Full text of DOI:10.1557/JMR.2002.0158

Identification of Cl and Na impurities in inclusions of a
vapor-grown CdTe doped with Zn and Cl
V. Corregidor and V. Babentsov
Departamento de Fisica de Materiales, Universidad Autonoma de Madrid, 28049 Madrid, Spain
M. Fiederle, T. Feltgen, and K. Benz
Kristallographisches Institut, Universitat Freiburg, Hebelstrasse 25, D-79104, Freiburg, Germany
E. Diegueza)
Departamento de Fisica de Materiales, Universidad Autonoma de Madrid, 28049 Madrid, Spain
(Received 9 October 2001; accepted 13 February 2002)
Morphology and analysis of composition of inclusions were done by secondary
electron microscopy and spatially resolved energy-dispersive analysis of x-ray on
semiintrinsic CdTe:Cl and CdTe:Zn:Cl crystals grown from the vapor phase by the
modified Markov technique and on undoped CdTe crystals grown from the melt by
the Bridgman method. In CdTe:Cl and CdTe:Zn:Cl crystals nonstoichiometric
inclusions of about 10–20 ␮m were found, which contain high concentrations of Cl
and Na impurities. The Cl is concentrated in small precipitates of 1–2 ␮m inside these
inclusions. After short-time low-temperature annealing (600 °C), the inclusions mostly
disappeared.
I. INTRODUCTION
main Cl-related defects are a shallow donor (ClTe) and
acceptor-like complex (ClTe–VCd).11,12 It is also known
that the nonuniformity of the Cl distribution leads to
deterioration of the resistivity uniformity,13 but so far a
limited study of Cl precipitation has been done14 and to
the best of our knowledge this problem is important both
for the melt- and vapor-grown CdTe materials.
Potential industrial application of room-temperature
x-ray detectors based on semiinsulating cadmium tellu-
ride requires the homogeneity of electrical properties and
low concentration of native defects and impurities.1 In
many cases Te-rich precipitates and inclusions enriched
with impurities are the main sources of inhomogeneity of
CdTe-based materials,2–4 which creates the problem
of resistivity control at the postgrowth heat treatment and
during the operational period.
This paper presents the results of the chemical analysis
of inclusions in CdTe:Cl and CdTe:Zn:Cl grown from
the vapor phase and CdTe grown from the melt for the
comparison following our previous study of Te-rich in-
clusions and the deviation from the stoichiometry in the
boundary regions.15–17 The annealing of the samples was
carried out by taking into account the possibility to re-
duce the concentration of inclusions15,16 and to increase
the Cl solubility in the bulk.11,18
As it was earlier shown,5,6 in CdTe crystals grown
from the melt, tellurium precipitates are formed from
native defects (VCd Tei) during the cooling process due to
the strong retrograde character of Te solubility in solid
CdTe, while the Te-rich inclusions are originated by melt
droplet capture on the growth surface. This phenomenon
leads to inhomogeneous distribution of impurities due to
segregation. Moreover, taking into account that the solu-
bility of many impurities in liquid Te is higher than in
solid CdTe, the concentration of residual impurities
in Te-rich defects can be considerably higher than in bulk
material.7
II. EXPERIMENTAL
CdTe:Cl, CdTe:Zn:Cl monocrystals were grown from
the vapor phase by the modified Markov technique10
under controlled vapor pressure PCd/PTe2 ס 2 with a
nominal concentration of chlorine of 5 × 1018 cm−3 and
zinc of 2.5 × 1018 cm−3 in the feed material added to the
melt as CdCl2 and ZnCl2. The crystals have high struc-
tural quality and are twin-free, with an average disloca-
tion density lower than 103 cm−2 and with a resistivity up
to 109 ⍀ cm, which was measured by the Van der Pauw
Commercially available semiintrinsic CdTe has al-
ways a high concentration of the doping impurity,
particularly chlorine,8–10 and it is well known that the
a)Address all correspondence to this author.
