3. Experimental Details
3.6 Scanning Electron Microscopy
Fig. 3.6 An overall view of (a) the thermal evaporation machine (b) the e-beam evaporation machine.
3.6 Scanning Electron Microscopy
Electron microscope utilizes an electron beam to produce a magnified image of the sample. There are three principle types of electron microscopes: scanning, transmission, and emission. In the scanning and transmission microscope, an e-beam incident on the sample produces an image while in the field-emission microscope the specimen itself is the source of electrons. Scanning electron microscopy (SEM) is similar to light microscopy with the exception that electrons are used instead of photons and the image photons and the image is formed in a different manner. A SEM consists
of an electron gun, a lens system, scanning coils, an electron collector, and a cathode ray display tube (CRT). The electron energy is typically 10-30 KeV for most samples, but for insulating samples the energy can be as low as several hundred eV. The use of electrons has two main advantages over typical microscopes: much larger magnifications are possible since electron wavelengths are much shorter than the photon wavelengths and the depth of field is much higher.
De Broglie proposed in 1923 that particles can also behave as waves. The electron wavelength λe depends on the electron velocity v or the accelerating voltage V as
V
The wavelength is 0.012 nm for V=10000 (v), a wavelength significantly below the 400 to 700 nm wavelength range of visible light – making the resolution of an SEM much better than that of an optical microscope.
The contrast in a SEM depends on a number of factors. For a flat, uniform sample the image shows on contrast. If, however, the sample consists of materials with different atomic numbers, a contrast is observed if the signal is obtained from the backscattered electrons, because the backscattering coefficient increases with the atomic number Z.
The secondary electron emission coefficient, however, is not a strong function of Z and atomic number variations give no appreciable contrast. Contrast is also influenced
feature is the sample topography. Secondary electrons are emitted from the top 10 nm or so of the sample surface. When the sample surface is tilted from the normal beam incidence, the electron beam path lying within this 10 nm is increased by the factor1/cosθ, where θ is the angle from the normal incidence (θ = 0° for normal incidence). The interaction of the incident beam with the sample increases with path length and the secondary electron emission coefficient increases. The contrast C depends on the angle as C tan()d .
For θ = 45° , a change in the angle of dθ = 1° produces a contrast of 1.75% while at 60°
the contrast increase to 3% for dθ = 1°.
The beam diameter of SEMs is in the range of 1 to 10 nm. Yet the resolution of e-beam measurements is not always that good. Why is that? It has to do with the shape of the electron-hole cloud generated in the semiconductor. When electrons impinge on a solid, they lose energy by elastic scattering and inelastic scattering. Elastic scattering is caused mainly by interactions of electrons with nuclei and is more probable in high atomic number materials and at low beam energies. Inelastic scattering is caused mainly by scattering from valence and core electrons. The result of these scattering events is a broadening of the original nearly collimated, well-focused electron beam within the sample.
The generation volume is a function of the e-beam energy and the atomic number
Z of the sample. Secondary electrons, backscattered electrons, characteristic and continuum X-rays, Auger electrons, photons, and electron-hole pairs are produced. For low-Z samples most electrons penetrate deeply into the sample and are absorbed. For high-Z samples there is considerable scattering near the surface and a large fraction of the incident electrons is backscattered. The shape of the electron distribution is
“teardrop”-shaped, as shown in Fig. 3.7. As Z increase (15<Z<40) the shape becomes more spherical and for Z>40 it becomes hemispherical. “Teardrop” shapes have been observed by exposing polymethylmethacrylate to an electron beam and etching the exposed portion of the material. Electron trajectories, calculated with Monte Carlo techniques, also agree with these shapes.
Fig. 3.7 Summary of the range and spatial resolution of backscattered electrons, secondary electrons, X-rays, and Auger electrons for electrons incident on a solid
The depth of electron penetration is the electron range Re, defined as the average total distance from the sample surface that and electron travels in the sample along a trajectory. A number of empirical expressions have been derived for Re. One such expression is
) 10 (
28 .
4 6 1.75 E cm Re
(3.2)
where ρ is the sample density (g/cm3) and E the electron energy (KeV).
