HAL Id: jpa-00226687
https://hal.archives-ouvertes.fr/jpa-00226687
Submitted on 1 Jan 1987
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
POSITION AND CHARACTER (D OR X) OF ENERGY STATES IN SHORT-PERIOD
(GaAs)m(AlAs)n SUPERLATTICES
J. Nagle, M. Garriga, W. Stolz, T. Isu, K. Ploog
To cite this version:
J. Nagle, M. Garriga, W. Stolz, T. Isu, K. Ploog. POSITION AND CHARACTER (D OR X) OF ENERGY STATES IN SHORT-PERIOD (GaAs)m(AlAs)n SUPERLATTICES. Journal de Physique Colloques, 1987, 48 (C5), pp.C5-495-C5-498. �10.1051/jphyscol:19875105�. �jpa-00226687�
JOURNAL DE PHYSIQUE
Colloque C5, supplBment au noll, Tome 48, novembre 1987
POSITION AND CHARACTER (l- OR X) OF ENERGY STATES I N SHORT-PERIOD (GaAs)m(AlAs)n SUPERLATTICES
J. NAGLE, M. GARRIGA, W. STOLZ, T. ISU'') and K. PLOOG
Max-Planck-Institut fiir Festkorperforschung, D-7000 Stuttgart 80, F.R.G.
Rdsumd - Nous avons realise la croissance par dpitaxie par jets mol6culaires d'une sBrie de superrBseaux de courtes pBriodes (GaAs)m(AlAsln incluant des Bchantillons avec m et n diffdrents. Des mesures basse temperature de photoluminescence, d'excitation de la photoluminescence et d'ellipsometrie ont dt6 effectudes pour Btablir la position et la nature des differents gaps dlectroniques.
A m - We have grown a series of ( G ~ A s ) ~ ( A ~ A s ) ~ short period superlattices by molecular beam epitaxy including samples with m = n and mfn. L o w temperature photoluminescence, excitation spectroscopy and ellipsometry measurements have been performed t o determine t h e positions and origins of t h e different electronic states.
T h e interest in ( G a A s ) r n ( A l A ~ ) ~ short-period superlattices (SPSL's) consisting of a periodic arrangement o f m monolayers GaAs and n monolayers AlAs with m,n<lO, has recently increased because t h i s material appears t o be a good candidate t o replace AlxGal-xAs alloys in electrical and optical devices. Following t h e pioneering work o f Gossard et a1 .', SPSL's have been successfully grown by molecular beam epitaxy ( M B E ) ~ , ~ and metalorganic chemical vapor deposition ( M O C V D ) ~ . However, t h e knowledge of the fundamental band gap of these n e w materials is still
unsatisfactory, although this information is a prerequisite t o their application in band-gap-engineered structures. In particular the t y p e (F, X o r L) of the fundamental gap is not well established up t o now. T h e e ~ p e r i m e n t a l ~ , ~ and theoretical6 reports are contradictory concerning (GaAs )m(AlAs)m SPSL 'S with m)2. W e discuss t h e MBE growth of a series of SPSL's together with various optical measurements which allow an unambiguous determination o f the position and type of t h e lower-lying electronic states in these materials.
T h e ( G a A s l m ( A 1 A ~ ) ~ SPSL's were grown on ( 0 0 1 ) oriented GaAs substrates. The growth procedure is described in detail in Ref.3 for t h e case of m=n. We use a growth temperature of 5500C '10oC as measured by an infrared pyrometer calibrated through the oxygen desorption temperature o f GaAs substrates (taken t o be 580oC). T h i s temperature is found t o be reproducible from r u n t o run when using the substrate cleaning technique described in Ref.8. The temperature is kept relatively low t o prevent interdiffusion o f Ga and A1 atoms between alternating layers. We use low growth rates of typically 0.2 um/hr for both GaAs and AlAs t o compensate for the reduced surface mobility o f t h e adsorbed atoms. At each interface t h e growth is
interrupted f o r 5 t o 1 0 s under constant As4 pressure. It has been shown9 for large-period GaAs/AlAs SL's grown under t h e same conditions t h a z a 10s growth interruption provides a sharpening o f t h e X-ray satellite peaks indicating a smoothening of t h e interface and that further increas-e o f t h e interruption t i m e does not lead t o better results.
