Ca corrosion test

Ca degradation test was commonly utilized to measure the water vapor transmission rate (WVTR) which was often named as the “Ca test”. [35, 36] This method was based on the corrosion of calcium films and subjected at a fixed temperature and humidity. A home built Ca test system whose detection limit is 5x10^-5g/m²/day as exhibit in Fig. 3-10 and the experimental temperature was keep at 25℃ and 100 % relative humidity (RH) condition.

The fabrication of Ca sensor follows the following sequences. A 100 nm silver film was deposited by e-gun evaporated on the flexible PET film (45 mm × 45 mm size, 125 µm thick) to serve as the four-point probe electrodes. Then, calcium layer of 400 nm was deposited by evaporation deposition as the detection layer for moisture reaction. The important chemical reactions of calcium with moisture are listed as following:

2Ca+H2O→2CaO (3-4) CaO+H₂O→Ca(OH)₂ (3-5)

Assuming a homogenous corrosion occurs when the moisture ingress through PET film to the calcium surface, the resistance of remaining Ca layer is measured in-situ by a four point probe method and LabVIEW software, and is expressed by

and is expressed by Eq. (3-6)

bh R ρ l

=

(3-6)

where ρ denotes the Ca resistivity, l is the length and b is the width of the Ca layer, and h is the thickness of the calcium film. The water vapor permeation rate, P is then calculated using the change rate of resistance according to Eq. 3-7 [36].

[ ]

electrodes, δ is the Ca density, and M is the molar mass of the indicated reagent. Ri and hi are initial values of R and h.

Fig. 3-10 Schematic diagram of a home-built WVTR system.

3.4.6 Scanning probe microscope (SPM)

SPM (Veeco Dimension 5000) was utilized to study the surface morphology and the roughness of PET after the CF4 plasma treatment. The scanning boundary was 5×5 µm² and the deviation of roughness was below 0.5 angstrom (A). The operation was carried out at room temperature and one atmosphere.

3.4.7 FE-SEM

Field-emission scanning electro spectroscopy (FE-SEM) (JEOL 6700F) was employed to investigate the morphology of regular pattern. The operation pressure was under 9.86×10^-5 torr and the emission current was set at 10µA.

3.4.8 n&k analyzer

The n&k analyzer 1200 in National Nano Device Laboratory (NDL) was used in this study to obtain the varied thickness of photo-resist and PET by O2 plasma etching during the fabrication of regular pattern. The experimental steps were listed below:

1. To scan standard sample as a baseline. (The wavelength ranged from 190 nm to 900 nm.) 2. To put sample wafer upside down on the n&k analyzer, and then scan again to collect the

typical experimental curve.

The experimental curve was fitted using Forouhi-Boomer Dispersion relation [64] to deduce the film thickness. The goodness of fit was all above 99% for the entire sample.

Chapter 4 Results and Discussion

4.1 Characteristics of fluorination on PET

4.1.1 Wetting property and morphology of fluorinated PET.

Hydrophobicity was the index to determine the wetting property, and the evaluation of hydrophobicity could be made through water contact angle measurements. The original contact angle of PET utilized in this thesis was about 64.5° while a significant increment to 107.7° after CF₄ plasma treatment for 5 minutes at power 60W which indicated the surface became hydrophobic (contact angle>90°) as shown in Fig. 4-1. To understand the dependence of treatment time and hydrophobic characteristic, we extended treatment time was performed from 5 to 15 minutes. Fig. 4-2 shown the contact angle versus different treatment time, the value seemed to be saturated at 106~107°. This result indicated that the PET surface with treatment time within 15 minutes had the same hydrophobicity.

Fig. 4-1 Comparison of contact angle between the (a) untreated PET, and (b) 60W 5 minutes CF₄ plasma treatment.

(a) (b)

0 5 10 15 60

70 80 90 100 110 120

contact angle(degree)

treatment time (min)

Fig. 4-2 The relation between plasma treatment time and contact angle.

