- Email: [email protected]
Charge leakage characteristic of glass substrate for liquid crystal display
Journal of
ELECTROSTATICS ELSEVIER
Journal of Electrostatics 35 (1995) 309 322
Charge leakage characteristic of glass substrate for liquid crystal display* Hideto Takahashi*, Katsumi Sato, Soithiro Sakata, Takao Okada Takasago Thermal Engineering Co., Ltd. 3150, Iiyama, Atugi, Kanagawa, 243-02 Japan
Received 13 August 1994; accepted 4 February 1995
Abstract The effect of moisture and surface contamination on the charge leakage characteristics of glass substrate for LCD was discussed. The surface resistivity decreased with a relative humidity increase; the charge attenuation rate increased with the relative humidity increase. Two kinds of glass substrate were positively corona-charged under the same condition: one kind immediately after UV/O 3 cleaning (cleaned sample) and the other left for about one month after UV/O 3 cleaning (uncleaned sample). The surface resistivity of the uncleaned sample was 10z, up to 105 times greater than that of the cleaned sample. The organic surface contaminant on the uncleaned sample was approximately three times greater than that on the cleaned sample. It was concluded that the attenuation rate of surface charge on glass substrate substantially depended on the relative humidity and also depended on the surface contamination caused by atmospheric hydrocarbon.
I. Introduction Dislocations and point defects which are mainly responsible for low yields of T F T - L C D originate in the adhesion of particles generated from the manufacturing equipment [1]. In the T F T - L C D manufacturing process, insulating glass is used as substrate material, which constantly repeats contact with and separation from a number of materials on the substrate carrier line. Therefore, the generation of static electricity and the charging of the substrate are inevitable, tending to invite the adsorption of particles and the dielectric b r e a k d o w n of the accumulated film. The electric charge on the surface of an object can be removed either by causing the charge to leak from the surface to the grounding side or by neutralizing the surface charge
* This paper first appeared, in Japanese, in the proceedings of the Institute of Electrostatics, Japan, Vol. 18, No. 4 in 1994. * Corresponding author. 0304-3886/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSD! 0 3 0 4 - 3 8 8 6 ( 9 5 ) 0 0 0 0 4 - 6
310
H. Takahashi et al./Journal of Electrostatics 35 (1995) 309 322
with air ions generated by corona discharge neutralizers. In this research, aiming at the removal of the surface charge of glass substrates by leaking, it was studied how the humidity of the ambient atmosphere and the organic contaminants on the glass surface would affect the leak characteristic of the electric charge of glass substrates.
2. Experimental 2.1. Glass substrates All the glass substrates used in the experiment were made of non-alkaline glass C O R N I N G # 7059, whose main characteristics are listed in Table 1.
2.2 Cleaning of glass substrates After ultrasonic rinsing with ultrapure water (17-18 Mf~ cm) for 60 min, both faces of the substrates were cleaned with a UV/O3 photochemical cleaner (Japan Storage Battery's DUV-25 x 4A) for 10min each. Hereinafter the series of cleaning will be referred to as "UV/O3", and glass substrates immediately after "UV/O3" will be referred to as "cleaned" ones, and those allowed to stand for one month in a clean room (Class 1 by US Federal Standard on particles of 0.5 lam diameter, 22.5 _+ 0.5 °C temperature, and 42 _+ 1% relative humidity) after "UV/O3", as "uncleaned" ones. Dry cleaning with UV/O3 photochemical cleaners is used in the T F T manufacturing process to remove organic contaminants on the surface which cannot be wholly removed by wet cleaning [-2].
2.3. Measurement of surface resistivity The surface resistivity of each glass substrate (80 × 80 x 1.1 ' mm 3) was measured with a high resistance meter ( H E W L E T T PACKARD's Model 4339A; main electrode: 50 mm~b; counterelectrode: 70 mm~b; guard electrode: 100 x 100 mm 2) under various humidity conditions in a chamber with thermo-hygrostat. First, the glass substrate was placed in the chamber at a temperature of 23 °C and at a relative Table 1 Specification of glass substrate (CORN1NG # 7059); Relative dielectric constant 5.84 Component
Wt%
SiO2 BaO B20 3 AI20 3 Others
49 25 15 10 1
H. Takahashiet al./Journal of Electrostatics 35 (1995) 309 322
31 I
humidity of 40%. The time of transition from these temperature-humidity conditions to the prescribed temperature-humidity conditions was 20 min. One hour after the atmosphere in the chamber reached the prescribed temperature-humidity conditions, the surface resistivity of the glass substrate was measured. The handling of the glass substrate was carried out wholly within the chamber with the atmospheres in and out of the chamber isolated from each other. Applied voltage to the glass substrate was at 500 V for 60 s in the measurement. 2.4. Measurement of surface charge attenuation of cleaned glass substrates
The effect of humidity on the charge attenuation rates of cleaned glass substrates was examined by the method shown in Fig. 1. Tetra-fluoroethylene (Teflon) sheets (2 x 2 x 0.1' m m 3) were put in four corners on a surface of a grounded indium tin oxide (ITO) film coated glass substrate (hereinafter called the I T O plate), and a glass substrate (50 x 50 x 1.1 t m m 3) was placed over it. The probe of the surface potentiometer ( M O N R O E ' s Model 244) was set 2 m m above the central part of the glass substrate. The electrode needle was brought into contact with the glass substrate surface, which was charged by impressing the needle with a DC voltage of about + 1 kV. At this time, there was a uniform distribution of surface potential values ow,'r the surface of the glass substrate. In this state, the electrode needle was grounded, and the attenuation of surface potential over time was measured and recorded. The
Recorder Surface [ potent lometer I
{Electrode[
Probe
I m,L W
DC
_]_
'T°/ Probe
Z5 +
Electr°del > • J_]~'"O
{ mm
Z5
20
5 I ass substrate
ZO 5
Fig. 1. Schematic diagram for measuring surface charge decay of glass substrate.
H. Takahashi et al. /Journal o f Electrostatics 35 (1995) 309 322
312
measurement was carried out in a chamber with thermo-hygrostat at a temperature of 22.5 °C and at relative humidities of 40%, 60% and 80%.
2.5. Evaluation of contamination of glass surface In order to evaluate the extents of contamination of the surfaces of cleaned and uncleaned glass substrates, the contact angles were measured with a contact angle meter (Kyowa Kaimen Kagaku's CA-S 150 type). Ultrapure water was used for liquid drops. The surface contamination was also assessed by X-ray photoelectron spectroscopy (XPS) (with FISONS Instrument's INSPECTOR). Binding energy values were referenced to gold (Au 4t7 at 83.93 eV) and copper (Cu 2p3 at 932.47 eV).
