LASER CLEANING OF NANO- TO MICROCONTAMINANTS
FROM CRITICAL SILICON SURFACES
S.I. Kudryashov,a) S. Shukla,b) S.D. Allena,b)
Department of Chemistry and Physics, Arkansas State University,
State University, AR 72467-0419, USA
b)
Department of Mechanical Engineering, University of Memphis,
Memphis, TN 38152-3180, USA
a)
Sub-micrometer (0.1-1 m) polystyrene and alumina particles were removed by single-shot pulsed laser radiation from Si wafer surfaces with
or without pre-deposited thin liquid layers. Nearly complete (>90%) single-shot laser cleaning has been achieved in combinations of polystyrene
and alumina particles with 2-propanol and water, respectively, while for
other combinations cleaning was absent or incomplete. Optical microscopy has revealed important transient microscopic interactions between
particles and Si substrates in the pre-deposited micron-thick liquid layers, resulting in some cases in strongly reduced particle-substrate coupling. Visualization results give insight into particle removal mechanisms relevant to our laser cleaning experimental conditions.
INTRODUCTION
The removal of sub-m particles adhered to surfaces poses a challenge to IC fabrication, space optics, high resolution and high power optics, large area displays, magnetic
storage devices and other critical surfaces (1-3). According to the IC fabrication industry’s current trends, the chip feature size is shrinking rapidly. In particular, for Moore’s
law (4) projecting doubling the number of transistors on an IC chip will be every two
years, the problem will only become more significant. As chip feature size decreases,
smaller particles can mask a larger portion of the pattern element. This will mean that
part of the chip will not be produced to specification and will lead to the rejection of the
whole chip. As a result, in the modern very large scale integration (VLSI) fabrication industry 75% of yield loss is due to particle contamination (3). Various cleaning techniques
(wet chemical cleaning, scrubbing, pressurized jets and ultrasonic processes) (5) currently used to clean critical surfaces, are limited to the removal of m-sized particles and
some of them can even damage the wafer. Thus, new cleaning approaches are needed to
safely remove sub-m or even smaller particles from critical surfaces.
The laser-assisted particle removal (LAPR) method was developed to clean critical
surfaces in a non-contact fashion without damage (1-2). In the 90s LAPR was predominantly used in its dry variant (dry laser cleaning, DLC) (6), which allows efficient removal of particles by “trampoline” and “hopping” accelerations occurring during transient thermal expansion of the laser-heated Si wafer (2) or the particles themselves (7), respectively. However, eventually DLC was recognized to have very limited opportunities
for cleaning of sub-m and, especially, sub-100nm particle contaminants for the two reasons: i) such small particles require higher particle accelerations for removal, provided
during DLC only at higher laser fluences at the expense of laser damage (e.g., melting) of
the substrate; ii) stronger re-deposition of smaller particles back on the substrate takes
place because of their reduced inertia and the effects of viscous drag and Brownian motion in ambient air (8). In order to extend DLC capabilities to removal of sub-100nm particles, different DLC modifications like ablative removal (9) and laser-induced plasma
(LIP) cleaning (10-11) were proposed to increase contaminant particle acceleration without related substrate damage. Also, to reduce particle re-deposition, DLC employs some
auxiliary tools based on thermophoresis or suction phenomena (12,13).
In contrast, steam laser particle cleaning technique (SLC) (1-2,6,8,14-18) has none of
the abovementioned crucial DLC limitations, exhibiting low size-independent (“universal”) particle cleaning threshold fluences and little or no indications of particle redeposition. In this cleaning process, chemically pure condensed vapor acts as an energy
transfer medium (ETM) between the laser energy and the Si substrate and/or particles,
forming either droplets or a uniform thin layer on the Si substrate depending on wetting
properties of the particular ETM. Then, the substrate of interest is irradiated by the pulsed
laser at a specified time after the ETM dosing pulse, causing the Si surface to undergo a
rapid temperature increase. When during the heating laser pulse the surface temperature
comes close to 0.9Tcr (where Tcr is the critical temperature of the low-boiling ETM) at the
threshold laser fluence FB, the nanometer-thick superheated and pressurized interfacial
ETM sublayer in the vicinity of the Si substrate undergoes sub-ns explosive boiling under
near-critical conditions, pushing the entire ETM layer to lift off from the substrate (19).
