The dual structure on a cicada wing has been revealed by G.Y. Xie et al. [3].
Though the structure could be replicated on PMMA films by Au deposition and nano-imprint, the size of the structure could not be made change. Furthermore, the size of the sample was limited by the original cicada wing. In this chapter, we will show how to fabricate dual structure on a silicon substrate. The whole processing steps are shown in Fig. 3-1.
Fig. 3-1 Processing steps.
3-1 ARC and PR Coating
First, we get thermal oxidized wafers from NDL (National Nano Device Laboratories). 60nm thick silicon dioxide is grown on each wafer. Prior to spin coating, RCA clean process should be applied. Next, anti-reflection coating (ARC) and photoresist (PR) are coated layer by layer by using a spin coater. The specific parameters of ARC and photoresist coating are shown in Table 3-1. ARC, HXRIC-16, is provided by Brewer Science Asia. Photoresist, Sumiresist PFI-34A2, is provided by Sumitomo Chemical.
The thickness of ARC layer should be controlled very accurately. The reflectance in the interface of ARC and PR changes with the variation of ARC thickness. In subsequent lithography process, the reflectance in the interface of ARC and PR has bad influence on profile of PR patterns.
If the reflectance in the interface of ARC and PR approaches zero, the sidewall of the PR pattern will be vertical and there will be no residue PR layer between the PR patterns and ARC. On the contrary, if the reflectance in the interface of ARC and PR is higher than 1%, there will be residue PR layer and there will be standing wave on side wall of the PR pattern. The residue PR layer made subsequent etching process unstable because it should take one more etching process to remove the residue PR layer before performing subsequent process. The residue PR layer not only needs one
more process step to remove it but also makes the height of the PR pattern lower since the original thickness of the PR layer is held constant by the coating process.
To achieve the lowest reflectance in the interface of ARC and PR, the most suitable thickness of ARC is calculated by thin film theory. As shown in Fig. 3-2, the lowest reflectance in the interface of ARC and PR can be achieved when the thickness of ARC is 128nm. Referring to the data sheet of ARC shown in Fig. 3-3 [23], it was found that the 128 nm thickness of ARC could be achieved by setting rotation speed of spin coater to 3200 rpm. In fact, the rotation speed of spin coater is set to be 5500 rpm. Due to storage condition and continual degradation of ARC, the parameters for ARC coating change day by day.
After ARC layer is coated on the wafer, the wafer is baked on a hot plate with temperature at 175 o C for 60 seconds. This process drives the solvent away and releases some residue stress in the structure of ARC film.
It is desired to have thicker PR layer in considering the process of pattern transfer to ARC and silicon dioxide. Pattern transfer to ARC and silicon dioxide is not the only factor in determining the thickness of the PR layer. The thickness of PR layer should be controlled to prevent the PR pattern from collapsing.
Fig. 3-2 Reflectance in the interface between PR and ARC layers.
Fig. 3-3 Spin speed curve of ARC [20].
The PR pattern is prone to collapse if the aspect ratio of the PR pattern was higher than 2. The thickness of PR layer should be larger than the thickness of ARC.
The thickness of the PR layer was determined to be 200 nm. Referring to the data sheet of the PR shown in Fig. 3-4 [24], it is found that 200 nm thickness of PR layer could be achieved by setting the rotation speed of spin coater to 4000 rpm. In our experiment, the rotation speed of PR coating is set to 4000rpm and the thickness of PR is 224 nm.
After PR is coated on wafer, the wafer is soft-baked on a hot plate at temperature of 110oC for 60 seconds. The baking process drives the solvent in PR away and makes the film property stable. The specific baking parameters are listed in Table 3-2. After the wafer is baked, the wafer is cut into 7 smaller chips for subsequent processing.
Fig. 3-4 Spin speed curve of PR [21].
Table 3-1：Coating parameters of ARC and photoresist.
1st step 2nd step
speed (rpm) time (seconds) speed (rpm) time (seconds)
ARC 500 18 5500 60
PR 500 10 4000 60
Table 3-2：Baking parameters of ARC and PR.
soft bake post bake hard bake
temperature time temperature time temperature time
(℃) (seconds) (℃) (seconds) (℃) (seconds)
ARC 175 60
PR 90 60 110 60 140 60
3-2 Two-beam Interference Lithography
3-2-1 Exposure Principle of Two-beam Interference
Lithography
Two-beam interference lithography is a simple method in fabricating large area periodical nano patterns. In comparison with E-beam lithography, two-beam interference lithography could be more economic and time-saving. As shown in Fig.
3-5, this is our whole exposure system.