e-mail: ernesto.dieguez@uam.es
J. Mater. Res., Vol. 17, No. 5, May 2002
© 2002 Materials Research Society
1069
http://journals.cambridge.org
Downloaded: 19 Mar 2015
IP address: 138.251.14.35

V. Corregidor et al.: Identification of Cl and Na impurities in inclusions of a vapor-grown CdTe doped with Zn and Cl
method.8,9 Undoped high-resistivity CdTe crystals con-
sidered as reference samples were grown by the Bridg-
man method as it has been reported elsewhere.19
Oriented (111) wafers of about 20 × 10 × 1.5 mm3 or
cleaved (110)-oriented samples were taken from the as-
grown ingot for comparison. The wafers were mechani-
cally polished (Al2O3 powder of 5 ␮m), followed by a
micropolish with 1.0 and 0.3 ␮m ␣-alumina in ethylene
glycol. Afterward, the samples were free etched in Br/
methanol solution (2% Br), followed by chemopolishing
with Br/ethylene glycol (0.5% Br) which gives a Te-free
surface as it was shown by Raman scattering analy-
sis.15,16 All samples were treated in the same way, prior
to being analyzed by energy-dispersive x-ray analysis
(EDAX) measurements. On the cleaved (110) surface
EDAX analyses were done without chemical treatment
of samples. These precautions were undertaken to avoid
the influence of the surface contamination by Al or Br
on the EDAX results.
The annealing processes were carried out in evacuated
and sealed off quartz ampules at 600 °C for 4 h under a
Cd vapor atmosphere. After annealing, the ampules were
withdrawn from the furnace and cooled rapidly in air.
Afterward, the samples were again chemically polished
to remove the surface layer at the depth of about 10 ␮m.
The samples were studied by EDAX with a Philips
XL30 equipment combined with secondary electron
microscopy (SEM) after deposition of a thin Au layer
(10 nm). Corrections for absorption, fluorescence, and
atomic number were performed with DX4i EDX soft-
FIG. 1. SEM image of precipitates in the (a) as-grown and (b) an-
nealed CdTe:Cl samples.
ware. Influence of the Au layer was taken into account by
comparison the EDAX signal with that from the low-
resistivity CdTe sample without Au coverage. The ho-
mogeneity of each sample was checked by SEM image
are accumulated in a high concentration (up to 30 at.%
for Cl and 10 at.% for Na). The measurements carried out
in other inclusions (columns 2A and 4A) show that the
Cd concentration is larger than the concentration of Te,
but this situation is not a general rule (see column 3A). In
this way, we can suggest that the small precipitates
would contain the CdCl2–CdTe phase, which was earlier
found by Magee et al.14 Furthermore, the Zn concentra-
tion detected in some inclusions up to a value of 3 at.%
(column 4A) is relatively high compared with the nomi-
nal doping of the feed of 2.5 × 1018 cm−3. For the com-
parison with these results, measurements of Cl, Na, and
Zn concentrations at a distance of 30 ␮m from the inclu-
sion are presented in column 5A, which practically re-
flect the concentrations in the bulk taking into account
that these data can be interpreted with some precaution
because they are at the limit of the EDAX sensitivity. All
data from the columns 1–5A correspond to the inclusions
revealed on the chemically treated (111) surface, while
data from column 6A reflect the case on the cleaved
(110) surface. Comparing these results, one must conclude
overview, followed by EDAX analysis in high spatial
resolution when some inclusion was found.
III. RESULTS AND DISCUSSION
Figures 1(a) and 2(a) present the SEM images of typi-
cal precipitates in as-grown CdTe:Cl and CdTe:Zn:Cl
samples concentrated inside of relatively big inclusions
of 10–15 ␮m. Such precipitates have either a seedlike or
oval shape and a length of about 1 ␮m. These precipi-
tates mainly disappear after the annealing process as it is
shown in Fig. 1(b) for the CdTe:Cl sample.
Figure 2(b) shows the detail of one of the oval-shaped
precipitates present in Fig. 2(a). The EDAX maps of Cl
and Na distribution on this precipitate are shown in
Figs. 2(c) and 2(d) with the same scale for presentation.