(''present address : Optoelectronics Technology Research Laboratory, Toyosato. Tsukuba.
Ibaraki 300-26. Japan
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19875105
C5-496 J O U R N A L DE PHYSIQUE
The growth rates of AlAs and GaAs are determined accurately prior to the growth from the period of the intensity oscillations of the diffraction spots in the RHEEO pattern. During growth of the SPSL's we observe a 2x4 surface reconstruction
characteristic for the
- L _ . _ . I . . . I . I . . I . . . . ,
50 1 0 0 150
T i m e Is1
Fig.1 RHEED o s c i l l a t i o n s r e c o r d e d during t h e g r o w t h o f t h e S P S L ' s ( s p e c u l a r r e f l e c t i o n a l o n g t h e <110, direction).
two-dimensional layer-by-layer growth mode. The RHEED intensity oscillations recorded during growth remain constant for the entire run indicating the excellent periodicity of the layer sequence and the stability of the growth conditions
(Fig.1). We did not rotate the substrate because the
synchronisation of the shutter opening time with an integer number of substrate rotations is not possible for all the samples without changing the temperature of the effusion cells. This induces a small inhomogeneity of the thickness and stoichiometry of the SPSL's across the wafer (typically a few percent for the relatively small substrates used in the present study).
Structural characterisation
The stuctural properties of all samples were systematically investigated using X-ray diffraction. A high-resolution double-crystal diffractometer was used to determine the average A1 composition of the entire epitaxial layer from the position of the n = 0 SL peak in the vicinity of the (004) Bragg reflection. A counter diffractometer allowed a quick finding of the satellite peaks associated with the SL around the (002) and (004) reflections. The intensity and sharpness of the satellite peaks confirm the smoothness of the interfaces, and the position of these peaks determine the period of the SPSL. Typical K-ray spectra are presented
in Ref.3 for the case of m=n SPSL's.
The result of the measurements are shown in Table 1 together with the nominal stucture of the samples. Considering the precision of these measurements (1%
absolute precision for the mean A1 composition and less than 1% relative precision for the period), it is relevant to indicate the measured m and n with two digits after the decimal point particularly for SPSL's with very small periods.
Sample nominal X-ray PL intrinsic PLE peak or E0 gap
(m-n) m n peak(eV) threshold(eV) (eV)
#5867 lOOX(7-7) 7.18 7.38 1.831 ( X ) 1.95
X5868 350X(2-2) 2.01 2.06 2.070 (X) 2.19 2.18
X5869 467X(2-1) 2.06 0.75 1.815 (I') 1.83 1.84
#5870 lOOX(7-7) 6.99 7.07 1.835 (X) 1.97 1.94
X5871 350X(3-1) 3.02 0.87 1.776 ( r ) 1.82
#5872 467X(1-2) 0.74 1.96 2.101 (X) 2.48 2.53
X5873 35OX(1-3) 0.69 3.08 2.127 (X) 2.59 2.60
15894 377X(11-6) 10.49 6.29 1.769 (I') 1.79 1.77 X5897 233X(4-2) 3.83 2.02 1.905 (l") 1.92
X5902 7OOX(1-1.5) 0.89 1.61 2.086 (X) 2.35 2.27
Table 1. Characteristics and optical transitions of the SPSL samples The E 0 gap values have been determined by ellipsometry.
O ~ t i c a l ~ r o o e r t i e s
Photoluminescence (PL) measurements were T = 4 6 K
made at T-4.6 K. Excitation was performed -
with the 476.2-nm line o f a K r C ion -
laser. The luminescence is detected using K
a lm-grating-spectrometer equiped with a 5 - #5868 - cooled photomultiplier. T h e
photoluminescence excitation (PLE)
measurements were performed using a 500W -
high-pressure Xe lamp filtered through a
double 0.60111-monochromator with the -
sample located 4 5 0 off-incidence and the (PLE)
detection 9 0 0 off-incidence. The typical
excitation power for P L E was 25 mW/cm2 -
with a spectral resolution o f 1 nm. Since
t h e spectrum o f the lamp is fairly flat
1
in t h i s region, w e did not correct our 1
spectra f o r the excitation density.
