The surface tension and contact angle ofmolten CdTe were measured by the sessile drop technique. .... etched with 2% bromineâmethanol solution, fol-.
Journal of Crystal Growth 100 (1990) 51—57 North-Holland
51
SURFACE TENSION AND CONTACT ANGLE OF MOLTEN SEMICONDUCTOR COMPOUNDS I. Cadmium telluride Rajaram SHEYT’Y, Raghuraman BALASUBRAMANIAN and William R. WILCOX Center for Advanced Materials Processing and Department of Chemical Engineering, Clarkson Universitt’, Potsdam, New York 13676, USA
Received 22 June 1989; manuscript received in final form 1 October 1989
The surface tension and contact angle of molten CdTe were measured by the sessile drop technique. The dependence of surface tension and contact angle on temperature and on the deviations from stoichiometry was determined. A computer program was used to calculate the properties from the sessile drop profile. The wetting behaviour of the melt was studied on the following surfaces: quartz, HF-etched quartz, sandblasted quartz, carbon-coated quartz and pyrolytic boron nitride. The surface tension of molten CdTe decreased with increasing temperature, but increased slightly with excess Cd. The degree of wetting increased in the following order: pyrolytic boron nitride, carbon-coated quartz, sandblasted quartz, HF-etched quartz and plain quartz. The contact angle decreased with increasing temperature and seemed to increase slightly with increasing Cd.
1. Introduction Cadmium telluride is a binary TI—VT semiconductor compound with application to electrooptic devices [1]. It is an important nuclear-detector material, particularly for gamma-ray spectrometry [4]. It is used as a substrate material for HgCdTe in infra-red detector applications [2,3]. A variety of techniques have been tried for the growth of CdTe, though the production of large, defect-free single crystal CdTe has met only limited success, The float zone process conducted at low gravity offers a number of advantages that could significantly enhance both crystal size and quality. Grayity-related problems tend to be effectively eliminated in a microgravity environment, and surface tension related effects assume prominence, Marangoni convection results when a surface tension gradient occurs along a free liquid surface, such as the molten zone in a float zone process or the free melt surface in the horizontal Bridgman technique. Since surface tension depends both on temperature and composition of the liquid, a
estimating Marangoni convection. In the float zone process, the molten zone is held by surface tension forces in opposition to gravity. Thus the surface tension of molten CdTe will be useful in estimating limits on the geometry (length and diameter) and stability of floating zones. The contact angle that a liquid forms on a solid surface indicates the degree of wetting of that surface by the liquid. The contact angle depends on the solid-melt and solid-vapor surface energies which are characteristic of the material. It also depends on the structural features of the solid surface like surface roughness and heterogeneity. The surfaces that would be of interest for the study of wetting behaviour of the CdTe melt are those that can be used as ampoule or crucible material for crystal growth of CdTe. In this study, the wetting behaviour of molten CdTe was investigated on clear fused quartz, HF-etched quartz, sandblasted quartz, carbon-coated quartz and pyrolytic boron nitride. The melting point of stoichiometric CdTe is 1092°C [4]. The vapor pressures of Cd and Te2
knowledge of surface tension and its dependence on temperature and composition is useful in
over a stoichiometric CdTe melt at 1092°C have been reported to be 0.8 atm and 0.4 atm respec-
0022-0248/90/$03.50 © Elsevier Science Publishers B.V. (North-Holland)
52
R. Shette ci cii.
/
Sur/ace tension and contact angle of tnolten semiconduit or compounds. I
tively [51.In order to suppress evaporation of CdTe and avoid tellurium precipitates, excess Cd is usually used in the growth ampoule.
Vapor
I
Though there are several methods for the measurement of surface tension and contact angle [6—8],the sessile drop method was adopted in this study as the most suitable technique in view of the difficulties mentioned above. The evaporation of CdTe was controlled by using Cd overpressure. The surface tension and contact angle were measured from the melting point of 1092°Cto about
V
Solid
a
d
dz
try was investigated by varying the amount of excess 1150°C. CdThe usedeffect in theof ampoule. deviation from stoichiome-
,o~
X ~Apex(R
1
=
R2)
To our knowledge, no attempt. with the excep-
b
madeoftoCdTe. molten tion the fill effort theOffisterov need led for by surface etOffisterov al. measured property [9], has only data been the of surface tension of molten CdTe, but did not report any quantitative results.