The transient multi-MPa pressure of the superheated ETM may be applied underneath the
contaminant particles as a “vapor piston” force (16), while the lifting-off ETM also exerts
a viscous “drag” force on the particles (8,14). Lifting off distances for sub-m ETM layers may approach sub-mm magnitudes (20), enabling the removal of contaminant particles from the surface sufficiently far away to prevent re-deposition irrespective of their
sizes. Importantly, FB, being a function of the ETM, is usually much lower than DLC
thresholds for sub-100nm model particle contaminants (fig.1) (21).
3
10
FDLC (mJ/cm2)
2
FCL(L,R) (J/cm )
Fmelt(Si, 248 nm)=0.55 J/cm2
FSLC(IPA, 248 nm)=0.22 J/cm
2
2
10
Rmin>0.03 m
1
10 -2
10
-1
10
R (m)
0
10
0.25
IPA FSLC
0.20
K(0.1m)=
1.1±0.2 K(0.25m)=
0.15±0.03
0.15
0.10
PS 0.1 m
PS 0.25 m
PS 0.55 m
0.05
0.00
K(0.55m)=
0.04±0.01
0.0
0.5
1.0
1.5
2.0
2.5
L (m)
Fig.1 (left). Experimental FDLC(R) curve for PS particles after this work (stars) and work
(21) (circles, triangles). The Si melting threshold, Fmelt=0.55 J/cm2 (22), is shown by the
upper horizontal line, while the vertical line shows the minimum radius of particles, Rmin,
which can be cleaned by DLC without melting of the Si substrate. The SLC threshold
with IPA ETM, FSLC=0.22 J/cm2, is given by the lower horizontal line for comparison.
Fig.2 (right). Cleaning thresholds for 0.1, 0.25 and 0.55-m PS particles as a function of
IPA ETM layer thickness L.
Despite of a number of successful SLC applications for removal of sub-100nm particles, its microscopic mechanisms have not yet been clarified depending on various aspects of the particle-ETM-substrate interaction. Deposition of a thin ETM layer plays an
important role during SLC as particle removal mechanisms may switch between DLC,
SLC or combined ones depending on wetting of particle contaminants by particular
ETMs (23) and their absolute thickness (8), while cleaning efficiencies may also change
due to reduction or enhancement of particle-substrate adhesion strength at presence of
specific ETMs (11). Therefore, comprehensive enlightening microscopic studies of the
ETM effects on particle-substrate coupling and removal are very promising for development of the SLC technique and its applications. In this work we report preliminary results
on visualization of microscopic interactions between sub-m inorganic and organic particles deposited on Si wafers and dosed with polar or non-polar ETMs of variable thickness or total ETM amount. Comparison of the visualization data with laser cleaning results is used to discuss the underlying particle removal mechanisms.
EXPERIMENTAL
Monodisperse polystyrene (PS) particles of radii, R, of 0.1, 0.25 and 0.55 μm (SurfCalTM grade, density PS1.05 g/cm3, rsd(R)1%, Duke Scientific Co.) were deposited on
commercial 0.25-mm thick Si(100) wafer samples with a nanometer-thick native oxide
layer from a suspension of PS particles in a water/ethanol mixture maintained at 55oC
using an airbrush (9,19). Typical particle densities and average aggregation numbers for
PS particles were about 104-105 cm-2 and 3-4 particles/cluster, respectively. Similarly,
small clusters of alumina particles of radii R=0.3-1 m (Buehler micro-polish) were deposited in the same way on similar Si wafers.
An ETM dosing system described elsewhere (19,20) was used to deposit ETM layers
on the Si substrates mounted on a three dimensional stage. ETM dosing was achieved by
using pressurized N2 gas with a triggered valve connected to a bubbler immersed in a
flask with pre-heated ETM. The vapor from the flask was directed on the Si surface using
a heated nozzle attached to the flask. The optimum distance between the doser nozzle and
Si substrate for uniform deposition of a thin 2-propanol (iso-propyl alcohol, IPA) layer or
layer of m-sized water droplets was 5 cm. The dosing conditions for these experiments
were N2 pressure of 0.7 bar and ETM temperatures of 40oC and 44oC for IPA and water,
respectively. The thickness of IPA layers was measured using optical interference of a
HeNe laser in the layers at an angle of 30o from the normal to the Si surface. The reflected beam was captured by a photodiode and its temporal interference fringes were used to
calculate the IPA layer thickness. The linear relationship between layer thickness, L, and
ETM (IPA, water) dosing time, tdose, has been established (19,20). In the case of water,
the deposited liquid layer consists of separate droplets as the native silicon oxide surface
film on the Si wafer has dewetting properties relative to water with an expected contact
angle of 20-45o (24). Deposition, laser removal and natural drying of a water layer were
monitored in real time by observing the optical reflectance/scattering of a HeNe probe
laser focused on the center of the irradiated area. Nearly Gaussian transversal distributions of water on the Si wafer were obtained at different tdose from the HeNe-laser scattering measurements of drying times across the dosing area. Absolute measurements of deposited water mass performed with CaSO4 absorbent (drierite) exhibit its linear increase
with increasing tdose, within the range of 0.1-1.0 s at the overall deposition rate of 0.007
g/s.