It is clearly seen that Cl is concentrated in the precipi-
tates, while Na is distributed uniformly throughout the
group of precipitates.
The results of the EDAX composition analysis made
in the inclusion shown in Fig. 2(b) are presented on
Table I, column 1A, where it can be seen that Cl and Na
1070
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 19 Mar 2015
http://journals.cambridge.org
IP address: 138.251.14.35

V. Corregidor et al.: Identification of Cl and Na impurities in inclusions of a vapor-grown CdTe doped with Zn and Cl
that the impurities revealed on the chemically treated
surface are substantially the same as the one which can
be seen on the cleaved surface, but the composition of
different inclusions can significantly differ being en-
riched with Cl or with Na in a different proportion. After
the annealing it was still possible to find inclusions, but
the presence of Cl, Na, and Zn atoms is lower pro-
nounced than in inclusions of as-grown samples (Table I,
columns 1–4B).
The data of the compositional analysis on inclusions
for undoped CdTe samples grown from the melt are
shown in Table II, for as-grown and annealed samples
FIG. 2. (a) SEM image of the inclusion in as-grown CdTe:Zn:Cl crystal. (b) Detail of the inclusion, which is marked in Fig. 2(a) by the arrow.
EDAX mapping of (c) Cl and (d) Na distribution in this inclusion.
TABLE I. EDAX data (at.%) in different inclusions of CdTe:Zn:Cl samples: columns 1A–4A for (111), 5A for an inclusion-free area, and 6A
for (110) surface.
As-grown
Annealed
Atom
1A
2A
3A
4A
5A
6A
1B
2B
3B
4B
Cd
Te
Zn
Cl
29.85
28.14
1.48
30.24
10.29
42.83
41.82
1.46
11.29
2.60
41.51
43.79
0.75
10.49
3.46
41.27
39.52
3.32
15.89
0.00
46.84
49.02
0.85
1.42
1.87
38.41
38.07
1.54
13.80
8.18
47.19
46.81
0.64
3.30
2.06
48.22
47.12
1.12
2.51
1.03
47.10
47.32
0.58
4.21
0.79
48.46
48.63
1.32
0.98
0.61
Na
TABLE II. EDAX data (at.%) in different inclusions of as-grown and annealed p-CdTe samples.
As-grown
Annealed
3B
1A
2A
3A
4A
5A
1B
2B
4B
5B
Cd
Te
Na
Si
50.67
48.14
0.00
48.45
47.32
0.59
49.65
47.06
1.25
48.01
50.67
0.00
49.45
48.32
0.59
52.04
46.37
0.78
51.10
47.32
0.53
50.47
48.00
0.58
49.67
48.22
0.66
49.31
48.91
0.49
0.57
2.04
1.04
0.51
1.04
0.50
0.63
0.40
0.91
0.81
Al
0.62
1.60
1.00
0.81
0.60
0.31
0.42
0.55
0.54
0.48
J. Mater. Res., Vol. 17, No. 5, May 2002
1071
http://journals.cambridge.org
Downloaded: 19 Mar 2015
IP address: 138.251.14.35

V. Corregidor et al.: Identification of Cl and Na impurities in inclusions of a vapor-grown CdTe doped with Zn and Cl
Inclusions with an average dimension of about
10–20 ␮m were found similar to those in the vapor-
grown crystal. Some enrichment of the inclusions with
Na, Si, and Al and deviations from the stoichiometry up
to 2 at.% of excess either of Cd or Te were registered at
the limit of EDAX error (columns 1–5A). After anneal-
ing, the Te-rich inclusions disappeared, while Cd-rich
inclusions still remained probably due to the influence
of the Cd atmosphere conditions (columns 1–5B). Nev-
ertheless, lower concentration of impurities such as Na,
Si, and Al was detected, while other usual impurities for
CdTe Bridgman growth like Fe, K, Ca, and Ag7 were
below the EDAX limit.
in the vapor-grown material is much lower than in the
feed, the influence of precipitates is larger. The esti-
mation made by Fiederle et al.13 shows that real con-
centration of electrically active Cl can be around 5 ×
1015 cm3, and in this way the inclusions would have
served for “self-cleaning” of the bulk crystal during the
cooling process, which is an important factor for obtain-
ing high-resistivity crystals.
