Femtosecond Laser-Induced Formation Of Submicrometer Spikes On A Semiconductor Substrate
20180182630 ยท 2018-06-28
Inventors
Cpc classification
H01L21/302
ELECTRICITY
H01L21/268
ELECTRICITY
H01L29/30
ELECTRICITY
Y10S438/94
GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
International classification
H01L21/268
ELECTRICITY
H01L21/02
ELECTRICITY
B23K26/12
PERFORMING OPERATIONS; TRANSPORTING
H01L21/67
ELECTRICITY
H01L21/225
ELECTRICITY
Abstract
The present invention generally provides semiconductor substrates having submicron-sized surface features generated by irradiating the surface with ultra short laser pulses. In one aspect a method of processing a semiconductor substrate is disclosed that includes placing at least a portion of a surface of the substrate in contact with a fluid, and exposing that surface portion to one or more femtosecond pulses so as to modify the topography of that portion. The modification can include, e.g., generating a plurality of submicron-sized spikes in an upper layer of the surface.
Claims
1-33. (canceled)
34. A substrate, comprising: a semiconductor substrate, wherein a surface layer of the semiconductor substrate exhibits a modified topography having a plurality of irregular features having an average height less than about 1 micrometer and an average width in a range of about 100 nm to about 500 nm.
35. The substrate of claim 34, wherein said irregular features have an average width in a range of about 100 nm to about 300 nm.
36. The substrate of claim 34, wherein said plurality of irregular features exhibit an average ratio of height to width in a range from 3 to 10.
37. The substrate of claim 34, wherein said plurality of irregular features exhibit an average separation between adjacent features in a range from about 500 nm to about 1 micron.
38. The substrate of claim 34, wherein said semiconductor substrate comprises a silicon substrate.
39. The substrate of claim 34, wherein said semiconductor substrate comprises an n-doped silicon substrate.
40. The substrate of claim 34, wherein said surface layer has a thickness in a range of about 20 nm to about 1 micrometer.
41. The substrate of claim 34, wherein said irregular features have an average height in a range of about 500 nm to less than about 1 micrometer.
42. The substrate of claim 34, wherein said features protrude above an original surface of the semiconductor substrate by a distance in a range of about 100 nm to about 300 nm.
43. The substrate of claim 34, wherein at least a portion of said plurality of irregular features extend from a base to a tip at a level above an original surface of the semiconductor substrate by a distance in a range of about 100 nm to about 300 nm.
44. A substrate, comprising: a semiconductor substrate, wherein a surface layer of the semiconductor substrate includes a plurality of protrusions of various heights, the plurality of protrusions characterized by an average width in a range of about 100 nm to about 300 nm and an average ratio of height to width in a range from 3 to 10.
45. The substrate of claim 44, wherein said protrusions exhibit height variations characterized by an average height in a range of about 500 nm to about 1 micrometer.
46. The substrate of claim 44, wherein said protrusions protrude above an original surface of the semiconductor substrate by a distance in a range of about 100 nm to about 300 nm.
47. The substrate of claim 44, wherein said plurality of irregular features exhibit an average separation between adjacent features in a range from about 500 nm to about 1 micron.
48. The substrate of claim 44, wherein said semiconductor substrate comprises a silicon substrate.
49. The substrate of claim 44, wherein said semiconductor substrate comprises an n-doped silicon substrate.
50. The substrate of claim 44, wherein said protrusions are disposed within a surface layer of the substrate having a thickness in a range of about 20 nm to about 1 micrometer.
51. A substrate, comprising: a semiconductor substrate, wherein a surface layer of the semiconductor substrate includes a plurality of protrusions of various heights, each of the plurality of protrusions characterized by an undulating side surface extending from a base to a tip, said plurality of protrusions having an average height less than about 1 micrometer, wherein said protrusions have a width in a range of about 100 nm to about 500 nm.
52. The substrate of claim 51, wherein said plurality of irregular features exhibit an average ratio of height to width in a range from 3 to 10.
53. The substrate of claim 51, wherein said plurality of irregular features exhibit an average separation between adjacent features in a range from about 500 nm to about 1 micron.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0019]
[0020]
[0021]
[0022]
[0023]
[0024]
[0025]
[0026]
[0027]
DETAILED DESCRIPTION
[0028] The present invention generally provides semiconductor substrates having surfaces that exhibit submicron-sized structures, and methods for generating such structures. In many embodiments, the submicron-sized structures are generated by irradiating a semiconductor substrate's surface with ultra short laser pulses (e.g., femtosecond pulses) while maintaining the surface in contact with a fluid (e.g., water). Exemplary embodiments of the invention are discussed below.
