**2. Principle and procedure**

52 Recent Advances in Nanofabrication Techniques and Applications

excited from ground state to metastable state. In metastable state, instead of immediately decay to ground state, the electron will stay for some time (long compare to usuall short lifetime-excited state) before it decays to its ground state. For example, the metastable helium atoms have natural lifetimes ≥ 20ms, which is much longer than its typical flight times. The energy stored in metastable noble gas atom can be used to create pattern in a SAM that act as resit. When a metastable noble gas atom strikes the SAM resist, the atom will release energy and it will go to ground state. The energy released will change the characteristic of the SAM resist in radius few angstroms from place where the atom strikes. The atom in ground state will not have any effect toward SAM resist because noble gas atoms are not chemically reactive. The locally change of characteristic of the SAM resist will enable us to do wet chemical etching. This new fabrication technique, which bridges the gap between the bottomup chemical self-assembly techniques and the modern top-down lithography, can overcome the intrinsic resolution limitations of traditional photolithographic techniques [10,19,20]. In principle, the atom lithography based on neutral MAB and SAM ideally meets the required conditions for sub-100nm fabrication [10]. MAB with a de Broglie wavelength of less than 0.1nm eliminates diffractive resolution limitations, and an ultrathin organic SAM resist with a thickness of 1~3nm ensures the sharpness of the edge profile within the resist. Moreover, the neutral metastable atoms are insensitive to the electric and magnetic fields, and the long-range inter-particle interactions are weak. Therefore, this novel method permits the direct and large area manufacturing of micro- and nanostructures on a silicon wafer, avoiding some inherent complications of electron-beam, ion-beam or photolithography. With these unique advantages, the atom lithography with neutral MAB and high-resolution SAM resist makes it possible to achieve nanolithography and provides a potential way in manufacturing structures at a large-

During the past fifteen years, considerable attention has also been given to the investigation of the mechanism of metastable atom lithography and microfabrication with different resolutions on various substrates, such as gold, silicon, silicon oxide, and mica [10,19-29]. Direct evidence of the emission of H+ and CHx+ from the SAM on the Au(111) surface under the irradiation of a metastable helium atom beam (He\*-MAB) at thermal energy supports the interpretation that the SAM is damaged through the interaction between the outermost surface of the SAM and the metastable atoms. This damage can cause a characteristic surface change from hydrophobic to hydrophilic or a molecular structural change from condensed to cross-linked, which consequentially affect the etching process in the fabrication of the micro- or nanostructure, resulting in the positive and negative pattern transfers, respectively. The typical micropatterning with sub-100 nm resolution on a silicon substrate covered with a SiO2 or Au layer has been achieved successfully [10,21,22,28,29]. Most of these studies are conducted through multistep etching processes using the cover layers as intermediate masks. For example, in the He\*-MAB lithography with dodecanethiol-SAM [10,21,22], a Au/Ti coating layer is used to create an intermediate mask on silicon substrate by the first etching process, and then, the intermediate mask pattern is transferred to silicon substrate by the second etching process. It is very interesting to directly transfer negative and positive patterns of SAM induced by He\*-MAB into a silicon substrate without

metal/oxide coating layer to fabricate arrays of micro- and nanostructures.

Recently, the He\*-MAB lithography of Si with SAM, which formed directly on silicon substrate by direct covalent linkage instead of metal/oxide coating layers, has been accomplished by our research team [19,20]. The latent image formed in SAM by He\*-MAB

scale based on micro- and nanoscale features.

### **2.1 The general principle and procedure**

Sheme 1. Shematic illustration of the principle and procedure for the frabrication of periodic arrays of Si(111)/Si(100) by He\*-MAB lithography(excited helium atom:He\*; ground-state helium atom: He).

The general principle and process of atom lithography using He\*-MAB and organosilane SAMs to create the arrays of silicon micro- and nanostructures on a Si substrate were illustrated in Scheme 1. By introducing a hydrogenation process into the pretreatment of silicon substrates (as described in (I) and (II) in Scheme 1), fine organosilanes SAMs was successfully formed as a resist on silicon surface under a controlled argon gas atmosphere at elevated temperature. The chemical modification of non-oxidized silicon surfaces utilizing monolayers could be achieved by neutralizing the silicon surfaces with alkyl chains through

Atom Lithography: Fabricating Arrays of Silicon Microstructures

temperature [19,20,30].

