10548 J. Am. Chem. Soc., Vol. 121, No. 45, 1999
Cohen et al.
PLmol are the PL intensity for the bare and molecule-adsorbed surface,
respectively, R′ is the sum of the semiconductor absorption coefficients
for the exciting and emitted light, and ∆D is the adsorption-induced
difference in the dead layer thickness.
These values are higher than those of benzoic acids onto the same
surface (∼2 × 104 M-1
)
40 or onto CdTe (1.3 × 103 M-1),45 for processes
best described by a one-site process.
For the dicarboxylic acids used here the FTIR spectra did not reveal
any spectral differences between adsorption of DHDC and DCDC. We
can speculate that the difference in the binding constant between DCDC
and DHDC is related to the strength of the orbital coupling with the
surface states and, thus, correlated with changes in Vs. However, more
information is needed to verify this issue.
Opto-Electronic Measurements. The semiconductor surface work
function was determined in a contactless, nondestructive manner by
CPD measurements, using a Kelvin probe arrangement (Delta-Phi
Elektronik).36 The CPD is defined as the work function difference
between the sample surface and an inert reference electrode, in this
case made of Au and with a known work function of 5.1 eV. All
measurements were conducted in the ambient inside a homemade
Faraday cage. Surface photovoltages were determined by photoinduced
changes in the CPD.
The SRV and Vs changes were correlated with changes in the surface
state positions relative to the band edges. The latter were extracted
from surface photovoltage spectroscopy (SPS) measurements. In SPS,
sharp slope changes in the surface photovoltage vs photon energy
curves, at sub-band gap photon energies, are attributed to the onset of
electron transitions between surface states and band edges.34-36
A
negatiVe change in slope implies electron excitation from a filled surface
state to the CB minimum, whereas a positiVe change in slope implies
electron excitation from the VB maximum to an unoccupied surface
state. At the band gap energy, another sharp slope change indicates
the onset of super-band gap absorption, which reduces Vs. A negative
slope change indicates a surface depletion layer of an n-type material
(bands bent upward toward the surface), whereas a positive slope change
indicates a surface depletion layer of a p-type material (bands bent
downward toward the surface).
High-intensity white illumination of ∼200 mW/cm2 was obtained
from a quartz tungsten-halogen lamp for the photosaturation experi-
ments. For SPS measurements, monochromatic illumination was
generated by passing light from a 600 W quartz tungsten-halogen lamp
(Oriel) through a grating monochromator (Model 270M, Jobin Yvon)
and an automated filter wheel (Model AB300, CVI) and focusing it on
the sample. Because the SPV induced by super-band gap illumination
was very large with respect to that induced by sub-band gap illumina-
tion, a strong attenuation of the second-order intensity of the mono-
chromator was required. Optical filters were therefore chosen such that
the second-order attenuation was at least 10 orders of magnitude. The
photon energy range was 0.6-2.0 eV (i.e., from sub-band gap to super-
band gap energies). The illumination intensity on the sample was lower
than 20 µW/cm2, guaranteeing a low injection level at all photon
energies. Due to sample sensitivity to intensity variations, noted
experimentally after replacements of gratings and filters, SPV spectra
were taken using a constant photon flux (except for the CdTe samples).
The constant photon flux was obtained by using a pyroelectric
photodetector as the input of a computerized feedback loop that
regulated the voltage supply to the lamp at each wavelength.
In some of the samples, the output of the constant photon flux setup
was of too low an intensity to observe sub-band gap transitions. In
those cases, a second SPS scan without a constant photon flux was
performed over the pertinent range of sub-band gap energies, where a
further increase of the illumination intensity was achieved by placing
a condensing lens before the sample.
Time-resolved photoluminescence measurements were performed
using the time-correlated single photon counting technique. The
excitation source was a Ti-Sapphire laser (Tsunami, Spectra Physics
Inc.) with an excitation wavelength of 730 nm (beam diameter: 1 mm;
pulse energy: 5.5 nJ; pulse duration: 1.2 ps). For CdSe, an excitation
wavelength of 386 nm (pulse energy: 0. 2 nJ), obtained by frequency
doubling the fundamental laser output, was used. The detection system
is based on a double subtractive monochromator (CM112, CVI) and
an MCP-Photomultiplier (R3809, Hamamatsu). Detection wavelengths
for the different semiconductors were 720, 920, and 870 nm for CdSe,
InP, and GaAs, respectively. The TRPL measurements for the CdTe
crystals were performed using a Nd:YAG pumped dye laser (excitation
wavelength: 600 nm; beam diameter: 1 mm; pulse energy: 2 nJ; pulse
duration: 1-2 ps). The detection wavelength was 820 nm. More details
on the experimental setup have been given elsewhere.46,47
Finally, the position of the surface states with respect to the vacuum
level was estimated by combining the photosaturation and surface
photovoltage (SPV) data with direct measurements of the surface work
function, obtained via the Kelvin probe technique.36 This allowed us
to estimate the surface state positions with respect to the positions of
the DCDC and DHDC LUMO levels quantitatively.
