This article was downloaded by: [University of Notre Dame Australia]

On: 08 May 2013, At: 06:09

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Advances in Physics
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/tadp20

Amorphous germanium III. Optical

properties
a a a
G.A.N. Connell , R.J. Temkin & W. Paul
a
Division of Engineering and Applied Physics, Harvard University,
Cambridge, Mass., 02138
Published online: 28 Jul 2006.

To cite this article: G.A.N. Connell , R.J. Temkin & W. Paul (1973): Amorphous germanium III. Optical
properties, Advances in Physics, 22:5, 643-665

To link to this article: http://dx.doi.org/10.1080/00018737300101359

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.
 [ 643 ]

Amorphous germanium
III. Optical properties t
B y G. A. N. CO~ELL, 1~. J. TEMKIN and W. PAUL
Division of E n g i n e e r i n g a n d A p p l i e d Physics, H a r v a r d U n i v e r s i t y ,
Cambridge, Mass. 02138
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

-2kBSTRACT
The optical properties of thick sputtered films ( ~ 30/x) of amorphous Ge,
grown with different substrate temperatures (0 ~<Ts ~<350°C), were obtained
between 0.05 and 4.5 eV by a combination of reflectance, transmittance and
ellipsometric measurements. The refractive index ab 0.15 eV decreases
monotonically with increasing Ts, or equivalently, with increasing density,
and is 4.13+_0.05 eV in the highest density films. The absorption edge is
approximately exponential (10~<c¢< 10~ cm -1) but shifts monotonically to
higher energy and increases in slope with increasing Ts. Similarly, ~he peak
in e~ grows by about 10% and shifts by about 0.15 eV to higher energies,
reaching a maximum of about 23 at 2.90+ 0"05 eV in the high density films.
The peak in the transition strength co2e~occurs at 4.2 ± 0.2 eV in all films, but
increases in magnitude with increasing Ts. The sum rules for neff(m) and
~o, elf (c°) are evaluated for ~o < 5 eV and vary monotonically with Ts. These
trends are neither compatible with Galeener's void resonance theory nor with
changes in the oxygen content of the films, determined by the examination of
abso13otion peaks at 0.053 eV and 0.09 eV. An explanation, suggested here
and expanded in I, is based on the observed changes in the structure of the
network and voids.

§ 1. INTRODUCTION
T h e o p t i c a l p r o p e r t i e s of a m o r p h o u s g e r m a n i u m p o t e n t i a l l y p r o v i d e a
w e a l t h of i n f o r m a t i o n f o r t h e s t u d y of t h e e f f e c t s of t o p o l o g i c a l a n d
q u a n t i t a t i v e d i s o r d e r (1) o n t h e e l e c t r o n e n e r g y s t a t e s of t h e c r y s t a l l i n e
f o r m s of g e r m a n i u m (2), (3). U n f o r t u n a t e l y , i t h a s p r o v e d d i f f i c u l t t o
d e f i n e u n i q u e l y t h e p r o p e r t i e s of t h e a m o r p h o u s p h a s e . V a r i a t i o n s of
t h e m e t h o d of p r e p a r a t i o n a n d t h e a n n e a l i n g h i s t o r y l e a d t o s i g n i f i c a n t l y
d i f f e r e n t v a l u e s of t h e i n f r a - r e d r e f r a c t i v e i n d e x (4)-(s), t h e p o s i t i o n a n d
s h a p e of t h e a b s o r p t i o n e d g e (5)-Cll) a n d t h e u l t r a - v i o l e t reflectivity(12).
These are accompanied by even more striking differences in the con-
d u c t i v i t y a n d i t s t e m p e r a t u r e d e p e n d e n c e (9), (18)-(~7) t h e e l e c t r o n s p i n
c o n c e n t r a t i o n ( 5 ) , (is) a n d t h e r m o p o w e r (19). I t is o f t e n p o s t u l a t e d t h a t
t h e s e d i f f e r e n c e s a r e a r e s u l t of t h e r e l a t i v e c o n c e n t r a t i o n s (5), (16) a n d
shapes(~0), (~1) of v o i d s i n t h e v a r i o u s s a m p l e s , b u t a d i r e c t p r o o f of t h i s
p o s t u l a t e h a s n o t b e e n g i v e n . C o n c o m i t a n t c h a n g e s i n t h e n a t u r e of t h e

W o r k s u p p o r t e d b y t h e Office of N a v a l Research, t h e N a t i o n a l Science

F o u n d a t i o n , a n d t h e Division of E n g i n e e r i n g a n d A p p l i e d Physics.

2U2
 644 G. A. N. Connell et al. on

local order, such as the bond length, the bond angle distribution and the
ring statistics, for example, have generally neither been sought experi-
mentally nor have their effects been considered. One reason, perhaps, is
the effort needed to perform precise structural and optical measurements
on the same samples.
We have undertaken an experimental correlation of the changes in the
electronic properties with the observed changes in the distribution of the
atoms and voids in amorphous Ge films prepared at different substrate
temperatures and consequently of different densities. In this paper we
describe the optical properties between 0 and 4-5 eV. In § 2 we estimate
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

the accuracy of various techniques for the determination of the optical

constants and describe the experimental procedures t h a t were used here.
In § 3 the results are then discussed, the sum rules are evaluated and their
implications for the changes in the density of states are examined.
Finally the main features of the new optical data are summarized in § 4.

§ 2. EXPERIMEIgTAL METHODS
2.1. General considerations
Several methods are available for the determination of the spectrM
dependence of the complex dielectric function ~(o))(-~1(o)) +iee(oJ)) but
their accuracy is not necessarily adequate to discern small differences
between the ~(o))'s of different samples. The Kramers-Kronig analysis
of the normal incidence reflectance from the air film interface, Rl(w ), has
been used for such differential experiments (12), but the suitability of the
technique cannot always be justified, for the following reason. The phase
angle 0 is related to R 1 by the expression
1 ~ d l n R l ( W ' ) l n ]o)'-o)
0(o))=~ o do)' ~ do)', (1)
such t h a t the derivative of In R 1 near o) influences 0(w) much more than
t h a t of In R 1 at remote points on the frequency scale (~). Nevertheless,
any regions in which d i n Rl(o))/do) is large, but which may be inaccessible
to measurement, can also make a large contribution. In practice, R1 is
measured in a limited range and the necessary but inexact extrapolations
for R 1 introduce errors into the calculation of 0(o)), which in turn limit
the resolution of the technique.
The importance of this problem in thin films, for which experimental
limitations in the vacuum ultra-violet and sample transparency in the
infra-red limit the spectral region in which R 1 can be measured directly,
m a y be ascertained from examination of the recent work on amorphous Ge
by Bauer and GMeener (BG) (12). R 1 was measured between 2 and l l eV
on films evaporated onto room temperature substrates and extrapola-
tions were made at low and high energies. For 2 ~<~o) ~<4 eV the extra-
polation at large o)', where both d In R(o)')/do)' and in ](o)'-w)(o)'+ o))l
are small, has little effect on 0(w), but at small w', where both
 Amorphous germanium 645