A,
Fig. 1. Sessile drop.
2. Experimental Both surface tension and contact angle can be determined from a single experiment by the sessile drop technique. The contact angle 0 (fig. Ia), is determined by the three interfacial energies: the liquid—vapor interfacial energy ~ (or the surface tension), the solid—liquid interfacial energy and the solid—vapor interfacial energy y~. The measurement of surface tension by the sessile drop technique is based on the application of the Laplace equation of capillarity [10]: “~
+
~_) =~P. 2
where y is the surface tension, R~and R-, are the principal radii of curvature of a liquid drop placed on a horizontal surface (see fig. Ib). and ~P is the pressure difference across the curved liquid—vapor interface. The Laplace equation can be transformed into a differential equation form (see, for example, ref. [11]), by simple geometrical considerations (as indicated in fig. 1b). Several schemes have been developed over the years for numerical solution of
this equation. Rotenberg et al. [11] developed a rigorous, user-oriented scheme to determine liquid—fluid interfacial tensions as well as contact angles from the shapes of axisymmetric menisci, i.e.. from sessile and pendant drops. A recent and improved version of this computer software called Axisymmetric Drop Shape Analysis. Profile (ADSAP) * was used in this work to calculate both surface tension and contact angle. The input data required for this program are the gravity, the density difference between the liquid and the vapor phases. the coordinate points on the sessile drop profile, and the distance between the drop base and the parallel reference axis. 2.1. Apparatus
The apparatus consisted of: (a) A cylindrical quartz ampoule (50.8 mm (2 inches) long and 28 mm ID) with quartz optical *
Acquired from Professor A W. Neumann of the Department of Mechanical Engineering, University of Toronto, Toronto. Canada.
R. Shetty et aL
/
Surface tension and contact angle of molten semiconductor compounds. /
Clear fused quartz
r— tu
Clear fused quartz tubing : 28s32 mm dia. and 2in. long
/
ing . 1 xl .2 mm dia. Sessile drop
53
monochrome microvideo camera and a NEC Multisync II color monitor.
Substrate
2.2. Procedure a
/ Ground
and
Details of the apparatus and procedure are given elsewhere [12 13]. A piece of CdTe of
polished optical......J
grade fused quartz discs
99.9999% purity (Il—VI Inc.) weighing approximately 1.6 g was cleaned with deionized water and etched with 2% bromine—methanol solution, fol-
QUARTZ AMPOULE
FILTER AND EXHAUST
ENCLOSURE
~j
TH~fOCOUPLES
TEMPERATURE AMPOULE ~________
I
CONTROLLERS SOURCE
under a vacuum of i0~ Torr. The sealed ampoule
-
L L
lowed by a wash with pure methanol. It was then washed with concentrated HCI and rinsed thoroughly with deionized water, then dried in air with methanol. A small amount of 99.9999% pure Cd (Johnson Matthey Inc.) was washed several times with deionized water and dried with methanol. After loading CdTe and Cd onto the desired surface in the cleaned and dried ampoule (fig. 2), the ampoule was flushed 6 or 7 times with a 10% H2—90% He mixture. The ampoule was then sealed
SPUT.TUBE FURNACE
Fig. 2. Schematic of apparatus.