Microscopic visualization of transient ETM layers on the particle contaminated Si wafer surfaces were performed at magnification of 400 using Mitutoyo WH microscope
equipped with a digital camera (Olympus 3030). Subsequent frames of dosed Si surfaces
with and without particles were taken at a rate 30 frames/s and then these movies were
analyzed using Windows Movie Maker graphics software to track spatiotemporal evolution of separate water droplets, particles or IPA layer frame by frame.
A 248-nm, 20-ns KrF excimer laser (Lambda Physik LPX 210) beam was focused by
a cylindrical lens onto the Si wafer with pre-deposited particles and ETM (8,9), providing
horizontal rectangular and vertical Gaussian laser fluence distributions on the wafer surface. The fluence magnitude was varied using color filters and measured by splitting off a
part of the beam to a pyroelectric detector. The excimer laser was fired 0.06 s after the
end of each liquid deposition step, allowing for the dosing jet to propagate between the
nozzle and the Si substrate surface for the 0.04 s delay, and was triggered, as well as the
gas valve, manually in a single-shot mode using a pulse generator.
LAPR experiments were performed at various tdose under ambient conditions. Cleaning thresholds, FCL, were measured by examining the laser-irradiated spots on contaminated Si samples illuminated by a while-light optical fiber light source using dark-field
optical microscopy. The white-light optical scattering from contaminant particles was
used to distinguish visibly sharp boundaries of each laser-cleaned spot from the surrounding un-irradiated, uncleaned area on these wafers.
RESULTS AND DISCUSSION
PS particles with IPA ETM
Optical microscopy visualization shows that IPA completely wets oxidized Si wafer
surfaces and stationary PS particles of different radii R, forming a homogeneous liquid
layer (with small elevations at particle positions at LR). As a result, in laser cleaning
experiments with IPA ETM for very small IPA layer thickness, 0LR, the cleaning
threshold FCL(R,L) is essentially the same as the DLC threshold FDLC(R), i.e., FCL(R,L) at
L=0, and equals 0.10 J/cm2 and 0.05 J/cm2 for 0.25- and 0.55-m PS particles (fig.2), respectively. At LR FCL(R,L) increases linearly vs. L above the DLC threshold with a
slope K(R)R-2 (see below). This tendency in FCL(R,L) may be related to quasi-dry laser
cleaning (QDLC), when the major removal forces are the DLC “trampoline” and “hopping” ones (2,7,9), while the ETM wetting these particles serves to retard their removal
by a viscous drag (Stokes) force. For 0.1-m particles the initial gradual increase in the
cleaning threshold with the increasing L from the DLC threshold of 0.14 J/cm2 was not
observed experimentally, as FCL(R,L) increases rapidly to a constant value of 0.220.04
J/cm2. This value corresponds to the “universal” SLC threshold, FSLC, in Fig.2 for the
248-nm laser wavelength and IPA ETM and is pretty consistent with the IPA explosive
boiling and lift-off threshold, FB(IPA)=0.170.02 J/cm2, measured in photoacoustic and
plume transmission studies (19,20).