d In R(co')/dco' and In ~)/(~' + are large, the e x t r a p o l a t i o n

makes significant contributions to 0(oJ). I n fig. 1 our calculation of
¢2(w ) is shown for these d a t a for two low e n e r g y extrapolations, the first
based on the d a t a of D o n o v a n et al. (DSBA) (1°) a n d t h e second on the d a t a
of Theye(6), whose films, n o t a b l y , exhibited similar reflectivity between
2 and 5 eV (23). Since the variation in different investigations of Rl(W )
at low energies is greatest for low T s films, the difference between the two
curves of fig. 1 m a y be t a k e n as a reasonable estimate of the maximum
variation of %(~o), due to t e r m i n a t i o n errors, for films m a d e a t all T~. The
p e a k location is d e t e r m i n e d to + 0.2 eV and its height to + 0.5. Although
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

these error estimates and corresponding ones for the other optical con-
stants are used in the r e m a i n d e r of this paper, when comparing d a t a from
K r a m e r s - K r o n i g analysis with ellipsometric data, it should perhaps be
n o t e d t h a t t h e y are a p p r o x i m a t e l y twice those e s t i m a t e d b y B G for films
m a d e at T s = 160°C.
The error m a y be minimized b y t h e calculation of R 1 a t low energies
from ~(~o), which is determined, for example, from reflection a n d trans-
mission d a t a on thin films. However, the reliability of t h e technique is

Fig. 1
22 L I t I [

aO

18

16
E2
14 (1)

12

I0

0 1 2 3 4 5
ENERGY [eV]
The imaginary p a r t of the dielectric function, calculated by the Kramers-
Kronig analysis, for the combined refleetivity data of Bauer and Galeener
(Ref. (12)) and 6ither (1) DSBA (Ref. (10)) or (2) Theye (Ref. (6)).
 646 G. A. N. Connell et al. on

t h e n d e t e r m i n e d b y the accuracy of R 1 a t high energies. I n order to avoid

surface oxidation and the resulting r e d u c t i o n in R~ (e~), all m e a s u r e m e n t s
m u s t be p e r f o r m e d in situ.
All of these e x p e r i m e n t a l difficulties h a v e been avoided in our work b y
the use of techniques in which g(~o) is d e t e r m i n e d explicitly at each energy.
I n the n e x t sections the details of these procedures are described.

2.2. g(w) from reflection and transmission measurements

T h e optical constants m a y be readily d e t e r m i n e d from reflection R and
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

transmission T m e a s u r e m e n t s in two limiting cases. T h e case of coherent

multiple reflections within the film has been described in detail b y
G r a n t (25). I t requires films with optical thicknesses of the order of the
w a v e l e n g t h of the incident radiation a n d therefore demands high a c c u r a c y
in the R a n d T m e a s u r e m e n t s to d e t e r m i n e the absorption coefficient
close to the absorption edge since here the absorptance is small. An
estimate of the limit of this t e c h n i q u e for this region is given b y
A~(A/~+AT)/t, where AR and A T are the errors in R and T res-
pectively and t is the film thickness. F o r a 2000 A thick film and AR
A T e 0 - 0 1 , A a ~ 1 0 0 0 cm -1.
A more accurate result m a y be obtained, for the same a c c u r a c y in T,
b y determining ~(oJ) from thick films, when the limit of incoherent
multiple reflections m a y be employed. I n this regime the t r a n s m i t t a n c e
of an absorbing film on a non-absorbing substrate can be written
T = (1 - R1)(1 - Re)(1 - R3) exp ( - at) (2)
1 - R e R 3 - exp ( - 2~t) [R1R e + R 1 R 3 - 2R1R~R3]'

where R 1, R 2 and R 3 are the reflectances of the air-film, film substrate

and s u b s t r a t e - a i r interfaces respectively (26). F o r ~ < 5 × 104 cm -~, the
e x t i n c t i o n coefficient/c near 1 eV is v e r y m u c h less t h a n the refractive
i n d e x n, so t h a t R~, R e and R 3 m a y be calculated directly from n and the
substrate refractive index ns (~7). The refractive index n m a y be deter-
m i n e d b y ellipsometry, the m e t h o d used here and described in § 2.3 for
~>500 cm -~, or from the oscillations in the reflectivity of thinner films
deposited on metallic substrates, the m e t h o d used when ~ ~<500 c m - L I n
this latter case, the wavelengths A of the m a x i m a and m i n i m a are given b y

A=--2t ( n 2 sine 0)~/~, (3)

m

where m is an order number, which is an integer for a m a x i m u m and a

half-integer for a minimum, and 0 is the angle of incidence. Since
(Am/2t)e>> sin s 0, the accuracy of the d e t e r m i n a t i o n of n is limited only b y
the a c c u r a c y of the m e a s u r e m e n t of t, and does not rely on an absolute
d e t e r m i n a t i o n of R. I t is easily shown t h a t this result is not affected b y
the presence of thin oxide layers. The a b s o r p t i o n coefficient ~ is thus
readily ~btained from eqn. (2). F o r A T ~ A t / t ~ O . O 1 an accuracy for
 Amorphous germanium 647

of about 50 cm -1 is obtained using 30/* thick films deposited on glass

substrates.

2.3. ~(~o) ]rom ellipsometric measurements

The ellipsometric method is described in detail by McCrackin eta/. (es).
Briefly, the surface of the sample, here assumed to be coated with a thin
oxide film, is illuminated at some large angle of incidence 0 with linearly
polarized light from a dielectric medium with refractive index n i. The
electric vector of the incident radiation may be resolved into components
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

parallel and perpendicular to the plane of incidence called, respectively,

the p and s components and having different amplitude reflectivities
~ p = rp exp (iAp) and A s-- r s exp (/As). The quantity actually measured
is the ratio
~ p / ~ s = t a n ¢ exp (iA) = (r~,/rs) exp [i(Ap - As) ],
here conventionally expressed in terms of a restored azimuth ¢ and a
differential phase change A. An analysis of the elliptieMly polarized
reflected light yields the values ¢o and A0, the restored azimuth and
differential phase of the uncoated sample. These in turn lead to values
for ~(~o) after substitution into the standard formulae (29) for an optically
isotropic film. (This last assumption is based on the structural studies of
II.)
E-A= c°s2 2¢0 -- sin2 2¢0 sin2 A° sin 2 0 tan 2 0 + sin e 0, (4)
ni 2 (1 + cos Ao sin 2¢o)~

EA = 2 sin 2¢0 cos 2¢0 sin Ao sin2 0 t a n 2 0. (5)

ni 2 (1 + cos Ao sin 2¢o) 2
The ellipsometric technique is extremely sensitive to thin surface
layers (s°) and has often been used for the study of the oxidation of clean
surfaces(a1), (83). The measurements must therefore be made either in situ
in high vacuum or on substrate-materiM interfaces in order t h a t the
measured values of ¢ and A be equal to ¢o and A0 and hence directly
usable in eqns. (4) and (5). The former creates experimental difficulties,
the latter necessitates completely strain-free transparent substrates for
reliable measurements, yet even then leaves the suspicion t h a t the optical
properties of the sample may be modified near the substrate-sample
interface. The alternative to these approaches is to correct the measured
values ¢ and A for the effect of the surface oxide layer.
Studies at room temperature of the oxidation of crystalline Ge, Si,
GaAs and A1(~1)-(aa) in the atmosphere indicate t h a t the growth of the film
is logarithmic with time with a rate of order 5 A/decade. After one day
of exposure, the variation of the thickness during the time of our experi-
ment, which is typically 8 hours, is consequently of order 0.2 A and m a y
therefore be neglected. This conclusion is justified since A(w) changes
experimentally by less than 0.1 o during the time of the experiment.
 648 G. A. N. Connell et al. on

