CHARACTERIZATION OF SURFACE STATES AT A SEMICONDUCTOR ELECTROLYTE INTERFACE BY ELECTROLYTE ELECTROREFLECTANCE SPECTROSCOPY
ABSTRACT Les états de surface aux interfaces semiconducteur-liquide sont caractérisés par électroréflectance électrolytique supra- et sub-bande interdite. Les états de surface peuvent se manifester, soit par des transitions optiques directes comme dans le cas de n - TiO2 - électrolyte aqueux, soit par leur effet sur la réponse du niveau de Fermi à de faibles variations du potentiel d'électrode comme dans le cas de monocristaux CdIn2Se4 dans des solutions de polysulfures. Supra bandgap and subband gap Electrolyte Electroreflectance is being used to characterize surface states at semiconductor liquid interfaces. The surface states can manifest themselves either through direct optical transitions as in the case of n - TiO2 - aqueous electrolyte interface or through their effect on the response of the Fermi level to small changes in the electrode potential as in the case of single crystal CdIn2Se4 in polysulfide solutions.
JOURNAL DE PHYSIQUE
Colloque CIO, supplément au n°12, Tome W, décembre 1983 page C10-163
CHARACTERIZATION OF SURFACE STATES AT A SEMICONDUCTOR ELECTROLYTE
INTERFACE BY ELECTROLYTE ELECTROREFLECTANCE SPECTROSCOPY*
M. Tomkiewicz and Wt Siripala
Department of Physios,
Brooklyn College of CUNY, Brooklyn, New York 11210,
Résumé - Les états de surface aux interfaces semiconducteur-liquide sont caractéri-
sés par électroréflectance électrolytique supra- et sub-bande interdite. Les états
de surface peuvent se manifester, soit par des transitions optiques directes comme
dans le cas de n - TiO„ - électrolyte aqueux, soit par leur effet sur la réponse
du niveau de Fermi à de faibles variations du potentiel d'électrode comme dans le
cas de monocristaux CdIn„Se, dans des solutions de polysulfures.
used to characterize surface states at semiconductor liquid interfaces. The sur-
face states can manifest themselves either through direct optical transitions as in
the case of n - Ti02 - aqueous electrolyte interface or through their effect on the
response of the Fermi level to small changes in the electrode potential as in the
case of single crystal CdT^Se/ in polysulfide solutions.
Supra bandgap and subband gap Electrolyte Electroreflectance is being
We will present here, two modes in which Electrolyte Electroreflectance (EER) is
being used to detect and characterize surface states at the semiconductor electro-
lyte interface. In the first mode, which is being demonstrated on single crystal
Ti02, direct optical transitions between the surface states and the conduction band,
are being observed. In the second mode, demonstrated here on single crystal
CdIn2Se4 in polysulfide solution, the surface states are responsible for Fermi level
pinning which quenches the EER signal.
Sub-Bandgap EER -. Ti02: The potential distribution at the Ti02 aqueous electrolyte
interface with particular emphasis on the surface states and their dependence on
various electrolytes was investigated in detail in our laboratory, using a variety
of techniques. These techniques include impedance spectroscopy
sub-bandgap photorepsonse spectroscopies (2) ; photoelectrochemistry with single and
double beam excitation") and EER™) . Three main groups of surface states were
identified: One state that tails from the conduction band edge and is primarily
responsible for the recombination of light generated minority carriers(3), second
state which resides 0.8 eV below the conduction band and is being controlled by
and a third state 1.3 eV below the conduction band which can
be observed only when an adsorbing anion penetrates into the inner Helmholtz
layer") and is responsible for catalysis of water oxidation. This last state is
the subject of the work that will be presented in this section. Whenever compari-
sons can be made, satisfactory agreement exist between UPS measurements under
UHV(5) an(j the in situ measurements. Details about the electrode, cell and the expe-
rimental techniques were previously publishedC--^) . The doping level of the TiC>2
Is 5 x 10l9/cm3, Fig. 1 shows the EER spectra of TiC^ in various electrolytes in
the sub-bandgap region. All spectra were taken at a potential which shows the maxi-
mum response. The broad peak is centered around 1.3 eV and the peak position does
not change with the electrolyte but the intensity of the peak does. Within the
halogen series it follows the same trend as the expected strength of adsorption,
if the latter is dominated by substitution of the hydration shell of the ionsW.
In all cases, the potential of the maximum response is 0.1 - 0.3 V more negative
than the corresponding flatband potential. We interpret these results by conclud-
ing that the sub-bandgap EER originates from optical transitions between filled
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19831032
JOURNAL DE PHYSIQUE
surface states, located 1.3 eV below the conduction band, and the conduction band.