flats at its ends, containing the substrate under study. The substrate was in the form of a rectangular piece, 50.8 mm X 25.4 mm (2 inches xl inch) and about 1 mm thick (fig. 2). (One experiment on carbon-coated quartz was performed in a bigger ampoule, 127 mm (5 inches) long, due to difficulty in keeping the carbon coating from oxidizing while sealing the optical flats to the ampoule.) (b) A three-zone, split-tube horizontal furnace, with temperature controllers, to melt the material and maintain the drop at the desired temperature. (c) An optical arrangement consisting of a Nikon FM-2 camera fitted with a teleconverter and a magnifying lens, and a high wattage lamp to iiluminate the drop. A Kodak 38A gelatin filter was used to filter the infrared radiation from the furnace. A schematic of the apparatus is shown in fig. 2. The digitization equipment consisted of a Jandel Scientific Image Analysis System with a Circon
was placed quartz boat holder inside the furnace and onthea substrate in the ampoule aligned horizontal using the reference line on was the camera. A drop of molten CdTe was formed by raising by furnace temperature to the melting point. Photographs of the drop were taken after allowing time for thermal equilibration (60 to 90 mm) at 5.6°C (10°F) intervals. The photographic negatives were digitized to obtain 300 to 400 points along the profile of each drop, whose base diameters were in the neighborhood of 10 to 12 mm. The coordinate points thus obtained along with the value of gravity, the density difference between the melt and the vapor, and the vertical coordinate of the drop base formed the input data for the program ADSAP, to compute surface tension and contact angle. Experiments were conducted on quartz microscope slides, HF-etched quartz slides, sandblasted quartz slides, carbon-coated quartz slides and on pyrolytic boron nitride. In experiments on quartz microscope slides the amount of Cd used in the ampoule was varied with a view to investigate the effect of small composition changes (stoichiometnc deviations) on the surface properties of the melt. Some experiments on quartz and boron
54
R. Shettv et al.
/
Surface tension and contact angle of molten semiconductor compounds. I ~ of Cd; 0.0433 25gg :O.O7 0.1047 0.1398 g
Amount
o
6*556
r
L
F
F
OF
F
F
.~‘18O
~
;~ 5.
_—.
00000; FFFF~
~
0
•
-
F
F
*
U
-
F
* °.~ 170 0 IX
a
U
C
iso
a.o~oo~~
1080
1090
1100
1110
1120
1130
1140
1100
1160
Temperature, Vc Fig. 4. Surface tension of molten C’dTe as a Function ut temperature, for different Cd overpressures.
where ‘y is the surface tension the temperature (in °C),and
dyn/cm). T
(in
is
amount of excess Cd in ampoule (g) Fig. 3. Sessile drops of molten CdTe: (a) on quartz at 1096°C, with excess Cd of — 0.1 g: (b) on pyrolytic boron nitride at 1098°C with excess Cd of — 0.07 g.
nitride coated quartz were also conducted earlier using the same apparatus but different digitization procedures and a different computer program [121. Typical drop photographs are shown in fig. 3.
p
=
vapor space in ampoule (cm .
.
.
,
The 95% confidence limits on the coelficient ol Tare —0.21 and —0.11 dyn/cm ‘o C: 95cr confidence limits on the coefficient of p are 930 and 2300 dyn cm2/g. The probabilities that the coeffi-
200
—-~-~-
-~--
---
-
Amount of Cd • 00901 o
3. Results
~.
0.0723 9
00000 .....
F
0 0699
6)190
3. 1. Surface tension
•
The surface tension of molten CdTe subject to different Cd overpressures is shown in fig. 4, as a function of temperature. These results were obtamed from experiments conducted on quartz microscope slide substrates. A regression analysis of the results with surface tension as the dependent variable and temperature and the amount of cxcess Cd in the ampoule as independent variables gave the following model: y
=
0.16T+ I600p
+
350,
~ 0 0 F
F
V
,~
0
F
0 170
0 0
160
‘“‘‘“
1080
~
1090
1100
1110
1120
I,,,~mthmur0
1130
1140
1150
1160
Temperature, C Fig. 5. Surface tension of niolien (‘dTe, from experiments on sandhlasied quartz.
