4
250 sccm H2
Growth temperature: 900°C Growth time: 20 minutes Pressure: 1 atm (760 Torr)
Figure 5.11 Image of CVD instrument (ETS) and process conditions INDUSTRIAL CENTURA TOOL
(a) (b)
Figure 5.12 Images of CENTURA tool
Table 5.4 The experimental conditions applying for CENTURA tool 1 slm CH4 (standard litre per minutes)
0,25 slm H2
Growth temperature: 900°C Growth time: 20 minutes
Pressure: 600 Torr (maximum pressure)
Table 5.4 shows the experimental conditions for catalyst and synthesis of carbon nanotubes which are applied in the CENTURA reactor. The optimal composition of the catalyst is used in this step.
Synthesis of carbon nanotube in the CENTURA reactor using the optimal catalyst
The results of produced CNTs inside the CENTURA reactor are characterized by SEM and Raman spectra to determine the properties of the product.
Fig 5.13 shows the SEM images of produced CNTs.
Figure 5.13 SEM images of as-grown CNTs with CENTURA tool
a
b
From the SEM images, it is clear that CNTs network is obtained. A close up image (fig 5.12b) shows the network tubes distributed uniformly over the whole substrate. These nanotubes are typically a few micrometers long.
To determine the properties and diameters of produced CNTs grown in the CENTURA instrument at 900°C with the optimal catalyst, we use Raman spectroscopy analysis.
2 0 0 3 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 0
1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 5 0 0 0 0 6 0 0 0 0
1 5 0 2 0 0 2 5 0 3 0 0
5 0 0 1 0 0 0 1 5 0 0 2 0 0 0 2 5 0 0
Intensity (arb. units)
F re q u e n c y (c m -1 )
Figure 5.14 Raman spectra of as-grown SWNT samples with theCENTURA instrument Fig 5.14 shows the Raman spectra of as-grown carbon nanotubes with the CENTURA instrument. The presence of high G-peak intensity (~ 56000 a.u.), very low D-peak intensity (~ 3500 a.u.) and several RBM peaks are the evidence of high quality of the single walled nanotubes.
The RBM modes found at 136, 144, 158, 166, 183, 211 cm-1 show that our SWNTs have diameters in the range of 1.1 to 1.8nm (inset figure). We can compare the calculated-values of diameter with Kataura plots to determine the properties of produced-SWNTs. Table 5.5 shows the properties of SWNTs. Note that the distribution of diameters is more narrower than one obtained with the EASYTUBE system.
From the frequency of RBM peak we can calculate the diameters distribution of obtained SWNTs by eq. 4.1. The conducting properties determined from Kataura plot is also presented on table 5.5.
Table 5.5 Nanotube diameters calculated from RBM peaks of SWNTs
ωRBM d (nm) CNTs properties
136 1.65 Semiconductor
144 1.56 Semiconductor
158 1.42 Semiconductor
166 1.35 Semiconductor
183 1.22 Metallic,
semiconducting
211 1.06 Metallic
262 0.85 Semiconducting
So, we succeeded in the process transfer from the laboratory CVD equipment (EST) towards the industrial CVD instrument (200mm compatible CVD CENTURA reactor). Using the optimal catalyst and the controlled process condition, we can obtain large quantities of high quality single walled carbon nanotubes. This result contributes to the large scale CVD synthesis of SWNTs on substrates.
5.3 Patterned growth of SWNTs for electro devices
The unique properties of carbon nanotubes offer extreme potential for various applications.
One of the most promising applications of nanotubes is their use in nano-electronics devices such as a field effect transistor, nanotube interconnects and nanosensors. The application of single walled carbon nanotubes in electronic devices system requires the controlled placement of nanotubes. So, developing controlled-synthesis methods to obtain well-ordered carbon nanotubes is important and a viable route to nanotubes based devices. Dai et al. [76, 179]
showed self-directed growth of suspended nanotube networks on silicon tower tops having a liquid-phase catalyst precursor by chemical vapor deposition (CVD). Also recently Homma et al. [181] demonstrated the fabrication of suspended carbon nanotube networks on 100 nm scale silicon pillar structures by simply depositing a catalyst film on the silicon substrate.
These are indeed effective ways to control the growth of carbon nanotubes. However, for the actual application of such self-assembled single-walled carbon nanotube networks, additional efforts to build highly dense and organized nanotube networks connecting all designed locations even on a large scale are required.
In order to determine the growth sites of the SWNTs on the substrate, a resist pattern is defined lithographically, the liquid catalyst material is brought onto the surface, calcinated, and the excess catalyst is then removed in the lift-off step. In the last sections we present a systematic study of the influence of selected catalyst, the CVD process parameters on the quality of as-grown single walled nanotubes. Here, we propose to obtain high-yield growth of single-walled carbon nanotubes networks between patterned catalyst structures.
5.3.1 The fabrication of the nanotubes devices:
The fabrications of the nanotube devices used in the experiments and described in this thesis require state of the art nanofabrication facilities and techniques. Roughly speaking, the fabrication process can be divided in four parts: (i) fabrication of markers; (ii) catalyst deposition/CNT growth; (iii) nanotube location and electrode fabrication, and (iv) room temperature characterization and sample bonding. All these steps are detailed here after.
Fabrication of markers
In all experiments the nanotubes are grown/deposited on top of oxidized silicon substrates.
The thickness of the thermally grown oxide is typically ~300 nm, and isolates the devices from the back gate. A set of markers is necessary to later locate the position of the nanotubes and for the fabrication of the electrodes. These include a set of electron beam lithography alignment markers (e-beam markers) and atomic force microscopy (AFM) markers. The fabrication of markers is depicted in fig 5.14.
Figure 5.15 Fabrication of markers and scanning electron microscopy images of AFM markers
Substrate lithography
Substrate
Substrate Metal evaporation
Substrate After lift-off