RESEARCH ARTICLE | OCTOBER 01 1961

Investigations on SnS 
W. Albers; C. Haas; H. J. Vink; J. D. Wasscher

J. Appl. Phys. 32, 2220–2225 (1961)

https://doi.org/10.1063/1.1777047


View Export
Online Citation

21 May 2025 12:27:18

JOURNAL OF APPLIED PHYSICS SUPPLEMENT TO VOL. 32, NO. 10 OCTOBER. 1961

Investigations on SnS
W. ALBERS, C. HAAS, H. J. VINK, AND J. D. WASSCHER
Philips Research Laboratories, N. V. Philips' Gloeilampenjabrieken, Eindhoven, Netherlands

The p, T, x diagram of the Sn-S system was determined especially in the region of the compound SnS.
The pressure of S2 in equilibrium with SnS and a liquid phase was found to extend over several decades up
to 25-mm Hg at the "Sn-rich" side, whereas at the "S-rich" side the S2 pressures in equilibrium with solid
SnS and a liquid phase lie between 25-mm Hg and 100-mm Hg. It was shown that the existence region of
solid SnS very probably lies entirely at the excess sulfur side. The hole mobility in a plane perpendicular
to the c axis, ",,90 cm2/v sec at room temperature, was proportional to 1'-2. 2 for higher temperatures. The
mobility in the direction of the c axis was about five times smaller. Reversible annealing effects were found
for temperatures above 200°C which could be explained by assuming association of neutral Sn vacancies.
Absorption measurements showed that the edge absorption is due to indirect transitions. The bandgap
was 1.08 ev at 300 0 K and 1.115 ev at 77°K. Interband transitions in the valence band were also found.
The effective charge of the atoms (e*=0.7co) and the effective masses of the holes in the three principal
crystal directions (ma*=mb*=0.20mo; m/""mo) were determined from reflection measurements in the
infrared. From these values and the value for the density of states mass obtained by means of the Seebeck
effect (ma*;:: O.95mo), the number of equivalent maxima of the valence band was found to be at least four.

I. INTRODUCTION connection is the study of the three-phase (solid-liquid-

vapor) equilibrium lines of the system.
I N view of the extensive research carried out on the
semiconducting compounds PbS, PbSe, and PbTe,
it seemed worthwhile to study somewhat more closely
The relation between the composition of the liquid,
the temperature, and the corresponding sulfur pres-
the compounds SnS and SnSe which so far have at- sure along these three-phase lines have been determined
tracted relatively little attention, the investigations of by a method already used for other substances. 4- 7 The
Asanabe and his co-workers1,2 on SnSe being the most T-x projection of this "liquidus" line 8 is shown in Fig. 2,
recent. An additional incentive for the study of SnS and together with some data of other investigators. The
existence of the compounds SnS and SnS2 is clearly