One can also roughly estimate the Cl distribution into
the bulk during the annealing process in the following
way. Using the calculation of the impurities diffusion
length from inclusions in CdTe7 and taking into account
the diffusion coefficient for Cl20
The comparison of the EDAX data for vapor-grown
CdTe:Zn:Cl with those for the reference CdTe
grown from the melt shows that the composition of in-
clusions differs depending on the growth method. Na is
present in very small concentrations in both cases, the
concentration being higher in the vapor grown sample
whereas Al and Si were only detected in the melt-grown
sample. This difference can be explained by taking into
account the growth temperature, which is higher for the
Bridgman method than for the modified Markov tech-
nique, and the quartz ampule used in the Bridgman
method, which can be a source of other additional im-
purities. In fact Al was registered only in the reference
sample, which means that the misinterpretation of the Al,
K, and Br L lines due to their overlapping in EDAX
analyses must be excluded due to the fact that all samples
were chemically treated in the same way.
One must take into account that the EDAX data re-
ported in this paper reflect the impurity concentrations in
the inclusions but not in the bulk. Nevertheless, using the
data of the impurity concentration present on inclusions,
one can estimate the role played for the segregation on a
given impurity. Considering as one example the case of
Cl segregation, the total quantity of Cl atoms (NCl) in
inclusions shown in Fig. 2 can be estimated in a similar
way to the calculation made in Ref. 4, using the formula
D (cm2 s−1) ס (0.071 to 2.4) exp(−1.6 eV/kT)
,
and the parameters of our annealing process (T ס
600 °C and t ס 4 h), the obtained diffusion length can be
as much as 30 ␮m. This means that the homogenization
of the Cl distribution can be really reached during the
annealing process.
The latest experimental results obtained with the low-
temperature photoluminescence study21 show an increase
of the Cl- and Na-related LT PL bands after the annealing
process due to the Cl and Na impurities diffusion into the
bulk. The earlier results of Saminadayar et al.11 and
Tanaka et al.18 are in good agreement with our results.
Furthermore, the confirmation of the Cl-rich inclusions
dissolution in CdTe:Cl during annealing can be also
found in Ref. 14. More work is currently being done on
the analysis of the kinetics of Cl-rich inclusion dissolu-
tion with the help of the spatially resolved cathodolumi-
nescence and Raman scattering methods, in the same
way as it was published for undoped CdTe.15,16
IV. CONCLUSIONS
Morphological SEM study of the typical inclusions in
semiintrinsic CdTe:Cl and CdTe:Zn:Cl samples grown
by the modified Markov technique was carried out com-
bined with the composition measurement of inclusions
by EDAX. Reference undoped Bridgman-grown CdTe
samples were used. In CdTe:Cl and CdTe:Zn:Cl crystals
typical inclusions of 10–15 ␮m containing Cl and Na
impurities were found. The Cl is concentrated in 1–2 ␮m
small seedlike or oval-shaped inclusions up to 30 at.%.
The annealing of the samples in Cd-vapor atmosphere
has been found to be fruitful to obtain more uniform
samples with a Cl concentration at the level of the EDAX
sensitivity.
N
Cl ס ViNNA␳Cl(1/ACl) ,
where Vi is the volume of the “i” inclusion, N is the
density of inclusions, NA is Avogadro’s number, ␳ is
Cl
the density of the CdCl2–CdTe phase, and ACl is the
relative atom mass of Cl. In consequence, taking into
account the shape and the size of the inclusions shown in
Fig. 2(a) and if the etch-pits density N ס 102–103 cm−3,
one can calculate the value of NCl as high as 1014–
1015 cm−3, while the nominal doping concentration in the
feed is 5 × 1018 cm−3. This means that an increase of N
up to 105–106 cm−3 is needed to make the concentration
of Cl atoms in inclusions comparable with the total Cl
concentration in the bulk, which in turn means that the
segregation of Cl can influence significantly the concen-
tration of Cl in the bulk. But if the concentration of Cl
ACKNOWLEDGMENTS
This work has been partially supported by the CICYT
under Project PNE-005/2001-C and by the ESA under
Project AO-99-035. V.B. is grateful to the fellowship
from the Spanish MEC (SAB-1998-0166).