**2.3 Exposure to MAB** 

Using Self-Assembled Monolayer Resist and Metastable Helium Beam 55

SAMs were formed as a resist on silicon surface by introducing a hydrogenation process into the pretreatment of silicon wafer and a controlled argon atmosphere at elevated

Sheme 2. The illustration of the mechanisam of forming SAMs on Si substrates.

The experimental setup for the sample exposure to He\*-MAB was depicted in Scheme 3. The Si substrate with organosilane SAMs was covered by a physical mask with TEM Cu grids. As shown in the inset of Scheme 3, six TEM grids were well distributed on a nickel sheet (with a thickness of 0.1mm and a square area of 1cm×2cm) and fixed over the opening holes with a diameter of 2mm using Ag conductive paste. The assembly was immediately inserted into the sample chamber of the MAB source (as shown in Scheme 3) and exposed to He\*- MAB for 5~120 min, producing latent patterns in organosilanes SAM on Si substrates. The exposure experiments were performed with the existing apparatus depicted in Scheme 3. The He\* source was operated in a dc discharge mode, and a stable pressure (~5×10-7 Pa for the ultimate pressure and ~3×10-2 Pa for the working pressure) of chambers was retained by turbo-molecule pump (T.M.P) vacuum systems. A transverse deflecting voltage of 600~800 V was applied along the beam line to eliminate ions or electrons in the beam, which otherwise damage the SAM during exposure. Through a skimmer opening of ~0.7 mm diameter, He\*-MAB was directed toward a rotatable thin tantalum (Ta) sheet sample holder at a distance of 26.5 cm. Three tantalum (Ta) plates were arranged triangularly around the rotation axis (shown in the inset of Scheme 3). Two of them were used for installing samples, and the remaining one was used for monitoring the intensity of the primary beam. The thickness of the nickel sheet defined the gap between physical masks and sample surfaces (~100um), avoiding the mechanical contact that may cause possible artifacts of SAMs. Since the mask almost covered the entire sample, the metastable flux was represented by the metastable-atom-induced electron emission from the Ta plate or from the sample and mask, which was the main content of the induced current. The typical values of the measured current on both the Ta plate and sample were 100 and 50 nA, which corresponded to an electron emission rate of ~1×1012 and ~2×1012 s-1·cm-2, respectively.

direct covalent linkage, i.e., silicon-carbon [30-32]. This provided an opportunity to directly use an ultrathin hydrophobic SAM with a thickness of 1~3nm as a super resist on silicon surfaces. The samples with the physical stencil masks were exposed to the irradiation of He\*-MAB at thermal energy, and the SAM was damaged through the interaction between the outermost surface of the SAM and the metastable atoms (as described in (III) in Scheme 1). When the metastable atoms hit the outermost surface of the SAM on Si substrate, they transferred their internal energy (~20eV) to the chemisorbed hydrocarbons of SAM molecules and caused chemical changes that might result in either the formation of a durable crossing-linking polymerization material or the loss of hydrophobicity of the SAM (as indicated in (IV) and (V) in Scheme 1). After exposure, the pattern was transferred into a latent pattern in SAM. The SAM exhibited the positive-tone and negative-tone sensitivity due to largely different dosage of MAB and the length of the alkyl chains of SAM molecule. The negative pattern implied the polymerization of SAM molecules induced by more He\*- MAB irradiation dose. Meanwhile, the positive pattern indicated the hydrophilic region of SAM induced by less He\*-MAB irradiation dose. Furthermore, the longer alkyl chains were easier to undergo cross-linking polymerization than the shorter alkyl chains under the He\*- MAB irradiation. Consequently, these could affect the etching process in the fabrication of the micro- or nanostructure (i.e., the hydrophilic exposure region was easier to etch using KOH solution than the hydrophobic unexposed region, whereas the cross-linking polymerization region was more durable to KOH etching than the hydrophobic unexposed region), resulting in the positive and negative pattern transfers (as described in (VI) and (VII) in Scheme 1), respectively.

