Chapter 2: Experiment
2.2 Experimental Procedures
2.2.3 Analysis
The morphologies of CNTs’ samples are characterized by scanning electron microscopy (SEM), the morphologies of pre-treatment catalyst are observed by atomic force microscopy (AFM). The finer internal structures of interface of CNTs and nano-sized catalytic materials were examined by high-resolution transmission electron microscopy (HRTEM), JEOL JEM-2000EX and X-ray energy dispersive spectroscopy (EDS) respectively.
Field emission characteristics of CNTs are measured with a parallel diode-type configuration in a high-vacuum chamber with the pressure of 5×10-6 torr. A glass substrate coated with indium tin oxide (ITO) and P22 phosphor (ZnS: Cu, Al) was used as the anode plate, and the distance between the cathode and the anode plate was set to be 150 μm. The emitting area was variable, which was determined by pillar-spacing and pattern-area.
Anode voltages is sweep-type from 0 V to 1000 V which are applied at intervals of 10 V by a source measure unit (Keithley 237) shown in Fig. 2-6 for the verification of field emission characteristics while the cathode was biased at 0 V.
2.3 Experiment Procedures
2.3.1 Experiment A: Optimum R/H Ratio of Different Inter-Pillar Spacing
We investigated field emission properties from a pillar arrays of aligned carbon nanotube (CNT) bundles, which are fabricated on a Si substrate by thermal chemical
vapor deposition (T-CVD). To explore the influence of the pillar arrays’ arrangement on its field emission, the ratio of inter-pillar distance (R) to pillar height (H), R/H shown in Fig.2-7[2.12], was investigated by changing H while maintaining R at 80 μm, 150 μm and 250 μm, respectively.
A circle arrays with 50 μm in diameter with three different inter-distances (80μm, 150μm and 250μm, respectively) using Co-Ti (40 Å )/Ti(10 Å ) /Al (100 Å ) as multilayer catalyst shown in Fig.2.8 are made on a Si substrate (1cm×1cm) by photolithography and magnetron sputtering. To form catalyst nanoparticles, the catalyst was pre-treated at 700
°C for 5 min in hydrogen at 50sccm. Subsequently, thermal CVD was carried out at a growth temperature of 700 °C maintained by flowing 60sccm ethylene (C2H4) gas diluted with nitrogen (N2) gas as carrier gas. The total experimental process of profile was shown schematically in Fig. 2-9.
Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) are used to characterize the synthesized pillar array of aligned CNT bundles. The field emission property is measured using a parallel diode type configuration at a pressure of
~10−6 torr. The gap between the anode plate and the CNT emitter was set at 150μm or 300μm.
2.3.2 Experiment B: Two-Step Growing Method for CNTs’ synthesis
In this section, circle patterns with 50μm in diameter using Co-Ti(40 Å )/Ti(10 Å ) /Al (100 Å ) multilayer catalyst and inter-pillar distance is 80μm shown in Fig.2.8-(a). is made on a Si substrate (1cm×1cm) by photolithography and magnetron sputtering. To form catalyst nanoparticles, the catalyst pattern was pre-treated at 700 °C for 5 min in hydrogen at 50sccm. Subsequently, thermal CVD was carried out at a growth temperature of 700 °C maintained by flowing 60sccm ethylene (C2H4) gas diluted with nitrogen (N2) gas as carrier gas and maintain 20mins for first step growth. Than flowed 1000sccm nitrogen (N2) to perch space, and pretreated again for holding on second step CNTs growth. The total experimental process of profile was shown schematically shown in Fig. 2-10.
2.3.3 Experiment C: Plasma Post-treatment for CNTs’ Pillar arrays
Aligned CNT pillar array samples are synthesized by thermal CVD (T-CVD) on the patterned silicon wafer. The patterns for circular of 50μm diameter are fabricated silicon substrate.
In this section, sample with pillar height about 30μm and inter-pillar distance about 80μm is post-treated by High Density Plasma Reactive Ion Etching System (HDP-RIE).
