10866
J. Phys. Chem. C 2010, 114, 10866–10872
Long-Term Stability and Electrical Performance of Organic Monolayers on Hydrogen-Terminated Silicon Richard T. W. Popoff, Hidehiko Asanuma, and Hua-Zhong Yu* Department of Chemistry, Simon Fraser UniVersity, Burnaby, British Columbia V5A 1S6, Canada ReceiVed: February 22, 2010; ReVised Manuscript ReceiVed: April 30, 2010
The long-term stability (gradual oxidation) of two structurally different organic monolayers, C6H5(CH2)3-Sit (C3Ph-Sit) and CH3(CH2)11-Sit (C12-Sit), covalently bonded to n-type silicon(111) was studied and correlated to the electrical performance of the thus formed mercury | monolayer | n-Si junctions. High-resolution XPS analysis of the O 1s region on freshly prepared samples identified physically adsorbed oxygen (Oad) and oxygen-containing species (SiOx) bound to silicon at 532.2 and 533.5 eV, respectively. When left under ambient conditions, both bands increased in intensity as time progressed, while no change in the carbon content (C 1s peak) was observed. More importantly, upon aging, there was a noticeable increase in the electrical current flow of both and junctions; a considerably larger increase occurred in the former sample, which has less oxygen content than that of the latter junction (based on the O 1s(O-Si)/Si 2p peak area ratios). These results suggest that the slow oxidation of molecularly modified silicon may, in fact, dictate the electrical performance of the thus formed molecular diode junctions. 1. Introduction The ability to tune the structural and electronic properties of a metal-semiconductor (MS) junction by making molecular modifications to its interface is an important advancement for applications in nanoscale electronic devices.1-4 Covalently bonding organic molecular films to semiconductors, such as silicon and gallium-arsenide (GaAs), before placing the metal contacts on top has been greatly focused upon for the past decade.5-9 The electrical properties of these molecular junctions change substantially when different organic monolayers are incorporated at the interfaces.10-15 If we are to utilize molecularly modified MS junctions, we must achieve competitive performance (including long-term stability) with current macroscale devices, while maintaining the cost and lifetime of traditionally used electronics. Covalently bonding an organic monolayer to the chosen semiconductor allows for more reproducible results over other methods of surface modification.5-9 On well-prepared samples, the overall dipole of a covalently bound monolayer can alter the electrical properties of the semiconductor,10-12 whereas varying the chain length of the adsorbate molecules can tune the electron tunneling behavior through the metal | monolayer |semiconductor or metal | monolayer | metal junctions.13,14 Charge-transport mechanisms, such as the classical thermionic emission model, have been used in an attempt to explain the behavior of electrons passing through molecularly modified MS junctions.10,11,13,16-19 However, these models do not appear to be complete as deviations from the ideal electrical transport mechanism are often found.10,11,13,17-20 The answers may lie in using different electrical transport models (thermionic, recombination, diffusion, or Ohmic) over different potential regions.1-3,11,19 It has been demonstrated previously that organic monolayers prepared on hydrogen-terminated silicon have minimal oxidization, and they are extremely stable even under harsh conditions;21-23 oxidation can be monitored using X-ray photoelectron spec* To whom correspondence should be addressed. Tel: (778) 782-8062. Fax: (778) 782-3765. E-mail:
[email protected].
