The Optics Laboratory
Group of Hans Hallen, Physics Department, North Carolina State University
When one tries to do chemistry on a self-assembled monolayer (SAM), one finds that the constraints on mobility and rotation imposed by the bonding to the surface can interfere with chemistry that normally happens with high yield in solution. This is especially true of wet chemistry methods. 'Dry' methods are less subject to this as the gas phase molecules are more free to adapt to the surface. We have observed this when oxidizing a vinyl-terminated SAM layer to a hydroxide to change the surface from hydrophobic to hydrophilic. Better SAM layer quality and single layer is easier to obtain with a hydrophobic top termination, especially when using a trichlorosilane attachment moiety for bonding to oxide surfaces. Thus, this is a common need when a hydrophilic surface is desired, or a patterned surface as is required for self-alignment at the whole wafer level or as part of a covalent nanoglue. High points are here, more detail is in the paper . A high-yield method has been developed for the production of hydrophilic, carboxyl-terminated alkylsiloxane monolayers on silicon using the ozonolysis and hydrolysis of 10-undecenyltrichlorosilane SAMs. Contact angles with water, a common measure of hydrophilicity, were brought to 0° on receding and approximately 16° on advancing, compared to 105° and 98° respectively before ozonation. Ellipsometry showed the presence of a full monolayer, 1 nanometer thick before ozonation, that decreases by about 0.2 nm during ozone treatment, largely due to the removal of a carbon atom in the process. This removed material coalesces as nanoparticles, observed with AFM imaging. The process has a surprisingly narrow ozone dose window, with excess ozonation resulting in complete layer removal. The carboxylate moieties allow further chemical modification of the surface in addition to the hydrophilic surface that can be produced by exposing the silicon dioxide substrate.
To understand the problem, contact angles of the vinyl surface, after chemical oxidation, and after ozonation are shown in Fig. 1. The nonzero angle of the chemically oxidized surface implies only about a 20% yield of the oxidation.
(a) (b) (c)
Figure 1. Droplets of water on a SAM with (a) vinyl termination before oxidation (~90-100° contact angle), (b) after permanganate/ periodate treatment (35° contact angle), and (c) on the surface of an ozone-treated 10- undecenyltrichlorosilane SAM after hydrolysis (0° receding and ~16°advancing contact angle).
The ozonation process is based on the fact that ozone preferentially attacks C=C double bonds compared to C-C single bonds. Thus, the vinyl group is oxidized before the SAM is attacked. Careful control of the dose is required to insure the surface termination is complete while the layer remains intact, as we show later with ellipsometry. The process is shown in Fig. 2. The oxidation process was performed in a Novascan PSD-UV ozone cleaner; the wafers were placed in the cleaner face up and exposed to ozone for the appropriate amount of time; the lamps were turned off and the cover removed upon completion of the time interval. The oxygen flow rate into the cleaner was 145 mL/s, which should provide an ozone concentration less than 100 ppm. Hydrolysis was performed by placing the wafers in deionized water in a petri dish and leaving them there overnight, or for the appropriate number of days for longer soaking trials.
Figure 2. Mechanistic pathways in the ozonolysis of a vinyl group. A primary ozonide (five membered ring) is formed by 1,3-dipolar cycloaddition and breaks down with cleavage of the vinylic carbon-carbon bond. The principal products are aldehydes, which are highly oxidizable, and carboxylic acids.
One way to measure the efficacy of the treatment is through water contact angle maesurements. Contact angles were measured by placing a drop of water on the surface, tilting it so one side would be advancing and the other receding, and then placing it on a stage, taking a picture (as in Fig. 1), and finding the angle on a computer. Unless they were just plasma or ozone cleaned, all samples were spin-cleaned with isopropanol before ellipsometry or contact angle measurements. We find (Fig. 3) that the contact angle does not fall immediately after ozonation unless the entire layer is removed (a), but do fall dramatically after hydrolysis, with adequate, but not excessive, ozonation (b).
Figure 3. Contact angles after hydrolysis but (a) before and (b) after hydrolysis. The 150 second exposure sample is optimal.
Another way to evaluate the process success os to observe the layer thickness. Ellipsometry is used to measure the thickness of the layers after ozonation and is shown in Fig. 4. No measurable changes in thickness are found during hydrolysis. We see an early change in thickness of about 0.2 nm. The fact that this is close to the length of a carbon bond suggests that this reflects the removal of the double-bonded carbon atom and its replacement with oxygen. Further ozonation results in a very rapid removal of the rest of the SAM layer, indicating efficient attacking of the substrate-first carbon bond by the ozone after all carbon double bonds are removed. Adequate process control on ozone concentration and exposure is required. Hydrolysis is also required, and the results are excellent.