The plasma treatment of as grown CNT pillar array was carried out in a plasma of a mixture of O2 and Cl2 gases generated by microwave power. The flow rates and treating time of O2 and Cl2 were shown in Table 2-1. The operating temperature, pressure, and process duration in the plasma treatment are room temperature , 10 mtorr, and 2、3、5、
10 min, respectively. The structural and chemical modifications of CNTs arrays before and after the plasma treatment are examined by a Hitachi model S-4700I scanning electron microscope (SEM) and Jobin Yvon model HR800 micro-Raman spectrometer, EDS, and TEM respectively. The field emission characteristics of the sample before and after plasma treatment were measured in a high vacuum chamber with a parallel diode-type configuration at a base pressure of ~10-6 torr.
2.3.4 Experiment D: Triode Structure for CNT Field Emission
Schematically illustrates the fabrication procedures of triode structure for CNT field emission. Heavily doped N-type silicon wafers are used as the substrate. A 5000 –Å -thick silicon dioxide is thermally grown on the silicon substrate at 1050oC. A 2000 –Å -thick Poly-Si layer is then deposited on the thermal oxide by pressure chemical vapor deposition (LPCVD) using a pure SiH4 gas at 620oC. The Poly-Si layer was further doped by POCl3 at 950oC for 20 min. An array of 100×100 circle patterns with the diameter of 6μm and inter-circle distance of 9μm are defined by photolithography and etching by Die-electric Material RIE 200L shown in Fig 2-12[2.13]. Then, the Poly-silicon layer is etched laterally by Die-electric Material RIE 200L to create a gap between the Poly-silicon gate and the CNTs shown in Fig 2-13. A catalytic layer of 50 Å cobalt-titanium in thickness is deposited on the 100 Å aluminum layer on the sample by sputtering. For deposition of the catalytic layer, the base and operating pressure of the sputter chamber are kept below 2×10-6 and 7.6×10-3 torr, respectively. The input RF power is 60W and the deposition rate was 0.1 Å /S for 100 Å aluminum layer. Then, the
input RF power for cobalt-titanium co-sputtering was kept 70W and 100W, respectively for 40 Å catalyst layer. Utilizing the original photoresist as a mask and lift-off technique, the circle pattern is transferred to the silicon substrate. Finally, CNTs are grown selectively on the iron layer by thermal chemical vapor deposition (T-CVD) system shown in Fig 2-12. The growth morphology of CNTs was observed by scanning electron microscopy (SEM).
Field emission properties of triode structure are measured in a high-vacuum environment with a base pressure of 1.0×10-6 torr. A glass plate coated with indium-tin-oxide (ITO) and phosphor is used as an anode and located 150μm above the sample surface. During the measurement, the device is in a common emitter configuration with the emitter grounded. The gate is applied with a voltage swept from 0 V to 80 V to obtain the current-voltage characteristics of the CNT triodes. The anode is applied with the constant voltage 800 V and the cathode maintain grounded, configuration shown in Fig 2-14.[2.14]
Fig.2.1 Simulation of the equipotential lines of the electrostatic field for tubes of different distances between tubes.
Fig.2.2 High Density Plasma Reactive Ion Etching System, HDP-RIE.[2.9]
Fig.2.3 schematic of a typical backlight unit.
Fig. 2-4 (a) Schematic picture and (b) photograph of thermal CVD.
Fig. 2-5 Process of CNTs synthesis (an example of CNTs growing 30min at 700℃.
Fig. 2-6 High vacuum measurement system.
Fig.2-7(a) SEM image of pillars of aligned CNT bundles grown by thermal CVD. (b) Cross-sectional SEM image of the pillar, showing that it is aligned perpendicular to the substrate surface and consists of high density CNTs[2.12]
50μm 80μm
1cm
1cm
(a)
50μm 150μm
1cm
1cm
(b)
50μm 250μm
1cm
1cm
(c)
Fig.2.8 Masks design: (a)-(c) show array of three different inter-distance for 80 μm、150 μm and 250 μm defined in 1 cm×1 cm area, respectively.