The mechanism for the contrast inversion, i.e., the competition between the loss of hydrophobicity and the cross-linking polymerization of SAM molecules determined the polarity of the pattern transfer process, has been confirmed previously [19,20]. In our experiments, the lithographic patterns were obtained in samples with a relatively large area of about 2 cm2. In these areas, the mask pattern was reproduced with high fidelity, and the repetition of lithographic patterns was consistent in different runs. In order to obtain better contrast patterns, we carefully optimized the experimental parameters. After etching, the resulting arrays fabricated on the silicon substrate were clearly observed under the optical microscope.

#### **2.2 Mechanism of forming SAMs on Si substrates**

The mechanism of forming SAMs on Si substrates was illustrated in details in Scheme 2. The p-type Si(111)/Si(110)/Si(100) wafers with a native oxide layer (p-type Si(111): Φ=100mm, thickness=380µm, ρ=10~20 Ω·cm; n-type Si(110): Φ=100mm, thickness=300µm, ρ=1~5 Ω·cm; p-type Si(100): Φ=100mm, thickness=300µm, ρ=10~20 Ω·cm) were used and cut into pieces of 10mm×20mm as Si substrates. The Si substrates were first rinsed using toluene, ethanol, and deionized water sequentially, then immersed into a 2vol.% aqueous HF solution for 5~10min to remove the native oxide layer and hydrogenate the silicon substrates, and finally dried. Protected from air by a flowing argon gas, the Si substrates were immersed into a 10~40mM organosilanes (octadecyltrichlorosilane, 18C-long-chain hydrocarbon molecules, ODTS; dodecyltrichlorosilane, 12C-long-chain hydrocarbon molecules, DDTS; hexycyltrichlorosilane, 6C-long-chain hydrocarbon molecules, HTS) toluene solution at room temperature for 24~50 hours, and then rinsed in toluene and dried at 100ºC for 30~80min. Thus, through direct covalent linkage, i.e., silicon-carbon, fine organosilanes SAMs were formed as a resist on silicon surface by introducing a hydrogenation process into the pretreatment of silicon wafer and a controlled argon atmosphere at elevated temperature [19,20,30].

Sheme 2. The illustration of the mechanisam of forming SAMs on Si substrates.

### **2.3 Exposure to MAB**

54 Recent Advances in Nanofabrication Techniques and Applications

direct covalent linkage, i.e., silicon-carbon [30-32]. This provided an opportunity to directly use an ultrathin hydrophobic SAM with a thickness of 1~3nm as a super resist on silicon surfaces. The samples with the physical stencil masks were exposed to the irradiation of He\*-MAB at thermal energy, and the SAM was damaged through the interaction between the outermost surface of the SAM and the metastable atoms (as described in (III) in Scheme 1). When the metastable atoms hit the outermost surface of the SAM on Si substrate, they transferred their internal energy (~20eV) to the chemisorbed hydrocarbons of SAM molecules and caused chemical changes that might result in either the formation of a durable crossing-linking polymerization material or the loss of hydrophobicity of the SAM (as indicated in (IV) and (V) in Scheme 1). After exposure, the pattern was transferred into a latent pattern in SAM. The SAM exhibited the positive-tone and negative-tone sensitivity due to largely different dosage of MAB and the length of the alkyl chains of SAM molecule. The negative pattern implied the polymerization of SAM molecules induced by more He\*- MAB irradiation dose. Meanwhile, the positive pattern indicated the hydrophilic region of SAM induced by less He\*-MAB irradiation dose. Furthermore, the longer alkyl chains were easier to undergo cross-linking polymerization than the shorter alkyl chains under the He\*- MAB irradiation. Consequently, these could affect the etching process in the fabrication of the micro- or nanostructure (i.e., the hydrophilic exposure region was easier to etch using KOH solution than the hydrophobic unexposed region, whereas the cross-linking polymerization region was more durable to KOH etching than the hydrophobic unexposed region), resulting in the positive and negative pattern transfers (as described in (VI) and

The mechanism for the contrast inversion, i.e., the competition between the loss of hydrophobicity and the cross-linking polymerization of SAM molecules determined the polarity of the pattern transfer process, has been confirmed previously [19,20]. In our experiments, the lithographic patterns were obtained in samples with a relatively large area of about 2 cm2. In these areas, the mask pattern was reproduced with high fidelity, and the repetition of lithographic patterns was consistent in different runs. In order to obtain better contrast patterns, we carefully optimized the experimental parameters. After etching, the resulting arrays fabricated on the silicon substrate were clearly observed under the optical