3. Results
3.1. Effect of humidity of leak characteristic of electric charge of cleaned glass substrate 3.1.1. Surface resistivity The surface resistivity (f]/[]) of the cleaned glass substrate was measured at various relative humidities (shown by the symbol [] in Fig. 2). The surface resistivity of a cleaned glass substrate decreased as relative humidity increased. When relative humidity increased by 20%, surface resistivity fell to approximately one-tenth of its initial value.
3.1.2. Attenuation speed of electrostatic charge Assuming Vo to be the initial surface potential of a charged material surface, potential V(t) after time t can be expressed as
V(t) = V o e x p ( - t/z),
(1)
z = CR,
(2)
where C is the electrostatic capacity (F/cm2), R is the surface resistivity (~/[]), and r is the relaxation time (s). The potential attenuation curve is characterized by the magnitude of the relaxation time r, and z increases in proportion to the surface resistivity R. Fig. 3 shows measurements of the electric charge attenuation over time for a charged, cleaned glass substrate under relative humidities of 40%, 60% and 80%. The black dots in the figure show results when the glass substrate surface was grounded; the whitened dots when the substrate surface was not grounded. The vertical axis shows logarithmically the rate of attenuation, V/Vo. Linear regression analysis was performed on these data up to the point at which the attenuation rate fell to one-tenth of its initial value.
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309-322
313
i 0 to mm
•
[] C~
lot6
i 0 ~4
D L.
Cleaned
[]
El
L.
0
[] 0
1 0 $o I0 ~
'
2'o
'
4b
~b
'
'
80
:
,oo
Relative humidity [%] Fig. 2. Effect of relative humidity on surface resistivity of glass substrate, The temperature was constant at 23' C.
L,...a
|0-0. 5
• • • O []
L. e--
l c i t-0
oN
e..+.a
#O*/,ffH/Grounded "-1 60~RH/Grounded 80~RH/Grounded 40~WNot grounded 60~H/Not grounded 80~RH/Not grounded
i01.s
10-2.0
0
.
.
.
.
. 10
.
.
20
Time [sec] Fig. 3. Time course of surface charge attenuation rate (V/Vo) of glass substrate immediately after UV/O3 cleaning.
314
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309 322
We found that whether the substrate surface was grounded or not, at a relative humidity of 60% or 80%, the correlation coefficient r 2 was greater than 0.98, and the potential attenuated more or less exponentially with respect to time. At a relative humidity of 40%, however, r z was less than 0.3, and potential did not attenuate as described by Eq. (1). Fig. 4 shows the relationship between surface resistivity and relaxation time for each of the three relative humidity values, where relaxation time represents the time required for the attenuation rate to reach 1/e = 0.37. Relaxation time at relative humidities of 60% and 80% is equivalent to r in Eq.(1). As the ambient humidity increased and surface resistivity decreased, charge attenuation speed increased whether the glass substrate surface was grounded or not. When the surface was not grounded, however, attenuation speed was some 3-10 times slower than when the surface was grounded. 3.1.3. M o i s t u r e adsorption o f glass substrates
The increase in the moisture adsorption of glass substrates with the rise in the relative humidity of the ambient atmosphere was measured by the following method. A mechanical microbalance of the electromagnetic force compensation type ( M E T T L E R ' s AT20, with the smallest graduation of 1 tag) was placed in a chamber which allowed for interception between the atmospheres inside and outside. Two glass substrates (50 x 50 × 1.1 t m m 3) were placed in the chamber. First in the usual clean
10 3
I! I I I
r---n
o o,2 L~f
[] 10 2
.ooo.oo.==o.oooo.
,.oooo..ooooo°o ......
o~,q
[]
Not grounded ! I o..oouoooooiooooo | ! ! ! !
oo
Grounded
r'-' 0
.........
!
!
101
.................
I.-=
............
t ................
~ ................
['
! t at $O~RH
I
!
!
[] I0 ~
iI t at 60~RH i • l t at 80~;RH i!
|0-I I{
101;
i 0 lz
1013
i014
Surface resistivity [fl/D] Fig. 4. Relationship between surface resistivity and relaxation time of charge attenuation in a glass substrate immediately after UV/O3 cleaning.
H. Takahashi et at./Journal of Electrostatics 35 (1995) 309-322
315
room atmosphere of 22.5 °C and 42% RH, the initial weights of samples 1 h after cleaning with UV/O3 were measured. Then a Petri dish containing saturated KCI solution was placed in the chamber, and the relative humidity was gradually raised by about 10% every half an hour. Each time the relative humidity in the chamber reached a prescribed level, the glass substrates were weighed. The glass substrates and the balance were handled with the atmospheres in and out of the chamber always isolated from each other. The weight increment, after compensation for buoyancy, during the relative humidity rise from 42% was assumed to be the moisture adsorption. The volume of moisture adsorption was converted to number of molecular layers by assuming an intermolecular distance of 3 A for the adsorbed moisture molecules. (The figure of 3 A was calculated based on macroscopic density, indicated by the number of water molecules per unit volume.) The increases in the number of aqueous molecule layers with reference to a relative humidity of 42% are traced in Fig. 5. It is estimated that the adsorbed water increases by about 2 molecular layers with every 10% rise in relative humidity.
3.2. Effect of surface contamination on charge leakage characteristics of glass substrate 3.2.1. Surface resistivity vs. charge attenuation speed The black dots (11) in Fig. 2 show surface resistivity for an "uncleaned" glass substrate left in a clean room for one month after cleaning. The surface resistivity of the uncleaned glass substrate was of the order of 10 z-10 s times greater than that of the cleaned glass substrate.
230
7
200
6
= ° ~
"~ ,_.,
150
=~
100
.3
.N~
o~
~
50 !
0
• 40
l 50
t 60
~ 70
0 80
Relative humidity [g] Fig. 5. Increase in adsorbed moisture on glass surface relative to the adsorbed moisture at 42% RH.