The period of grating pattern which is fabricated by two-beam interference lithography can be expressed by
The chip coated with ARC and PR was placed on a stage and then the exposure was performed. As mentioned above, the PR recorded the dosage distribution of the two-beam interference image which was spatially sinusoidal.
Fig. 3-5 Exposure system of two-beam interference lithography.
After development of the exposed chip, PR grating pattern was formed on the chip. To achieve two dimensional PR patterns, the chip was exposed with half exposure dosage and the chip was rotated ninety degrees to perform another exposure with half exposure dosage.
The PR in our experiment was positive type. There was a threshold of exposure dosage for the positive PR to be washed out in the development. The PR with exposure dosage higher than the threshold will be washed away in development.
The PR pattern can be made to be posts array or holes array by changing the total exposure dosage. For example, we assume the dosage distribution of the interference pattern as shown in Fig. 3-6 and Fig. 3-7, and the exposure threshold of the PR is 40 (arbitrary unit, the same with the unit of the color bar in Fig. 3-6 and Fig. 3-7). The dosage distribution shown in Fig 3-6 will result in PR pattern to be posts array after development because only the deep blue part of the PR in Fig. 3-6 remains. The dosage distribution shown in Fig. 3-7 will result in holes array of PR pattern after development because the red part of the PR in Fig. 3-7 will be washed away. As shown in Fig. 3-8 and Fig. 3-9, the PR pattern of posts array can be made by exposing higher total dosage and the PR pattern of holes array can be made by exposing lower total dosage.
Fig. 3-6 Two-beam interference image with higher exposure dosage.
Fig. 3-7 Two-beam interference image with lower exposure dosage.
Fig. 3-8 SEM image: PR pattern of posts array (total exposure dosage: 58mJ).
Fig. 3-9 SEM image: PR pattern of holes array (total exposure dosage: 42mJ).
In our experiment, higher total exposure dosage was chosen to make the final PR pattern to be posts array. Various total exposure dosage can be achieved by changing the exposure time. As shown in Fig. 3-5, the shutter which was driven by computer can control the exposure time very precisely. After exposed by using the two-beam interference lithography system, the chip was baked on the hot plate at temperature of 110oC for 60 seconds. After post baked, the chip was developed by EPD-1000 for 60 seconds. EPD-1000 was the standard developer of our PR which was provided by Everlight Chemical Industrial Co. The development was composed of 2.38%
Tetramethylammonium hydroxide. The top view and side view of the posts array PR pattern are shown in the SEM images, Fig. 3-8 and Fig. 3-10. The period of our PR pattern was 270nm.
Fig. 3-10 SEM image: side view of PR pattern of posts array.
3-3 Pattern Transfer to ARC and Silicon Dioxide by
RIE
After the PR pattern of posts array has been achieved, we started to transfer the pattern from PR layer to ARC and silicon dioxide. Silicon dioxide was used to be the mask of subsequent ICP process. The ICP process will be discussed in the following section.
In considering the geometry and scale of our pattern, anisotropic etch was needed. RIE (Samco RIE 10NR) in this step was used.
The first step of dry etching process in our experiment is to transfer pattern from PR to underlying ARC layer. ARC, a species of polymer, is usually etched by oxygen-based RIE in dry etching process. Here, CF4 (25sccm) and O2 (6sccm) were used as the feeding gases of RIE to transfer pattern from PR to ARC layer. In our experiment, the power of RIE was set to 150W and the pressure is set to 1.3 Pa.
The mechanism of CF4/O2 RIE to etch polymer was discussed by A. M. Wrobel et al. [22]. The oxygen atoms caused aromatic ring opening and increased carbonyl (C
= O) bond concentration. They found addition of CF4 to O2 dramatically increased the etch rate. First, even small amounts of CF4 (<10%) could result in oxygen atomic concentration. Secondly, atomic fluorine was produced by dissociation of the CF4
molecule. Fluorine atoms could abstract hydrogen atoms from the polymer, thus creating radical sites on the surface which readily undergo reactions with oxygen atoms.
The hydrogen abstraction by fluorine ions is sufficiently exothermic to cleave carbon-carbon bonds. Combined with carbon-oxygen bonds formation with oxygen atoms, this eventually leads to the observed volatile reaction products. When the ration of CF4 in the mix CF4/O2 gases increases to a certain level, the polymer surface is passivated by the formation of chemically stable fluorocarbon groups which strongly impede the etching process.
This phenomenon may lead to some good property in our experiment because of the vertical side wall of the PR pattern can be protected by the fluorocarbon layer.