21 May 2025 12:27:18

SnSe is the fact that, where as PbS, PbSe, PbTe, and
SnTe crystallize in the NaCI structure, the compounds indicated. The maximum melting points of these com-
SnS and SnSe have an anisotropic crystal structure. It pounds are 881.5° and ",=,870°C, respectively. At the
was thought that, in view of the great number of cubic Sn-rich side of SnS there is a large region of liquid
semiconductors presently under investigation, it would immiscibility extending down to the quadruple point at
be interesting to study a strongly anisotropic one. 860°C, where solid SnS, a liquid phase having a com-
position close to SnS, a second liquid phase rich in Sn,
II. CRYSTAL STRUCTURE and a vapor phase are in equilibrium.
The relation between the S2 pressure and the tem-
According to the investigations of Hofmann,3 SnS and
perature along the various three-phase lines is shown in
SnSe have an orthorhombic structure that may be
Fig. 3. It is seen that the S2 pressures in equilibrium
described as pseudotetragonal (Fig. 1). Every Sn atom
with SnS and SnS2 at their maximum melting points are
is surrounded by three S atoms as nearest neighbors
25 mm Hg and about 40 atm, respectively. From these
(drawn lines). These three bonds form angles with each
two graphs a plot of PS 2 vs composition of the liquid
other which are nearly 90°. Therefore in general one can
can be constructed for the three-phase equilibria (Fig.
say that the bonding type, apart from its ionic part, is
4). The existence of the liquid immiscibility gap is
covalent of the p3 type.
clearly demonstrated by the horizontal part of the
Many of the semiconductors now under investigation
graph at the Sn-rich side of SnS. From the foregoing, it
show a bonding that, in its covalent part, is of the Sp3
is clear that at the Sn-rich side of solid SnS the pressure
type. This difference in type of bonding is another
of S2 in equilibrium with SnS and a liquid phase extends
reason for studying these substances. Fig. 1 also shows
over several decades up to 25 mm Hg, whereas at the
clearly that SnS has a layer structure with double layers
S-rich side the corresponding S2 pressure extends over a
perpendicular to the c axis.
very limited range, viz. from 25 mm Hg to 100 mm Hg.
III. p, T, x DIAGRAM OF Sn-S The S2 pressure as a function of temperature of
another three-phase equilibrium, viz. solid SnS at the
In order to grow single crystals under controlled S-rich side of its existence region, a solid phase of a
conditions it is necessary to have a knowledge of the
p, T, x diagram of the system Sn-S, particularly in the
4 J. Bloem, Philips Research Repts. 11, 273 (1956).
5 J. Bloem and F. A. Kroger, Z. physik. Chem. (Frankfurt) 7
region of the compound SnS. Of special interest in this 1 (1956). '
6 J. van den Boomgaard and K. Schol, Philips Research Repts.
1 S. Asanabe and A. Okazaki, Proc. Phys. Soc. (London) A73, 12, 127 (1957).
824 (1959). 7 D. de Nobel, Philips Research Repts. 14,357,430 (1959).
2 S. Asanabe, J. Phys. Soc. Japan 14, 281 (1959). 8 W. Albers and K. Schol, Philips Research Repts. (to be
, W. Hofmann, Z. Krist. 92, 161 (1935). published).
2220
 I N V EST I GAT ION SON SnS 2221

higher S content (SnS2), and vapor, is indicated by the 1300

dotted line in Fig. 3 but has not been determined ex- ...................
perimentally. The highest S2 pressure (100 mm Hg) of 1200 ."

~r~Sa
s"s
this three-phase equilibrium is given by the quadruple r
point at 740°C in Fig. 3, where a liquid phase is also
present, this point being the lowest temperature (and
(OC) 1100

t 1000
1,:::::
u
highest S2 pressure) where the three-phase equilibrium, 8et.5 I I : I ·two melts? J
900
solid SnS (at the S-rich side of its existence region),
liquid, and vapor is possible. !rszot~~~.:~~
I. \
The limited range of equilibrium S2 pressures at the II
.,.- \.
sulfur-rich side of the three-phase equilibria of SnS is 700
+ 740±3.5'C\
II
\
,,
unusual compared to substances like PbS,4,5 CdTe 7 and I I
I I
600 I
CdS.9 II
I I I
II
With regard to the solidus line of the three-phase 500 I I
II
equilibria, i.e., the composition of solid SnS as a function II
I I
of temperature and corresponding S2 pressure, it must II
II
be borne in mind that the existence region of solid com- I I
II
300 I,
pounds is often very small. 5, 7 ,10 It has thus often proved
to be impossible to determine the compositions of the 'XrooM..O:-'<-x---
Z32"C
I,"
I,
boundaries of the existence region of the solid by purely II
"
100
chemical means. However, if measurements of the Hall I I
II
I I
effect can be interpreted in terms of excess or defect
concentrations of the constituents, it offers a very sensi- 9(} 100
tive, though indirect, method of determining the com- S

position of the solid along the three-phase equilibria. A FIG. 2. T-x projection of the three-phase equilibria in the
necessary condition of the ability to do this is the phase diagram of the system Sn-S, XL giving the composition of