H. Takahashi et al./Journal of Electrostatics 35 (1995) 309 322
316
By the method as shown in Fig. 6, cleaned and uncleaned glass substrates, each placed in contact with a grounded ITO plate, were corona-charged in the positive polarity. Their surface charge densities, 5 min later, after the charging were measured. The measured surface potentials were converted into surface charge densities by the following equation: Vs = V b -
(3)
V = crd/e~eo,
where a is the equivalent surface charge density on the upper surface of glass substrate (C/m2), Vs is the difference in potential corresponding to the thickness of the glass substrate (V), Vb is the measured surface potential (V), V is the difference in potential between the probe and the glass surface ( = 0 V), er is the relative dielectric constant of glass ( = 5.84), e0 is the permittivity of air ( = 8.85 x 10 -12 F/m), and d is the glass substrate thickness ( = 1.1 x 10 -3 m). Measurements were repeated five times at 25 points on the substrate surface in an atmosphere of 23 °C and 43% RH. Table 2 shows the average surface charge density for the entire 25 measurement points. While virtually no charge was observed with the cleaned glass substrate, which had an average surface charge density of the order of 10 -8 C/m 2, a strong charge was observed for the uncleaned glass substrate, which had an average surface charge density of the order of 10 5 C/m z.
d ~r
'E:O
-
/,LI Electrode (170)
"-------A_
~ d
-V,~
V
! Sample (Glass)
hi
Electrode (Probe)
Fig. 6. Method for measuring surface charge density.
Table 2 Surface charge density of glass substrate measured at 5 min after corona-charging Glass substrate Cleaned a Uncleaned b
Charge density (Coulomn/m z) < 4.7 × 10 -8 4.3 x 10- 5
a Immediately after cleaning with ultrasonic rinsing and U V / O a. b Left for about one m o n t h after the cleaning.
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309-322
317
3.2.2. Surface contamination o f glass substrates T a b l e 3 shows the m e a s u r e d c o n t a c t angles of the surfaces of cleaned a n d u n c l e a n e d glass substrates. T h e c o n t a c t angles of u n c l e a n e d s u b s t r a t e s were evidently g r e a t e r t h a n those of the cleaned ones. T h e surfaces of cleaned s u b s t r a t e s e x p o s e d to the clean r o o m a t m o s p h e r e for m a n y h o u r s were found to be m o r e h y d r o p h o b i c t h a n they were i m m e d i a t e l y after the cleaning. F u r t h e r , the results of analysis of the elemental c o m p o s i t i o n of g l a s s s u b s t r a t e surfaces by X P S are s h o w n in T a b l e 4. T h e region to a d e p t h of a b o u t 80 A from the surface was analyzed. In this analysis, the ratio a m o n g O, St, AI a n d Ba, which are the original c o n s t i t u e n t elements of the glass substrates, was s u b s t a n t i a l l y identical to that specified by their m a n u f a c t u r e r as s h o w n in T a b l e 1. Besides these elements, C was also detected. T h e q u a n t i t y of C a t o m s a d h e r i n g to the glass s u b s t r a t e surfaces was a b o u t three times greater on the u n c l e a n e d surfaces t h a n that on the cleaned ones. X P S analysis was also p e r f o r m e d on the surface of a glass s u b s t r a t e that h a d been e x p o s e d to a clean r o o m a t m o s p h e r e for 15 d following U V / O 3 cleaning. Fig. 7 shows the resulting C(ls) spectra. M o s t of the c a r b o n a d h e r e d to the glass s u b s t r a t e surface
Table 3 Contact angles on glass surfaces measured with ultra-pure water Glass substrate
Contact angle (degree)
Cleaned" Uncleaned b
4 70 + 2 (S.D.)
a Immediately after cleaning with ultrasonic rinsing and UV/O 3. b Left for about one month after the cleaning.
Table 4 Results of glass surfaces analysis with XPS Element
C O Si A1 Ba B
Atomic% measured with XPS Cleaned
Uncleaned
4.8 62.9 (66)a 23.8 (25) 5.7 (6) 2.8 (3) n.d.b
15.9 55.7 (66) 21.3 (25) 4.2 (5) 2.9 (3) n.d.
a Values in parentheses show atomic % when carbon is excluded. b n.d. = not detectable.
Atomic % calculated from Table 1
67 19 5 5 5
318
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309- 322
Binding
energy
(eV) o,-,~
a:
C-C
284.93
a
=
b: C-O
286.23
.~
c: C=0
287.30
~-
d: 0=C-0
288.73
>,,
e: C03
289 " 53
~"
b~
[[~
;
~
C
e,,,
;
;
295
;
."
;
;
I
290
I
I
I
!
I
I
I
285
I
I
!
:
;
;
.
280
Binding energy leVY Fig. 7. XPS spectra of the C (ls) peak for the glass substrate exposed to clean room atmosphere for 15 d after UV/O3 cleaning. was organic substance containing hydrocarbon radicals, carbonyl groups, and hydroxyl groups, and no peaks originating in metal carbide or graphite were detected. A slight peak originating in carbonic acid was detected, but this is thought to be the result of trace amounts of carbonate formed by Ba in the glass substrate. 3.2.3. Adhered carbon vs. surface resistivity Fig. 8 shows changes over time in volume of adhered organic matter (A), contact angle (B), and surface resistivity at 40% RH (C) of a UV/O3-cleaned glass substrate exposed in a clean room atmosphere. The volume of adhered organic matter was represented as the ratio of carbon atoms to silicon atoms, the latter of which is an element found in the glass substrate. Contact angle and adhered organic matter increased over time. As surface contamination caused by the readhesion of organic matter proceeded, surface resistivity increased. When a UV/O3-cleaned glass substrate was exposed in a clean room atmosphere for 24 h, surface resistivity increased to approximately 10 times the value immediately after cleaning. The variations over time of the contact angle and the surface resistivity of UV/O3cleaned glass substrates when exposed to a purified clean room air atmosphere cleared of organic matter were also assessed. The purified clean room air was prepared by removing impurities from the clean room air by molecular sieving until the total organic carbon concentration was reduced to 50 ppb or below. A UV/O3-cleaned glass substrate was immediately placed in a 1 1 Telfon exposure chamber, and bathed in a 1 1/min flow of purified clean room air. Contact angle and surface resistivity were
H. Takahashi et al. /Journal o]"Electrostatics 35 (1995) 309- 322
319
O° 7 ' o
0.6
0.5, L.
0.4
o.3,
\L~
o.2, O.1 i
5°
e
40
~
20
,
.
.
l
.
':
A
.