/ Surface
R. Shettv et al.
tension and contact angle of molten semiconductor compounds. /
190
55
90
Amount of Cd: FFFFF 0
0
° ~
I
~ 0.0724 g 0.0609
ccxx:o: 0.0743 g
F
F
F
*****:
Amount of Cd~ 93R151* I 00433
*00 0’
*
•••~• I 0.0725 mXX)~0.1C)47
•
6)
0
2) 0 F
*
*
C• o
*
0
F
F
00
F
~
0
00
*
F
0
*
‘FFF~ 0.1356
~V
F
I) ‘
*5.
• 6~
F
C
*
~
0
o0
F
00
•
0’
F
.
0 6)170 0 10
1070
.~~1
F 0
F •
C 0
0
F 0
F
C
F
160
,~u :l~,l I
1090
~
ilOO
I
I
1110
i120
~
60
~
1130
Temperature,
1140
1150
1160
I::,,,, I
,).).).)OUVLV)
1060
1090
C
=
I
,:~V~’’ I ~
1110
1120
I
I ~
1130
Temperature,
Fig. 6. Surface tension of molten CdTe, from experiments on pyrolytic boron nitride.
cient of T is negative and the coefficient of p is positive are both 99.99%. Figs. 5 and 6 show the surface tension of CdTe obtained from experiments on sandblasted quartz and PBN respectively. A regression analysis combining the surface tension data from all experiments on different surfaces showed very little vanation of surface tension with surface type (variation of the same order as scatter). Most of the scatter in the data can be attributed to photography. The difficulty was in obtaining a sharply defined drop profile and in locating the drop base accurately; these were important to obtain data that accurately reproduced the sessile drop profile. The standard errors in surface tension associated with the experiments were of the order of 0.8 to 2.2 dyn/cm in 180 dyn/cm. At the melting point, surface tension varied from 177 dyn/cm with excess Cd in the ampoule of p 2.3 X i0~ g/cm3 (0.07 g in 31 cm3) to 185 dyn/cm with excess Cd of p 4.5 x l0~ g/cm3 (0.14 g in 31 cm3). The surface tension decreased by about 10 dyn/cm in going from the melting point of -~-1092°Cto1150°C.
,l,,ll,l
1100
1140
Fig. 7. Contact angle of molten CdTe on quartz.
function of temperature. The degree of wetting increased with increasing temperature. The degree of wetting increased for the different surfaces studied in the following order: PBN, carbon-coated quartz, sandblasted quartz, HF-etched quartz and plain quartz. The contact angle at the melting point (with an excess Cd of p 2.3 x 1O~~ g/cm3) was approximately 132° on PBN, 108° on carbon-coated quartz, 105°on sandblasted quartz, 90°on HF-etched quartz, and 83°on plain quartz. That the sandblasted quartz is less wetted than plain quartz seems to agree with the generally held
Amount of Cd
102
0.1129 g 0.0693 g 0.0736 g
FFFFF: iXSX3J: *****:
90
0 *
F~0
0
0
0
F
,
**
F
*
0
0
* ‘~
*
_________________________________ 60 ,,,,,,II 1080 1090
The contact angle of molten CdTe on the different surfaces is shown in fig. 7 to fig. 11, as a
1160
°C
=
3.2. Contact angle
1150
I uuII:I,,lI[VVV,,
1100
1110
I
V~!!!~!I
1120
Temperature,
1130
1140
1150
1160
°C
Fig. 8. Contact angle of molten CdTe on HF-etched quartz.
56
/
R. Shetti et al. 110
Surface tension and contact angie of mo/ten semiconductor compounds. I 140
~-
Amount of Cd Amount of Cd: FFFFF 0.0901 g coo’xo 0 0723 g
F °
cr11 20
cc000
20 ~130
F
Ii)
‘C
* •
* 0
*
1*
61
~
F
*
*
F
F
oc
F
F
F
F
~~120
~
80
*
F -
F
00743 g
F
*
F
00809
F
‘~
F
-
-
00724 g
*****
0 0699 g
~‘~.***
*
FFFFF
cc.