troscopy (XPS) through analysis of the high-resolution Si 2p and O 1s regions.23-26 For example, Boukherroub et al. have shown that n-alkyl monolayers on silicon can prevent the deterioration of the silicon wafers.21 Nevertheless, the long-term stability of these organic monolayers has not yet been extensively explored. In this work, using surface characterization measurements, such as water contact angle and ellipsometry, in addition to direct XPS measurements, we are aiming at a better understanding of the physical property changes of organic monolayers prepared on oxide-free silicon as they age from exposure to ambient conditions. Aging of the sample may lead to dramatic changes in the electrical properties of thus prepared mercury (Hg) | monolayer | n-silicon(111) diode junctions and can provide further insight into their long-term performance and viability for integration into nanoelectronics. 2. Experimental Section 2.1. Materials. All chemicals used were of ACS reagent grade. Tetrahydrofuran (THF), trifluoroacetic acid, and 1,1,1trichloroethane were used as received. 1-dodecene and allyl benzene were distilled in the presence of sodium and underwent five freeze-pump-dry cycles. 2.2. Sample Preparation. Attachment of 1-dodecene and allyl benzene molecules to the silicon surface is shown in Scheme 1. Silicon(111) wafers (0.5-5.0 Ω · cm, n-type phosphorus doped, Virginia Semiconductor Inc.) were cut into pieces (1 × 2.5 cm) and cleaned by sonication in ethanol (95%) for 30 min, rinsed with deionized water (18.3 MΩ · cm), and then placed in a 3:1 solution of concentrated H2SO4 and 30% H2O2 (30%) at 90 °C for 1 h. Caution: such a mixture is typically named as “Piranha Solution”, which is Very reactiVe and should be handled with extreme care. After copious washing with deionized water, the samples were etched using deoxygenated NH4F (40% aqueous solution) to remove the native oxide and obtain hydrogen-terminated silicon (H-Sit, Scheme 1a). Etched wafers were then transferred into Schlenk tubes containing deoxygenated 1-dodecene or allyl benzene and sealed under
10.1021/jp101595w 2010 American Chemical Society Published on Web 05/26/2010
Organic Monolayers on Hydrogen-Terminated Silicon
J. Phys. Chem. C, Vol. 114, No. 24, 2010 10867
SCHEME 1: Preparation of Molecularly Modified Silicon Samplesa
a (a) Removal of the native oxide by etching with 40% NH4F solution to obtain hydrogen-terminated silicon (H-Sit). (b) Reaction of H-Sit samples with anhydrous 1-dodecene or allyl benzene at 175 °C for 4 h, to yield high-quality monolayers of CH3(CH2)11-Sit and C6H5(CH2)3-Sit, respectively.
Ar. The monolayers on silicon were produced by heating at 175 °C for 4 h (Scheme 1b). The modified silicon was then washed with 1% trifluoroacetic acid in THF and 1,1,1trichloroethane and dried under a stream of N2. This procedure yielded high-quality (CH2)11CH3-Sit and (CH2)3C6H5-Sit (henceforth, referred to as C12-Sit and C3Ph-Sit, respectively) samples. For the long-term stability tests, the samples were stored in the dark under ambient conditions (19-22 °C, 30-45% relative humidity). All samples were washed again with copious amounts of 1% trifluoroacetic acid in THF and 1,1,1-trichloroethane and dried with N2 prior to measurement. 2.3. Surface Characterization. An AST Optima contact angle apparatus was used to make wetting measurements using a horizontal light beam to illuminate the water droplet. Contact angles were determined on three independent samples (three to six different spots per sample) with a 1.5 µL drop of deionized water. A UVISEL spectroscopic ellipsometer (Horiba Jobin Yvon) was used to take ellipsometric measurements over a spectral range of 300-700 nm (incident angle of 70°). The thickness of the organic film was determined according to the Cauchy absorbent mathematical model.10,27 X-ray photoelectron spectra were acquired using a Kratos Analytical, Axis ULTRA apparatus equipped with a DLD detector. The chamber pressure was kept at 10-9 Torr during data acquisition. A monochromatic X-ray source of Al KR was used to acquire each spectrum with an analysis area of 700 × 300 µm2. Hybrids of Gaussian and Lawrencian statistical models were used to analyze peak areas in the high-resolution spectra. 2.4. Electrical Characterization. Electrical characterizations were performed using an Autolab electrochemical analyzer (PGSTAT30, Eco Chemie BV, The Netherlands) in a Faraday cage. An Ohmic contact was made on the back of the modified silicon wafer using a eutatic InGa paste before placing it on a copper block. A gastight syringe filled with mercury was used to make a metal top contact with the silicon sample. Mercury droplets protruding from the tip of the syringe were monitored using a digital video microscope (model DM143, Micro-Optic
Figure 1. (a) Static water contact angle and (b) ellipsometric thickness of C12-Sit and C3Ph-Sit monolayers on silicon as a function of time. Ellipsometric data were analyzed with a bilayer model (an organic film on top of a silicon substrate) without considering any oxide growth.
Industrial Group Co., Hong Kong) with a 40× objective lens as they were brought into contact with the silicon samples. The contact areas were measured using the Motic Images 2000 software provided by the manufacturer. The diameters of the contacts ranged from 500 to 800 µm (uncertainty