The effect on ¢o and A 0 of a thin t r a n s p a r e n t layer on the sample surfaces

was first discussed b y D r u d e (s4) and later b y several authors(35), (s6). W e
use the a p p r o a c h developed b y L u c y (37) in which the only a p p r o x i m a t i o n
is the dropping of higher terms in dl/A , where d 1 is the thickness of t h e
oxide layer. T h e error t h a t this causes in the d e t e r m i n a t i o n of ¢o and A 0
is t h e n of the order of the error in t h e m e a s u r e m e n t of ¢ and A, i.e. 0.01 °
and 0.03 ° respectively(a0), and therefore is a d e q u a t e for our purposes.
T h e relationships between ¢ and ¢o and A and A o are t h e n as follows for
ni=l :
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

¢o = ¢ + ~ sin ¢0, (6)

A o = A - b, (7)
where

a= r~ + s~ ] \ n~ s ] sin 0 tan 0,

O,
%

= n ~ - k 2 ; fi= 2nk,
p=y--1; q=y-tan 20; r = p q - f i 2; s=fi{p+q),

a n d n 1 a n d d 1 are respectively the refractive index a n d thickness of the

surface oxide layer.
T h e average refractive index of t e t r a g o n a l GeO~ (3s) is used for nl(w)
as suggested b y the work of Gambino et al. (ss), the m e a s u r e m e n t s of
Archer (sl), the presence of an absorption peak a t 720 cm -1 in o x y g e n a t e d
a m o r p h o u s Ge films and the insolubility of the film in water. H o w e v e r ,
t h e use of the refractive index of glassy GeO2 (4°) does n o t lead to a n y
significant changes in the results.
T h e thickness of the film must n o w be found. I n the region of the
a b s o r p t i o n edge, n(w)>~lc(w)~0 and A0z~r, such t h a t for hi= 1
lim a = 0, (8)
k-~0

limb=4~-~--ffAfflsinOtano~nle-l~ y(n12-- ~), (9)

cos 2¢. \ 2
lim e l = n s . . . . ~ / sin s 0 t a n s 0 + s i n s 0. (10)
~-~0 1 - sin 2¢0 ]

F r o m eqns. (6) and (8) it is seen t h a t in this region the measured values ¢
can be s u b s t i t u t e d directly for ~b0 in eqn. (10), as long as multiple reflections
within the a m o r p h o u s film m a y be neglected. Consequently, for a 30 t~
t h i c k film, n(w) m a y be f o u n d w h e n 5 × l0 s ~<a ~<5 × 10 ~ cm -~. A t this
 Amorphous germanium 649

stage the values of n(¢o) are used to determine e(¢o) from the transmission
measurements, and these in turn uniquely define a value of d 1 in eqns. (6)
and (7). The error in the determination of d 1 in this way, caused b y t h e
error in the measurement of A, is easily estimated from eqn. (9) and is
roughly 0.2 •. I t will be noted t h a t this sensitivity is similar to the
stability of the oxide film during the experiment.
~(~) m a y now be evaluated. Its approximate value is first found by
substitution of ¢ and A in eqns. (4) and (5). ¢0 and A0 are then calculated
from eqns. (6) and (7) and ~(~) redetermined. This process is iterated
until the absolute values of the differences between the n t h and (n + 1)th
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

values of el(~o) and e2(¢o) are each less than 0.1. The convergence and
accuracy of the technique were established from test measurements on
crystalline Ge.
Jungk has also used an ellipsometric technique to determine ~(¢~) of
evaporated films of amorphous Ge (41) but did not allow for an oxide layer
in his calculations. Although the trends in his results will be roughly
correct, the absolute values of g(o)) and the finer details of its variation
with sample preparation cannot be obtMned from his work.
Finally, we have considered the effects of surface roughness on the
determination of f(o)) by the ellipsometrie technique. A simple extension
of Beckman's work (~) to include second-order corrections to the amplitude
refleetivities t h a t are a result of local surface gradients gives for specular
reflection

where ~ ( 0 ) and ~j°(O) are respectively the amplitude refleetivities of the

j-wave ( j - s or p) for rough and smooth surfaces and S and c~ are res-
pectively the standard deviations of the local slope and loeM displacement
from the mean plane. I t follows immediately t h a t the dependence of
~.(0) on ~ drops out in ellipsometry and t h a t for films with small values of
S, characteristic of smooth glass substrates (a°), the measured ratio, ~ p / ~ ,
is equal to ~p°/Ns°.

2.4. Experimentalprocedure
The details of the preparation and physical and chemical properties of
the amorphous Ge used in this work are given in I. Three types of
samples were used. Samples, approximately 30/~ thick, were deposited
simultaneously on glass substrates for transmission and ellipsometric
measurements and on Be substrates for infra-red refleetivity measure-
ments and for the structure measurements of I I . On the other hand, the
variation of ~(w) with temperature and pressure, determined in earlier
studies (5), (~), was obtained from samples removed from metM substrates.
The similarity of the samples made on different substrates during the same
T s run was examined by measuring the resistivity which, as explained in
 650 G . A . N . Connell et aI. on

I, is sensitive to substrate temperature Ts, and by measuring the X-ray

diffraction in the vicinity of the first and second diffraction peaks. The
effects of the different substrate materials on the structure of the films at a
particular T s were undetectable and therefore the optical and structural
changes should be closely correlated.
The infra-red reflectivity and transmission of the samples were obtained
with standard techniques. On the other hand, the ellipsometry was
performed using the novel features of the split-beam system described by
Smith (44) and based on the design of Archard et al. (aS). Full details of its
operation and excellent sensitivity are given in Smith's paper. Using
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

photomultipliers (RCA C31025C) and PbS cells (Infratron Industries

Type 2, Waltham, Mass.) it was operable in the spectral range from 0.5 eV
to 5.0 eV. An important feature of the system is t h a t the angle of
incidence in all of the experiments is constant. In contrast, the angle of
incidence is varied at each energy in Jungk's technique (as), thus making it
impossible to correct for the oxide film by the method outlined in § 2.3.

§ 3. RESULTS
3.1. The optical constants below the absorption edge
The variation of the refractive index at 0-15 eV with the substrate
temperature during deposition is shown in fig. 2. Below about 150°C, n

Fig. 2
I I [

4.3

:=.

c~

::>-
4.2

I
~ 4.1

4.00 I i I
100 200 300
SUBSTRATE TEMPERATURE Ts (°C)

The dependence of the refractive index at 0.15 eV on substrate temperature.