Photoresistance Silicon substrate
Mask(pillar-like pattern)
PR development
Spatter catalyst
Lift off
Exposure
Pretreatment
CNTs growth (e)
(f)
(g)
(h)
(j)
Fig.2.9 Fabrication flow diagrams (a) ~ (j). (f) Co-Ti / Al (2nm-3nm/10nm) catalyst by sputtering system, (h) pretreatment with H2 (50 sccm), and (j) CNTs
(1)
Fig.2-10 (1) Process of two steps growing for CNTs’ synthesis.(2) Fabrication flow diagrams (a) ~ (e). (d) pretreatment in H2 at 50 sccm again , and (e)
second step CNTs growth method under C2H4 atmosphere.
Silicon substrate
PR development
Spatter catalyst
Lift off by acetone
CNTs growth
CNTs treated after high density plasma (a)
(b)
(c)
(d)
(e)
(f)
Fig.2.11 Fabrication procedures for CNTs treated by plasma etching.
Table 2-1 Sample1 and sample2 represent the post-treatment process in oxygen and chlorine plasma. (R.T=room temperature)
Gas flow rate ICP power BIAS power O2 Cl2 Time Substrate temperature Pressure (W) (W) (sccm) (sccm) (min) (oC) (mtorr) Sample1 600 20 40 0 2 R.T 10 600 20 40 0 10 R.T 10 Sample2 600 20 30 10 2 R.T 10 600 20 30 10 3 R.T 10 600 20 30 10 5 R.T 10 600 20 30 10 10 R.T 10
Photoresist
Carbon Nanotubes N+ Poly-Si Gate
Thermal Oxide
Si(100)
Photoresist Mask
Catalytic Metal Layer (a)
(b)
(c)
(d)
(e)
(f)
(g)
Fig 2-12 Schematic representation of fabrication procedures of triod structure.
Fig 2-13The under etching condition of N-type Poly-silicon by Die-electric Material RIE 200L[2.13]
Fig 2-14 The aligned nanotubes in the bunch have variable heights and may protrude through the gate opening, as shown in the figure, or lie beneath the gate electrode plane.[2.14]
Va＞ Vg
Vg＞ 0
VC =0 Anode
Gate
Cathode
Chapter 3:
Results and Discussion
3.1 Analysis of Catalyst
3.1 Analysis of Catalyst
Prior to CNTs’ growing, the choose of catalyst is very important. According to our group’s researches, there are many advantages of CNT field emission characteristics by using novel co-deposition of catalyst. The most obvious characteristic is substantial increment of reliability.
In the Fig 3.1.7, better uniform roughness and nanoparticle diameter are obtained in cobalt-titanium co-deposition case.[3.1] We have speculated that this phenomenon is caused by decreasing the diameter of nano-sized particles, therefore, finer particle provide higher activity and lower melting temperature. The AFM top view is the easiest method for making a comparison and it will clearly show the particle size and roughness, as shown in Fig 3.1.7.
We compare catalysts of tri-components with catalysts of bi-components, with or without aluminum between electrode and other catalyst layers. Subsequently, we can easily find out the differences. The fluctuation in catalyst surface shows that simple and repeated nano-sized particle when pretreatment of bi-components catalyst without 100 Å Al buffer film, on the other hand, the fluctuation in catalyst surface is complex and additional curvature under those nano-sized particles when using the tri-components catalyst with 100 Å Al layer. As this result, we knew that Al plays a role of providing an additional curvature for surface of catalyst and quite increasing roughness mean square (RMS). The advantage of increasing RMS is raising the density of catalyst particle on the same top view area as shown in Fig 3.1.2, and then we can obtain better density of CNTs’
pillar after the growing step.