For quantitative description of the thin IPA layer effect on QDLC thresholds,
FQDLC<FSLC, damping of initial “trampoline/hopping” lift-off velocities, V(FQDLC(R,L)),
of the PS particles at the Si/IPA interface due to the Stokes force in IPA, fS(R)=
6IPAR(dZ/dt), has been considered. The simple expression of the second Newton’s law,
d2Z/dt2=fS(R)/(4PSR3/3), for 1D PS particle motion in the viscous IPA environment
along a normal, Z, to the Si substrate surface has been solved for the clean area boundary,
corresponding to the laser fluence of FQDLC(R,L), and the initial conditions Z(t=0)=0,
dZ/dt(t=0)=V(FQDLC(R,L)). It was also assumed that the particle lift-off velocities at the
Si/air (in the DLC case) or IPA/air (in the QDLC case) interface should be equal to nonzero value V(FDLC(R)) to prevent recontamination of the Si substrate due to Brownian
motion of the removed PS particles in air (8). Then, the initial lift-off velocity,
V(FQDLC(R,L)), relaxes due to the Stokes force in the IPA layer of the thickness L as
  t ( L)  V ( FQDLC ( R, L)) τ IPA ( R)  ( L  R)
 
, [1]
V ( FDLC ( R))  V ( FQDLC ( R, L)) exp 
τ IPA ( R)
 τ IPA ( R) 
where the (L-R) term corresponds to the “effective” IPA layer thickness accounting for
the PS particle center-of-mass position above the substrate surface, IPA(R)=2PSR2/9IPA
is the characteristic lift-off velocity relaxation time and IPA(293 K)2.410-3 Pas is the
IPA viscosity (25). According to Eq.(1), the initial lift-off velocity V(FQDLC(R,L)) at the
Si/IPA interface must be higher than the intermediate lift-off velocity V(FDLC(R)) at the
IPA/air interface by (L-R)/IPA(R) term to account for the Stokes deceleration in IPA. The
expression for V(FQDLC(R,L))=V(FDLC(R))+(9IPA/2PSR2)(L-R), obtained by substituting IPA(R) in the equation above, and the corresponding experimental linear fits in
fig.2, FQDLC(R,L)=FDLC(R)+K(R)(L-R), have the same functional form for LR. This fact
is consistent with theoretical predictions that lift-off particle velocity should be linearly
proportional to laser fluence for the “trampoline” and “hopping” DLC mechanisms
(2,7,9). The slopes of the FCL(R,L) curves in fig.2, K(0.55 m)=0.040.01 and K(0.25
m)=0.150.03, at L>R exhibit R-2 dependence predicted for V(FQDLC(R,L)). For 0.1m particles a slope K(0.1 m)=1.10.2 was predicted for 0.1 mL0.2 m and
FFSLC, calculated as an average of K(0.25m)[(0.25m)/(0.1m)]2=0.90.2 and
K(0.55m)][(0.55m)/(0.1m)]2=1.20.3, in reasonable agreement with the experimental data presented for these particles in fig.2. Importantly, the K(R) dependence
shows that at F<FSLC viscous drag deceleration of lifting-off particles by the stationary
ETM increases stronger for smaller particles. Conversely, at FFSLC the lifting off ETM
layer will produce higher viscous “drag” accelerations for smaller stationary particles adhered to the substrate. This means that the SLC “drag” mechanism may be even more
effective for smaller particles and this suggestion is supported by experimental and theoretical results of previous studies (15,18).
The SLC “drag” effect on particle contaminants can be described as follows. First,
one can consider lift-off of the entire IPA liquid layer taking place after explosive boiling
in its interfacial sublayer of thickness, Ldep(IPAmin)1/2, of several nanometers (for IPA
being the IPA thermal diffusivity), heated during a cleaning laser pulse by thermal conduction from the hot Si substrate until F-dependent onset of IPA explosive boiling at min
(19,20). The corresponding lift-off velocity, Ulift, reads (19,20)
U lift  Cl
V ( F )  V0 Ldep ( F )
,
V0
L
[2]
where Cl and V0 are the IPA sound velocity and molar volume under ambient conditions,
V(F) is the fluence dependent molar volume of the explosively boiling IPA vapor/droplet
mixture, steeply increasing at FFB (19,20). The lifting off liquid ETM layer produces a
steady-state flow around a stationary spherical particle which can be characterized by the
Reynolds number, Re=IPAlUlift/IPA, where IPA is the IPA mass density and l is a typical length in the flow problem under the study (in the case lR). Substituting Ulift from
Eq.(2) to the expression for the Reynolds number, one can evaluate ReR/L for characteristic parameters of low-boiling liquids predominantly used in SLC [103 kg/m3,
10-3 Pas and Cl103 m/s, Ldep10-9 m, and (V(F)-V0)/V01-2 in the temperature range
of 0.92TcritTTcrit (26)]. At low Reynolds numbers, say Re1 or less (R/L1), one can
ignore the inertial momentum transfer from a IPA flow moving at the velocity Ulift to the
particle in comparison to the surface shear (fluid “friction”) and the cleaning force is the
Stokes (“drag”) one, fS(R)=6IPARUlift. This force acts on the spherical particle adhered
to a flat solid substrate by an attractive van der Waals force, fadh(R)=ARx, where the constant “A” accounts for the strength of this interaction (Hamaker constant) and the characteristic particle-substrate distance, z0, while the parameter “x“ changes from 2/3 to 1 at
transition from elastic to plastic or none particle deformation (16). Therefore, the requirement for SLC particle removal, fadh(R)fS(R), at R/L1 reads
AR x  6IPACl
V ( F )  V0 Ldep ( F )
R.