The mechanism of forming SAMs on Si substrates was illustrated in details in Scheme 2. The p-type Si(111)/Si(110)/Si(100) wafers with a native oxide layer (p-type Si(111): Φ=100mm, thickness=380µm, ρ=10~20 Ω·cm; n-type Si(110): Φ=100mm, thickness=300µm, ρ=1~5 Ω·cm; p-type Si(100): Φ=100mm, thickness=300µm, ρ=10~20 Ω·cm) were used and cut into pieces of 10mm×20mm as Si substrates. The Si substrates were first rinsed using toluene, ethanol, and deionized water sequentially, then immersed into a 2vol.% aqueous HF solution for 5~10min to remove the native oxide layer and hydrogenate the silicon substrates, and finally dried. Protected from air by a flowing argon gas, the Si substrates were immersed into a 10~40mM organosilanes (octadecyltrichlorosilane, 18C-long-chain hydrocarbon molecules, ODTS; dodecyltrichlorosilane, 12C-long-chain hydrocarbon molecules, DDTS; hexycyltrichlorosilane, 6C-long-chain hydrocarbon molecules, HTS) toluene solution at room temperature for 24~50 hours, and then rinsed in toluene and dried at 100ºC for 30~80min. Thus, through direct covalent linkage, i.e., silicon-carbon, fine organosilanes

(VII) in Scheme 1), respectively.

**2.2 Mechanism of forming SAMs on Si substrates** 

microscope.

The experimental setup for the sample exposure to He\*-MAB was depicted in Scheme 3. The Si substrate with organosilane SAMs was covered by a physical mask with TEM Cu grids. As shown in the inset of Scheme 3, six TEM grids were well distributed on a nickel sheet (with a thickness of 0.1mm and a square area of 1cm×2cm) and fixed over the opening holes with a diameter of 2mm using Ag conductive paste. The assembly was immediately inserted into the sample chamber of the MAB source (as shown in Scheme 3) and exposed to He\*- MAB for 5~120 min, producing latent patterns in organosilanes SAM on Si substrates. The exposure experiments were performed with the existing apparatus depicted in Scheme 3. The He\* source was operated in a dc discharge mode, and a stable pressure (~5×10-7 Pa for the ultimate pressure and ~3×10-2 Pa for the working pressure) of chambers was retained by turbo-molecule pump (T.M.P) vacuum systems. A transverse deflecting voltage of 600~800 V was applied along the beam line to eliminate ions or electrons in the beam, which otherwise damage the SAM during exposure. Through a skimmer opening of ~0.7 mm diameter, He\*-MAB was directed toward a rotatable thin tantalum (Ta) sheet sample holder at a distance of 26.5 cm. Three tantalum (Ta) plates were arranged triangularly around the rotation axis (shown in the inset of Scheme 3). Two of them were used for installing samples, and the remaining one was used for monitoring the intensity of the primary beam. The thickness of the nickel sheet defined the gap between physical masks and sample surfaces (~100um), avoiding the mechanical contact that may cause possible artifacts of SAMs. Since the mask almost covered the entire sample, the metastable flux was represented by the metastable-atom-induced electron emission from the Ta plate or from the sample and mask, which was the main content of the induced current. The typical values of the measured current on both the Ta plate and sample were 100 and 50 nA, which corresponded to an electron emission rate of ~1×1012 and ~2×1012 s-1·cm-2, respectively.

Atom Lithography: Fabricating Arrays of Silicon Microstructures

**2.4 Etching processing and pattern transferring** 

Scheme 1).

around 100 nm.

**2.5 Characterization** 

(Si+2OH⎯+2H2O→SiO2(OH)22

electron microscope operated at 30 keV.