,
B
,
C
2
N
,
O
'
IOO
200
Exposure time [hour] Fig. 8. Time dependence on (A) the surface atomic C/Si ratio; (B) contact angle; and (C) surface resistivity
at 40% RH of the glass substrate exposed to (I) clean room air, or to ([2]) purified air, after UV/O3 cleaning. measured outside the c h a m b e r in the clean r o o m atmosphere with 4 0 % RH. W h e n the substrates were left in the purified clean r o o m air, neither the contact angle nor the surface resistivity increased noticeably as shown by the symbol [] in Fig. 8.
4. Discussions
4.1. Effect of humidity In water, trace a m o u n t s of water molecules spontaneously disassociated into H + and O H - ions. The transfer of electric charge in water is a result primarily of the
320
H. Takahashi et aL /Journal of Electrostatics 35 (1995) 309-322
transfer diffusion of water molecules (H3O+, H9O+, etc.) that have associated with protons [3]. The layers of surface-adsorbed water increase with the rise in relative humidity. Aqueous molecules in lower layers, which are closer to the surface, are prevented from readily moving by their adsorption to the surface, but the ones in upper layers presumably move more readily. The decrease of surface resistivity with a rise in relative humidity seems attributable to an increase of more mobile aqueous molecules in the upper layers. As shown in Fig. 5, the quantity of water adsorbed by glass substrates increased with the rise in the relative humidity of the ambient atmosphere. This was also confirmed by Kawasaki [4], who represented the relationship between the number of water adsorption layers (the coating rate in the case of a single molecular layer or less) and the surface electric conduction by the following equation:
log(i/io) =/30,
(4)
where i is the conducted current when the number of adsorption layers is 0, io is the conducted current when there is no adsorption, and/3 is a constant of proportion. The findings that the surface resistivity of cleaned glass substrates decreased exponentially with a rise in relative humidity (Fig. 2) and that the attenuation of the charge of cleaned glass substrates accelerated with a rise in relative humidity (Fig. 3) are not in contradiction with the theory of Kawasaki. Nakagawa et al. I-5] measuring with fluoric anhydride the quantity of surfaceadsorbed water when electropolished stainless steel surfaces were exposed at 25 °C to argon gases having moisture contents of 1.5%, 2% and 3% (respectively corresponding to 48%, 64% and 96% RH at 25 °C), reported that there were about 40, 90 and 125 layers, respectively, of adsorbed water. The increments of adsorbed water due to a humidity rise according to our measurement are smaller by a one-digit order than the findings ofNakagawa et al. The reason for this difference will be a subject of future study. The attenuation of the surface potentials of cleaned glass substrates, whether the surface was grounded or not, accelerated with the increase in relative humidity. The leak of the charge occurring when the glass substrate surface was not grounded presumably took place via the air [6] or the tetrafluoroethylene sheets. For glass substrates immediately after cleaning with UV/O3, a certain amount of charge leak can be expected by raising the relative humidity of the ambient atmosphere even if the surface is not grounded. However, faster de-electrification (about 10 times as fast at 40% RH) is made possible by grounding the glass substrate surface.
4.2. Effect of surface contamination The surface resistivity of glass substrates decreased with the rise in relative humidity whether they were cleaned or uncleaned (Fig. 2). However, the values of surface resistivity were 102 to 105 times greater for uncleaned substrates than that for the cleaned (Fig. 2). The attenuation of the surface charge density was significantly faster for cleaned glass substrates than for the uncleaned (Table 2). These findings can be
H. Takahashi et al./Journal of Electrostatics 35 (1995) 309 322
321
explained on the basis of the measurement of contact angles and surface analysis by XPS (Fig. 8) as follows. Even if organic contaminants are removed once from the glass substrate surface by cleaning with UV/O3, the exposure of the surface to the clean room atmosphere for many hours again invites the contamination of the surface with hydrocarbon components deriving from the atmosphere. The contamination raises the surface resistivity and thereby inhibits the leakage of the surface charge. The organic matter adhering to the surface, as it enlarges the contact angle, is likely to consist of organic compounds having such hydrophobic groups as the alkyl group or the benzene ring. In water dissolving compounds having hydrophobic groups, the hydrogen bond among aqueous molecules directly around the hydrophobic groups becomes stronger than in pure water, and restricts molecular movements (hydrophobic hydration) I-7]. The organic matter adhering to the glass substrate surface presumably causes the hydrophobic hydration among adsorbed aqueous molecules and obstructs the migratory diffusion of aqueous molecules. As a result, the organic contamination of glass surface inhibits the movements of electric charges on the surface and increases the surface resistivity.
5. Conclusion The effect of moisture and surface contamination on the charge leakage characteristics of glass substrate for LCD was discussed. The surface resistivity decreased with a relative humidity increase. The amount of organic surface contaminant, most of which has been removed once with UV/O3, increased with the lapse of time after the cleaning owing to the readhesion of organic matter deriving from the clean room atmosphere, and the surface resistivity also increased with the lapse of time. The surface charge on glass substrates can be removed through grounding within a few seconds under the following conditions: the charged surface should be grounded as soon as possible after cleaning with UV/O3 while being exposed to a high level of relative humidity in clean rooms, or it should be grounded at any time after the cleaning while being exposed to a purified atmosphere cleared of organic matter.
Acknowledgements We would like to thank Dr. B.V. Crist of Hakuto Co., Ltd., for helpful suggestions and comments on XPS analysis of glass samples.
References [1] M. Morimoto,1993LiquidCrystal Process Technologies,SemiconductorWorld SpecialIssue (1993) 39. [2] H. Itoh, 1993 LiquidCrystal Process Technologies,SemiconductorWorld Special Issue (1993) 138.
322
H. Takahashi et aL /Journal of Electrostatics 35 (1995) 309 -322
[3] D. Eisenberg and W. Kauzmann, The Structure and Properties of Water (transl. S, Seki and T. Matsuo), p. 229, Misuzu Shobo (1975) (in Japanese). [4] K. Kawasaki, Ohyo Butsuri (J. Appl. Phys. Japan) 27 (1958) 216. [5] Y. Nakagawa, H. Izumi and T. Ohmi, Proc. 19th Symp. ULSI Clean Technology, Tokyo (1993) pp. 41-51. [6] Y. Murata and N. Masui, Proc. 1986 Annual Meeting of The Institute of Electrostatics Japan, Tokyo (October 1986) pp. 123-124 (in Japanese). [-7] S. Kondo, T, Ishikawa and I. Abe, Kyuchaku No Kagaku (Science of adsorbtion), Maruzen (1991) pp. 26-30 (in Japanese).