Partial crystallization occurs at 375°C.
 Amorphous germanium 651

is changed slowly b y T~ variation, b u t between 150°C a n d 300°C it is

reduced b y a b o u t 5 % to a value of 4.13 + 0.05. F o r substrate t e m p e r a -
tures greater t h a n 350°C the films are partially or wholly crystallized.
The dependence of n and other film properties on T~ in this way suggests
t h a t film g r o w t h is controlled b y an a c t i v a t e d process, a possibility which
is considered more fully in I.
The v a r i a t i o n of n w i t h annealing t e m p e r a t u r e , T a, of films sputtered (5)
and e v a p o r a t e d (4), (5) onto room t e m p e r a t u r e substrates is similar to n
versus T s. T h e variation of n with T s of e v a p o r a t e d films, however, is
less well defined. T h e y e has f o u n d t h a t the refractive index at 0.5 eV
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

decreases f r o m a b o u t 4.40 when T~ is 25°C to 4.25 when T~ is 300°C, which

compares f a v o u r a b l y with our values of 4.48 w h e n T s is 25°C and 4.25
when T s is 300°C (6). Similar results were o b t a i n e d b y Wales et al. (~).
D o n o v a n et al., on the other hand, f o u n d t h a t all of their amorphous
films had values of n within 0-1 of the crystalline value at 0.5 eV. This
contrasts with observation of changes in the position of t h e absorption
edge above 10 ~ em -1 t h a t are similar to those f o u n d in the present
study(a0), (11).
Below 0.1 eV, absorption bands are seen in some of our films and two
typical b u t e x t r e m e cases are shown in fig. 3. I n the u p p e r example a
weak absorption is observed at 0.07 eV, recognized b y the oscillation in
n(~o). A l t h o u g h the absolute error in each point on a particular curve is
greater t h a n the amplitude of this oscillation, owing to the 1 °/o error in
the thickness measurement, the relative error of the points for a single
film is considerably less. An oscillator fit to n(w) (47) gives an absorption
coefficient at 0.07 eV of 200 + 100 cm -1, which m a y be c o m p a r e d with t h e
absorption bands a t this energy, observed b y T a u c et al. for materials
m a d e in several different ways (4s). I n their work the strength of the
absorption varies from a b o u t 5 × 103 cm -1 in electrolytic material to a b o u t
5 × 102 cm -1 in e v a p o r a t e d and s p u t t e r e d films a n d it seems, for this reason,
t h a t the b a n d is not an intrinsic p r o p e r t y of the network. Some contri-
b u t i o n to the absorption m a y arise from t w o - p h o n o n absorption, b u t such a
contribution should p r o b a b l y be less t h a n or of order of the one-phonon
absorption, which is a b o u t 200 cm -1. On the o t h e r hand, o x y g e n m a y be
eliminated as a cause of this absorption since there is no companion
absorption a t 0.113 eV (cf. the discussion later in this section). Other
impurities, such as carbon, h y d r o g e n or w a t e r (OH), are therefore p r o b a b l y
responsible for the bulk of the absorption in electrolytic material, b u t t h e
possibility of t w o - p h o n o n absorption c a n n o t be eliminated as an explana-
tion for the a b s o r p t i o n in our samples.
The absorption at 0.09 eV, seen in the lower example, is always observed
in t a n d e m with an absorption at 0.05 + 0.005 eV, b u t e x p e r i m e n t a l limita-
tions p r e v e n t an accurate d e t e r m i n a t i o n of n and ~ a t the l a t t e r energy.
Nevertheless, the occurrence of these absorptions as a pair allow us to
confirm the hypothesis of Tauc et al. t h a t their presence is a result of Ge-O
vibrational modes(4S). H e x a g o n a l and glassy GeO 2 have strong absorptions
 652 G . A . N . Connell et al. o n

at 0.069 eV and 0.113 eV (49), such t h a t oxygen with this local con-
figuration m a y be ruled out. On the other hand, tetragonal GeO 2
has strong absorptions at 0.053 eV and 0.09 eV (49), firmly indicating t h a t
oxygen occurs with this local order in sputtered germanium films. I t
should be noted t h a t in tetragonal Ge02, the Ge-Ge separation is 2.86 A,
which disagrees with the value of 3-35 ~ assumed by Moss et al. (5°) to
explain their structural results for Ge films with large oxygen content.
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

Fig. 3
I J t

4.30 - - (I)

x
w
tm
z

.~ 4.20
t--

b_
LLI
n,"

4.10 -

i I i I J
0.08 0.10
ENERGY [eV]
The refractive index near 0.1 eV for an oxygen-free film (1) and an oxygenated
film (2). The substrate temperatures were 25°C and 325°C respectively.

The oxygen concentration in this film may be calculated by the techni-

que of Kaiser et al.(51). The absorption coefficient ~(o~)is first found by an
oscillator fit to n(~o), giving a maximum absorption for this film of (650 _+
50) cm -1 at 0.09 eV. A K - K analysis gives a similar result. The
oscillator strength may then be determined, and, remembering t h a t each
oxygen is strongly bonded to only two Ge atoms although actually three-
fold coordinated(52), an oxygen concentration of 0.5 at. % is calculated.
The oxygen concentration in our films ranges therefore from less than
0.1 at. % to 0.5 at. %, and in I the cause of this variation is suggested.
Tauc et al.(~5) have reported t h a t some films, in particular those pre-
pared by electrolytic deposition, had an absorption at 0-23 eV. We have
not, however, detected an absorption band at this energy in our sputtered
films, although electrolytic material t h a t we examined showed an absorp-
tion at 0.23 eV of 1000 cm -1. We must conclude t h a t such absorption in
electrolytic films is not intrinsic to the network and again probably stems
from the presence of carbon or hydrogen.
 Amorphous germanium 653

3.2. The absorption spectrum

The energy dependence of the absorption coefficient of films made at
0°C and 350°C is shown in fig. 4 by curves 1 and 2 respectively. I n
both cases the edge is roughly exponential for absorption coefficients
below about 5 x 103 em -1 with ~(w) ~ % exp hoJ/Es, b u t a steepening and a
shift to higher energy occurs in the h o t film. Values of E s are given in
the table in § 4. Above about 5 x 103 cm -1, the absorption is no longer
exponential.
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

Fig. 4

°/

102 I
0.2 0.4 0,6 0.8 1.0 1.2 1.4
ENERGY l e V I
The absorption edge between 0.3 and 1.4 eV ; (l) Ts=25°C ; (2) Ts=350°C ;
(3) T s = 25°C ; T~ = 150°C for 100 hours (Ref. (5)) ; (4) Clark's evaporated
material; Ts=25°C (Ref. (8)); (5) Theye's evaporated material;
Ts=25°C ; T~=300°C (Ref. (6)) ; (6) Theye's evaporated material ;
Ts=300°C (l~ef. (6)) ; (7) crystalline Ge (gel. (54)). In curves (1) and
(2), transmission and ellipsometry data are represented respectively by
dots (. • .) and circles (© © o).