Furthermore, we focus the advantages of titanium (Ti) element in CNTs’ growth. To
verify the effect of the Ti layer, because Co and Ti surface energy are very nearly [3.2], so that cobalt film turns into nanoparticles with uniform diameter during pre-treatment process. This phenomenon is shown in Fig 3.1.3. Adhesion is evidently improved is also found in our experiment. As for CNT films on Ti, in the high temperature due to the formation of conductive TiC , a barrier on the CNT–Ti junction is removed (as exhibited in Fig 3.1.5). Electrons can pass through this junction without obstacle. So, during the whole field emission, electrons need to overcome CNT-vacuum barrier only, as shown in Fig 3.1.6. Therefore, a very small voltage will result a considerable electron emission. So, titanium (Ti) layer can improve uniformity and strengthen adherence between CNTs and substrate for CNTs’ growth [3.3]. The evident is shown in Fig 3.1.4.
3.2 Finding Optimum R/H Ratio of Different Pillar Spacing
3.2.1 Effect of carbon source flow rates on the morphology of CNTs’
growth
It is indispensable to reduce the turn on field and threshold field to achieve practically applicable field electron emitters that operate at lower power consumption. It has been reported that field emission can effectively enhanced for aligned CNTs as field emitters when the ratio of distance between neighboring nanotubes to the height of each individual CNT is about 2 [3.4]. However, the optimum rule (R/H=2) for CNTs as electron field emitters has not been realized.
In our researches, different R/H ratios have been designed to study the relation between R/H ratios and the field emission characteristics, and it is essential to grow longer aligned CNTs’ pillar with larger inter- pillar distance to meet the R/H ratios we need. However, growth of longer CNTs beyond 100μm is a little bit difficult. Therefore, it is very important to study the relation between the length of CNTs′ pillar and the growth condition including growth time, flow rate of carbon source, etc.
The length of the aligned CNTs pillar as a function of growth time is shown in Fig 3.2.4(a). It is observed that there is a CNTs′ length limit of approximately to 130μm. In
this case of the aligned CNTs′ pillars grown on the aluminum substrate, the aligned CNTs can′t be lengthened even if prolongs the growth time beyond 90 min. When we increase the flow rate of C2H4 to 135sccm, its growth rate is about 2μ m/min faster than that at 60sccm at 700oC in growth process. But the length of CNTs′ pillar does not match our prediction. Fig 3.2.4 (b) shows the longest CNTs′ pillar is about 50μ m , even if prolongs the growth time beyond 40min. It is probably due to too much amorphous carbon putting on catalytic nanoparticles. Too much carbon source to close crystalline graphite sheets of the wall and encapsulate nanoparticle at the closed tip is another probable reason.
According to this result, we try to reduce the flow rate of C2H4 from 135sccm to 60sccm.
Thus, moderate carbon source flow rate is indispensible to grow longer CNTs′ pillar.
Besides, the tendency that CNTs′ growth rate is linear with the time until the length limit is reached is also observed from Fig 3.2.4. Therefore, moderate flow rate of carbon source as well as growth time is required to precisely control once we want to grow the specified length of CNTs′ pillar. According to our experimental results, we can easily grow longer length of CNTs′ pillar to meet the required R/H ratios we need.
3.2.2 Effect of arrangement of pillar array on its field-emission characteristic
Three circle patterns with different inter-pillar distance (80μ m,150μ m and 250μ m ) are designed to get different R/H ratios. In the previous section, we have succeeded to find the suitable growth condition to grow longer length of CNTs′ pillar to meet the required R/H ratios we need.
The aligned CNTs pillar is characterized by high-resolution scanning electron microscopy (SEM). Aligned CNTs pillar that is perpendicular to the substrate surface is obtained due to J.van der Waals force between CNTs. Fig.3.2.1 ~ Fig.3.2.3 show SEM images of an aligned CNTs pillar grown on silicon substrate with aluminum buffer layer.