Figure 4. Ellipsometric measurements of layer thickness show the physical results of some of the chemistry involved with oxidation and removal of layer with over-treatment.
AFM images were taken of the layers after several different amounts of treatment. Figure 5 shows images from the qualitatively different ozonation stages. The stage that leave the layer intact (excepting the 0.2 nm loss discussed above) are shown in Fig. 5(a). Surface roughness is about 0.6 nm. Much of this is presumably due to the roughness of the oxide layer under the SAM, since the ellipsometric data suggests a fairly uniform monolayer. Nanodots in the 2-7 nm size range are also observed. The density of these dots is higher in the Fig. 5(b), layer removed, image, and a few dots larger than 10 nm are observed, as expected with the longer ozonation time. We have verified that the nanodots result from the layer treatment, since they do not occur on samples that do not have SAM layers. It is likely that the nanodots are formed from formaldehyde polymers or other reaction products. To test this hypothesis, we calculated the volume of material in the nanodots, and calculated the thickness it would account for if spread evenly across the surface. In particular, a threshold placed in the histogram, chosen at the sharp drop in the histogram (also ~10% of the points above it), separated the nanodots from normal surface roughness. The histogram was used to sum the nanodot volume, and the height corresponding to the histogram peak value was used as the background. The thickness increase if the nanodots were spread evenly was found to be 0.21 nm for Fig. 5 (a). This is very close to the ellipsometric 0.2 nm thickness decrease, and gives credence to the reaction product source of the nanodots. The sample with the destroyed SAM layer, Fig. 5(b), has a higher roughness (~1 nm) and significantly more nanodots at the few nm level, and several at the ~40 nm size range. The thickness increase if these nanodots were evenly spread on the surface is 0.32 nm, significantly less than the ~1 nm layer thickness, indicating that much of the layer has been volatized.
The chemistry of these transformations appears to be a standard ozonolysis-hydrolysis reaction, with aldehyde functions on the monolayers being oxidized to carboxylate functions over the course of the soaking time, as might be expected, and interchain peroxide or other linkages commonly found after ozonolysis being hydrolyzed. The aldehyde branch of the reaction seems to be favored, since hydrolysis is required to produce the very small contact angles obtained. Previous studies with ozone exposure of similar species would suggest the production primarily of aldehyde and carboxylic acid functions in similar amounts11. The one-carbon species released into the water, if released as formaldehyde, would likely be expected to polymerize in water, and that is a plausible explanation for the objects found on the layer in the AFM images. Formate or other one-carbon compounds are almost certainly produced as well in some concentration. Several reactions are available to ozone that could result in destruction of the carbon chain at higher ozone doses, such as attack at carbons adjacent to a double bond or to a heteroatom, or simply at carbon-hydrogen or carbon-carbon bonds in the alkyl chain.
Figure 5. AFM images of wafers subjected to ozonolysis/hydrolysis treatment. (a) an intact, functionalized layer (receding contact angle ~ 0˚, ellipsometry reveals little layer loss). It ends up on the surface in small (3 nm high, ~30-50 nm across) dots, ~30/µm2. The volume of the dots is close to the volume lost. (b) An excessively ozonated wafer (ellipsometry reveals loss of organic layer). The bumps are bigger, ~10/µm2 at 6-7 nm high, and 50/µm2 of the smaller ones. The volume of the dots is ~twice above, not enough to be the entire layer. Ozone volatizing is likely.
In conclusion, an effective method has been developed that significantly improves the production of monolayers on silicon with carboxylate moieties, achieving contact angles of 0° on receding and around 16° on advancing, and with ellipsometry showing a uniform monolayer. The reaction shows the characteristics of a fairly standard Criegee mechanism. This method is likely to be useful for a variety of applications either requiring a hydrophilic substrate, or a carboxylate moiety for further reaction, such as biological applications, nano-orifice coating, or capillary-force driven processes. The ozone exposure, however, requires tight process control, and some reaction products are left as nanodots on the surface.
More info is in the papers.
NC State University | Physics | Optics Home
Copyright ©2000-2017, Hallen Laboratory, NCSU, Raleigh, NC. www.physics.ncsu.edu/optics
Comments or questions? Hans_Hallen@ncsu.edu