**3. Experimental achievements** 

Using Self-Assembled Monolayer Resist and Metastable Helium Beam 57

The bar widths and the hole diameters define the pitch sizes of the patterns. Even though the image of a pattern was in micrometer range, a sharp edge in the tens-of-nanometers range for the resulting microstructures of silicon indicated a potential of nanosized pattern in our experiments. In this sense, actually, the spatial resolution of the patterning in our

After exposure, the samples with latent patterns in SAM were taken out from the vacuum chamber and rapidly dipped into an etching solution in which a magnetic bar was used for stirring, and were etched for 5~60min at room temperature in an aqueous solution of 0.1M KOH, rinsed by deionized water, and then dried in air. By this single step KOH etching process, the latent image formed in SAM by He\*-MAB passing through a stencil was directly transferred into the underlaying Si substrate (as described in (VI) and (VII) in

For the wet etching of Si substrate in KOH solution, the anisotropy of etching rate of silicon have been studied in detail over the passed 20 years. Especially, in the case of the photolithography in MEMS and IC industry, the anisotropy of etching rate of silicon has been investigated completely, and the anisotropy can induce obvious differences of the spatial resolution of patterning of Si along different crystal orientation. The etching rates for the silicon wafer in the KOH solution are known to depend strongly on its orientation

experiments, the anisotropic etching process of silicon occurred in the same way, and strongly affected the patterning results including edge step resolution. The sharpest step edge resolution was only 41nm with 1 pixel and the average step edge resolution was

AFM images of the samples were taken by a multimode Nanoscope IV atomic force microscope (Veeco Metrology Group, Santa Barbara, CA), operated at ambient conditions. For all images we recorded the retrace direction of the tip using a scan angle of 0º or 90º. Substrates decorated with organosilanes SAMs were imaged in tapping mode using silicon cantilevers (NanoWorld, Neuchâtel, Switzerland) with a spring constant of 42 N m-1. All images were recorded at a rate of 1.0 Hz, and with a pixel resolution of 512. SEM inspection of patterned Si substrates and the masks were carried out with a JSM-6500/SG scanning

Up to now, we have developed a process successfully, and realized a He\*-MAB lithography of Si(111)/Si(110)/Si(100) wafer substrates with SAMs, where the SAMs was formed not on metal/oxide coating layers but directly on silicon substrate. The process of atom lithography using He\*-MAB and organosilane SAMs to create the arrays of silicon micro- and nanostructures on a Si substrate were illustrated in Scheme 1. The first step was the formation of very fine organosilanes (ODTS, DDTS and HTS) SAMs as a resist on Si(111)/Si(110)/Si(100) surface by introducing a hydrogenation process into the pretreatment of silicon wafer and a controlled argon gas atmosphere at elevated

<sup>⎯</sup>+2H2(g); etching ratio: {110}>{100}>>{111}) [33]. In our

experiments was not ascribed to TEM grid mask but to the edge resolution.

Sheme 3. Shematic setup of the sample exposure to MAB for atom lithography.

During exposure He\*-MAB impacted the surface of SAM formed on silicon substrate after the beam was patterned in transverse direction by the TEM grids. Atoms with kinetic energy smaller than 1eV do not pass freely through materials. So, the physical mask (i.e., TEM grids) with thickness in the order of sub-micrometers was enough to pattern the He\*- MAB. Our experimental results showed good reproducibility with a high fidelity of the patterns of the TEM copper grids and TEM holey carbon grids (shown in figure 1), as well as a good repetition of the lithographic patterns in different runs.

Fig. 1. (a) SEM image of TEM Cu grid used as a mask (Note: the pitch size of patterns is 12.5 um, 5.5 um for wires, and 7.0 um for spaces) and (b) SEM image of TEM holey carbon grid used as a mask (Note: The diameters of the circular holes are about 1, 1.4, and 2µm, the bar widths range from 0.5 to 4µm, the oval holes with a dimension of 8х2µm and 4х1 µm).

Two kinds of physical masks (i.e., TEM Cu grids and TEM holey carbon grids) were used in our experiments. The SEM images and parameters of the TEM grids were shown in figure 1. The bar widths and the hole diameters define the pitch sizes of the patterns. Even though the image of a pattern was in micrometer range, a sharp edge in the tens-of-nanometers range for the resulting microstructures of silicon indicated a potential of nanosized pattern in our experiments. In this sense, actually, the spatial resolution of the patterning in our experiments was not ascribed to TEM grid mask but to the edge resolution.