Ellipsometric measurements w e r e performed 18 1 9 2.0 2 1 2.2 2.3 2.4 2 5 I I I I I I I at T z 3 0 K using an automatic rotating
analyser spectro-e1 lipsometer .'l The ENERGY (eV) change in the band gap energy between Fig.2 Photoluminescence(PL) and
excitation spectra(PLE) of t w o indirect T-4.6 K and T z 3 0 K is expected t o be only SPSL's. T h e P L s p e c t r a are recorded a f e w mev s o that it is possible t o using t h e Kr+ 476-nm line with a power directly compare the different results. density of 70 u/cmZ.
It is important t o note that the optical
measurements were all made o n the same samples that w e r e characterised with X-ray diffraction. The results o f t h e measurements are presented in T a b l e 1 together with t h e structural data. The position o f the intrinsic luminescence peak is deduced f r o m t h e evolution o f the PL spectrum with excitation intensity between 2 5 mW/cm2 and 7 0 W/cmZ. With the term 'intrinsic' we exclude impurity related and phonon assisted transitions but we include bound excitons s o that we may
underestimate t h e corresponding gap by an amount equal t o t h e binding energy o f t h e exciton plus any additional binding energy.
S o m e samples(label1ed 'l in Table 1) exhibit a strong luminescence and an excitonic peak in P L E with moderate Stokes shift similar t o t h e case of usual quantum wells.
The states involved in these optical transitions are obviously states.
The other samples (labelled X) exhibit a large shift between the luminescence and t h e PLE threshold. We attribute t h e luminescence t o X states mainly localized in t h e AlAs layers5,", and t h e PLE threshold t o t h e onset of the r transition. This interpretation is confirmed by the determination o f t h e E 0 gap by ellipsometry which gives the first gap having a significant absorption, i.e. the r gap at k=O.
Taking into account the typical errors o f the determination of the PLE threshold ('10 meV) and E 0 gap ('20 meV), t h e agreement is very good f o r t h e two
measurements of Egr. Fig.2 shows t h e PL and P L E spectra for the (GaAs)z(AlAs)Z and (GaAs)7(AlAs)7 samples. The luminescence associated with t h e l" states is directly observable in these indirect-gap samples. The other lower lying energy peaks which are not labelled are probably due t o phonon assisted transitions riding over an impurity related background. These peaks are not caused by slightly different periods present in the sample since their excitation spectra are similar and
indicate t h e same absorption threshold. It should be noted that the 'indirect' samples d o not always exhibit a weak luminescence.
In Fig.3 w e compare our results with a simple Kronig-Penney (KP) model. Following Danan et al. l' w e calculated separately the energles o f the .confined r and X
states using t h e following parameters f o r GaAs: Egr=1519 meV, E g x = 1 9 8 3 meV, mer=0.0657(m0), mhh-0.377, m e x l = l .3, mext=0.23, and f o r AlAs: E g r = 3 1 1 3 meV, Egx=2228 meV, mera0.228, m h h d . 4 7 8 , meX1'1.1, mext=0.19 . T h e curves are calculated with a 5 5 8 m e V valence band-offset corresponding t o a 65% offset parameter and include a simple non-parabolicity correction for t h e r electrons.
T h e agreement is surprisingly good for t h i s calculation that neglects interaction and mixing o f X and l" states. Furthermore, as pointed out by 1hmi4, the Xz state in t h i s model (where z represents t h e growth direction) is always located at lower energy than the Xxy state due to t h e lower transverse effective m a s s at t h e X
C5-498 JOURNAL DE PHYSIQUE
8."
5 10 15
m (number of monolayers 1
Fig.3 Comparison between experimental results for ( G ~ A S ) ~ ( A ~ A ~ ) ~ SPSL's and the Kronig-Penney model (for parameters see text). The m=4 and m-15 points are from Ref.13. The arrows indicate the position of the A10.5Ga0.5As alloy band gaps.