(a) 80μm inter-pillar distance case
Fig.3.2.1 (a) shows the aligned CNTs bundle with a diameter of 50μ m and a height of 4.5μ m. The R/H value of 18.18 is obtained. Fig.3.2.1 (b) shows the aligned CNTs bundle with a diameter of 50μ m and a height of 8μ m. The R/H value of 11.2 is
obtained. Fig.3.2.1 (c) shows the aligned CNTs bundle with a diameter of 50μ m and a height of 10μ m. The R/H value of 8 is obtained. Fig.3.2.1 (d) shows the aligned CNTs bundle with a diameter of 50μ m and a height of 10.8μ m. The R/H value of 7.8 is obtained. Fig.3.2.1 (e) shows the aligned CNTs bundle with a diameter of 50μ m and a height of 13.2μ m. The R/H value of 6.06 is obtained. Fig.3.2.2 (f) shows the aligned CNTs bundle with a diameter of 50μ m and a height of 22.3μ m. The R/H value of 3.6 is obtained. Fig.3.2.1 (g) shows the aligned CNTs bundle with a diameter of 50μ m and a height of 53μ m. The R/H value of 1.5 is obtained.
( b ) For 150μm inter-pillar distance case
Fig.3.2.2 (a) shows the aligned CNTs bundle with a diameter of 50μm and a height of 14μm. The R/H value of 10.7 is obtained. Fig.3.2.2 (b) shows the aligned CNTs bundle with a diameter of 50μm and a height of 20μm. The R/H value of 7.5 is obtained.
Fig.3.2.2 (c) shows the aligned CNTs bundle with a diameter of 50μm and a height of 26μm. The R/H value of 6 is obtained. Fig.3.2.2 (d) shows the aligned CNTs bundle with a diameter of 32μm and a height of 14μm. The R/H value of 4.7 is obtained. Fig.3.2.2 (e) shows the aligned CNTs bundle with a diameter of 50μm and a height of 47μm. The R/H value of 3.2 is obtained. Fig.3.2.2 (f) shows the aligned CNTs bundle with a diameter of 50μm and a height of 60μm. The R/H value of 2.5 is obtained. Fig.3.2.2 (g) shows the aligned CNTs bundle with a diameter of 50μm and a height of 82μm. The R/H value of 1.87 is obtained.
( c ) For 250μm inter-pillar distance case
Fig.3.2.3 (a) shows the aligned CNTs bundle with a diameter of 50μm and a height of 17.9μm. The R/H value of 14 is obtained. Fig.3.2.3 (b) shows the aligned CNTs bundle with a diameter of 50μm and a height of 23μm. The R/H value of 10 is obtained.
Fig.3.2.3 (c) shows the aligned CNTs bundle with a diameter of 50μm and a height of 49μm. The R/H value of 5.1 is obtained. Fig.3.2.3 (d) shows the aligned CNTs bundle with a diameter of 50μm and a height of 80μm. The R/H value of 3.1 is obtained.
Fig.3.2.3 (e) shows the aligned CNTs bundle with a diameter of 50μm and a height of 100μm. The R/H value of 2.5 is obtained.
From the SEM images in the above section, patterned CNT pillars with different
R/H ratios we need have been fabricated successfully using thermal chemical vapor deposition (CVD) method. Therefore, it is easily to study the effect of R/H ratios on the field emission characteristics of CNTs′ pillar arrays.
3.2.3 Effect of different CNTs′ pillar arrays with different R/H ratios on the field-emission characteristic
In our experiment, we get high density, vertical, long length and aligned CNTs pillar structure with different inter-pillar distance. Then, the relation between R/H ratios and field-emission characteristics from the CNT pillar array are studied. Fig.3.2.5(a) shows field emission current density (J) as a function of applied electric field (E) for CNT pillar arrays with heights of 4.5μm, 8μm, 10μm, 10.8μm, 13.2μm, 22.3μm and 53μm (R/H=18.18, 11.2, 8, 7.8, 6.06, 3.6,and 1.5,respectively). We find that the best field-emission characteristic is obtained from the pillar array with R/H=8 (H=10μm) and it will be discussed later. Fig3.2.5(c) shows the Fowler–Nordheim (FN) plots corresponding to Fig3.2.5(a) .The occurrence of field emission is confirmed by the linearity of the FN plot. Fig 3.2.6(a) shows the turn-on field (Eturn-on) obtained from Fig.