Materials. Unintentionally p-doped (specific resistivity: 30 Ω‚cm)
CdTe(111) single crystals were obtained from II-VI Inc., USA, and
were In-doped to n ) (1-5) × 1017. Details of the doping procedure
have been published elsewhere.37 The results of various measurements
performed on n-CdTe differed significantly between two different
manufacturing batches, even though these two batches were processed
and measured under the same conditions. These two variations are
referred to below as “type I” and “type II”. All other crystals were p-
or n-doped as purchased. These included the following: n-CdSe(0001)
(7 × 1015 cm-3, Cleveland Crystals, USA); n-GaAs(100) (3 × 1018
cm-3, AXT, USA); n-InP(100) (4 × 1016 cm-3, Crystacomm, USA);
and p-GaAs(100) (2 × 1017 cm-3, ITME, Poland).
Surface Treatments. CdTe and CdSe samples were first mechani-
cally polished with a 0.05 µm alumina suspension, whereas no such
polishing was required for the GaAs and InP samples. All samples
were subsequently etched chemically. Etching procedures of the CdTe,
CdSe, and GaAs surfaces are described in refs 38-40, respectively.
InP surfaces were etched according to the GaAs procedure. All etching
treatments were performed on freshly cut substrates. Back (Ohmic)
contacts to the samples were obtained using a eutectic (In,Ga) alloy.
Chemisorption of the dicarboxylic acids was performed immediately
after etching by overnight immersion of the samples in a 2.5 mM
molecular solution in acetonitrile (HPLC and spectroscopic grade). The
samples were then rinsed with a pure acetonitrile solution for 10 s to
remove excess unbound molecules, resulting in a surface coverage of
about one monolayer, verified by FTIR.41-43 Longer or additional
acetonitrile rinse did not change the surface coverage.
Using adsorption isotherms derived from FTIR spectra and from
special electrical measurements, Vilan et al. found binding of DCDC
and DHDC onto GaAs(100) to be best described by a two-site process,
with binding constants of 3 × 106 and 3 × 105 M-1, respectively.43,44
(34) Gatos, H. C.; Lagowski, J. J. Vac. Sci. Technol. 1973, 10, 130-
135.
(35) Lagowski, J. Surf. Sci. 1994, 299/300, 92-101.
(36) Kronik, L.; Shapira, Y. Surf. Sci. Rep. In press.
(37) Lyahovitskaya, V.; Kaplan, L.; Goswami, J.; Cahen, D. J. Cryst.
Growth 1999, 197, 106-112.
(38) Bruening, M.; Moons, E.; Yaron-Marcovich, D.; Cahen, D.; Libman,
J.; Shanzer, A. J. Am. Chem. Soc. 1994, 116, 2972-2977.
(39) Bruening, M.; Moons, E.; Cahen, D.; Shanzer, A. J. Phys. Chem.
1995, 99, 8368-8373.
(40) Bastide, S.; Butruille, R.; Cahen, D.; Dutta, A.; Libman, J.; Shanzer,
A.; Sun, L.; Vilan, A. J. Phys. Chem. 1997, 101, 2678-2684.
(41) Cohen, R.; Bastide, S.; Cahen, D.; Libman, J.; Shanzer, A.;
Rosenwaks, Y. Opt. Mater. 1998, 9, 394-400.
(42) Bastide, S. Unpublished results, Weizmann Institute of Science.
(43) Vilan, A. Chemical Modifcation of the electronic Properties of GaAs
Surface. M. Sc. thesis, Weizmann Institute of Science, Rehovot, 1996.
Intensity-resolved PL measurements were performed in a nitrogen
atmosphere using either a HeNe (Model 80, Melles-Griot) or an Ar+
(Innova 90, Coherent) laser, with excitation wavelengths of 633 (HeNe),
514 (Ar+), and 458 nm (Ar+). The PL signal of CdSe at 720 nm was
collected by a CCD array (Instaspec II, Oriel).
(44) Vilan, A.; Ussyshkin, R.; Gartsman, K.; Cahen, D.; Naaman, R.;
Shanzer, A. J. Phys. Chem. B 1998, 102, 3307-3309.
(45) Yaron-Marcovich, D. Studies of the Adsporption Mechanism of
Polar Ligands on the Surface of Chalcogenide Semiconductors. M. Sc. thesis,
Weizmann Institute of Science, Rehovot, 1993.
(46) Liu, A.; Rosenwaks, Y. Submitted for publication.
(47) Rosenwaks, Y.; Shapira, Y.; Huppert, D. Phys. ReV. B 1991, 45,
9108-9119.