The annealing of low-temperature sputtered and evaporated films has

the same qualitative effect(5), (6). For example, curve 3 is the absorption
edge of a low-temperature film t h a t h a d been annealed in vacuo at 150°C
for 100 hours (5). Similarly, curve 4 is the absorption for room temperature
evaporated material, determined b y Claxk (s) b u t often confirmed b y other
workers(5), (~), (9). I t should be pointed out, perhaps, t h a t DSBA,
Chopra and Bahl (Sa) and Theye(S) have reported t h a t films produced
on room temperature substrates b y slow evaporation in a good v a c u u m
and with a large substrate-to-source distance have a sharp edge below
 654 G . A . :N. Connell et al. on

1000 cm -1, situated at about 0-55 eV. Above 1000 cm -1 the edge is
similar to Clark's. However, in a recent paper, Connell and Lewis (26)
suggested t h a t this evidence for a sharp edge in low temperature eva-
porated films m a y in fact be an artifact resulting from either experimental
or caleulational errors. This discrepancy can only be cleared up satis-
factorily by the analysis of unadjusted I~T measurements on films roughly
2 t~ thick. Upon annealing, Theye found t h a t the ' Clark edge ' sharpened
and shifted to higher energies, reaching curve 5 in the fully annealed state.
Theye has also examined the effects of substrate temperature during
sample preparation (e>. For a 300°C substrate, the highest temperature
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

before erystMlization, the absorption edge is represented by curve 6. The

absorption edge of a film prepared at 300°C had earlier been measured by
Donovan et al. (n). The edge in this case was identical to our result,
curve 2, for ~w >~0-7 eV but at this energy decreased to zero in approxi-
mately the same way as the crystal, indicated in fig. 4, by curve 7 (s4).
However, we suspect t h a t the latter behaviour may be the result of the
same systematic error t h a t is present in their room temperature data.
The dependence of the absorption edge of sputtered films on the sub-
strate temperature during deposition is therefore i n quantitative agree-
ment within experimental error with the results on evaporated films.
I t has been suggested t h a t oxidation might account for the variation
in the edge strength and location (s°), (ss). This hypothesis may be tested
for two limiting eases.
First, the effect of oxygen incorporation in the fully coordinated
(' ideal ') network can be estimated by the method outlined by Philipp in
his work on amorphous SiO x compounds (s6). The absorption edge in the
various forms of GeO 2 occurs at about 6 eV (ss), (40), such t h a t in GeO x
compounds the absorption below this energy must be associated with the
presence of Ge-Ge bonds. Since there are no competing optical processes
in this spectral region, the absorption, especially near threshold, should
afford a semi-quantitative measure of the density of these bonds. Philipp
showed t h a t the minimum configuration of Si atoms t h a t exhibited the
optical structure of amorphous Si had the dimensions of a Si-(Si4) tetra-
hedron. Thus, in the absorption tail, the absorption coefficient should be
directly proportional to the number of such tetrahedra, but the onset of
absorption should remain at a fixed energy.
The number of such tetrahedra m a y be calculated as follows. For n
germanium atoms there are 2n sites which may contain oxygen. When a
site is filled, Ge-O bonds are formed, and when it is empty, a Ge-Ge bond.
Thus the probability t h a t a site is empty in GeOx is ( l - x / 2 ) and the
fraction of Ge tetrahedra t h a t are Ge-(Gea) tetrahcdra is ( 1 - x/2) 4. The
absorption coefficient near the absorption threshold should therefore
vary as ( 1 - x / 2 ) 4.
The reduction of the absorption coefficient, without change in the
energy at which absorption begins, is clearly seen in Philipp's work as
the oxygen concentration is increased. In contrast, the onset of absorp-
 Amorphous germanium 655

tion in a m o r p h o u s Ge appears to m o v e to higher energy with annealing

or with increasing substrate t e m p e r a t u r e . F u r t h e r m o r e , an oxygen
c o n c e n t r a t i o n of 0-5 at. °/o would lead to a fractional decrease of the
absorption only of order ( 1 - 0 . 0 0 5 / 2 ) 4 z 0 . 9 9 . This is a n insignificant
p a r t of the observed reduction of a b o u t an order of magnitude.
I n the second m e t h o d the effect of o x y g e n i n c o r p o r a t i o n in the partially
coordinated n e t w o r k can be e s t i m a t e d b y the m e t h o d used b y Conncll and
P a u l (5). I-[ere it is assumed t h a t u n p a i r e d electrons (' dangling bonds ')
p u t states between t h e valence and conduction bands. F u r t h e r , it is
assumed t h a t o x y g e n reacts with Ge atoms having dangling bonds and
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

therefore removes states f r o m the b a n d gap. The shift of the absorption

edge t h e n corresponds to t h e r e m o v a l of these states.
The equivalent n u m b e r of electrons involved in transitions below
E 0 = 0-9 eV in the low density material can be e s t i m a t e d from a sum rule
and is a b o u t 2 × 1020 cm -a. This of course is consistent with the maxi-
m u m o x y g e n contamination. T h e oxygen, however, is r e m o v i n g states
associated with n e t w o r k imperfections and as such is acting as a compen-
sating agent. If the o x y g e n incorporation is responsible for the shift in
the absorption edge, t h e n the properties of the slightly o x y g e n a t e d material

Fig. 5
I I

106

lO5
'E

I--

CD
i,
U_

~ 10 4

Z
o

io 3

]0 ~ I I
0 I 4
ENERGY [eV]

The absorption edge between 0.4 and 5 eV; - - . - - : Ts---25°C;

Ts=350°C ; . . . . : DSBA (Ref. (10)).
 656 G . A . N . Connell et al. on

may more closely represent those of an ideal fully coordinated network.

We do not believe, however, t h a t this is the correct interpretation of the
data.
In summary, oxygen contamination can only lead to the observed
changes in the absorption edge when acting as a compensator for network
imperfections. In I I it was found t h a t the coordination number increases
with densification which of course would remove the states associated with
dangling bonds without appeal to contamination. We also find t h a t the
changes in the optical properties are not confined to the absorption edge
and this behaviour may be closely correlated to the structural changes.
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

Finally, we find t h a t trends in the optical properties scale monotonically

with density and not with oxygen concentration. All of these observa-
tions, taken together, indicate t h a t oxygen contamination is not the
controlling factor for the optical properties of the films.
In fig. 5 the absorption coefficients of films made at 25°C and 350°C
are shown for energies up to 5 eV, clearly showing the non-exponentiM
region at higher energies. I t appears t h a t at the highest energies, the
absorption coefficient approaches a maximum in both films but the
maximum value increases roughly as the density. In the highest density
film the absorption coefficient is (1.4 + 0.1) × l06 em -1 at 4.5 eV.
The position of the absorption energy gap cannot be defined with great
precision for exponential edges. However, in the non-exponentiM region,
the absorption spectrum can be described by

which defines an absorption energy gap ~0 (57). In our previous work on

the influence of annealing upon cold samples, this definition was used,
and values of E0 in the range 0.9 eV to 1.0 eV were obtained for several
kinds of samples after anneMing(~). In fig. 6 a plot of ]~e21/2 versus/~oJ
is made for the 25°C and 350°C samples of fig. 4. I t is seen t h a t the zero
intercept moves from 0.7 eV in the cold film to 0.9 eV in the hot film, in a
way t h a t characterized the annealing studies. Therefore, increasing the
substrate temperature during deposition and raising the annealing tem-
perature appear to have very similar effects both on the absorption tail (6)
and on the states deeper in the valence and conduction bands.