3.2.5(a) as a function of R/H. Eturn-on is defined to be the field required to produce J=10 μA/cm2.The optimal value of R/H giving a value of minimum Eturn-on is approximately 8.
which corresponds to H=10μm. When R/H is smaller than 8, the field-emission characteristic is poor, which is likely to be due to the field screening effect. However, when the R/H is greater than 8, Eturn-on increases, and the reduction of field enhancement owing to a small H is conjectured.Fig3.2.6(b) shows R/H versus the field enhancement factor (β), [3.12] which is estimated from the slope of the FN plots in Fig. β is estimated experimentally by setting the work function of the CNTs to be 5 eV. The obtained values of β for the CNT pillar array with R/H values of 8, 11.2, 6.06 and 1.5 are 61060, 54540, 29075 and33031, respectively. The best field enhancement factor is obtained at R/H~8.
In the case for inter-pillar distance of 150μm, the relation between R/H ratios and field-emission characteristics from the CNT pillar array are studied. Fig3.2.7 shows emission current density (J) as a function of applied electric field (E) for CNT pillar arrays with heights of 14μm, 20μm, 26μm, 32μm, 47μm, 60μm and 82μm (R/H=10.7,7.5,6, 4.7,3.2,2.5and1.87,respectively). We also find that the best
field-emission characteristic is obtained from the pillar array with R/H=4.7 (H=32μm), and it will be discussed later. Fig.3.2.7(c) shows the Fowler–Nordheim (FN) plots corresponding to Fig3.2.7(a). The occurrence of field emission is confirmed by the linearity of the FN plot. Fig.3.2.8(a) shows the turn-on field obtained from Fig.3.2.7(b) as a function of R/H. The optimal value of R/H giving a value of minimum Eturn-on is approximately 4.7, which corresponds to H=32μm. When R/H is smaller than 4.7, the field-emission characteristic is poor, which is likely to be due to the field screening effect.
However, when the R/H is greater then 4.7, Eturn-on increases, and the reduction of field enhancement owing to a small H is conjectured.Fig.3.2.8(b) shows R/H versus the field enhancement factor (β), [3.12] which is estimated from the slope of the FN plots in Fig 3.2.7(c).The obtained values of β for the CNT pillar array with R/H values of 6, 4.7and 1.87 are 52907, 97345 and 42901, respectively. The best field enhancement factor is obtained at R/H~4.7.
In the case for inter-distance of 250μm, the relation between R/H ratios and field-emission characteristics from the CNT pillar array are studied. Fig.3.2.9(a) shows emission current density (J) as a function of applied electric field (E) for CNT pillar arrays with heights of 18μm, 23μm, 49μm, 80μm and 100μm (R/H=14,10,5.1, 3.1,3.2,2.5and1.87,respectively). We also find that the best field-emission characteristic is obtained from the pillar array with R/H=3.1 (H=80μm), as discussed later. Fig.3.2.9(c) shows the Fowler–Nordheim (FN) plots corresponding to Fig.3.2.9(a) The occurrence of field emission is confirmed by the linearity of the FN plot. Fig.3.2.10(a) shows the Eturn-on
obtains from Fig.3.2.9(b) as a function of R/H. The optimal value of R/H giving a value of minimum Eturn-on is approximately 3.1, which corresponds to H=80μm.When R/H is smaller then 3.1, the field-emission characteristic is poor, which is likely to be due to the field screening effect.However, when the R/H is greater than 3.1, Eturn-on increases, and the reduction of field enhancement owing to a small H is conjectured. Fig.3.2.10(b) shows R/H versus the field enhancement factor (β), [3.12] which is estimated from the slope of the FN plots in Fig.3.2.9(c). The obtained values of β for the CNT pillar array with R/H values of 10, 3.1and 2.5 were 58223, 89427 and 44792, respectively. The best field enhancement factor is obtained at R/H~3.1.
Based on these results and measurements, we find not all of CNT pillar arrays
Based on these results and measurements, we find not all of CNT pillar arrays