21 May 2025 12:27:18

the liquid phase. The composition of the vapor phase has been
certainty that no chemical impurities are present in a omitted. Our data are indicated by the circles.
concentration sufficiently high to make the interpre-
tation of the Hall measurements in terms of excess of
one of the constituents doubtful. Therefore, in order to
try to determine the solidus line, pure SnS is needed.
This was obtained by a double distillation at 960°C in
vacuum of SnS prepared from the purest Sn and S
available. Spectrochemical analysis of this purified SnS
c showed that the concentration of all detectable elements
was below the limits of detectabilitywith the exception of
r Pb, which was still present at a concentration of 7.1016
at.jcms.
Sometimes the distillation itself gave crystals of
sufficient size to carry out electrical or optical measure-
ments; sometimes it was necessary to zone melt the
material in a horizontal boat to get single crystal bars
- Cleavage plane out of which samples could be cut.
Small crystals of pure SnS obtained in this way were
brought into equilibrium at various temperatures with
the pressure of S2 corresponding to the three-phase
equilibrium at those temperatures. The S2 pressures
needed were established either by the two-temperatures
_b furnace method,5-8 where the sulfur pressure in ques-
OSnin lA- ®SninJA- tion must be known, or by the elegant method of
Brebrick and Allgaier,lO where exact knowledge of the
O Sin 1/4 ~ Sin J/+ sulphur pressures is not necessary. Hall measurements
FIG. L Crystal structure of SnS according to Hofmann, drawn carried out on those crystals after quenching from the
lines indicating bonds that are at least partly covalent. temperature of equilibrium to room temperature showed
that the two methods mentioned above did not give
9 F. A. Kroger, H. J. Vink, and J. van den Boomgaard, Z,
physik. Chern. (Leipzig) 203, 1 (1954).
results entirely consistent with each other, especially
10 R. F. Brebrick and R. S. Allgaier, J. Chern. Phys. 32, 1826 with regard to the dependence of the carrier concentra-
(1960). tion on the S2 pressure. This question is still under study.
2222 ALBERS, HAAS, VINK, AND WASSCHER

-1200 1000 800 660

TEMPERATURE ('Ie)
400 300 200 mentioned, this result must be considered with some
reserve.

IV. ELECTRICAL PROPERTIES

Hall-effect measurements have been carried out on
a great number of crystals. Resistance measurements
have also been carried out.
The Hall mobility perpendicular to the c axis f.i.l =
(8/37r)R ll (jl, where (jl=((ja(jb)!, was measured as a
function of temperature from 600 0 K down to 20 0 K
(Fig. 5). At high temperatures the hole mobility is
proportional to T-2.2, indicating lattice scattering. The
low-temperature branch could be explained by assuming
impurity scattering of the Conwell-Weisskopf type and
a concentration of charged centers of the order of 3 .1017

I I ! III1 I
I
1000 ~~m~lfl'!§~3~Sm
t== 22
_r-
l_lLp'-o-+-_+_t-t--+-t-t-t-'<\.~ I--t-- o!
FIG. 3. The S2 pressure, as a function of temperature, over
various three-phase equilibria in the system Sn-S.

However, for both methods p-type conductivity was

found with a carrier concentration of about 1--3.1018

21 May 2025 12:27:18

holes per cmS at both boundaries of the existence region
of SnS. In view of this result, it seems reasonable to
assume that the existence region of solid SnS at the
temperatures concerned lies entirely at the excess sulfur
side. However, with regard to the inconsistencies just

2
I
I tl.tomelts?
-,,"---1---~~
5
~ 2
(mmHg) 10" fJ I I \
FIG. 5. The Hall mobility of the holes in SnS perpendicular

t 2
5 'fih I
1r-%B5;/+L;/+G 1
I
\ to the C axis [VtaJtb)! ] as a function of temperature.

,I L .J at./cm3• This concentration agrees well with the number

5 Ls+G~..sSns +L+
I' 'I 'I
1)"'l of acceptor levels (4.1017 at./ cm3) found for this crystal

5 J +
ss!.s SSn~;/ +Lz~G-
by Hall measurements.
With regard to the anisotropy of the crystal it was
I two melts ,t ~SSnS+L2+G found that the hole-mobility at room temperature
a
10 fC- t"1~
SSnS7L, +Lz+G differs for the different crystal axes, viz. f.i.a/f.i.c= 5.5±O.5
5 and f.i.a/f.i.b= 1.15±O.1, if J.!a and f.i.b are the mobilities
1I
~ \ measured in the directions of the two nearly equivalent
1;=SSnS+L,+o
L directions of the pseudo-tetragonal structure and f.i.c is
5
;I
the mobility measured in the direction perpendicular to
10c1 the a,b plane (a<b<c).
5 To calculate the concentration p of the free holes from
SSI!+L.+G. .
zl 0 the Hall effect, the well-known relation
1
1(1'0 10 :;0 3(J 4() 50 60 70 80 00 m (1)
s,., -"L(af. ,*,5) 5
was used. It is of course doubtful whether this simple
FIG. 4. The S2 pressure, as a function of the composition of the
liquid phase, over the various solid-liquid-vapour three-phase relation is entirely valid over the range of temperatures
equilibria in the system Sn-S. used.
 I N V EST I GAT ION SON SnS 2223