3.3. The optical constants above the absorption edge

To facilitate the comparison with other work on amorphous Ge, the
variation with T s of the normal incidence reflectivity Rl(w), cMculated
from ~(w), is shown in fig. 7. In the range from 2 eV to 4.5 eV, Rl(w )
increases as T s increases, but the change is less dramatic than t h a t found
by BG in evaporated films. Although Rl(~O) of high density sputtered
material (Ts=350°C) is within experimental error the same as t h a t of
evaporated material (Ts=160°C) prepared by BG, Rl(w ) of the low
density sputtered material (T s = 25°C) is significantly larger than t h a t of
the low density evaporated material (Ts=25°C). This difference is
 Amorphous germaniu~n 657

Fig. 6
I l f

14

12

I0 - h~

~ e-
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

-co 6--

4--

2-

o ", Y ° °/] I I
0 1 2 3
ENERGY leVI

h~oE,,t"2(,)) for films made with T~=25°C (. • -) and T s = 3 5 0 ° C (o o o).

Fig. 7
I 1 ~[ I I

0,6
~ (4)
(2)
(i)

0.5

E 04
>-
t-
~ 0.5
w
la_
~ 0.2

O.T --

0 l I I I I
0 1 2 3 4 5
ENERGY (eV)
The normal incidence reflectivity (1) T s = 2 5 ° C ; (2) Ts=350°C ; (3) Bauer and
Galeener ; T~=25°C (Ref. (12)) ; (4) Bauer and Galcener ; Ts=160°C
(Ref. (l 2)).

A.I'. 2X
 658 G. A. N. Connell et al. on

consistent with our own ellipsometric measurements on thin evaporated

material made near room temperature under similar conditions (2a).
E v a p o r a t e d and sputtered material therefore appears to be very similar
when made at elevated temperatures, at least as far as their absorption
coefficients above about 1000 em 1 and their reflectivities between 2 eV
and 5 eV are concerned. On the other hand, evaporated material,
prepared near room temperature, is less strongly absorbing below 1 eV
than similarly prepared sputtered material, a trend t h a t is apparently
continued at higher energy judging b y the relative sizes of Rl(OJ).
The variations with T s of the real and imaginary parts of the complex
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

refractive index, n(w) and k(~o), are shown in figs. 8 and 9 respectively.
With densification, the natural frequency of the oscillation increases,
judged by the shift of n(w) above 2 eV. This in t u r n is responsible for
the decrease in n(0), as was originally suggested b y Connell and Paul (5).
k(~o) reaches a broad maximum at about 3.8 eV in the 25°C film, b u t
narrows and shifts to higher energy in the high density film. However,
the maximum remains at a considerably lower energy than the highest
maximum in crystaJline Ge. The maximum value of k(oJ) of 3.2 in the
Ts=350°C film is in good agreement with the maximum of 3.2+ 0.1,
measured b y BG in evaporated films made at 160°C. The agreement for
the position of the peak is less satisfactory, b u t we attribute this to the
difficulty of determining its precise location in the broad maximum of
k(w), when experimental reflectance data must be extrapolated to perform
a K r a m e r s - K r o n i g analysis. On the other hand, their maximum k of
2.8 + 0-1 in evaporated films made at 25°C is significantly less t h a n our
result of 3.1 for low density sputtered films. The reason for this difference
m a y be associated with the very different void networks. Thin films,
pervaded with fissures as prepared b y Donovan and I-Ieinemann (Ss), are
expected to be more transparent and less reflecting than films with
isolated small voids. They would therefore be less absorbing, as indicated
in fig. 9, b u t would have a similar natural frequency, as indicated by
figs. 8 and 9.
In figs. 10 and 11 respectively, the variations with T s of El(OJ) and E~(w)
are shown. To p u t these trends on a more quantitative basis, it is con-
venient at this point to introduce the Lorentzian model for optical
dispersion(aT). In this

~(~) = 1~ ~.2 (12)

w02 - oJ2 - iy~o'
where w~2 = 4~rnee/m, n is the number of charges per unit volume, oJo is the
natural frequency and y is a damping or line width term. Clearly, the
energy at which e l = 1 is ?/oJ0 and ~%~=%~a~7co 0. The dispersion of
amorphous Ge is not exactly described by eqn. (12), b u t first approxima-
tions for co0 and ~% and their dependence on T~ can be obtained.
F r o m el(oJ), we estimate t h a t ?/w0~3.3 eV, which may be compared
with the mean separation of the valence and conduction bands in crystal-
 Amorphous germanium 659

Fig. 8
l E i i

5
E
X
hA
C5
Z 4

hA
\
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

>
F-
(D 3
<
¢Y
b_
hA
05
2-

I I I I I
0 l 2 3 4 5
ENERGY (eV)
The refractive index ; .... T~ = 25°C ; - - - - : T s = 350°C ;
-- DSBA (Ref. (10)).

Fig. 9

31 ' , ' ~ [
t //'"
~- ~1- !/"
~- I !I'
12)

Z
0.9_
o~- l~- /,//

~-
I-
X I 1/I

o
0 1 2 3 4 5
ENERGY (eV)

The extinction coefficient ; -- • -- : T~= 25°C ; -- • T~=350°C ;

--- : DSBA (Ref. (10)).

A.P. 2 Y
 660 G.A.N. Connell et al. o n

Fig. 10

25 ~ ~ ' ~ '

e,
152"
\':~
~,,~\
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

10-

0
• %,,.

-5- N.~

t l r I I
0 l 2 5 4 5
ENERGY [eV]

The real part of the dielectric function ; - - • - - : Ts=25°C ;

T~=350°C ; - - - : DSBA (Ref. (10)).

line Ge of 3.5 eV (59). U p o n densification, an increase of a b o u t 0.2 eV

occurs. F r o m E2(w), it appears t h a t 7 ~ coo, such t h a t 1/cop~ (22 x 3.3~) 1/~
16 eV. This is in excellent, b u t fortuitously close, agreement with the
electron energy loss measurements of Zeppenfeld and R a e t h e r on eva-
p o r a t e d a m o r p h o u s Ge (G°>.
T h e shift of the energy at which e2(w ) is a m a x i m u m , / ~ , with increasing
T s corresponds to the increase in coo of the L o r e n t z i a n oscillator. This
has n o t been observed b y other workers, possibly as a result of the experi-
m e n t a l inaccuracies cited above. The m a x i m u m of 23 at 2.90 eV in high
density material is similar to t h a t in high t e m p e r a t u r e e v a p o r a t e d films.
H o w e v e r , the m a x i m u m heights in low density s p u t t e r e d and e v a p o r a t e d
material are significantly different. The relative values indicate a reduc-
tion of the effective strength of the dispersion b y a b o u t 11%. This
conclusion is also reached b y scaling the k(co) curves for the two materials.
I n I we examine the variations of g(co) in more detail and compare our
results with the changes in the structural properties. The trends in coo
of the L o r e n t z i a n oscillator are explained b y variations in the m e a n gap
/~s separating the valence and conduction states t h a t result from changes
in the co-ordination n u m b e r and bond angle distribution. Finally, a new
t e c h n i q u e is described which permits the valence b a n d density of states
to be o b t a i n e d from g(co).
 Amorphous germanium 661

Fig. 11
L l t t I

22

2O

!
16-- I
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

I
!
14-- !
e2 !
!
~2- / \
/ , ,
lO - - 25

8 22

6 2t

4 20
b

2 19 2.15 51.0 \ 5.5

t
0 ,,~' I I I I
0 l 2 5 4 5
ENERGY leVI

The imaginary p~rt of the dielectric function ; - - - - - : T 25°C •

- - : Ts=350°C ; - - - : DSBA (Ref. (10)).