To learn more about the morphology of CF₄ plasma modified PET, the scanning probe microscopy (SPM) had been then used in this work by evaluating the surface area 5×5 µm². Fig. 4-3 showed the surface morphology of untreated PET with height scale 20 nm and the roughness was only 0.34 nm. The roughness increased to 1.01 nm after CF4 plasma treatment for 15 minutes at power 60W. The variation in surface roughness of treated PET as a function of treatment time was exhibited in Fig. 4-4. Revealed by literature studies, the CF4

plasma modification was described as the sum of two mechanisms: ion etching and fluorination, and the two reactions seemed to be competitive and parallel. The ion etching action had been suggested to be mostly due to ion bombardment which will roughen the surface. [65]

Fig. 4-3(a) The surface roughness of untreated PET, and (b) 15 minutes treatment.

(a) (b)

0 5 10 15 0.2

0.4 0.6 0.8 1.0

surface roughness(nm)

treatment time(min)

Fig. 4-4 The variation in surface roughness as a function of treatment time.

As mention in previous chapter, contact angle would be affected with roughness factor.

However, this phenomenon didn’t appeal in our case because the increment of surface roughness was only 0.67 nm. The slight increase in roughness may attribute to the pure CF4 as the gas source. Several researches concluded that the addition of other gas source will enhance the ion etching process. [66, 67]

4.1.2 Chemical structure and depth of plasma fluorination zone

The chemical structure of plasma fluorinated PET was justified by the resolution of the X-ray Photoelectron Spectroscopy (XPS) spectra. Typically the XPS spectra of PET C 1s had three main peaks shown in Fig. 4-5 and table 4.1.1 C–C/C–H at 284.6eV, C–O at 286.4 eV, O=C at 288.7 eV. After CF₄ plasma treatment for 60w 5min, there was a dramatic change exhibited on the XPS spectra which could be attributed to formation of fluorinate carbon functionalities: CF at 289.3 eV, CF2 at 291.1 eV and CF3 at 293.1 eV.[68,69] These results indicated that both the C-C、C-H and C=O group were all likely effected by fluorination. As shown in O (1s) spectra in Fig 4-6 and table 4-2, C=O bond at decreased in the proportion while C-O bond increased relatively. An extra peak at 534.7 eV was assigned as C(O)F. This

peak didn’t appear in the C 1s spectra due to the relative low concentration of C(O)F_X group, and the peak position was within the CF₂ and CF₃ boundary.[70,71] Energy provided by RF power generated the fluorine F* and CFn (n =1, 2, 3) radicals which were easily reacted with PET. The mechanism of plasma etching and surface fluorination was proposed as follows.

Substitution of hydrogen usually initiated with hydrogen abstraction: R-H+F*→R*+HF.

Chain scission in ester group also leaded to the formation of radicals. The fluorine radical generated by CF4 plasma subsequently undergoes fluorination reactions by addition of different fluorinated radical. XPS spectra indicated that 26% of PET component were affected with fluorinated radical for treatment power at 60W for 5 min, and this number evaluated from different treatment time shown within a range 24~26%. This message revealed that the degree of surface fluorination were almost the same during the treatment time 5 to 15 minutes.

The depth of the XPS measurement was about 5 nm in this study, and the component ratio within the detected zone were almost the same for different treatment time. Since the increase in roughness was unobvious during the plasma treatment, the similar hydrophobicity on the surface of PET revealed by contact angle measurement was controlled by the degree of fluorination. Furthermore, we used XRR to quantify the fluorination zone, and the results were arranged in table 4-3. After the CF4 plasma treatment, XPS showed the surface of PET formed a “teflon like” layer. For 5 minutes treatment, a dense layer formed on the top of PET with thickness 22 nm while the density varied from 1.42 to 1.68 g /cm³. The weight increment by fluorination is due to replacement of hydrogen by fluorine in the fluorinated layer. [46,72]

Literature research indicated that the density of polymers was increased greatly under fluorination and usually varied over the range 1.6~2 g /cm³. [46] Corresponding with the extended treatment time, the depth of the fluorinated zone increased from 22 nm to 47 nm as the treatment time varied from 5 to 15 minutes. XRR results revealed that the fluorination wasn’t limited to the surface directly exposed to plasma. [46,73]

Fig. 4-5 The C 1s spectra of (a)untreated PET(b) CF₄ plasma treated PET with 60w for 5min.