Chapter 4
Fabrication and Characterization of
Si-nanotips Array
Chapter 4 Fabrication and Characterization of Si-nanotips Array
4.1 Introduction
Antireflection techniques play an important role for many optoelectronic devices, such as solar cells, displays, and light sensors.1-4
One of popular methods to suppress the reflection is surface coatings. And another one is to use a SWS as an antireflection layer. The SWS is the surface-relief grating with the period smaller than the light wavelength, behaves as an antireflection surface.5-8 However, there are problems of surface coating such as thermal mismatch, adhesiveness, and stability. So an alternative to surface coatings is the surface-relief structures with subwavelength structures (SWS).
The SWS can solve the above problems of surface coatings due to the same material constitutions as the substrate. According to previous reports, they show that the reflectivity can be suppressed for a wide spectral bandwidth. In this work, we use relative ion etching with Ag nanopaticles as a mask to fabricate Si-nanotips with SWS surface to suppress the reflection. We will discuss the different antireflection characteristics for the SWS surface obtained from different RIE times.
4.2 Experiment
The fabrication procedure of Si nanotips array with SWS surface is schematically
shown in Fig.4.1. Firstly, p-type (100) Si wafer with a resistivity of 10 ohm was cut into 1×0.5 cm2 used as the substrate. Before the fabrication, the Si substrate was ultrasonically cleaned with acetone, methanol and DI water for 3 min, and Piranha Clean process (SPM) composed of H2SO4/H2O2 (4:1) was used to clean organic substance under 130℃ for 10 min. The wafers were then dipped in buffer oxide etching (BOE, 1:6) to remove the native oxide, and they were cleaned with DI water.
sputter Ag film and
anneal to form Ag nanoparticles
dry etching by RIE Si wafer
Removal of residual Ag Si-nanotips
sputter Ag film and
anneal to form Ag nanoparticles
dry etching by RIE Si wafer
Removal of residual Ag sputter Ag film and
anneal to form Ag nanoparticles
dry etching by RIE Si wafer
Removal of residual Ag sputter Ag film and
anneal to form Ag nanoparticles
dry etching by RIE Si wafer
Removal of residual Ag Si-nanotips
Fig. 4.1 Schematic diagram for the fabrication procedure.
Secondly, an Ag thin film about 10 nm was deposited onto the p-Si wafer by using DC sputtering technique (JEOL, JFC-1600 auto fine coater). Then, they were put into a furnace and annealed for 10 min at different temperatures of 300, 400, 500, and 600℃, to form Ag nanoparticles which will serve as the etching mask in the following process. The Si wafer was then etched by using reactive ion etching (RIE) equipment (Oxford plasma 80 plus) at the CF4 flow rate of 40 sccm, the pressure is 25 mtorr, and the RF power is 100 W. To generate Si nanotips array with different heights, etching times of 10-40 min were used. After dry etching by RIE, the substrate was dipped in HNO3 (10%) solution in order to remove the residual Ag nanoparticles. Finally, the sample was immersed in BOE (1:6) to remove silicon oxide, which was formed by HNO3 clean step.
4.3 Results and discussion
Figure 4.2 shows the etching mask of Ag nanoparticles after the Ag film was annealed at different temperatures of 300, 400, 500, and 600℃. Under 300℃ condition, it shows the Ag film can not form Ag nanoparticles, because of the temperature is not high enough to change the Ag film into Ag nanoparticles. Under 400℃ condition, the formed Ag nanoparticles show anomalous shape instead of circular. Because the temperature is not high enough, silver intermolecular cohesive force is not big enough.
From this reason, the Ag nanoparticles possess an anomalous shape.