A typical log p versus liT curve obtained in this way °c"--

from Hall measurements is shown in Fig. 6. It can be 600 500 4(JO 300 2DO
seen that even at 20 0 K the concentration of holes shows
only a small tendency to drop. This seems to indicate

r:
5
that the activation energy of the corresponding level is
J
virtually zero. At higher temperatures there is a slight
increase in the concentration of holes as calculated by
formula (1). This could be caused either by the fact
1\
that (1) is not valid over the entire range of tempera-
tures or by an incipient ionization of another deeper
1
lying level. To decide between these two possibilities
Hall measurements at higher temperatures are needed. 5 \

However, when measured at higher temperatures the

Hall effect changed with time in a complicated way, 1\
reminiscent of the similar effects found for SnSe by ~
Asanabe. 1 ,2 0

o~
V. EFFECTS OF ANNEALING to"
This effect of the Hall effect changing with time at
5
higher temperatures was thought to be interesting
enough for closer investigation. This was done in the
following way. At a series of temperatures lying be-

•
) 10'6
lJ.8 1.2 1.4 t.6 1.8 2 2.2 U

21 May 2025 12:27:18

f'::.
............ ---- '? (OJ(-~
t-- t-- FIG. 7. The concentration of free holes in SnS measured at room
..... temperature, after quenching, as a function of the temperature of
annealing.
10"
o 10 50
associates at lower temperatures:
FIG. 6. The concentration of free holes in SnS, as determined
by the Hall effect, as a function of temperature. 2h'+2V Sn';.::::(V SnV Sn)X, (2a)
or
tween 230-S00°C the crystals were allowed to "anneal" (2b)
for a time sufficient to establish equilibrium over the with " x, and x denoting the effective charges with
entire crystal. Then the crystals were rapidly cooled to respect to the crystal of the positive hole, the negative
room temperature, and the concentration of free holes Sn vacancy, and the neutral associate. Such an associ-
at room temperature measured by means of the Hall ation is not improbablell because one way of visualizing
effect. The result is shown in Fig. 7. It can be seen that such a (V Sn V Sn) center is to consider it as a S2 molecule
when the concentration of holes measured at room dissolved in the SnS crystal.
temperature is plotted logarithmically against the re- Now, assuming that the (V Sn V Sn) center neither acts
verse of the absolute temperature of annealing, a as a donor nor as an acceptor, and taking into account
straight line is obtained. The temperatures of annealing that the concentration of holes nearly equals that of
could not be chosen higher than about 500°C because the nonassociated V Sn centers, for high degrees of
for these temperatures the rate of cooling was very
probably not high enough to prevent some degree of Transmisston
annealing from taking place during the cooling also. The
experiments further showed that at the temperatures
i SnS
4. ~d=O'()24.cm
used, the crystal does not change its composition, 3
V
..- - 17"K
........

""
because the experiments could be repeated several times ./ \
:2
going up and down the temperature scale without
evincing any change in the Hall effect at room temper- 1 /V
1/
3OO"K
\ 1\
ature for a given annealing temperature. 0
o 0.:2
\ \
Starting from the assumption, mentioned in Sec. III,
that solid SnS always contains excess sulfur, these FIG. 8. Transmission of SnS as a function of photon energy.
phenomena could be explained by the formation of 11 F. A. Kroger (private communication).
2224 ALBERS, HAAS, VINK, AND WASSCHER