3.4. Sum rules

The effective n u m b e r of free electrons per a t o m , n ~ , c o n t r i b u t i n g to
the optical p r o p e r t i e s u p to ~n e n e r g y / h % is g i v e n b y

- - - n°~-- I ~oc~(~o)&o, (13)

0

where N is t h e a t o m density. Similarly, the effective dielectric constant,

%, ~ , p r o d u c e d b y i n t e r b a n d transitions in this r a n g e m a y be w r i t t e n
0

%, ~ = 1 + co-lea(co) d~o. (14)

F r o m a p l o t of %, e~ versus ?gw0 it is therefore possible to e s t i m a t e which

transitions m a k e the l~rgest contributions to t h e static dielectric constant,
More i m p o r t a n t for the present purposes, however, is t h a t f r o m these d a t a
useful c o m p a r i s o n s of the valence a n d c o n d u c t i o n states in a m o r p h o u s and
crystalline materiM c~n be made.
 662 G.A.N. C o n n e l l et a l . o n

Fig. 12
I I i I 1

10

'E 8

0
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

o 4--

2-- ~!..,..-"
I
0
0 l 2 3 4- 5
ENERGY (eV)

T h e effective n u m b e r of electrons i n v o l v e d i n t r a n s i t i o n s u p to e n e r g y ~ w ;
__.__ : Ts'=25°C ; - - : T s = 3 5 0 ° C ; . . . . : c r y s t a l (Ref. (61)).

Fig. 13
I I I I

16-- .f
/
14--

12--

lO-

go,eft
8--

6--

4--

2-

0 I I I I
0 1 2 3 4
ENERGY (eV)

T h e c o n t r i b u t i o n to t h e s t a t i c dielectric c o n s t a n t of t r a n s i t i o n s u p to e n e r g y hw ;
• o
.... : T~=25°C , - - : Ts=350C.
 Amorphous germanium 663

A plot of Nna~ versus hw 0 is s h o w n in fig. 12 for low a n d high density

a m o r p h o u s m a t e r i a l a n d for crystalline Ge (61). More states are s i t u a t e d
within 2.5 eV of the b a n d edges in a m o r p h o u s films, a l t h o u g h a r e d u c t i o n
in their n u m b e r occurs with densification. This, of course, is in a g r e e m e n t
w i t h the valence a n d conduction b a n d densities of states t h a t h a v e b e e n
deduced b y D o n o v a n a n d Spicer (6~), L e y et al. (6a) a n d B r o w n and I~ustgi (64).
At the highest e n e r g y this t r e n d has been reversed. H e r e the effective
n u m b e r of electrons i n v o l v e d in transitions u p to 5 eV increases as the film
density increases. This is opposite to the b e h a v i o u r , n o t e d b y B G a n d
s u m m a r i z e d b y Galeener(65), of n a t (12), b u t m a y not differ from the
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

b e h a v i o u r of n a f (5). T h a t is, we find b y K r a m e r s - K r o n i g analysis of

B G ' s d a t a t h a t the t r e n d in n a t (5) depends on t h e precise f o r m of the low
a n d high e n e r g y reflectance extrapolations.
A plot of Co, e~ versus hoe0 is shown in fig. 13. T h e r a t i o of effective
dielectric c o n s t a n t s of the low a n d high d e n s i t y films a t 4.5 eV is 1.10,
which c o m p a r e s well w i t h the r a t i o of the s q u a r e of t h e r e f r a c t i v e indices
of 1.09. Of course, these should n o t be equal until s a t u r a t i o n of t h e d-
b a n d transitions. T h e %, ~r plot is f u r t h e r evidence of the g r e a t e r density
of states in t h e g a p region in a - G e relative to c-Ge. F o r example, a t
4 eV the r a t i o of %, err to % is 0.78 in high d e n s i t y a - G e , b u t only 0.68 in
c-Ge. I n o t h e r words, larger contributions are m a d e b y the larger
n u m b e r of lower e n e r g y states in a - G e .

§ 4. SUMMARY
Table. Salient features of the optical constants of amorphous and crystalline germanium

Amorphous Crystal
T s = 25°C T s = 350°C

Es eV 0-13 0"10
n(0.1) 4.33 + 0"05 4"13 _+0"05 4"00
E0 eV 0"70 _+0.02 0"90 _+0"02 0-67(0.80)t
E2 eV 2.74 _+0"05 2"90 _+0"05
Eg$ eV 3.13 3"34 3"50
Nne~f(4"5 ) electrons/cm ~ 7"40 x 1022 7"84 x 102~ 8"28 x 10~§
nerr(4"5):~ electrons/atom 1"83 + 0"1 1.82 + 0.1 1-86§
~0, at(4"5) 15"7 14-3 11.6§
1In (On/~p)T ][ bar -1 ( - 0 . 8 + 0 . 2 ) x 10 .6 - 1"0 x 10 .6
(3Eo/~P)TII eV/bar (3"5 +_0"5) x 10 .6 5"0(14"2) x lO-6t
(3Eo/~T)p¶ eV/K (--4"5_+0.2) x 10 4 --3.7(--4.5) x 10-4~
(~Eo/OT)v¶ eV/K (--4"0_+0"3) x l0 -4 ( -- 2.9 _+0.2) x 10 -a

The values and coefficients of the smallest gap and the direct gap are given, the latter
in parentheses. :~ See I. § gel. (61). [l Ref. (5). ¶ R e f . (43).

The salient features of our new optical d a t a are p r e s e n t e d in n u m e r i c a l

f o r m in the table. T h e d a t a on t h e t e m p e r a t u r e a n d pressure d e p e n d e n c e
 664 G . A . N . Connell et al. on

of the absorption edge and refractive index were obtained and published
by us previously but are included here for completeness (5), (aa). The
notation used in the table is that used earlier in this paper.

ACKNOWLEDGMENTS
We thank Javier Tejeda for his help in the design, building and evalua-
tion of the ellipsometer, David MaeLeod for a wide variety of technical
assistance and Albert Boudreau for constructing much of the optical
systems. We also acknowledge helpful discussions with Marc Kastner,
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

Adam Lewis and Denis Weaire.