Table 4-1 The component ratio of untreated and modified PET based on C 1s spectra.

C-C, C-H

C-O C=O C-CFn CF2 CF3

Bonding energies, eV

284.6 286.4 288.7 289.3 291.1 293.1

Untreated PET 46% 27.8% 26.2% － － －

60W 5 min 33.5% 24.1% 16.2% 8.1% 12.5% 5.6%

60W 10 min 33.8% 22.9% 15.8% 5.8% 13.2% 8.5%

60W 15 min 36.3% 22.6% 16.2% 7.5% 11.5% 5.9%

Fig. 4-6 The O 1s spectra of (a) untreated PET (b) CF4 plasma treated PET with 60w for 5min.

(a) (b)

(a) (b)

Table 4-2 The component ratio of untreated and modified PET based on O 1s spectra.

Table 4-3 Density and depth of fluorination zone measured by X-ray Reflectivity (XRR).

Fluorination

4.2 Moisture solubility and diffusivity of fluorinated layer.

As mentioned in last section, the fluorination of the PET achieved non-polar or low-polar surface detected by contact angle. To further understand the hydrophobic characteristic, the fundamental properties such as moisture diffusivity and solubility of fluorinated PET were investigated. Quartz crystal microbalance (QCM) was used to measure moisture solubility and diffusivity of fluorinated layer. Fig 4-7 showed the transient moisture absorption curve of untreated PET with moisture uptake 1.86 wt.%, surprisingly decreased to 1.41 wt.% of PET after 15 min CF4 plasma treatment. Both of the two cases reached to the equilibrium level within 500 seconds under environment condition set at 25℃ and 100% relative humidity (RH). The moisture uptake as a function of plasma treatment time was arranged in Fig. 4-8.

The moisture uptake shows different tendency with the contact angle whose value kept at 106~107°, the continuously decreased in moisture uptake corresponded with the thicker fluorination zone on the PET surface. With the lower composition of polar group such as C=O demonstrated with XPS analysis, the non-polar group bond increased relatively and attributed

to the lower moisture solubility of fluorinated PET. To considerate a Teflon, the solubility of moisture on Teflon surface is around 10^-3 mol/mol of polymer repeat unit, which translated into a moisture uptake of only 0.01 wt.%. [74] We attempted to quantify the diffusion coefficients of PET films based on equilibrium moisture absorption isotherms analyzed with diffusion model developed by Crank and Park: [75, 76]

( )

2.1×10^-11m²/sec. The decrease in diffusion coefficient may attribute to the fluorinated layer on the surface of PET. It’s difficult to quantify the diffusion coefficient of fluorinated layer individually due to the ultra-thin thickness of fluorination zone. So we introduced the equation discussed by Meares [77] and attempted to evaluate the diffusion coefficient of fluorinated layer roughly.

D=D0exp[-γC] (4-2) Where D0: limit diffusion coefficient

γ: the plasticization coefficient C: the local permeant concentration

The parameter involved in the equation was based on the following assumption.

1. We assumed that the absorbed water vapor concentration on the fluorinated surface was directly proportional the hydrophobic composition. The XPS analysis indicated 26% of

PET molecules are replaced by fluorinated group.

2. The plasticization coefficient was constant for the untreated and fluorinated PET.

Taking the parameter into calculation, the diffusion constant of fluorination layer was 3×10^-11m²/sec. The actual diffusion constant of the fluorination layer must lower than this number because we had not taken the density factor into consideration. Its quite hard to further quantify the diffusion constant since the chemical structure of the fluorinated layer was different from the original PET, the higher density of the fluorinated layer may caused by lots of possibilities such as lower free volume or higher molecular weight. Moreover, different chain orientation and conformation also affected the diffusion constant. [39] However, we can consider the simplest condition by assuming the higher density of fluorination layer was just caused by lower free volume to simulate the true condition. Fujita et al. considering the dependence of the free volume and diffusion constant described as following equation: [78]

exp[ ]

The diffusion constant decreased to 8.7×10^-12m²/sec after we took the density factor into consideration. Generally speaking, the variation of the solubility and diffusion constant between the untreated PET and fluorinated PET were caused by the change of the surface affinity (polarity) and thus by integration of fluorine groups and CF_X compounds, giving a more difficult access to water molecules.