(a)
(b)
Fig. 4.2 Scanning electron microscope images of Ag nanoparticles deposited on Si wafer at different temperatures (a) 300℃ (b) 400℃ (c) 500℃ (d) 600℃
(c)
(d)
At 500℃, the nanoparticles size of ranges from about 10 nm to 100 nm, and the distribution is even. When the temperatures increase to 500℃ and 600℃, the Ag intermolecular cohesive force is enough to cause the Ag film to contract the circular pellet. Under 600℃ condition, the diameter and distance of Ag nanoparticles were both larger than 500℃ condition. From the above discussion, we choose the thermal dewetting of 500℃ to process the dry etching by using RIE. Figure 4.3 shows the scanning electron microscopy (SEM) images for the fabricated Si nanostructures with different etching times. With increasing of the RIE time, the dimensions of the nanotips decrease and the grooves of the nanotips grow deep, which is shown in table 4.1 and figure 4.4. The heights of Si nanotips were measured to be 143, 315, 361 and 402 nm for etching times of 10, 20, 30 and 40 min, respectively. As the RIE times increased, the nanotips become concentrated, and the appearance of Si nanotips varied from flat to sharp. From figure 4.3, we discovered that the Si nanorods were formed when the etching times have the values of 10 and 20 min. With increasing RIE times to 30 and 40 min, the Si nanotips were formed, and the Si-nanotips have a conical profile. To analyze the light trapping effect of Si-nanotips array, the reflectivity was measured. Figure 4.5 exhibits the reflectivity of wavelength dependence from 350~750 nm for Si substrate, Si-nanotip with etching time of 10, 20, 30, and 40 min, and the corresponding reflectivity is 45%, 8.6%, 6.2%, and 2.8%, respectively.
(a)
(b)
Fig. 4.3 Morphology of Si-nanotips array fabricated by RIE with different times (a)10min (b) 20 min (c) 30 min (d) 40 min
(c)
(d)
Fig 4.4 The grooves deep of Si nanotips array in different RIE times.
Table 4.1 Detailed list of Si nanotips grooves deep.
Etching time 10min 20min 30min 40min
Average depth 143nm 315nm 361nm 402nm
10 15 20 25 30 35 40
100 150 200 250 300 350 400
Depth (nm)
Etching time (min)
Fig 4.5 Reflectivity for Si nanotips with different RIE times.
300 400 500 600 700 800
0.0
It indicates that with the increasing of RIE time, the reflectivity is decreased. We summarized the results in Fig 4.6 and table 4.2.
The Si nanotips could be treated as an excellent antireflection layer due to the effective medium is a gradient profile. According to Eqs. 4.1 and 4.2, the Fresnel reflection can be suppressed to a very low value. From above results, the Si nanostructure possesses an excellent antireflection characteristic, which is very attractive in many optoelectronic applications, including solar cells, light emitting diodes (LEDs), and photodiodes.
Fig. 4.6 The average reflectivity for Si nanotips with different RIE time in visible light region.
Table 4.2 Detailed list of Si nanotips reflectivity average in visible region.
Summary
In summary, a simple lithographic approach is brought forward for SWS antireflection, in which the metal island films are used as shadow masks. Ag film at different temperatures forms different morphologies, which can be used as etching masks. The Si nanptips array with SWS surface could be treated as a good antireflection layer. The best antireflection of our Si-nanotips is fabricated by RIE for 40min and the deep grooves and periodic are about 400 nm and 300 nm, respectively. The Fresnel reflection can be suppressed to as low as 2.8%.
Etching time(min) 0 min 10min 20min 30min 40min
References
1. C. C. Striemer and P. M. Fauchet, Appl. Phys. Lett. 81, 2980 (2002).
2. T. Beyer and M. Tacke, Appl. Phys. Lett. 73, 1191 (1998).
3. A. Gombert, W. Glaubitt, K. Rose, J. Dreibholz, B. Bläsi, A. Heinzel, D. Sporn, W.
Döll, and V. Wittwer, Sol. Energy 68, 357 (2000).
4. M. Rajteri, M. L. Rastello, and E. Monticone, Nucl. Instrum. Methods Phys. Res. A
444, 461 (2000).
5. Y. Kanamori, M. Sasaki, and K. Hane, Opt. Lett. 24, 1422 (1999).
6. P. Lalanne and G. M. Morris, Nanotechnology 8, 53 (1997).
7.M. E. Motamedi, W. H. Southwell, and W. J. Gunning, Appl. Opt. 31, 4371 (1992).