region is due to indirect transitions. From this it may be

concluded that the maximum of the valence band and
the minimum of the conduction band occur at different
values of the crystal momentum. A further analysis
showed the bandgap to be 1.080 and 1.115 ev at 300° and
nOK, respectively. Absorption due to free carrier
absorption was observed as well. This absorption
increases almost linearly with the concentration of holes
and is proportional to the square of the wavelength A
in a large spectral region, starting at about A= 5/J,. In
FIG. 9. The square root of the absorption coefficient as a addition an absorption band was found at nOK at 0.8
function of the photon energy. ev (absorption starting at 0.4 ev). This absorption, too,
was proportional to the hole concentration. It is thought
association, for the concentration of holes measured at that this absorption band is due to interband transitions
room temperature after quenching the following is of holes from the highest valence band to a valence band
obtained: lying about 0.4 ev lower.
p= (N/2K)114, (3a) Reflection measurements have also been carried out
or in the infrared as a function of wavelength between 2
p= (2N/K)1/6, (3b) and 25).!.15 These measurements were done at different
K being the reaction constant of reaction (2) and N the temperatures and different concentrations of free holes,
total number of Sn vacancies, associated and non- varying from p < 1014 cm-3 (Sb-doped), to p = 1.1· 1019
associated. cm-3 (Ag-doped). Figure 10 gives two typical curves of
If now log p is plotted against l/T, as in Fig. 7, 1 of the reflection coefficient as a function of wavelength for
the dissociation energy of reaction (2) is found from the a SnS crystal containing a concentration p of holes equal
slope, neglecting the temperature dependence of the to p= 1.06.1019 cm-3 • The results were carefully ana-
pre-exponential factors or, if one assumes that a V Sn lyzed with regard to the contributions of lattice vibra-

21 May 2025 12:27:18

center can take up two electrons from the valence band, tions of a polar crystal, and of the free carriers, taking
t of that dissociation energy. into account the fact that the crystal structure of SnS
From Fig. 7 1.4 ev or 2.1 ev, respectively, can be is rhombic. Therefore, sometimes polarized infrared
calculated for this energy. light was used. An analysis of the lattice vibration part
of the complex refractive index gave the following
VI. SEEBECK EFFECT results: The extrapolated value of the index of refraction
lto (A=0)=3.6±0.1, effective charge on the atoms
Over an extensive range of carrier concentrations the e*=O.7eo, dielectric constant E= 19.5±2. The values of
thermoelectric power a at room temperature could be 110 and e* are essentially the same for the three principal
described12 by the well-known formula: axes in the crystal. The value found for the dielectric
a= k/q[r+ 2+ In (N./ p)], (4) constant, could be confirmed by capacitance measure-
ments of layers of high-ohmic SnS (E= 19± 1). More-
with er(ma*)!= 1.5mo!. The highest reasonable value of over, the refractive index found in this way could be
r is t (polar scattering below the Debye temperature), corrobated by an analysis of interference fringes in thin
which would give a value 0.95 mo for the density of plates of high-ohmic SnS for wavelengths lying between
states effective mass ma*. This is a lower limit for m/". 4 and lOjl, also leading to a value of 110= 3.6.
The contribution to the complex refractive index due
VII. OPTICAL MEASUREMENTS
to free carriers was analyzed in order to obtain the
Infrared transmission measurements were carried out la
at n° and 300 0 K as a function of wavelength from
A= 1).! to A= 10).!. Some of the results have already been
given in a previous paper.l4 Figure 8 gives curves of
the transmission versus photon energy for SnS having
about 5.1017 free holes per ems. In the region of the
absorption edge, the square root of the absorption
coefficient turned out to be proportional to the photon
energy (Fig. 9). This means that absorption in this
12 Our thanks are due Dr. F. N. Hooge of this Laboratory for
kindly carrying out these experiments. FIG. to. Reflection coefficient of SnS as a function of wavelength
13 W. Albers, C. Haas, H. Ober, G. R. Schodder, and J. D. (p= 1.06.1019 cm-3).
Wasscher (to be published).
14 W. Albers, C. Haas, and F. van der Maesen, J. Phys. Chern. 15 C. Haas and M. M. G. Corhey, J. Phys. Chern. Solids
Solids 15, 306 (1960). (to be published).
 I N V EST I GAT ION SON SnS 2225

effective ma~s of the holes. It was found that the effec- In Sec. VII a lower limit of 0.95 mo was found for the
tive mass in the plane perpendicular to the c axis is density of states effective mass md*. Now one has
approximately isotropic and has a value ma*::::::::mb*=
(5)
0.20 mo. The effective mass parallel to the c axis is much
larger and was estimated to be m/""'mo. where N is the number of equivalent maxima of the
It is remarkable that the ratio me*/ rna * is almost valence band. Substitution of the various masses in this
equal to the ratio /1-a//1-c. From this it can be concluded equation shows that N must be at least 4. The maximum
that, although the effective masses differ considerably of the valence band therefore certainly does not lie at
for the two crystal directions, the relaxation time is K = 000. A value N = 4 is not inconsistent with the
isotropic. crystal symmetry.