l~E~ERENCES
(1) WEAraE, D., and THORPE, M. F., 197], Phys. Rev. B, 4, 2508.
(2) O~TE~BURGER,I. B., and HENDERSON,D., 1972, llthlnt. Con]. on Physics
o] Semiconductors, Warsaw, p. 464.
(3) JOANNOPOULOS,J. D., and CoheN, M. L., 1973, Phys. Rev. B, 7, 2644.
(4) WALES,J., LOWITT, G. J., and HILL, 1%.A., 1967, Thin Solid Films, 1, 137.
(5) CONNELL,G. A. N., and PAUL, W., 1972, J. Non-Cryst. Solids, 8-10, 215.
(6) THEYE, M. L., 1970, Optics Commun., 2, 329; 1971, Mat. Res. Bull., 6,
103.
(7) GLASS,A. M., 1965, Can. J. Phys., 43, 1068.
(8) CLARK,A. H., 1967, Phys. Rev., 154, 750.
(9) CHITTICK, 1%. C., 1970, J. Non-Cryst. Solids, 3, 255.
(I0) DONOVAN, T. M., SI'ICER, W. E., BENNETT, J-. M., and ASHLEY, E. J.,
1970, Phys. Rev., B, 2, 397.
(II) DONOVAN, T. M., ASHLEY, E. J., and SPleEn, W. E., 1970, PhysicsLett., A,
32, 85.
(12) BAUEa, 1%. S., and GALEENE~, F. L., 1972, Solid St. Commun., 10, 1171.
(13) WALLEr, P. A., and Jo~scI~Ea, A. K., 1968, Thin Solid Films, 1, 367.
(14) WALLE¥, P. A., 1968, Thin Solid Films, 2, 327.
(15) GRmo~ovIcI, R., 1968, Mat. Res. Bull., 3, 13.
(16) CAMPnAus]~N,D. L., CONNELL,G. A. N., and PAUL, W., 1972, J. Non-Cryst.
Solids, 8-10, 223.
(17) STUKE, J., 1970, J. Non-Cryst. Solids, 4, 1.
(18) BRODSK¥, !Vi.H., and TITLE, R. S., 1969, Phys. Rev. Lett., 23, 581.
(19) BEYEI~, W., and STUKE, J., 1972, J. Non-Cryst. Solids, 8-10, 321.
(20) GXLE~ER, F. L., 1971, Phys. Rev. Lett., 27, 421.
(21) GALEENER, F. L., 1971, Phys. Rev. Lett., 27, 1716.
(22) STEaN, F., 1963, Solid St. Physics, i5, 299.
(23) This was ascertained by ellipsometry measurements on films kindly given
us by Dr. Theye.
(24) PHILIPP, H. 1%, 1972, J. appl. Phys., 43, 2835.
(25) GRANT,P. M., 1965, Tech. 1%eport HP-14, Harvard Univ. GI~A~T,P. 1V[~,
and PAUL, W., 1966, J. appl. Phys., 37, 3110.
(26) CONN]~LL,G. A. N., and LEwis, ADAM, 1973, Phys. Stat. Sol. B, 60, 251.
(27) HEAVENS, 0., 1955, Optical Properties o/ Thin Solid Films (New York :
Academic Press), p. 49.
(28) McCRACKIN, F. L., PASSAGLIA, E., STI~OM:BERG, l~. R., and STEINBERG,
H. L., 1963, J. Res. NBS A, 67, 363.
(29) BORN, M., and WOLF, E., 1959, Principles o/ Optics (Great Britain:
Pergamon Press, Inc.), p. 617.
 Amorphous germanium 665
(30) BENNETT, ]7I. E., and BE~bTETT, J. M., 1967, Phys. Thin _Films, 4, 1.
(31) A~CHER, •. J., 1958, Phys. Rev., 110, 354.
(32) BER~I~G, P. H., HASS, G., and MADDE~, ]~. P., 1960, J. opt. Soc. Am., 50,
586.
(33) LUK~S, F., 1972, Sur]. Sci., 30, 91.
(34) ])RUDE, P., 1900, Lehrbuch der Optik (Leipzig), p. 206.
(35) TRONSTAD,L., 1935, Trans. Faraday Soc., 31, 1151.
(36) BoR~, M., 1933, Optik (Berlin : Verlag, Julius Springer), p. 36.
(37) Lvc~. F. A., 1948, J. chem. Phys., 16, 167.
(38) I~I~CK, C. L., 1973, Ph.D. Thesis (MIT), unpublished.
(39) ALESSANDRII~I,E. I., GAMBI~O, P~. J., and KUPTISS, J. D., 1972, Thin
Solid Films, 11, 415.
Downloaded by [University of Notre Dame Australia] at 06:09 08 May 2013

(40) PAJASOVA,L., 1969, Czech. J. Phys. B, 19, 1265.

(41) JV~GK, G., 1971, Phys. Stat. Sol. B, 44, 239.
(42) BECKMAN,P., and SI"IZZICHI~O,A., 1963, The Scattering o] Electromagnetic
Waves/tom Rough Sur/aces (New York : Macmillan Co.), pp. 70-98.
(43) CO,NELL, G. A. N., 1972, Phys. Star. Sol. B, 53, 213.
(44) SMITH, N. V., 1969, Phys. Rev., 183, 634.
(45) ARCHARD,J. F., CL:EGG,P. L., and TAYLOR, A. M., 1952, Proc. phys. Soc.
B, 65, 758.
(46) JU~GK, G., 1970, Phys. Stat. Sol. A, 3, 965.
(47) CHRISTY,P~. W., 1972, Am. J. Phys., 49, 1403.
(48) TAuc, J., ABRAHAM,A., ZALLE~, 1~., and SLADE, M., 1970, J. Non-Cryst.
Solids, 4, 279.
(49) LIPPINCOTT, E. R., VA~ V~LKE~BVnG, A. V., WEre, C. E., and BU~TING,
E. N., 1958, J. Res. NBS, 61, 61.
(50) Moss, S. C., F L Y ~ , P., and BAvna, L. O., 1973, Phil. Mag., 27, 441.
(51) KAISER, W., KECK, P. H., and LANGE, C. F., 1956, Phys. Rev., 101, 1264.
(52) WYCKOFr, R. W. G., 1963, Crystal Structures (New York : Interseienee),
Vol. 1, p. 250.
(53) CnOPRA, K. L., ~nd BAHL, S. K., 1970, Phys. Rev. B, 1, 2545.
(54) DAsg, W. C., and NEWMAn, R., 1955, Phys. Rev., 99, 1151.
(55) K~OTEK, M. L., and DO,OVAl, T. M., 1973, Phys. Rev. Lett., 30, 652.
(56) PHILIPP, ]:[. 1~., 1971, J. Phys. Chem. Solids, 32, 1935.
(57) TAVC, J., GRIGOl~OWCI,R., and VA~cu, A., 1966, Phys. Stat. Sol., 15, 627.
(58) DO~OVAN,T. M., and HEINEMAbT~, K., 1971, Phys. Rev. Lett., 27, 1794.
(59) See I in this series.
(60) Z~t"~E~n~D, K., and I~AETH~R, H., 1966, Z. Phys., 193, 471
(61) P~HeP, H. R., and EHa]~Rw~C~, H., 1967, Semiconductors and Semimetals,
edited by R. K. Willardson and A. C. Beer (New York : Academic Press),
Vol. 3, p. 93.
(62) DO,OVAl, T. M., and Se~cE~, W. E., 1968, Phys. Rev. Lett., 21, 1572.
(63) L~.¥, L., KOWALCZYK, S., POLLAK, R., and SmE~Y, D. A., 1972, Phys.
Rev. Lett., 29, 1088.
(64) BROW~, F. C., and I~US~G~, 0. P., 1972, Phys. Rev. Lett., 28, 497.
(65) GALEE%E~,F. L., 1972, llth Int. Con/. on Phys. o] Semiconductors, Warsuw,
p. 523.