Fig. 4-7 The moisture absorption of untreated PET and CF4 plasma treatment for 15 min.

0 2 4 6 8 10 12 14 16

1.0 1.2 1.4 1.6 1.8 2.0 2.2

Moisture absorption(wt%)

Treatment time(min)

Fig. 4-8 Dependence of the moisture uptake and plasma treatment time.

Fig. 4-9 The transient curve and the fitting curve of untreated PET.

Fig. 4-10 The transient curve and the fitting curve of CF4 plasma modified PET for 15 min treatment.

0 5 10 15

Fig. 4-11 The variation of diffusion constant for different treatment time.

4.3 WVTR of fluorinated PET

To understand water vapor permeation rate (WVTR) of a polymer was a crucial factor in this study. Although the hydrophobic features of fluorinated PET has been discussed in previous section, contact angle only revealed the interfacial chemistry between a PET surface and a water drop; the true moisture permeation behaviors of a polymer and a fluorinated layer were still unrevealed.

Hence, the calcium degradation test was used to measure the WVTR of untreated and modified PET. The principle of WVTR measurement is based on in-situ calcium differential resistance following the reaction shown in chapter 2.2. First of all, the deviation of the experiment attributed to the moisture ingress from the O-ring was lower 1×10^-4 g/m²-day, and then the WVTR of a virgin PET with thickness of 125 µm was evaluated 2.7 g/m²-day shown in Fig 4.3.2. For the plasma treated cases, Modified PET with 5 min CF4 plasma treatment, exhibits a similar permeation curve, but with a lower permeation rate 1.6 g/ m²-day. Further increased treated time to 15 min, the WVTR of modified PET reduced to 0.43 g/m²-day. In Fig. 4.3.2, the initial stage with little resistance change based on the percentage of calcium

degradation during specific period was presumed the time required for the absorption of the water vapor and diffusion through the polymer composite until reach an equilibrium state.⁷⁹ This delay phenomenon was not clearly observed in the transient moisture absorption curve in QCM test. The reason was the much thinner of thickness in QCM sample with the dimension less than 0.4 µm. Assuming the tested PET without the micro-defect such as crack and pores, we can refer the dissolution/diffusion model as the water vapor permeated through the PET.

This mechanism is generally considered to be a three-step process. In the first step the water vapor was absorbed by the membrane surface on the upstream end. This is followed by the diffusion of the water vapor molecules through the polymer matrix. Finally, the water vapor molecular desorbed from the other side of surface and reacted with the calcium layer. Base on the process described above, the permeability following the ideal solution–diffusion model can be defined as:

DS

P= (4-4) where D= diffusion coefficient

S=solubility

Surface modification with fluorinated groups leads to decrease in the equilibrium water uptake due to the formation of “Telfon like” layer at the feed side while the dense fluorocarbon zone affects the diffusivity. Permeation rate is then effectively altered with the modified PET. To discuss the permeation rate of fluorination layer, we regarded the fluorinated PET as two lamination stack: the fluorocarbon layer and virgin PET. Then we can be obtained by the following equation: [80,81]

n

permeability of layers 1,2....n and L _i are the thickness of the layers 1,2....n. For a composite

of two layers,1 and 2, the series resistance model relates the overall permeation rate of the composite membrane to the permeation rate of each layer. If subscript 1 denotes the fluorocarbon zone and subscript 2 denotes the unmodified PET layer. The permeation rate of fluorocarbon zone is only 10^-3~10^-4 g/m²-day, less than one-thousandth of bare PET acquired from the equation (4-5). Although the thickness of fluorinated layer was less than 50 nm in this study, the extremely low WVTR of fluorinated layer suggested that the barrier property of PET can be improved by simple CF4 plasma treatment.