V0
L
[3]
It shows that for non-deformable or plastically-deformed particles (x=1) SLC cleaning
threshold FSLC is R-independent, i.e. “universal” (15), but may be increased by increasing L up to 4 m or using the increasing and decreasing fluence dependences of V and
Ldep, respectively, in the Eq.(3). In contrast, for elastically-deformed particles with x=2/3
SLC threshold FSLC is size-dependent, but may be kept constant over a broad R range
adjusting R/L ratio.
PS particles with water ETM
Our visualization studies show that water, as a polar ETM, when deposited onto Si
surfaces contaminated with PS particles, forms separate droplets instead of a uniform
layer, while PS particles inside each water droplet are completely detached from the Si
surface floating toward droplet boundaries during the initial period about 0.03 s after the
dosing pulse or forming particle clusters at much longer times. This behavior of water
exhibits its strong dewetting properties relative to both oxidized Si substrates and PS particles. As a result, FCL(R)FDLC(R) over the entire tdose range (fig.3), indicating that the
cleaning mechanism involved is presumably DLC on dry spots between water droplets or
in shallow water near droplets boundaries.
Alumina particles with water and IPA ETMs
According to microscopic visualization data, both water and IPA ETM are wetting
ionic alumina particles which do not move after ETM deposition, while act as nucleation
FDLC
0.2 m
0.1
0.5 m
1.1 m
0.0
0.0
0.3
tdose
0.6
(s)
0.3
FDLC
FCL (J/cm2)
FB
0.2
2
FCL (J/cm )
centers for water droplets. In a particular case of, nearly complete cleaning of alumina
particles from the water-dosed Si wafer has been observed with the “universal” cleaning
threshold, FCL(R1 μm)=0.190.01 J/cm2, coinciding with the explosive boiling and liftoff threshold of water, FB(water)=0.200.04 J/cm2 (20), at all tdose used (fig.4). This may
mean that good wetting of ionic alumina particles by water provides effective coupling of
SLC “vapor piston” and/or Stokes forces to these particles. Surprisingly, when IPA ETM
was applied to the alumina particles, no significant cleaning was observed at laser fluences up to 0.25 J/cm2, i.e., well above FB(IPA)=0.170.02 J/cm2, and this issue is requiring further detailed investigation.
FB
0.2
0.1
0.0
0.9
0.0
0.3
0.6
0.9
tdose (s)
Fig.3 (left). Cleaning thresholds for 0.1, 0.25 and 0.55-m PS particles as a function of
water ETM dosing time tdose.
Fig.4 (right). Cleaning thresholds for alumina particles (R=0.3-1 m) vs. water ETM dosing time tdose. Explosive boiling threshold of water on Si surfaces, FB, is given for comparison with the SLC thresholds.
CONCLUSIONS
Our experimental results have shown 90% efficiencies for steam laser cleaning of 0.11 m model contaminant polystyrene and alumina particles from surfaces of commercial
Si wafers, when appropriate non-polar (2-propanol) and polar (water) ETMs capable of
wetting the corresponding particles have been applied. The viscous “drag” force from
wetting ETM liquids was shown to be significant during SLC, especially for smaller particles, providing size-dependent deceleration/acceleration of the particles at laser fluences
below/above the explosive boiling thresholds of the corresponding ETMs. In case of nonwetting ETMs only DLC “trampoline” and/or “hopping” mechanisms were effective for
quasi-dry removal of particles.
ACKNOWLEDGEMENTS
This work was supported by NSF CTS Grant # 0218024 under technical direction of
Dr. Triantafillos J. Mountziaris.
REFERENCES
1. K. Imen, S.J. Lee and S.D. Allen, Appl. Phys. Lett., 58, 203 (1991); S.J. Lee, K. Imen
and S.D. Allen, Microelectron. Eng., 20, 145 (1993).