2.4
X l n t r l n s ~ c PL peak IT:46K1
+ Threshold or excltonlc peak m PLE IT=46K)
Ell~psometrrc E. gap
lT=30Kl 2.3
16
01 0 2 03 04 05 0 6 0 7 Q8 0 9 AVERAGE COMPOSITION IXAO
(GaAs), (ALAS), SL T = 4.6 K
- '\,
'\ x Intrinsic P L peak
Fig.4 Comparison between experimental band gaps of several ( G ~ A S ) ~ ( A T A S ) ~ SPSL's and the calculated band gaps of the corresponding alloys. The black dot indicates the PLE excitonic peak for a ternary alloy sample grown under similar conditions.
Threshold in P L E
- KP calculation
&Ev = 5 5 8 m e V (off set : 65 'l0
1.7 -
point. Fig.4 shows that the alloy approximation seems t o be an adequate model in t h e c a s e o f very short periods (m+n<6) and predicts the position and t y p e o f t h e gap quite accurately. W e used t h e expression o f C a s e y and Panish" f o r t h e evaluation o f the F gap and t h e following expression f o r t h e X gap o f t h e alloy:
E x(eV) = 1.986 + 0.113 x ~ 1 + 0.129 xA12. Band gap properties from these very stort period SL's seem t o be close t o those o f the ternary alloy. However, other interesting properties o f these materials d o exist, like t h e suppression o f t h e alloy disorder o r t h e lifting of t h e valence band degeneracy.
gonclusion
We have grown and characterised a series of ( G ~ A s ) ~ ( A I A s ) ~ SPSL's. RHEED oscillation data and X-ray diffraction measurements give evidence of the good quality o f the grown layers. We determine t h e band gaps o f these structures using PL, P L E and ellipsometry. In particular, t h e (GaAs)z(AlAs)2, (GaAs)%(AIAs)g, (GaAs)d(AlAs)4 SPSL's are clearly shown t o be indirect in contradiction with various previous experimental and theoretical works. The 'l and X gaps o f t h e SPSL's with m+n(6 are very close t o those o f t h e corresponding alloys. F o r SPSL's with m+n)6 a Kronig-Penney calculation including a separate evaluation o f confined l
' a n d X states gives a surprisingly good estimate of t h e band gap. - . - References
1 A.C.Gossard, P.R.Petroff, W.Wiegmann, R.Dingle,and R.Savage, Appl. P h y s Lett. B , 3 2 3 ( 1 9 7 6 )
N.Sano, H.Kato, M.Nakayama, S.Chika,and N.Tereauchi, Jpn. J. Appl. Phys.
=,1640(1984)
T.lsu, D.S.Jiang,and K.Ploog, Appl. Phys. A B , 7 5 ( 1 9 8 7 )
A.Ishibashi, Y.Rori, N.Itabashi,and N.Watanabe, J. Appl. Phys. s , 2 6 9 1 ( 1 9 8 5 ) E.Finkman, M.D.Sturge,and M.C.Tamargo, Appl-Phys-Lett. B , 1 2 9 9 ( 1 9 8 6 )
T.Nakayama and H.Kamimura, J. o f the Phys. Soc. o f Japan =,4726(1985) M.A.Gel1, D.Ninno, M.Jaros,and D.C.Herbert, Phys. Rev. B 2 , 2 4 1 6 ( 1 9 8 6 ) K.Ploog and A.Fischer, Appl. Phys. Lett. 3 , 1 3 9 2 ( 1 9 8 6 )
L.Tapfer and K.Ploog, Phys. Rev. B a , 5 5 6 5 ( 1 9 8 6 )
M.Garriga, R.Cardona, N.E.Christensen, P.Lautenschlager, T.Isu,and K.Ploog, (to be published in Phys. Rev. B)
G.Danan et al., Phys. Rev. B E , 6 2 0 7 ( 1 9 8 7 )
H.C.Casey and M.B.Panish in Heterostructure Lasers, Part.A (Academic, New York, 1978). pp.192-3
D.S.Jiang, K.Kelting, T.Isu, H.J.Queisser,and K.Ploog, ( t o be published in J.Appl.Phys.)
J.lhm, Appl.Phys.Lett. =,1068(1987)