[0029] With reference to a flow chart 10 of
[0030] The modification of the surface topography can include generating submicron-sized features in an upper surface layer of the substrate. For example, the submicron-sized features can include a plurality of microstructured spikes, e.g., columnar structures extending from the surface to a height above the surface.
[0031] In general, the fluid is selected to be substantially transparent to radiation having wavelength components in a range of about 400 nm to about 800 nm. Further, the thickness of the fluid layer is preferably chosen so as to ensure that it would not interfere with the laser pulses (e.g., via excessive self-focusing of the pulses) in a manner that would inhibit irradiation of the substrate surface. While in this embodiment water is selected as the fluid, in other embodiments other fluids, such as alcohol or silicon oil, can be employed.
[0032] In some embodiments, at least a portion of the substrate can be placed in contact with an aqueous solution having an electron-donating constituent. For example, a solution of sulfuric acid can be applied to at least a portion of the substrate followed by irradiating that portion with short pulses (e.g., femtosecond pulses) to not only cause a change in surface topography in a manner described above but also generate sulfur inclusions within a surface layer of the substrate.
[0033] Referring to a flow chart 24 of
[0034]
[0035] The apparatus 32 further includes a harmonic generation system 38 that receives the amplified pulses and doubles their frequency to produce 140-microjoule, 100-femtosecond second-harmonic pulses at a wavelength of 400 nanometers. The harmonic generation system can be of the type commonly utilized in the art. For example, it can include a lens 38a for focusing the incoming pulses into a doubling crystal 38b to cause a portion of the incoming radiation to be converted into second-harmonic pulses. A dichroic mirror 38c can direct the second-harmonic pulses to a lens 40, and a beam stop 38d can absorb the portion of the radiation that remains at the fundamental frequency.
[0036] The lens 40 focuses the second-harmonic pulses onto a surface of a sample 42 (e.g., a silicon wafer) disposed on a 3-dimensional translation system 44 within a vacuum chamber 46. A glass liquid cell 48 is coupled to the stage over the sample so as to allow a sample surface to have contact with the fluid (e.g., water) contained within the cell. The three-dimensional stage allows moving the sample relative to the laser pulses for exposing different portions of the surface to radiation. The vacuum chamber can be utilized to pump out air bubbles in the fluid. Alternatively, the processing of the sample can be performed without utilizing a vacuum chamber.
[0037] To illustrate the efficacy of the teachings of the invention and only for illustrative purposes, submicrometer-sized silicon spikes were generated in surface layers of silicon wafers submerged in water by irradiating those surfaces with 400-nm, 100-fs laser pulses. For example, a Si (111) wafer was initially cleaned with acetone and rinsed in methanol. The wafer was placed a glass container, such as the container 48 described above, that was filled with distilled water and mounted on a three-axis stage. The silicon surface in contact with the water was irradiated by a 1-KHz train of 100-fs, 60-microjoule pulses at a central wavelength of 400 nm generated by a frequency-doubled, amplified Ti:Sapphire laser, such as that described above. A fast shutter was utilized to control the number of laser pulses incident on the silicon surface. The laser pulses were focused by a 0.25-m focal-length lens to a focal point about 10 mm behind the silicon surface. The pulses traveled through about 10 mm of water before striking the silicon surface at normal incidence. The spatial profile of the laser spot at the sample surface was nearly Gaussian characterized by a fixed beam waist of about of 50 microns. To correct for chirping of the laser pulses in the water and to ensure minimum pulse duration at the silicon surface, the pulses were pre-chirped to obtain the lowest possible damage threshold at that surface. The results, however, did not depend strongly on the chirping of the laser pulses.
[0038] During sample irradiation, the irradiated sample surface was monitored with an optical imaging system having a spatial resolution of about 5 microns. It was observed that irradiation cause formation of micrometer-sized water bubbles at the silicon-water interface. After a single pulse, two or three microbubbles were generated; after irradiation with trains of laser pulses thousands of bubbles were generated. It was also observed that some bubbles at times would coalesce to form larger ones, which would adhere to the silicon surface. These larger bubbles were removed by shaking the cell.