ELECTROSTATICS ELSEVIER
Journal of Electrostatics 35 (1995) 309 322
Charge leakage characteristic of glass substrate for liquid crystal display* Hideto Takahashi*, Katsumi Sato, Soithiro Sakata, Takao Okada Takasago Thermal Engineering Co., Ltd. 3150, Iiyama, Atugi, Kanagawa, 243-02 Japan
Received 13 August 1994; accepted 4 February 1995
Abstract The effect of moisture and surface contamination on the charge leakage characteristics of glass substrate for LCD was discussed. The surface resistivity decreased with a relative humidity increase; the charge attenuation rate increased with the relative humidity increase. Two kinds of glass substrate were positively corona-charged under the same condition: one kind immediately after UV/O 3 cleaning (cleaned sample) and the other left for about one month after UV/O 3 cleaning (uncleaned sample). The surface resistivity of the uncleaned sample was 10z, up to 105 times greater than that of the cleaned sample. The organic surface contaminant on the uncleaned sample was approximately three times greater than that on the cleaned sample. It was concluded that the attenuation rate of surface charge on glass substrate substantially depended on the relative humidity and also depended on the surface contamination caused by atmospheric hydrocarbon.
I. Introduction Dislocations and point defects which are mainly responsible for low yields of T F T - L C D originate in the adhesion of particles generated from the manufacturing equipment [1]. In the T F T - L C D manufacturing process, insulating glass is used as substrate material, which constantly repeats contact with and separation from a number of materials on the substrate carrier line. Therefore, the generation of static electricity and the charging of the substrate are inevitable, tending to invite the adsorption of particles and the dielectric b r e a k d o w n of the accumulated film. The electric charge on the surface of an object can be removed either by causing the charge to leak from the surface to the grounding side or by neutralizing the surface charge
* This paper first appeared, in Japanese, in the proceedings of the Institute of Electrostatics, Japan, Vol. 18, No. 4 in 1994. * Corresponding author. 0304-3886/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSD! 0 3 0 4 - 3 8 8 6 ( 9 5 ) 0 0 0 0 4 - 6
310
H. Takahashi et al./Journal of Electrostatics 35 (1995) 309 322
with air ions generated by corona discharge neutralizers. In this research, aiming at the removal of the surface charge of glass substrates by leaking, it was studied how the humidity of the ambient atmosphere and the organic contaminants on the glass surface would affect the leak characteristic of the electric charge of glass substrates.
2. Experimental 2.1. Glass substrates All the glass substrates used in the experiment were made of non-alkaline glass C O R N I N G # 7059, whose main characteristics are listed in Table 1.
2.2 Cleaning of glass substrates After ultrasonic rinsing with ultrapure water (17-18 Mf~ cm) for 60 min, both faces of the substrates were cleaned with a UV/O3 photochemical cleaner (Japan Storage Battery's DUV-25 x 4A) for 10min each. Hereinafter the series of cleaning will be referred to as "UV/O3", and glass substrates immediately after "UV/O3" will be referred to as "cleaned" ones, and those allowed to stand for one month in a clean room (Class 1 by US Federal Standard on particles of 0.5 lam diameter, 22.5 _+ 0.5 °C temperature, and 42 _+ 1% relative humidity) after "UV/O3", as "uncleaned" ones. Dry cleaning with UV/O3 photochemical cleaners is used in the T F T manufacturing process to remove organic contaminants on the surface which cannot be wholly removed by wet cleaning [-2].
2.3. Measurement of surface resistivity The surface resistivity of each glass substrate (80 × 80 x 1.1 ' mm 3) was measured with a high resistance meter ( H E W L E T T PACKARD's Model 4339A; main electrode: 50 mm~b; counterelectrode: 70 mm~b; guard electrode: 100 x 100 mm 2) under various humidity conditions in a chamber with thermo-hygrostat. First, the glass substrate was placed in the chamber at a temperature of 23 °C and at a relative Table 1 Specification of glass substrate (CORN1NG # 7059); Relative dielectric constant 5.84 Component
Wt%
SiO2 BaO B20 3 AI20 3 Others
49 25 15 10 1
H. Takahashiet al./Journal of Electrostatics 35 (1995) 309 322
31 I
humidity of 40%. The time of transition from these temperature-humidity conditions to the prescribed temperature-humidity conditions was 20 min. One hour after the atmosphere in the chamber reached the prescribed temperature-humidity conditions, the surface resistivity of the glass substrate was measured. The handling of the glass substrate was carried out wholly within the chamber with the atmospheres in and out of the chamber isolated from each other. Applied voltage to the glass substrate was at 500 V for 60 s in the measurement. 2.4. Measurement of surface charge attenuation of cleaned glass substrates
The effect of humidity on the charge attenuation rates of cleaned glass substrates was examined by the method shown in Fig. 1. Tetra-fluoroethylene (Teflon) sheets (2 x 2 x 0.1' m m 3) were put in four corners on a surface of a grounded indium tin oxide (ITO) film coated glass substrate (hereinafter called the I T O plate), and a glass substrate (50 x 50 x 1.1 t m m 3) was placed over it. The probe of the surface potentiometer ( M O N R O E ' s Model 244) was set 2 m m above the central part of the glass substrate. The electrode needle was brought into contact with the glass substrate surface, which was charged by impressing the needle with a DC voltage of about + 1 kV. At this time, there was a uniform distribution of surface potential values ow,'r the surface of the glass substrate. In this state, the electrode needle was grounded, and the attenuation of surface potential over time was measured and recorded. The
Recorder Surface [ potent lometer I
{Electrode[
Probe
I m,L W
DC
_]_
'T°/ Probe
Z5 +
Electr°del > • J_]~'"O
{ mm
Z5
20
5 I ass substrate
ZO 5
Fig. 1. Schematic diagram for measuring surface charge decay of glass substrate.
H. Takahashi et al. /Journal o f Electrostatics 35 (1995) 309 322
312
measurement was carried out in a chamber with thermo-hygrostat at a temperature of 22.5 °C and at relative humidities of 40%, 60% and 80%.
2.5. Evaluation of contamination of glass surface In order to evaluate the extents of contamination of the surfaces of cleaned and uncleaned glass substrates, the contact angles were measured with a contact angle meter (Kyowa Kaimen Kagaku's CA-S 150 type). Ultrapure water was used for liquid drops. The surface contamination was also assessed by X-ray photoelectron spectroscopy (XPS) (with FISONS Instrument's INSPECTOR). Binding energy values were referenced to gold (Au 4t7 at 83.93 eV) and copper (Cu 2p3 at 932.47 eV).