JOURNAL OF APPLIED PHYSICS SUPPLEMENT TO VOL. 32, NO. 10 OCTOBER, 1961

Investigations on Silicon Carbide

H. J. VA~ DAAL, C. A. A. J. GREEBE, W. F. KNIPPENBERG, ANB H. J. VINK
Philips Research Laboratories N. V. "Philips' Gloeilampenjabrieken, Eindhoven, Netherlands

Measurements of Hall effect and resistivity up to 1300 0 K on p-type hexagonal SiC showed an acceptor
level for aluminium of 0.27 ev at zero donor concentration and a not yet identified acceptor level of 0.39 ev.
The spin multiplicity of this unknown center appears to be four times smaller than that of the aluminium
center, so that we may conclude that this unknown center in non-ionized state has paired electrons. Taking
a temperature dependence of the level depths proportional to that of the bandgap, the density-of-states
effective mass of the holes amounts to 0.59 mo. The Hall mobility shows at high temperatures the same
temperature dependence as that ascribed to scattering of holes by optical phonons. Assuming that optical

21 May 2025 12:27:18

phonons really come into effect, the behavior of the Hall mobility in the temperature range from 1300 0 to
300 0 K can be explained taking also into account the effect of scattering by acoustical phonons and charged
impurities. By a study of I-V characteristics of grown junctions in ",SiC and also by applying Roosbroeck-
Shockley's theory to the spectral distribution of the p-n luminescence under forward bias, inhomogeneities
were found over the junction area. By means of pyrolysis of gaseous compounds of Si and C pure crystals
(4X2X2 mmS) of "cubic" /3SiC were obtained. With the aid of polarized light the existence of a skeleton of
a hexagonal twinning system was found in these crystals, the cubic SiC filling up the pores of this skeleton
structure.

INTRODUCTION assuming a p-i-n structure of the diode. With regard to

the p-n luminescence it appeared that in the case of
Igiven
N this paper a short review will be given of current
investigations on silicon carbide. Therefore what is
here is an interim report on several investigations
light having photons of an energy higher than the band-
gap all phenomena could be explained by applying
still going on. This being so, no rounded-off picture of Roosbroeck-Shockley's theory. In both cases, however,
the various problems can be expected. Nevertheless it the existence of inhomogeneities over the junction area
seemed worthwhile publishing some results obtained so had to be assumed. There was some indication that
far. Three different problems will be dealt with. In the some of these inhomogeneities had to do with the
first section an account is given of Hall effect measure- polytypism of SiC.
ments carried out on p-type hexagonal SiC prepared in The last part of the paper (from Sec. 5 onwards)
the Lely furnace from technical grade material using deals with the problem of preparing pure SiC by thermal
argon gas either pure or doped with AlCh, AI, and CCI 4• reduction, on a hot wire, of gaseous compounds of Si
The concentration of centers not added deliberately is and C. It was found that crystals of a reasonable size
rather high. Nevertheless it is believed that some con- (4X2X2 mmB) of "cubic" SiC could be obtained with
clusions might be drawn with regard to the properties a large efficiency by the "hot-wire" method, if only the
of an acceptor center due to aluminium. Other data too right conditions were applied. A remarkable feature of
could be obtained in this connection. these "cubic" crystals is the fact that they consist
Quite a different problem is treated in Secs. 2,3, and 4. actually of a "sponge-like" matrix of thin colorless
Grown p-n junctions in a SiC were studied. A brief platelets of anisotropic hexagonal SiC, intersecting each
description of their preparation is given, followed by a other at definite angles, filled up with the isotropic
discussion of their forward characteristics as a function cubic material. Some results of chemical analysis are
of temperature. The p-n luminescence emitted from also given, indicating that these crystals are very pure
these junctions under forward bias was also studied. indeed. Yet the crystals showed n-type conductivity
The I-V characteristics could in principle be explained with a concentration of free electrons of n= 1017 cm- B.