We successfully decreased the WVTR of PET films without by simple plasma treatment. Because we didn’t need to deposit the barrier layer, the reliability issues wouldn’t appeal here. The dissolution/diffusion model indicated that the lower WVTR value was contributed to the lower solubility and diffusivity in fluorinated layer. From the discussions before, this method may be limited by the formation of thicker fluorinated layer and the saturate of hydrophobicity even we extended the CF₄ plasma treatment time. To broaden the scope of the research, we would enhance the water-repellency characteristic of PET surface.

It’s suggested that such requirement could be attained by roughening the hydrophobic substance as discussed in section 2.5.

Fig. 4-12 Water vapor permeation rate measurement of untreated PET compared with the moisture ingress from O-ring.

Fig. 4-13 Water vapor permeation rate measurement of untreated PET film and CF₄ plasma treated-PET films.

4.4 The fabrication and WVTR of ultrahydrophobic structure

Fluorination was put into use to improve the WVTR of PET as discussed previously due to the lower water solubility and diffusion constant of fluorocarbon layer. To broaden the scope of the research, we would enhance the water-repellency characteristic of PET surface.

It’s suggested that such requirement could be attained by roughening the hydrophobic substance as discussed in section 2.5.

4.4.1 Selectivity of photoresist and PET during O

2

plasma etching

Lithography process described in section 3.3 defined the shape of pattern, dry etching was then carried out by O2 plasma. Fig. 4-14 showed the relation between the etching time and thickness variation of photoresist detected by n&k analyzer. The etching rate of photoresist was 144 nm/min while the etching rate for PET was 234 nm/min. Selectivity defined as the ratio of photoresist etching rate to PET etching rate was 1.625.

Selectivity=

t photoresis of

rate etching

PET of rate

etching (4-6)

Fig. 4-14 Variation of thickness and etching time for photoresist.

Fig. 4-15 Dependence of thickness and etching time for PET.

4.4.2 Configuration of the regular pattern

Lithography procedure illustrated in section 3.3 had been carried out and followed by the O₂ plasma which would dry etched the photoresist and uncovered area of PET. Fig. 4-16 was the top view and bird's-eye view of the regular pattern without any O₂ plasma treatment.

Lithography process translated the mask information such as the shape and propagations of the pattern to the photoresist. Hence, the configuration of regular pattern here would be accepted to be pillar-like. The height of the pillar shown in Fig 4-16 was about 1.63 µm which depended on the thickness of photoresist, and the diameter of the circular area was 2.2~2.3µm.

We executed the O2 plasma treatment until the photoresist was completely cleaned up. At the same time, PET out of the protection of photoresist also adopted the etching process. Along with the plasma treatment time for 700 second, the contours of the regular pattern were still pillar-like. Due to the faster etching rate of PET, we observed that the height of the pillar increased from 1.63µm to about 1.96µm and the diameter of the pattern was 1.8µm. In summary, we successfully applied the lithography process to fabricate the order regular

pattern, and the original objective to increase the roughness was also satisfied. The final configuration of the pattern maintained pillar-like during the etching process, and its dimension was within the micrometer classification.

Fig. 4-16 The configurations of the regular pattern without O₂ plasma etching (a) top view for large area,(b) the diameter of circular pattern, and (c) bird's-eye view.

Fig. 4-17 The configurations of the regular pattern with O2 plasma etching for 700 seconds with (a) bird's-eye view, and(b) the cross-section view.

4.4.3 Contact angle measurement for fluorinated regular pattern

The same CF₄ plasma treatment for 15 minutes was executed on regular pattern and we measured the contact angle of fluorinated regular pattern. Fig. 4-18 showed the contact angle for the surface which combined the fluorination and regular pattern was 134.7°. This result indicated the surface characteristic had entered the ultrahydrophobic ambits. As discussed in section 2.5, both Wenzel and Cassie- Baxter theories discussed the importance of roughness factor as we concerned the ultrahydrophobic characteristic.[55, 56] However, R. N. Wenzel

assumed that the space between the protrusions on the surface is filled with the liquid while

assumed that the space between the protrusions on the surface is filled with the liquid while

在文檔中 CF4 電漿處理對可撓式聚對苯二甲酸乙二酯基材水氣穿透率之影響 (頁 44-0)