Addition of CF4 to the etching gases lead to inhibitor layer on the whole surface of the PR pattern. Since the ions bombarded the samples mostly in the vertical way, the side wall which is protected by the fluorocarbon layer is not etched. On the contrary, the etching gas chemistry of pure oxygen leads to certain critical dimension loss without inhibitor layer on the side wall of the PR pattern.
The second step of our dry etching process is to transfer ARC pattern to silicon dioxide layer. In this step CHF3 (30 sccm) was chosen as the feed gas chemistry in RIE. Referring to some text book about semi-conductor technology, we used CHF3 to
etch silicon dioxide. In this step the pressure was set to 1.3 Pa and the power of RIE was set to 100W. The etching time was 260 seconds. Then the ARC pattern was etched by RIE with O2 gas. This process reduced the ARC posts to smaller sharp tips.
The final ARC and silicon dioxide patterns are shown in Fig. 3-11.
Fig. 3-11 AFM image of ARC and silicon dioxide pattern.
3-4 Pattern Transfer from ARC and Silicon Dioxide
Pattern to Silicon Substrate by ICP-RIE
In this step ARC and silicon dioxide pattern was transferred to silicon substrate to form dual structure by ICP. The etching tool we used in this step was Elionix EIS-700 Inductively Coupled Plasma Etching system. The ICP system was shown in Fig. 3-14. This ICP system is located in the Institute of Physics, Academia Sinica. The ICP recipe is shown in Table 3-4. Initially negative tapered structure was fabricated by recipe 1 which was provided by the manager of this ICP system. Negative tapered pillar structure means the diameter of the topper part is larger than the diameter of the base. To make the slope of the pillar from negative to positive, process time for passivation was increased. When the process time in recipe 3 was increased to 9 seconds, the structure became positive tapered. The influence of passivation time on profile control is shown in Fig. 3-13. The results fabricated by ICP recipes 1, 2, and 3 in Table 3-4 are shown in Fig. 3-13 (a), (b), and (c). Addition of oxygen in etching gas also helps to fabricate positive tapered structure. The addition of oxygen can lower SF6 ratio in the etching gas flow and the etching rate is suppressed. The influence of oxygen flow rate on profile control is shown in Fig. 3-14. The results fabricated by ICP recipes 1, 2, and 3 in Table 3-4 are shown in Fig. 3-14 (a), (b) and
(c).The dual structures were fabricated by recipe 2, 6, and 9. These dual structures are shown in Fig. 3-15, Fig. 3-16, and Fig. 3-17 Dual structure C respectively. The upper part of the structure was shorter than the lower part because silicon dioxide was a stronger mask than ARC. The side wall of the structure shown in Fig. 3-15 is more vertical than the side wall of the structure shown in Fig. 3-16 because the structure is prone to be vertical by using recipe 2.
Fig. 3-12 ICP system located in the Institute of Physics, Academia Sinica.
Fig. 3-13 The influence of passivation time on profile control. Passivation time：(a) 6s (b) 7.5s (c) 9s.
Fig. 3-14 The influence of oxygen flow rate on profile control. Oxygen flow rate：(a) 0sccm (b) 2sccm (c) 7sccm.
Fig. 3-15 Dual structure A.
Fig. 3-16 Dual structure B.
Fig. 3-17 Dual structure C.
3-5 Tunable Ratio of the Heights between Upper Part
and Lower Part of the Structure
In Section 3-4, the dual structure with shorter upper part is fabricated. In this section, the process for fabricating dual structure with shorter height of the lower part will be described.
As shown in Fig. 3-1, one more RIE process was performed after O2 etching process. In this step, the etching gas was CHF3 and the recipe was the same as the recipe used to transfer pattern from ARC to silicon dioxide. The etching time of this step was 45 seconds. After this RIE step, ICP process was performed to transfer pattern into silicon substrate. The fabrication results are shown in Fig. 3-18 and Fig.
3-19. By performing one more CHF3 RIE step, the dual structures with shorter lower part were fabricated because part of the uncovered Si dioxide mask was etched.
Fig. 3-18 Dual structure D.
Fig. 3-19 Dual structure E.
Table 3.4 ICP-RIE parameters.
3-6 Surface Modification
The dual structures were surface modified by ICP-RIE. The surface modification process was performed by using the 11th recipe of ICP-RIE which is listed in Table 3.4. The deposited fluorocarbon film was hydrophobic and had a static water contact angle of 109°.
3-7 Summary
In this chapter, the fabrication process of dual structures is detailed. The overall diameter of the structures can be modified by initial exposure dosage of two-beam interference. The sharp tips of the dual structures can be formed by using O2 RIE to reduce the ARC posts pattern to small sharp tips pattern. The height of the lower part of dual structure can be reduced by one more CHF3 RIE process after O2 RIE process.