2. A.C. Tam, W.P. Leung, W. Zapka and W. Ziemlich, J. Appl. Phys., 71, 3515 (1992).
3. International Technology Roadmap for Semiconductors, International SEMATECH
(http://public.itrs.net/) (2003).
4. G.E. Moore, Electronics, 38, 8 (1965).
5. M. Itano, I. Kawanabe and F. Kern, Jr., in Particles on Surfaces 3: Detection, Adhesion and Removal, K.L. Mittal, Editor, p.35, Plenum press, New York (1991); A.A.
Busnaina, I. Kashkaoush, F. Kern Jr. and R. Kunesh, ibid., p.217; O.V. Abramov,
High-Intensity Ultrasonics: Theory and Industrial Applications, Gordon & Breach,
Newark (1998); M.H. Jones and S.H. Jones, Wet-chemical Etching and Cleaning of
Silicon, Virginia Semiconductor, Inc., Fredericksburg (2003).
6. B. Luk’yanchuk (Ed.), Laser Cleaning, World Scientific, Singapore (2002).
7. J.D. Kelley, M.I. Stuff, F.E. Hovis, G.J. Linford, SPIE Proc., 1415, 211 (1991).
8. S.I. Kudryashov and S.D. Allen, SPIE Proc., 5713 (2005, in print).
9. S.I. Kudryashov and S.D. Allen, J. Appl. Phys., 92, 5159 (2002).
10. J.M. Lee and K.G. Watkins, J. Appl. Phys., 89, 6496 (2001); C. Cetinkaya, R.
Vanderwood and M. Rowell, J. Adhesion. Sci. Technol., 16, 1201
(2002);
I.
Varghese and C. Cetinkaya, J. Adhesion. Sci. Technol., 18, 795 (2004).
11. Z.B. Wang, M.H. Hong, B.S. Luk’yanchuk, Y. Lin, Q.F. Wang and T.C. Chong, J.
Appl. Phys. 96, 6845 (2004).
12. A.C. Engelsberg, SPIE Proc., 3274, 100 (1998).
13. X. Wu, E. Sacher, and M. Meunier, J. Adh. Sci. Technol., 75, 341 (2001); F. Zheng,
Adv. Colloid Interface Sci., 97, 255 (2002).
14. S.I. Kudryashov and S.D. Allen, Method and apparatus for removing minute particle(s) from surface, Provisional US patent (2002).
15. M. Mosbacher,V. Dobler, J. Boneberg and P. Leiderer, Appl. Phys. A, 70, 669 (2000).
16. Y.F. Lu, W.D. Song, M.H. Hong, Y.W. Cheng and T.C. Chong, Tribology Intern.,
33, 329 (2000); Y.W. Zheng, B.S. Luk’yanchuk, Y.F. Lu, W.D. Song, Z.H. Mai, J.
Appl. Phys., 90, 2135 (2001).
17. F. Lang, M. Mosbacher and P. Leiderer, Appl. Phys. A, 77, 117 (2003).
18. K.M. Smith, M.Y. Hussaini, L.D. Gelb and S.D. Allen, Appl. Phys. A, 77, 877 (2003).
19. S.I. Kudryashov and S.D. Allen, J. Appl. Phys., 95, 5820 (2004).
20. S.I. Kudryashov, S.D. Allen, J. Appl. Phys., 93, 4306 (2003); S.I. Kudryashov, S.D.
Allen, Appl. Phys. A, 79, 1737 (2004).
21. M. Mosbacher, M. Bertsch, H.-J. Muenzer, V. Dobler, B.-U. Runge, D. Baeuerle, J.
Boneberg and P. Leiderer, SPIE Proc., 4426, 308 (2002).
22. S.I. Kudryashov and S.D. Allen, J. Appl. Phys., 92, 5627 (2002).
23. P. Neves, M. Arronte and A.M. Botelho do Rego, Appl. Phys. A, 74, 191 (2002).
24. W.H. Lawnik, U.D. Goepel, A.K. Klauk, and G.H. Findenegg, Langmuir, 11, 3075
(1995) and references therein.
25. I. Grigor’ev and E. Meilikhova, Fizicheskie Velichini, Energoatomizdat, Moscow
(1991).
26. V.P. Skripov, E.N. Sinitsyn, P.A. Pavlov, G.V. Ermakov, G.N. Muratov, N.V. Bulanov, and V.G. Baidakov, Thermophysical Properties of Liquids in the Metastable State,
Gordon and Breach, New York (1988).