3. Results
3.1. Effect of humidity of leak characteristic of electric charge of cleaned glass substrate 3.1.1. Surface resistivity The surface resistivity (f]/[]) of the cleaned glass substrate was measured at various relative humidities (shown by the symbol [] in Fig. 2). The surface resistivity of a cleaned glass substrate decreased as relative humidity increased. When relative humidity increased by 20%, surface resistivity fell to approximately one-tenth of its initial value.
3.1.2. Attenuation speed of electrostatic charge Assuming Vo to be the initial surface potential of a charged material surface, potential V(t) after time t can be expressed as
V(t) = V o e x p ( - t/z),
(1)
z = CR,
(2)
where C is the electrostatic capacity (F/cm2), R is the surface resistivity (~/[]), and r is the relaxation time (s). The potential attenuation curve is characterized by the magnitude of the relaxation time r, and z increases in proportion to the surface resistivity R. Fig. 3 shows measurements of the electric charge attenuation over time for a charged, cleaned glass substrate under relative humidities of 40%, 60% and 80%. The black dots in the figure show results when the glass substrate surface was grounded; the whitened dots when the substrate surface was not grounded. The vertical axis shows logarithmically the rate of attenuation, V/Vo. Linear regression analysis was performed on these data up to the point at which the attenuation rate fell to one-tenth of its initial value.
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309-322
313
i 0 to mm
•
[] C~
lot6
i 0 ~4
D L.
Cleaned
[]
El
L.
0
[] 0
1 0 $o I0 ~
'
2'o
'
4b
~b
'
'
80
:
,oo
Relative humidity [%] Fig. 2. Effect of relative humidity on surface resistivity of glass substrate, The temperature was constant at 23' C.
L,...a
|0-0. 5
• • • O []
L. e--
l c i t-0
oN
e..+.a
#O*/,ffH/Grounded "-1 60~RH/Grounded 80~RH/Grounded 40~WNot grounded 60~H/Not grounded 80~RH/Not grounded
i01.s
10-2.0
0
.
.
.
.
. 10
.
.
20
Time [sec] Fig. 3. Time course of surface charge attenuation rate (V/Vo) of glass substrate immediately after UV/O3 cleaning.
314
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309 322
We found that whether the substrate surface was grounded or not, at a relative humidity of 60% or 80%, the correlation coefficient r 2 was greater than 0.98, and the potential attenuated more or less exponentially with respect to time. At a relative humidity of 40%, however, r z was less than 0.3, and potential did not attenuate as described by Eq. (1). Fig. 4 shows the relationship between surface resistivity and relaxation time for each of the three relative humidity values, where relaxation time represents the time required for the attenuation rate to reach 1/e = 0.37. Relaxation time at relative humidities of 60% and 80% is equivalent to r in Eq.(1). As the ambient humidity increased and surface resistivity decreased, charge attenuation speed increased whether the glass substrate surface was grounded or not. When the surface was not grounded, however, attenuation speed was some 3-10 times slower than when the surface was grounded. 3.1.3. M o i s t u r e adsorption o f glass substrates
The increase in the moisture adsorption of glass substrates with the rise in the relative humidity of the ambient atmosphere was measured by the following method. A mechanical microbalance of the electromagnetic force compensation type ( M E T T L E R ' s AT20, with the smallest graduation of 1 tag) was placed in a chamber which allowed for interception between the atmospheres inside and outside. Two glass substrates (50 x 50 × 1.1 t m m 3) were placed in the chamber. First in the usual clean
10 3
I! I I I
r---n
o o,2 L~f
[] 10 2
.ooo.oo.==o.oooo.
,.oooo..ooooo°o ......
o~,q
[]
Not grounded ! I o..oouoooooiooooo | ! ! ! !
oo
Grounded
r'-' 0
.........
!
!
101
.................
I.-=
............
t ................
~ ................
['
! t at $O~RH
I
!
!
[] I0 ~
iI t at 60~RH i • l t at 80~;RH i!
|0-I I{
101;
i 0 lz
1013
i014
Surface resistivity [fl/D] Fig. 4. Relationship between surface resistivity and relaxation time of charge attenuation in a glass substrate immediately after UV/O3 cleaning.
H. Takahashi et at./Journal of Electrostatics 35 (1995) 309-322
315
room atmosphere of 22.5 °C and 42% RH, the initial weights of samples 1 h after cleaning with UV/O3 were measured. Then a Petri dish containing saturated KCI solution was placed in the chamber, and the relative humidity was gradually raised by about 10% every half an hour. Each time the relative humidity in the chamber reached a prescribed level, the glass substrates were weighed. The glass substrates and the balance were handled with the atmospheres in and out of the chamber always isolated from each other. The weight increment, after compensation for buoyancy, during the relative humidity rise from 42% was assumed to be the moisture adsorption. The volume of moisture adsorption was converted to number of molecular layers by assuming an intermolecular distance of 3 A for the adsorbed moisture molecules. (The figure of 3 A was calculated based on macroscopic density, indicated by the number of water molecules per unit volume.) The increases in the number of aqueous molecule layers with reference to a relative humidity of 42% are traced in Fig. 5. It is estimated that the adsorbed water increases by about 2 molecular layers with every 10% rise in relative humidity.
3.2. Effect of surface contamination on charge leakage characteristics of glass substrate 3.2.1. Surface resistivity vs. charge attenuation speed The black dots (11) in Fig. 2 show surface resistivity for an "uncleaned" glass substrate left in a clean room for one month after cleaning. The surface resistivity of the uncleaned glass substrate was of the order of 10 z-10 s times greater than that of the cleaned glass substrate.
230
7
200
6
= ° ~
"~ ,_.,
150
=~
100
.3
.N~
o~
~
50 !
0
• 40
l 50
t 60
~ 70
0 80
Relative humidity [g] Fig. 5. Increase in adsorbed moisture on glass surface relative to the adsorbed moisture at 42% RH.