8. S. J. Wilson and M. C. Hutley, Opt. Acta 29, 993 (1982).
Chapter 5
Highly Sensitive
Photodetector Based on n-ZnO/p-Si-nanotips
Heterostructure
Chapter 5 Highly Sensitive Photodetector Based on n-ZnO/p-Si-nanotips Heterostructure
5.1 Introduction
ZnO has been widely studied due to its unique optical properties in the ultraviolet (UV) region, with its wide band gap (Eg=3.2eV) and large exciton binding energy1-14 (about 60meV). However, the high quality p-ZnO is difficult to obtain2, owing to its self-compensation effect and ionic nature. Therefore, there exists a large obstacle to fabricate good quality ZnO p-n junction, which has been served as the basic building block for most optoelectronic devices. To overcome this difficulty, the combination with other p-type materials, such as Si, GaAs and GaN, may provide an excellent alternative.
Silicon plays the most important role in current semiconductor industry, due to its low cost and mature technology. If the fabrication process of a newly desinged device is comparable with the existing Si technology, it should be very useful for the practical application.
Recently, Si nanostructures have attracted a great deal of attention due to their potential applications in many different areas, such as light emitting diodes, solar cells, photodetectors, and lasers10-14. Among all these possibilities, Si nanotips array represents one of the most promising candidates for optoelectronics devices, because of the excellent antireflection characteristics arising from its light trapping capability.
In this work, we combined both of the advantages of Si-nanotips and ZnO material to fabricate n-ZnO/p-Si-nanotips heterojunction photodiodes. It is found that the newly designed photodiode can serve as a highly sensitive detector with a wide rang of spectrum covering from infared to ultraviolet radiation. Our study shown here may pave a key step for the future development of many optoelectronic devices based on one-dimensional semiconductor nanostructures.
5.2 Experiment
P-type (100) Si with a resistivity of 10 ohm was used as the substrate. Before the fabrication, the Si substrate was ultrasonically cleaned with acetone, methanol and DI water for 3 min, and Piranha Clean process (SPM) composed of H2SO4/H2O2 (4:1) was used to clean organic substance under 130℃ for 10 min. The wafers were then dipped in buffered oxide etch solution (BOE 1:6) to remove the native oxide. Finally, they were cleaned with DI water and dried with nitrogen. .
After the above cleaning process, an Ag thin film about 10 nm was deposited onto the p-Si wafer by using a DC sputtering system. Then, they were put into a furnace and annealed under 500℃ for 10 min, to form Ag nanoparticles, which will serve as the etching mask in the following process. The Si wafer was then etched by using reactive ion etching (RIE) technique (Oxford plasma 80 plus) at the CF4 flow rate of 40 sccm for 40 min, the pressure is 25 mtorr, and the RF power is 100 W. After dry etching by RIE,
the substrate was dipped in HNO3 (10%) solution in order to remove the residual Ag nanoparticles. Finally, the sample was immersed in BOE (1:6) to remove silicon oxide, which was formed by HNO3 clean step.
ZnO films were deposited on Si-nanotips and Si-planar substrates in a e-beam evaporator system by using 0.5 cm thick ZnO ingots with 0.6 cm diameter (99.9%
purity, from Well-Being Enterprise CO.). The chamber base pressure was evacuated to 5×10-6 torr, and the substrate temperature was controlled at 250℃ with working
pressure of 10-5 torr and deposited rate of 1 A /S. After deposition, the sample was o annealed in furnace at 550℃ for 30 min to activate the ZnO films. The thickness of ZnO film was approximately 200 nm.
Ti/Au bilayer metal contacts with a thickness of 12 nm/120 nm were deposited on the ZnO films with a shadow mask by thermal evaporation at room temperature. The sample was followed by rapid thermal annealing process at 600℃ for 30 s in N2
ambient. Al metal contact with a thickness of 100 nm was deposited by thermal evaporation and annealed at 550℃ for 10 min. Figure 5.1 summarizes the schematic plot describing the fabrication of n-ZnO/p-Si-nanotips photodiodes.