[0039]
[0040] The chemical composition of the uppermost 10 nm of the silicon surface layer having the spikes was determined by employing X-ray photoelectron spectroscopy (XPS). The XPS spectra showed that this layer is composed of about 83% SiO.sub.2 and about 17% silicon. The wafer was etched in 4% HF for about 15 minutes to removed the SiO.sub.2 layer (about 20 nm in thickness) while leaving the underlying unoxidized Si intact. A comparison of the electron micrographs of the spikes before etching (
[0041] To study the development of the spikes, silicon samples were irradiated with different numbers of laser pulses.
[0042] Referring to
[0043] The above silicon spikes prepared in water are one to two orders of magnitude smaller than spikes that can be generated in a silicon substrate exposed to laser pulses in presence of a gas, such as those described in the above-referenced copending patent applications. This remarkable size difference suggests different formation mechanisms for the two types of spikes.
[0044] Without being limited to any particular theory, it is noted that when a 400-nm laser pulse interacts with the silicon surface, most of the light is absorbed by a silicon layer tens of nanometers thick near the silicon-water interface. The absolution of intense light in such a thin silicon layer can excite a plasma at the silicon-water interface, which can then equilibrate with the surrounding water and silicon, leaving behind a molten silicon layer on the surface. The molten layer can solidify before the next laser pulse arrives. Due to the high temperature of the plasma, some of the water can vaporize or dissociate, thereby generating bubbles at the silicon-water interface. The large bubbles that were observed after irradiation in the above experiments remain in the water for days, thus suggesting that they consist primarily of gaseous hydrogen and oxygen rather than water vapor.
[0045] Again, without being limited to any particular theory, several possible mechanisms can be considered by which the bubbles may produce the wave-like structures shown in
[0046] Roughness on the silicon surface can cause an uneven absorption of the laser pulse energy across the surface. The resulting non-uniform temperature of the surface can produce a random arrangement of bubbles. Silicon-water has a contact angle more than 45, making a gaseous layer between the silicon and water unstable and leading to the formation of bubbles. The vaporization and dissociation of the bubbles can remove thermal energy from the molten silicon surface just below the bubbles, causing the surface to cool rapidly. Because the surface tension of liquid silicon decreases with increasing temperature, the surrounding hot liquid silicon flows toward the cooled region, deforming the surface. This deformation can then excite a circular capillary wave at the liquid-silicon surface. Superposition of ripple-structures caused by multiple laser pulses can then produce the randomly distributed submicrometer beads that appear after 20 laser pulses (See
[0047] As noted above, the morphology and sizes of the above spikes generated in a silicon surface by exposing it to femtosecond laser pulses while in contact with water can be different than those observed for spikes generated by irradiating a silicon surface with femtosecond pulses in presence of a gas, such as SF.sub.6. The early stage of submicrometer spike formation in water can be different from that in gaseous SF.sub.6, while the later stages can be similar. In SF.sub.6, straight submicrometer-sized ripple structures first form on the silicon surface, then coarser, micrometer-scale ridges form on top of (and perpendicular to) the ripples. Next, the coarsened layer breaks up into micrometer-sized beads, and finally the beads evolve into spikes through etching. In both SF.sub.6 and water, the length scale of the final structures is set by the arrangement of beadlike structures that form after roughly 10-20 pulses, and this length scale appears to be determined by capillary waves in the molten silicon. The much smaller size of the spikes formed in water is likely to be due to a difference in capillary wavelength in the two cases.
[0048] The molten silicon layer is expected to solidify much faster in water than in SF.sub.6, the thermal conductivity and heat capacity of liquid water are much greater than those of gaseous SF.sub.6. The dispersion relation for capillary waves in a shallow layer of molten silicon indicates that decreasing the lifetime of the molten layer should also decrease the longest allowed capillary wavelength. Using a simple model that neglects the effects of ablation and cooling by heat transfer to the environment to calculate the lifetime and depth of the liquid layer, it was found that the longest allowed capillary wavelength is about 1 micron. Because the lifetime is certainly reduced by the flow of heat to the surrounding water in the experiments presented above, the longest allowed wavelength should be less than 1 micron, which is in agreement with submicrometer spike separation observed here.
[0049] In some embodiments, rather than utilizing a fluid, a solid substance having an electron-donating constituent (e.g., a sulfur powder) is disposed on at least a portion of a surface of semiconductor substrate, e.g., a silicon wafer. That surface portion is then irradiated with one or more pulses having pulse widths in a range of about 50 fs to about 500 fs so as to generate a plurality of inclusions containing the electron donating constituent in a surface layer of the substrate.
[0050] Those having ordinary skill in the art will appreciate that various changes can be made to the above embodiments without departing from the scope of the invention.