H. Takahashi et al./Journal of Electrostatics 35 (1995) 309 322
316
By the method as shown in Fig. 6, cleaned and uncleaned glass substrates, each placed in contact with a grounded ITO plate, were corona-charged in the positive polarity. Their surface charge densities, 5 min later, after the charging were measured. The measured surface potentials were converted into surface charge densities by the following equation: Vs = V b -
(3)
V = crd/e~eo,
where a is the equivalent surface charge density on the upper surface of glass substrate (C/m2), Vs is the difference in potential corresponding to the thickness of the glass substrate (V), Vb is the measured surface potential (V), V is the difference in potential between the probe and the glass surface ( = 0 V), er is the relative dielectric constant of glass ( = 5.84), e0 is the permittivity of air ( = 8.85 x 10 -12 F/m), and d is the glass substrate thickness ( = 1.1 x 10 -3 m). Measurements were repeated five times at 25 points on the substrate surface in an atmosphere of 23 °C and 43% RH. Table 2 shows the average surface charge density for the entire 25 measurement points. While virtually no charge was observed with the cleaned glass substrate, which had an average surface charge density of the order of 10 -8 C/m 2, a strong charge was observed for the uncleaned glass substrate, which had an average surface charge density of the order of 10 5 C/m z.
d ~r
'E:O
-
/,LI Electrode (170)
"-------A_
~ d
-V,~
V
! Sample (Glass)
hi
Electrode (Probe)
Fig. 6. Method for measuring surface charge density.
Table 2 Surface charge density of glass substrate measured at 5 min after corona-charging Glass substrate Cleaned a Uncleaned b
Charge density (Coulomn/m z) < 4.7 × 10 -8 4.3 x 10- 5
a Immediately after cleaning with ultrasonic rinsing and U V / O a. b Left for about one m o n t h after the cleaning.
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309-322
317
3.2.2. Surface contamination o f glass substrates T a b l e 3 shows the m e a s u r e d c o n t a c t angles of the surfaces of cleaned a n d u n c l e a n e d glass substrates. T h e c o n t a c t angles of u n c l e a n e d s u b s t r a t e s were evidently g r e a t e r t h a n those of the cleaned ones. T h e surfaces of cleaned s u b s t r a t e s e x p o s e d to the clean r o o m a t m o s p h e r e for m a n y h o u r s were found to be m o r e h y d r o p h o b i c t h a n they were i m m e d i a t e l y after the cleaning. F u r t h e r , the results of analysis of the elemental c o m p o s i t i o n of g l a s s s u b s t r a t e surfaces by X P S are s h o w n in T a b l e 4. T h e region to a d e p t h of a b o u t 80 A from the surface was analyzed. In this analysis, the ratio a m o n g O, St, AI a n d Ba, which are the original c o n s t i t u e n t elements of the glass substrates, was s u b s t a n t i a l l y identical to that specified by their m a n u f a c t u r e r as s h o w n in T a b l e 1. Besides these elements, C was also detected. T h e q u a n t i t y of C a t o m s a d h e r i n g to the glass s u b s t r a t e surfaces was a b o u t three times greater on the u n c l e a n e d surfaces t h a n that on the cleaned ones. X P S analysis was also p e r f o r m e d on the surface of a glass s u b s t r a t e that h a d been e x p o s e d to a clean r o o m a t m o s p h e r e for 15 d following U V / O 3 cleaning. Fig. 7 shows the resulting C(ls) spectra. M o s t of the c a r b o n a d h e r e d to the glass s u b s t r a t e surface
Table 3 Contact angles on glass surfaces measured with ultra-pure water Glass substrate
Contact angle (degree)
Cleaned" Uncleaned b
4 70 + 2 (S.D.)
a Immediately after cleaning with ultrasonic rinsing and UV/O 3. b Left for about one month after the cleaning.
Table 4 Results of glass surfaces analysis with XPS Element
C O Si A1 Ba B
Atomic% measured with XPS Cleaned
Uncleaned
4.8 62.9 (66)a 23.8 (25) 5.7 (6) 2.8 (3) n.d.b
15.9 55.7 (66) 21.3 (25) 4.2 (5) 2.9 (3) n.d.
a Values in parentheses show atomic % when carbon is excluded. b n.d. = not detectable.
Atomic % calculated from Table 1
67 19 5 5 5
318
H. Takahashi et al. /Journal of Electrostatics 35 (1995) 309- 322
Binding
energy
(eV) o,-,~
a:
C-C
284.93
a
=
b: C-O
286.23
.~
c: C=0
287.30
~-
d: 0=C-0
288.73
>,,
e: C03
289 " 53
~"
b~
[[~
;
~
C
e,,,
;
;
295
;
."
;
;
I
290
I
I
I
!
I
I
I
285
I
I
!
:
;
;
.
280
Binding energy leVY Fig. 7. XPS spectra of the C (ls) peak for the glass substrate exposed to clean room atmosphere for 15 d after UV/O3 cleaning. was organic substance containing hydrocarbon radicals, carbonyl groups, and hydroxyl groups, and no peaks originating in metal carbide or graphite were detected. A slight peak originating in carbonic acid was detected, but this is thought to be the result of trace amounts of carbonate formed by Ba in the glass substrate. 3.2.3. Adhered carbon vs. surface resistivity Fig. 8 shows changes over time in volume of adhered organic matter (A), contact angle (B), and surface resistivity at 40% RH (C) of a UV/O3-cleaned glass substrate exposed in a clean room atmosphere. The volume of adhered organic matter was represented as the ratio of carbon atoms to silicon atoms, the latter of which is an element found in the glass substrate. Contact angle and adhered organic matter increased over time. As surface contamination caused by the readhesion of organic matter proceeded, surface resistivity increased. When a UV/O3-cleaned glass substrate was exposed in a clean room atmosphere for 24 h, surface resistivity increased to approximately 10 times the value immediately after cleaning. The variations over time of the contact angle and the surface resistivity of UV/O3cleaned glass substrates when exposed to a purified clean room air atmosphere cleared of organic matter were also assessed. The purified clean room air was prepared by removing impurities from the clean room air by molecular sieving until the total organic carbon concentration was reduced to 50 ppb or below. A UV/O3-cleaned glass substrate was immediately placed in a 1 1 Telfon exposure chamber, and bathed in a 1 1/min flow of purified clean room air. Contact angle and surface resistivity were
H. Takahashi et al. /Journal o]"Electrostatics 35 (1995) 309- 322
319
O° 7 ' o
0.6
0.5, L.
0.4
o.3,
\L~
o.2, O.1 i
5°
e
40
~
20
,
.
.
l
.
':
A
.