The EER signal can be detected only under conditions i n which significant portion
of the potential drop can modulate these transitions.
is independent of the nature of the adsorbed species, but the intensity is strongly
dependent on the electrolyte. In t h i s sense the states must be intrinsic t o the
semiconductor and independent on the nature of the adsorbates. When the surfaceis
fully solvated and the Helmholtz is free of adsorbed species, the surface states
w i l l interact with the solvent t o s h i f t their energies toward the band edges.
the adsorbed species substituting for the solvent, the states w i l l resume their
intrinsic energy, provided that the adsorption i s purely electrostatic with neg-
ligible orbital interactions between the surface states and the adsorbate. Quanti-
tative comparison between the density of these states and ( N ~ ) ~ / ~
the UPS work(5) that identify these states as defect states.
The position of the EER peak
give support to
FEN1 Level Pinning - CdIn2Seq:
signal is given by: (7)
In the low field regime the electroreflectance
where L(fiw) is a spectral lineshape function, Vsc - t h e modulated voltage across the
space charge layer, ND - the density of the ionized donors, E, - the s t a t i c permit-
t i v i t y and e the electronic charge. The potential drop across the space charge
layer can be expressed as:
where U is the electrode potential and Ufb the flatband potential.
the electrode potential will be divided between the space charge layer and the
Helmholtz layer according t o
where 4I0 is the potential drop across the Helmholtz layer.
tential drop across the Helmholtz layer due t o a change i n the number of ionized
surface states is given by ( 9 ) :
The change i n the po-
where CH is the capacitance of the Helmholtz layer and N i S - the density of the
ionized surface states. Combining (I), ( 2 ) , (3) and (4) w i l l result in:
Cu . du
ahy change i n the electrode potential w i l l result i n corresponding change i n the
potential drop across the space charge layer, while 5 dNgS = 1 represent the
= 0 will represent the condition when the Fermi level is unpinned and
condition i n which the Fermi level is completely pinned and the EER signal w i l l be
reduced t o zero.
dNSs can be directly determined from the
potential dependence of the EER signal.
Since changes i n the flatband potential
With K known, - -
originate from the changes i n the potential drop across the Helmholtz layer i.e.,
6@ = 6Ufb it follows from Eq. (4) and (5) that the changes i n Ufb with the applied
potential can be directly evaluated from EER without assuming any energy distribu-
tion of the surface states. Figure 2 gives the variation with potential of the EER
signal of single crystal CdIn2Se4 i n polysulfide solution with native surface and
with photoetched surface. Full experimental details about t h i s system and detailed
analysis of the EER spectra w i l l be published elsewhere(lO). Figure 3 shows the
variations of Ugh with the applied potential that were calculated from the data i n
Fig. 2 by the method that was outlined here.
pinning is evident under reverse bias conditions. W e have evaluated K here, by
assuming that a t the potential where the signal is maximum there is no pinning.
Detailed analysis of these curves, assuming Gaussian distribution of the surface
states, strongly suggest that the pinning is due t o a complete monolayer of ad-
sorbed sulfur ions. (10)
Considerable degree of Fermi level
' 1 1 2 ' lj4 ' 1 1 6 ' lh '>lo
Photon Energy (eV)
' 212 ' 214'
Figure 1 - Sub-band-gap EER spectra for n - Ti02 i n different electrolytes. Modu-
lating voltage is 200 mV peak t o peak.
Figure 2 - Variation of the EER signal with the potential for CdIn2Se4 i n aqueous
Photoetching (photon energy = 2.0 eV).
Modulating voltage = 0.2 V peak t o peak.
(a) - Before Photoetching (photon energy = 1.8 eV) (b) - After
Figure 3 - Shift i n the flatband potential with the potential for CdIn2Seq i n
aqueous polysulfide, before (a) and after (b) photoetching.
JOURNAL DE PHYSIQUE
References - 1.
2. SIRIPALA W. and TOMKIEWICZ M.; J. Electsoch. Soc. 2,
3. SIRIPALA W. and TOMKIEWICZ M.; J. Electroch. Soc. 130, 1062 (1983).
4. SIRIPALA W. and TOMKIEWICZ M. ; Phys. Rev. Lett. 50, 4 L 1 3 (1983).
5. HENRICH V. E. AND KURTZ R. L.; Phys. Rev. g 3,
6. BOCKRIS J. O'M and E D D Y A. K. N.; Modern Electrochemistry (Plenum, N. Y. 1970
7. ASPNES D. E.; Surf. Sci. 37, 418 (1973).
8 . PETTINGER B., SCHOPPEL H. R., POKOYAMA
9. MYAMLIN V. A. and PLESKOV Y. V.; "Electrochemistry of Semiconductor" (Plenum,
N. Y. 1967) Ch. 1.
10. TOMKIEWICZ M., SIRIPALA W. and TENNE R.; J. Electroch. Soc. (1933) - in press.
* This work was supported by the Solar Energy Research Institute.
TOMKIEWICZ M.; J. Electroch. Soc. 126, 2220 (1979).
T. and GERISCHER H. ; Ber. Bun-Gesell 78,