1072
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 19 Mar 2015
http://journals.cambridge.org
IP address: 138.251.14.35

V. Corregidor et al.: Identification of Cl and Na impurities in inclusions of a vapor-grown CdTe doped with Zn and Cl
12. B.K. Meyer, W. Stadler, D.M. Hoffman, P. Omling, D. Sinerius,
and K.W. Benz, J. Cryst. Growth 117, 656 (1992).
13. M. Fiederle, C. Eiche, M. Salk, R. Schwarz, and K.W. Benz,
J. Appl. Phys. 84, 6689 (1998).
14. T.J. Magee, J. Peno, and J. Beans, Phys. Status Solidi A 27, 557
(1975).
REFERENCES
1. Y. Eisen and A. Shor, J. Cryst. Growth 184/185, 1302 (1998).
2. J.B. Mullin and B.W. Stragham, Rev. Phys. Appl. 12, 267 (1988).
3. H.G. Brion, C. Mewes, J. Hahn, and U. Schaufele, J. Cryst.
Growth 134, 281 (1993).
4. P. Rudolph and M. Mu¨hlberg, Mater. Sci. Eng. B16, 8 (1993).
5. P. Rudolph, A. Engel, I. Schentke, and Grochocki, J. Cryst.
Growth 147, 297 (1995).
6. M. Laasch, T. Kunz, M. Fiederle, C. Eiche, W. Joerger, G. Kloess,
and K.W. Benz, J. Cryst. Growth 174, 696 (1997).
7. L. Pautrat, N. Magnea, and J.P. Faurie, J. Appl. Phys. 53, 8668
(1982).
15. N.V. Sochinskii, V.N. Babentsov, and E. Die´guez, in Lumines-
cence and Related Properties of II–VI Semiconductors, edited by
D.R. Vij and U. Pal (Nova Science Publisher, Inc., Commack,
NY, 1998), Chap. VI, p. 247.
16. N.V. Sochinskii, M.D. Serrano, E. Die´guez, F. Agullo´-Rueda,
U. Pal, J. Piqueras, and P. Ferna´ndez, J. Appl. Phys. 77, 2806
(1995).
8. M. Fiederle, D. Ebling, C. Eiche, D.M. Hofman, M. Salk,
W. Stadler, K.W. Benz, and B.K. Meyer, J. Cryst. Growth 138,
529 (1994).
9. T. Kunz, M. Laasch, J. Meinhaardt, and K.W. Benz, J. Cryst.
Growth 184/185, 1005 (1998).
10. W. Joerger, M. Laasch, T. Kunz, M. Fiederle, J. Meinhardt,
K.W. Benz, K.W. Benz, K. Scholz, and G. Miller-Vogt, Cryst.
Res. Technol. 32, 1103 (1997).
11. K. Saminadayar, J.M. Francou, and J.L. Pautrat, J. Cryst. Growth
72, 236 (1985).
17. P.S. Duta, C. Marin, E. Die´guez, and H.L. Bhat, J. Cryst. Growth
160, 207 (1996).
18. A. Tanaka, Y. Masa, and M. Kawashima, J. Cryst. Growth 117,
271 (1992).
19. N.V. Sochinskii, M.D. Serrano, V.N. Babentsov, N.I. Tarbaev,
J. Garrido, and E. Die´guez, Semicond. Sci. Technol. 9, 1713
(1994).
20. D. Shaw and E. Watson, J. Phys. C 17, 4945 (1984).
21. V. Corregidor, V. Babentsov, J.L. Castan˜o, M. Fierdele, T. Feltgen,
K. Benz, and E. Die´guez (to be published).
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 19 Mar 2015
1073
http://journals.cambridge.org
IP address: 138.251.14.35