,
B
,
C
2
N
,
O
'
IOO
200
Exposure time [hour] Fig. 8. Time dependence on (A) the surface atomic C/Si ratio; (B) contact angle; and (C) surface resistivity
at 40% RH of the glass substrate exposed to (I) clean room air, or to ([2]) purified air, after UV/O3 cleaning. measured outside the c h a m b e r in the clean r o o m atmosphere with 4 0 % RH. W h e n the substrates were left in the purified clean r o o m air, neither the contact angle nor the surface resistivity increased noticeably as shown by the symbol [] in Fig. 8.
4. Discussions
4.1. Effect of humidity In water, trace a m o u n t s of water molecules spontaneously disassociated into H + and O H - ions. The transfer of electric charge in water is a result primarily of the
320
H. Takahashi et aL /Journal of Electrostatics 35 (1995) 309-322
transfer diffusion of water molecules (H3O+, H9O+, etc.) that have associated with protons [3]. The layers of surface-adsorbed water increase with the rise in relative humidity. Aqueous molecules in lower layers, which are closer to the surface, are prevented from readily moving by their adsorption to the surface, but the ones in upper layers presumably move more readily. The decrease of surface resistivity with a rise in relative humidity seems attributable to an increase of more mobile aqueous molecules in the upper layers. As shown in Fig. 5, the quantity of water adsorbed by glass substrates increased with the rise in the relative humidity of the ambient atmosphere. This was also confirmed by Kawasaki [4], who represented the relationship between the number of water adsorption layers (the coating rate in the case of a single molecular layer or less) and the surface electric conduction by the following equation:
log(i/io) =/30,
(4)
where i is the conducted current when the number of adsorption layers is 0, io is the conducted current when there is no adsorption, and/3 is a constant of proportion. The findings that the surface resistivity of cleaned glass substrates decreased exponentially with a rise in relative humidity (Fig. 2) and that the attenuation of the charge of cleaned glass substrates accelerated with a rise in relative humidity (Fig. 3) are not in contradiction with the theory of Kawasaki. Nakagawa et al. I-5] measuring with fluoric anhydride the quantity of surfaceadsorbed water when electropolished stainless steel surfaces were exposed at 25 °C to argon gases having moisture contents of 1.5%, 2% and 3% (respectively corresponding to 48%, 64% and 96% RH at 25 °C), reported that there were about 40, 90 and 125 layers, respectively, of adsorbed water. The increments of adsorbed water due to a humidity rise according to our measurement are smaller by a one-digit order than the findings ofNakagawa et al. The reason for this difference will be a subject of future study. The attenuation of the surface potentials of cleaned glass substrates, whether the surface was grounded or not, accelerated with the increase in relative humidity. The leak of the charge occurring when the glass substrate surface was not grounded presumably took place via the air [6] or the tetrafluoroethylene sheets. For glass substrates immediately after cleaning with UV/O3, a certain amount of charge leak can be expected by raising the relative humidity of the ambient atmosphere even if the surface is not grounded. However, faster de-electrification (about 10 times as fast at 40% RH) is made possible by grounding the glass substrate surface.
4.2. Effect of surface contamination The surface resistivity of glass substrates decreased with the rise in relative humidity whether they were cleaned or uncleaned (Fig. 2). However, the values of surface resistivity were 102 to 105 times greater for uncleaned substrates than that for the cleaned (Fig. 2). The attenuation of the surface charge density was significantly faster for cleaned glass substrates than for the uncleaned (Table 2). These findings can be
H. Takahashi et al./Journal of Electrostatics 35 (1995) 309 322
321
explained on the basis of the measurement of contact angles and surface analysis by XPS (Fig. 8) as follows. Even if organic contaminants are removed once from the glass substrate surface by cleaning with UV/O3, the exposure of the surface to the clean room atmosphere for many hours again invites the contamination of the surface with hydrocarbon components deriving from the atmosphere. The contamination raises the surface resistivity and thereby inhibits the leakage of the surface charge. The organic matter adhering to the surface, as it enlarges the contact angle, is likely to consist of organic compounds having such hydrophobic groups as the alkyl group or the benzene ring. In water dissolving compounds having hydrophobic groups, the hydrogen bond among aqueous molecules directly around the hydrophobic groups becomes stronger than in pure water, and restricts molecular movements (hydrophobic hydration) I-7]. The organic matter adhering to the glass substrate surface presumably causes the hydrophobic hydration among adsorbed aqueous molecules and obstructs the migratory diffusion of aqueous molecules. As a result, the organic contamination of glass surface inhibits the movements of electric charges on the surface and increases the surface resistivity.
5. Conclusion The effect of moisture and surface contamination on the charge leakage characteristics of glass substrate for LCD was discussed. The surface resistivity decreased with a relative humidity increase. The amount of organic surface contaminant, most of which has been removed once with UV/O3, increased with the lapse of time after the cleaning owing to the readhesion of organic matter deriving from the clean room atmosphere, and the surface resistivity also increased with the lapse of time. The surface charge on glass substrates can be removed through grounding within a few seconds under the following conditions: the charged surface should be grounded as soon as possible after cleaning with UV/O3 while being exposed to a high level of relative humidity in clean rooms, or it should be grounded at any time after the cleaning while being exposed to a purified atmosphere cleared of organic matter.
Acknowledgements We would like to thank Dr. B.V. Crist of Hakuto Co., Ltd., for helpful suggestions and comments on XPS analysis of glass samples.
References [1] M. Morimoto,1993LiquidCrystal Process Technologies,SemiconductorWorld SpecialIssue (1993) 39. [2] H. Itoh, 1993 LiquidCrystal Process Technologies,SemiconductorWorld Special Issue (1993) 138.
322
H. Takahashi et aL /Journal of Electrostatics 35 (1995) 309 -322
[3] D. Eisenberg and W. Kauzmann, The Structure and Properties of Water (transl. S, Seki and T. Matsuo), p. 229, Misuzu Shobo (1975) (in Japanese). [4] K. Kawasaki, Ohyo Butsuri (J. Appl. Phys. Japan) 27 (1958) 216. [5] Y. Nakagawa, H. Izumi and T. Ohmi, Proc. 19th Symp. ULSI Clean Technology, Tokyo (1993) pp. 41-51. [6] Y. Murata and N. Masui, Proc. 1986 Annual Meeting of The Institute of Electrostatics Japan, Tokyo (October 1986) pp. 123-124 (in Japanese). [-7] S. Kondo, T, Ishikawa and I. Abe, Kyuchaku No Kagaku (Science of adsorbtion), Maruzen (1991) pp. 26-30 (in Japanese).