The morphology of Si-nanotips was recorded by scanning electron microscopy (SEM) images using a JEOL JSM 6500 system. Photoluminescence (PL) spectra were performed at room temperature and excited by a 325nm He-Cd laser beam. The quality
of ZnO films was further investigated by Raman scattering measurements (Jobin-Yvon T64000 system) using 325nm He-Cd laser beam. Dark current and photocurrent were measured for n-ZnO/p-Si-planar and n-ZnO/p-Si-nanotip photodiodes by using Keithley equipment. Xe-arc lamp (Oriel Optical system, 450W) was used as the light source, and a monochromator covering the range from 350-650nm. The photodiodes were illuminated at normal incident.
Fig. 5.1. Schematic plots showing the fabrication process and the device structure of n-ZnO/p-Si-nanotips photodiode.
ZnO films and contact deposited on the sample
Al
ZnO Ti/Au sputter Ag film and anneal
to form Ag nanoparticles
dry etching by RIE
Si-nanotip Si wafer
ZnO films and contact deposited on the sample
Al
ZnO Ti/Au sputter Ag film and anneal
to form Ag nanoparticles
dry etching by RIE
Si-nanotip Si wafer
ZnO films and contact deposited on the sample
Al
ZnO Ti/Au sputter Ag film and anneal
to form Ag nanoparticles
dry etching by RIE
Si-nanotip
ZnO films and contact deposited on the sample sputter Ag film and anneal
to form Ag nanoparticles
dry etching by RIE
Si-nanotip
sputter Ag film and anneal to form Ag nanoparticles
dry etching by RIE sputter Ag film and anneal to form Ag nanoparticles sputter Ag film and anneal to form Ag nanoparticles
dry etching by RIE
Si-nanotip
Si wafer
5.3 Results and discussion
Figure 5.2 (a) shows the etching mask of Ag nanoparticles after the Ag film was annealed at 500℃. We can see that the size of nanoparticles ranges from about 10 nm to 100 nm. The morphology of Si-nanotips by RIE treatment is shown in Fig. 5.2 (b). The Si-nanotips have a conical profile and the groove is around 400 nm deep. The distance between Si-nanotips is approximately 300 nm. To analyze the light trapping effect of Si-nanotips array, the reflectivity was measured. Figure 5.3 (a) exhibits the reflectivity of wavelength dependence from 350~750 nm, which shows that for polished Si substrate and Si-nanotips array, the reflectivity is 45% and 2.8%, respectively. It indicates that the Si-nanotips array possesses an excellent antireflection characteristic, which is very attractive for many optoelectronic applications.
Figures 5.2 (c) and (d) show the SEM pictures of ZnO films on Si-nanotips.
Obviously, the whole Si-nanotips array was covered quiet well with ZnO films. The optical characteristics and the stoichiometric formation of ZnO films were investigated by Raman scattering, transmittance, and photoluminescence measurements. As shown in Fig. 5.3 (b), a strong and narrow peak was observed at 575 cm-1, which corresponds to the LO mode of ZnO. It indicates the good crystalline quality of the ZnO films.
Fig. 5.2. (a) Scanning electron microscope image of Ag nanoparticles deposited on Si wafer (b) Morphology of Si-nanotips array (c) ZnO films deposited on Si-nanotips (d) Detailed micrograph of ZnO films deposited on Si-nanotips.
(a) (b)
(c) (d)
A typical transmission spectrum of ZnO films deposited on glass is shown in Fig.
5.3 (c).The transmittance results show the normal transmittance is more than 80% in the 400~700 nm region, which evidently indicates that the high optical quality of the ZnO film. An abrupt decrease at 380 nm and a little variation at near 500 nm were observed, which correspond to the absorption of the band edge at around 3.26 eV and midgap states at around 2.5 eV, respectively. As shown in Fig. 5.3 (d), base on the photoluminescence spectrum of ZnO films deposited on Si nanotips, an obvious ultraviolet emission peak was observed at 380 nm, again which corresponds to the band gap emission of ZnO. The broad visible light emission centered near 500 nm, corresponds to defects emission.
In the inset of Fig. 5.4 (a) ,the current-voltage curves are linear for a pair of Ti/Au contact on n-ZnO (right) and Al on Si (left), respectively, which ensure the ohmic contact behavior for the metal-semiconductor interfaces used in our study. After ensuring ohmic contacts on both n-ZnO and p-Si, the I-V curves for both n-ZnO/p-Si-planar and n-ZnO/p-Si-nanotips photodiodes were measured in a dark room.
As shown in Fig. 5.4 (a), a typical rectifying and diode-like behavior was observed for n-ZnO/p-Si-nanotips (solid curve) and n-ZnO/p-Si-planar (dashed curve) samples. The rectifying behaviors of forward/reverse current ratio are 176 and 348 for n-ZnO/p-Si-nanotips and n-ZnO/p-Si-planar at ±1V, respectively. And the leakage
current of n-ZnO/p-Si-planar and n-ZnO/p-Si-nanotips at the reverse bias of -4V are 1.9×10-5 Acm-2 and 3.2×10-5 Acm-2, respectively. The turn-on voltage of both photodiodes has a similar value of about 0.5V.
300 400 500 600 700 800
0.0 0.2 0.4 0.6 0.8 1.0
Reflectivity
Wavelength (nm)
Silicon substrate Si-nanotips Si/ZnO
Si-nanotips/ZnO
(a)
400 500 600 700 800
Raman intensity ( a.u . )
Raman shift ( cm-1 )
(b)
300 400 500 600 700 800
0 10 20 30 40 50 60 70 80 90 100
T ran smitt a n ce ( % )
Wavelength (nm)
(c)
350 400 450 500 550 600 650
PL intensity ( a.u . )
Wavelength (nm)
(d)
Fig. 5.3. (a) Optical reflectance spectra of bulk Si (rectangular curve), Si-nanotips array (triangle curve), n-ZnO/p-Si-nanotips (rhombus curve) structure and n-ZnO/p-Si-planar (circle curve) structure. (b) Raman spectrum of ZnO films deposited on p-Si nanotips (c) Optical transmittance of ZnO films deposited on glass (d) Photoluminescence of ZnO films deposited on p-Si-nanotips.
-4 -3 -2 -1 0 1 2
-0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 -10
Fig. 5.4. (a) The current-voltage curves of n-ZnO/p-Si-planar (dashed curve) and n-ZnO/p-Si-nanotips (solid curve) photodiodes measured in a dark room (insert:
current-voltage curve of Al contact on p-Si (left); current-voltage curve of Ti/Au contact on n-ZnO (right)) (b) Spectral photoresponsivities of n-ZnO/p-Si-planar (dashed curve) and n-ZnO/p-Si-nanotips (solid curve) photodiodes.
350 400 450 500 550 600 650
0.01 0.1 1
Responsivity (A/W)
Wavelength (nm)
p-Si-nanotips/ZnO
p-Si-planar/ZnO
(b)
The following is the most significant study of the present work. We measure the spectral photoresponsivities for both n-ZnO/p-Si-nanotips and n-ZnO/p-Si-planar photodiodes. The dashed curve and solid curve in Fig. 5.4 (b) correspond to the results of n-ZnO/p-Si-planar and n-ZnO/p-Si-nanotips photodiodes, respectively. For the n-ZnO/p-Si-planar photodiodes, the photoresposivities at wavelength between 400 and 650 nm with an average 0.23 AW-1 at -4V were obtained, which is similar to previous reports11. Quite interestingly, at the same wavelength region, the photoresponsivity
The following is the most significant study of the present work. We measure the spectral photoresponsivities for both n-ZnO/p-Si-nanotips and n-ZnO/p-Si-planar photodiodes. The dashed curve and solid curve in Fig. 5.4 (b) correspond to the results of n-ZnO/p-Si-planar and n-ZnO/p-Si-nanotips photodiodes, respectively. For the n-ZnO/p-Si-planar photodiodes, the photoresposivities at wavelength between 400 and 650 nm with an average 0.23 AW-1 at -4V were obtained, which is similar to previous reports11. Quite interestingly, at the same wavelength region, the photoresponsivity