*
•0
*
o
*
70 1080
1090
1100
1110
1120
1130
1140
5C
10 ~ 1090
1160
1150
Temperature, Fig. 9. Contact angle of molten CdTe on sandhlasted quartz.
view that roughening a surface increases the contact angle on that surface. Hydrofluoric acid also
1100
1110
1120
1130
Temperature,
1140
°C
1150
1160
Fig. 11. Contact angle of molten (•dTe on p~roIv1ic boron nitride.
The 95% confidence limits on the coefficient of Tare —0.23 and 0.17 deg/°C: 95% confidence limits on the coefficients of p are — 2400 and
induces roughness on quartz and that may be the reason for the slightly higher contact angle on I-IF-etched quartz compared to plain quartz. Fig. 7 shows the contact angle on quartz with varying amount of excess Cd in the ampoule. A regression analysis gave the following model:
1640 deg cm3/g. The probabilities that the coefficient of T is negative and the coefficient of p is negative are both 99.99%. The standard errors in contact angle measurement due to photograph\ and subsequent digitization ranged from 0,4° to
0=
2.4°.
0.2T—2000p+310,
—
where 0 is the contact angle in degrees. 110
Acknowledgements This research was supported by the Center I’or
Development of Commercial Crystal Growth in
Amount of Cd. ~: 00716
*
Space, the New York State Science and Technology Foundation and Clarkson University. We are thankful to Professor A.W. Neumann of the Uni-
*
26
*
6)toou
*
versity of Toronto for providing a sophisticated
-
*
computer software for use in this work, We are grateful to the Il—VI Inc. for donation of high
* *
purity CdTe and for carbon-coating the quartz microscope slides, and to Dr. Michael Roberts of the Biology Department at Clarkson Universit\ for providing the Image Analysis System facilities.
I
I
90
80
1080
I !I-,HN,]!!,I,W[
1090
1100
I 11111!!] I ,IV![I!I!II,I!
1110
1120
1132
1140
1150
1160
Temperature, °C Fig. 10. Contact angle of molten CdTe on carhon•coated quartz (second experiment with 0.1782 g of Cd in anspoule of 3). volume 78 cm
References [1[ F.V. Wald. Res. Physique App!. 12 l977~272. [2] CT. Elliott. in: Handbook on Semiconductors. V C. Hilsum (North•Holland, Amsterdam. 1981).
01. 4. Fd.
R. Shetty et al.
/ Surface
tension and contact angle of molten semiconductor compounds. /
[31J. Schmitz, H. Walcher and
J. Baars, in: Materials Technologies for Infra-Red Detectors, SPIE Proc. 659 (1986) p.
137.
[41 K. Zanio,
Ed.. Semiconductors and Semimetals, Vol. 13,
Cadmium Telluride (Academic Press, New York. 1978). [5] MR. Lorenz, J. Phys. Chem. Solids 23 (1962) 939. [6] J.F. Padday, in: Surface and Colloid Science, Vol. 1, Ed. E. Matijevic (Wiley—Interscience, New York, 1969) p. 101. [71RE. Johnson and RH. Dettre, in: Surface and Colloid Science, Vol. 2, Ed. E. Matijevic (Wiley—Interscience, New York, 1969) p. 115. [8] A.W. Neumann and R.J. Good, Surface and Colloid Sci• ence, Vol. 11, Eds. R.J. Good and R.R. Stromberg (Plenum. New York, 1979).
[91 A.A.
57
Offisterov. A.V. Vanyukov, V.V. Golubtsov, MN.
Dubrovin and AM. Sokolov, Zavodskaya Laboratoria 40 (1974) 266. [101 N.K. Adam, The Physics and Chemistry of Surfaces (Ox• ford University Press. Oxford, 1946). [111 Y. Rotenberg, L. Boruvka and A.W. Neumann, J. Colloid Interface Sci. 3 (1983) 169. [121 R. Balasubramanian. Surface Properties of Molten Cadmium Telluride, MS Thesis, Clarkson University. Potsdam, NY (1988). [131 R. Shetty, Surface Tension and Contact Angle of Cadmium Telluride and Gallium Arsenide Melts, MS Thesis, Clarkson University, Potsdam. NY (1989).
