Supplemental information for:
Towards Metal-Organic Insulator-Semiconductor Solar Cells, based on Molecular Monolayer Self-assembly on n- Si
Rotem Har-Lavan, Izhar Ron, Florent Thieblemont, David Cahen[*]
Experimental
Alkoxy and alkoxy-thiol monolayers self-assembly
Samples were prepared from single side polished Si (100) with a nominal resistivity of 1-10 Ω.cm (Virginia Semiconductor), cleaved into ~10 x 30 mm2 pieces. The Si pieces were first rinsed with organic solvents, blown dry with nitrogen, cleaned in piranha solution (concentrated H2SO4 : 30% H2O2, 7:3 v/v) at 90°C for 1 hour, rinsed with deionized water and etched in 2% HF for 1 min to get an H-terminated surface. The piranha treatment, the rinsing and the etching were done twice, to assure reproducibility. After the surface was dried, the freshly etched samples were immersed in a solution of commercially available C12H25OH (dodecanol, 98.5% pure) and SH-C11H22OH (1,11-mercapto-undecanol, 97% pure), both from Sigma. This solution was previously cleaned of water traces by the addition of a drying agent (chlorotrimethylsilane 10% v/v), degassed by a few freeze/thaw cycles under vacuum and kept under inert atmosphere. The reactants were kept at 90°C for 16 hours. After the reaction the samples were cleaned, first by copious amounts of ethanol and then by immersion in boiling dichloromethane to remove physisorbed molecules from the monolayer.
Adsorbing Au nano-particles (Au NPs).
Small 5 x 5 mm samples were cleaved from the larger SAM-covered Si pieces, with minimal delay after monolayer adsorption. Those samples were then immersed again in boiling dichloromethane to remove Si dust and particles, blown dry by a nitrogenflow, and left in the dark in a toluene suspension of tetraoctylammonium bromide-capped Au nanoparticles (NPs), 3-5 nm in diameter, for 5 hours to form a very dense Au NP monolayer on the molecules (see Fig. 1S a, below). The toluene suspension was chosen because the thiol end group of the molecules is easily oxidized by aqueous suspensions and thus will result in poor coverage of the sample surface. After adsorption the samples were thoroughly rinsed with toluene and blown dry with a nitrogen flow.
Electro-less Au plating.
Substrates covered with a monolayer of Au NPs were immersed in 6 ml of aqueous 0.4 mM hydroxylamine hydrochloride and 0.1% HAuCl4·3H2O. All glassware was thoroughly cleaned with aqua regia. The solution was stirred to ensure the formation of a homogeneous Au film. The samples changed color from the typical one of polished Si to that of gold. SEM measurements showed that while for both a 1 min and 3 min plating bath, clear enhancement of the NPs seeds was observed, the film that grows was not yet continuous and, therefore, cannot serve as electrode. After 5 min of plating the film appeared to be continuous, but still porous, i. e., suitable as electrode (Fig. 1s b). While longer plating time increased film density, it also increased electrode reflectivity and decreased its transparency.
The thickness of the resulting Au films was characterized using Dektak stylus profilometer. Profilometry shows that, except for the 1 min film, all films grew as Au columns ~70 nm high, independent of the plating time. Because longer plating times increased the column width this explains the observed increase in film density. After plating, the samples were rinsed with deionized water and blown dry under nitrogen stream.
To define spatially the top electrodes on the same Si sample we used a polypropylene tip to draw a thin line by firmly wiping NPs off the surface, prior to the electro-less plating. While originally the definition lines could not be seen, they became visible in the Au plating bath, where the pads turned golden and the lines remained transparent.
Hg electrode structures
Initial and exploratory I-V measurements of the molecular junctions were performed using a Hg (99.9999% purity) drop on the molecular SAM, using a controlled growth hanging Hg drop (HMD) electrode apparatus (BAS, USA). The optically determined geometric contact area between the Hg drop and the monolayer is typically 0.2 mm2 (0.5 mm diameter).
Back contacts.
All n-Si samples were contacted on the back by applying In-Ga eutectic, after scratching the surface with a diamond knife.
Temperature-dependent I-V measurements.
We used a liquid Hg drop top electrode in an MMR variable temperature measurement system. Here the sample is in a Faraday cage-like sample chamber and can be contacted mechanically by micro-manipulators. After sample placement, the sample chamber was pumped down to 20 mTorr. The sample was then first heated (to 330 K) for ~ 10 min to clean it and to prevent ice clogging upon cooling. The sample holder was cooled by the MMR Joule-Thomson cooler, using a computerized control system. An additional cooling finger was placed in the chamber to prevent water condensation on top of the sample. The electrical contact was made to the back of the semiconductor, using the refrigerator cooling finger connectors. The Hg top electrode was contacted by a micro-manipulator. Current-voltage characteristics were measured, first upon cooling, then upon heating and then again upon (re)cooling, to assure junction stability and reliability of the experimental results. The back contact of the junction was grounded.
Electrical measurements
Both room temperature hanging Hg drop electrical measurements and temperature-dependent measurements in the MMRsetup were recorded using a Keithley 6430 voltage source / sub-femto-ammeter.
Current-voltage measurements under illumination
Except where noted otherwise, photovoltaic measurement were conducted using a MicroManipulator manual probe station and Keithley 6487 voltage source / picoammeter. The manipulator probe was extended with 30 μm diameter Au wire, which, because of its mechanical properties, ensures electrical contact to the top Au film without damaging it or penetrating the molecular insulator. Illumination was by a Bausch&Lomb white microscope light that was previously measured to produce 25 mW/cm2 using a calibrated Si photodiode. The junction's open-circuit voltage lies already in its tunneling-limited region as was observed by the temperature dependent current-voltage measurements. It is probably because of that fact, that any increase in the illumination intensity decreases the device's fill factor while leaving the open-circuit voltage almost unchanged.
Fig. 2s show dark and photovoltaic measurements taken under tungsten-halogen incandescent lamp illumination producing 40 mW/cm2, 65 mW/cm2, 85 mW/cm2 and 100 mW/cm2. The device's fill-factor drops with increasing illumination and was measured to be 36%, 31%, 28% and 25% respectively. The open-circuit voltage increases slowly with illumination and was recorded to be 490, 500, 510 and 520 mV, respectively.
Scanning Electron Microscopy
Electro-less plated Au films were imaged in a Zeiss ULTRA high-resolution field emission scanning electron microscope under an accelerating voltage of 3-5 keV.
Figure 1s.: SEM micrographs of a very dense Au NPs monolayer bound to the thiol-terminated monolayer (a), and Au film, grown using electro-less plating with 5 min plating time. The film appears to be electrically continuous, but still porous and semi-transparent (b).
1
[*]Corresponding author: