Chapter 1 Introduction
1.4 Thesis Organization
In chapter 1, a brief overview of LEDs and related developments were introduced.
In chapter 2, the fabrication process flow of LEDs, experimental recipes, and device parameter extraction methods will be described.
In chapter 3, we will show and discuss the characteristics of ITO film, ITO/p-GaN interface, and LED device, including sheet resistance, transmittance, specific contact resistance, I-V and L-I characterization.
In chapter 4, a short conclusion is addressed.
Finally, future works as well as suggestion for further research are given in chapter 5
Chapter 2 Experiment
2.1 The Process Flow of LEDs
The GaN-based LED epi-wafers used in this study were all grown on a 2-inch-diameter c-face sapphire (0001) substrates by metal-organic chemical vapor deposition (MOCVD) system. Details of the growth procedures could be found elsewhere [30, 31]. The epitaxial structure is consisted of 4-μm-thick n-type GaN layer, 0.1-μm-thick InGaN-GaN multiple-quantum-wells (MQWs), and 0.1-μm-thick p-type GaN layer. The carrier concentrations of the p-type and n-type GaN were 5×1017 and 3×1018 cm-3, respectively. The as-grown samples were rapidly thermal annealed (RTA) at 800 for Mg℃ -type activation. All wafers with a peak wavelength at 465 nm were chosen and cut into four pieces for different processes. One piece of each wafer was prepared for the conventional LEDs with Ni/Au (4nm/4nm) contact structure as a reference sample. The other pieces were prepared as experimental samples with ITO contacts of different thicknesses and annealed at different temperatures.
GaN-based LEDs (300 × 300 μm) with different conducting layers were fabricated. Figure 2-1 shows the cross section and process flow of LEDs. By means of photoresist lithography, the mesa region could be defined after removing photoresist.
The surface of p-type GaN layer was partially etched by an inductance-coupled plasma (ICP) etcher until the n-type GaN layer was exposed. Then these samples were immersed in H2SO4 : H2O2 and NH4OH : H2O solutions to remove photoresist mask and residue chemical oxides. Next, ITO was deposited on samples by E-beam evaporator at 300 in oxygen ambience.℃ The ITO source consisted of 95% In2O3 and 5% SnO2. The thicknesses of ITO were 60 nm, 180 nm and 300 nm with different
oxygen flow rates, as listed in table 2-1. The base pressure was 5×10-6 torr and working pressure was 2.2×10-4 torr and 4.5×10-4 torr when oxygen flow rate was 15 sccm and 30 sccm, respectively. ITO was also deposited on double-side-polished GaN wafers and quartz substrates for transmission studies. The Ni/Au (4nm/4nm) contact layer was deposited by E-beam evaporator with a base pressure of 2×10-6 torr. After lithography process, these ITO and Ni/Au contact samples were wet etched by ITO etchant and KI : HCl solution, respectively. Then, all samples with different thicknesses of ITO films were subsequently annealed in nitrogen ambience for 10 minutes at 400 , 500 , and 600 , respectively. As for the ℃ ℃ ℃ Ni/Au contact sample, it was annealed at 540 in nitrogen ambience to achieve the optimal ohmic contact to ℃ p-GaN. Finally, Cr-Au (0.1 μm/1.4 μm) was used as the n-type contact layer, p- and n-bonding pads. Cr was used for the CBL, as mentioned above.
After fabrication processes, the sheet resistance was measured using a four-point probe. In order to truly exhibit the effects of absorption, internal reflection and interference of the conducting films on GaN-based LEDs, the p-i-n GaN with a double-side-polished sapphire substrate was used in the transmittance measurement.
The transmission spectra of the deposited films were measured by a Hitachi U3010 dual-beam spectrophotometer. Scanning electron microscope (SEM) was used to observe the grain size of ITO with different thicknesses.
Test keys consisting of transmission line model (TLM) patterns were used for calculating the specific contact resistance of the conducting layers on p-GaN and sheet resistance of p-GaN. After the sheet resistance and specific contact resistance of all samples were known, we created a calculation for checking the behavior of current distribution and calculating the current spreading length.
The current-voltage (I-V) characteristics were measured at room temperature using a HP4156A semiconductor parameter analyzer with a voltage source. The
relationship between LED output power and injection current (L-I) was also be measured by the integrated sphere detector from top of the devices.
The detailed fabrication process flow is listed as follows:
1. As-grown GaN wafer on (0001) sapphire substrate.
2. Initial cleaning (H2SO4 : H2O2).
3. Mask #1: Define mesa region.
4. Dry etching by ICP-RIE.
5. Photoresist removing and cleaning (H2SO4 : H2O2, NH4OH : H2O).
6. TCL evaporation (ITO with 60~300 nm and Ni/Au with 4/4 nm).
7. Mask #2: Define TCL region.
8. Etching TCL and removing photoresist.
9. TCL alloy for 10 minutes in furnace (400~600 for ITO and ℃ 540 for ℃ Ni/Au).
10. Mast #3: Define pad region.
11. P/N bonding pad metal evaporation (Cr-Au with 0.1/1.4 μm).
12. Removing metal by lift-off procedure.
13. P/N bonding pad alloy at 510 for 5 minutes in furnace℃ for Ni/Au samples.
2.2 Methods of Measurements and Analysis 2.2.1 The Transmittance of ITO Film
Before every optical transmittance measurement by the Hitachi U3010 spectrophotometer, the wavelength calibration had been done to ensure the credibility of measured data. The transmittance of each film was normalized with respect to air, as shown in Fig. 2-2. The light passed through ITO film to the air, as same as the actual situation.
2.2.2 Transmission Line Model (TLM)
The transmission line model (TLM) was used for calculating the specific contact resistance of the conducting layers on p-GaN and sheet resistance of p-GaN. The pad size was 280×75 μm and the spacings between the pads were 10, 20, 30, 40, and 65 μm. A simple diagram of TLM pattern is shown in Fig. 2-3. We found the total resistance between any two contacts to be
(
T where RT and RS is the total resistance and contact resistance of p-GaN, and we hadused the approximation (The transfer length
S c
T R
L = r , rc is the specific contact
resistance). The total resistance was measured for various contact spacings d and RT
was plotted as function of d as illustrated in Fig. 2-4. Three parameters could be extracted from such a plot. The slope Δ(RT)/ Δ(d) = RS/Z led to the sheet resistance with the contact width Z independently measured. The intercept at d = 0 was RT = 2RC giving the contact resistance. The intercept at RT = 0 gives –d = 2LT, which in turn could be used to calculate the specific contact resistance with RS known from the slope of the plot.
2.2.3 Calculation of Current Spreading Length
After obtaining the sheet resistance of ITO film and the contact resistivityof ITO/p-GaN, we calculated the current distribution and defined current spreading length by creating an equivalent LED circuit [32-35]. The current distribution can be denoted by this equation
( ) ( )
are the electrical resistivities and thicknesses of p, n, and transparent layer. Variable x represents the distance from p-pad. Denoting Ls, current spreading length, as the length where the current density has dropped to the 1/e value of the current density at the p-pad, i.e., J(Ls)/J(0) = 1/e, yields
We will check if the current spreading length corresponds to device luminance later.
2.2.4 Current-Voltage Characteristic Measurement
The current-voltage (I-V) characteristics were measured at room temperature using a HP4156A semiconductor parameter analyzer with a voltage source. Ten devices of each sample were measured in order to choose a most reliable one to serve as the result. The relationship between LED output power and injection current (L-I) was also be measured by the integrated sphere detector from top of the devices. Each ITO sample was compared and normalized with Ni/Au sample which were cut from the same wafer to ensure a reliable result.
2.2.5 The Introduction of Scanning Electron Microscope (SEM)
Electron microscopy takes advantage of the wave nature of rapidly moving electrons. Where visible light has wavelengths from 4,000 to 7,000 Angstroms, electrons accelerated to 10,000 KeV have a wavelength of 0.12 Angstroms. Optical
microscopes have their resolution limited by the diffraction of light to about 1000 diameters magnification. Electron microscopes, so far, are limited to magnifications of around 1,000,000 diameters, primarily because of spherical and chromatic aberrations. Scanning electron microscope resolutions are currently limited to around 25 Angstroms, though, for a variety of reasons. A schematic diagram of a typical SEM is shown in Fig.2-5.
The scanning electron microscope generates a beam of electrons in a vacuum.
That beam is collimated by electromagnetic condensor lenses, focussed by an objective lens, and scanned across the surface of the sample by electromagnetic deflection coils. The primary imaging method is by collecting secondary electrons that are released by the sample. The secondary electrons are detected by a scintillation material that produces flashes of light from the electrons. The light flashes are then detected and amplified by a photomultiplier tube.
By correlating the sample scan position with the resulting signal, an image can be formed that is strikingly similar to what would be seen through an optical microscope. The illumination and shadowing show a quite natural looking surface topography.
There are other imaging modes available in the SEM. Specimen current imaging using the intensity of the electrical current induced in the specimen by the illuminating electron beam to produce an image. It can often be used to show subsurface defects. Backscatter imaging uses high energy electrons that emerge nearly 180 degrees from the illuminating beam direction. The backscatter electron yield is a function of the average atomic number of each point on the sample, and thus can give compositional information.
Scanning electron microscopes are often coupled with x-ray analysers. The energetic electron beam - sample interactions generate x-rays that are characteristic of
the elements present in the sample. Many other imaging modes are available that provide specialized information.
2.2.6 The Introduction of HP4156A
HP4156A is an electronic instrument for measuring and analyzing the characteristics of semiconductor devices. This one instrument allows us to perform both measurement and analysis of measurement results.
HP4156A has four highly accurate source/monitor units (SMUs), two voltage source units (VSUs), and two voltage measurement units (VMUs). The HP4156A is designed for Kelvin connections and has high-resolution SMUs (HRSMUs), so HP4156A is especially suited for low resistance and low current measurements. We can measure voltage values with a resolution of 0.2μV by using the differential measurement mode of VMUs.
HP4156A can perform stress testing. That is, can force a specified dc voltage or current for the specified duration.
Also, we can force ac stress by using pulse generator units (PGUs), which are installed in HP41501A SMU/Pulse Generator Expander. The HP41501A is attached to HP4156A, and can be equipped with a ground unit (GNDU), high power SMU (HPSMU), two medium power SMUs (MPSMUs), or two PGUs.
HP4156A can print and store, in addition to performing measurement and analysis. We can store measurement setup information, measurement data, and instrument setting information on a 3.5-inch diskette inserted into the disk drive of HP4156A. And we can print the setting information and measurement results on a plotter or printer that is connected to HP4156A.
HP4156A can be controlled by an external controller via HP-IB by using remote control commands. These commands are based on Standard Commands for
Programmable Instruments (SCPI), so we can easily develop measurement programs.
HP4156A has internal HP instrument BASIC, so we can develop and execute measurement programs by using the HP4156A only, without using an external controller.
Chapter 3
Results and Discussion
3.1 Properties of Films
3.1.1 Sheet Resistance of ITO Film
Figure 3-1 shows the sheet resistance of 180 nm and 300 nm thick ITO films which were evaporated on quartz substrates at two oxygen flow rates. When the flow rate is 30 sccm, we can find the sheet resistance decreasing after annealing at higher temperature. On the other hand, if the oxygen flow rate is 15 sccm, sheet resistance will have a maximum value after 500 annealing℃ . The appearances are similar on different thicknesses of ITO films. It is well known that the oxygen content in ITO films is a critical control parameter of the resistivity of the film. Conduction is partially a result of unfilled oxygen vacancies, so a high incorporation of oxygen into the film will result in few vacancies and a highly resistive film [36]. When annealing at higher temperature, the oxygen atoms may easily evaporate to the air. It will increase the amount of oxygen vacancies and attribute to a better conductivity of ITO film. Figure 3-2 shows the relationship between ITO film thickness and the sheet resistance. By any means, the sheet resistance decreases while increasing the thickness of ITO film due to a higher conductivity, as our preconception. The measured sheet resistances of all ITO films are listed in table 3-1.
3.1.2 Scanning Electron Microscope (SEM) Analysis
Figure 3-3 shows SEM pictures of ITO films with different thicknesses. ITO films were observed to be polycrystalline and showed a similar crystal structure to that of undoped In2O3. The grain size is about 70 nm, 120 nm, and 160 nm for 60 nm, 180 nm, and 300 nm ITO film, respectively. These pictures indicate that the grain size
of the films increased with growing film thickness. In the thin film growth, it is generally observed that the grain size increases with increasing film thickness [37, 38].
The larger grain size results in a lower density of grain boundaries, which behave as traps for free carriers and barriers for carrier transport in the film [39]. Hence, an increase in the grain size can cause a decrease in grain boundary scattering, which leads to an increase in the conductivity. It may be one reason of why thicker ITO film got a smaller sheet resistance.
3.1.3 Transmittance of ITO Film
Figure 3-4 shows the optical transmittance as a function of wavelength of the evaporated ITO film on GaN substrate with different annealing temperatures. In this figure, the transmittance of each film was directly compared with air. By varying the annealing temperature, the transmittances are almost the same. However, the curves were long-wavelength shifted after annealed at higher temperature. It means an increase of the refractive factor of ITO film. The transmittance of different ITO film thicknesses are showed in Fig. 3-5. For comparison, the most commonly used Ni(4 nm)-Au(4 nm) contact layer that was annealed at 540℃ in nitrogen ambience was also examined. The transmittances are about 80% at a wavelength of 465 nm of all thicknesses of ITO films. Compared with Ni/Au film, transmittance of ITO has an enhancement of about 20%.
3.2 Interfacial Characteristics
3.2.1 Specific Contact Resistance of ITO or Ni/Au Layer on p-GaN
Figure 3-6 shows the current-voltage (I-V) characteristics of Ni/Au (4 nm/4 nm) and ITO (300 nm) contacts on p-GaN after annealing at temperatures of 400-600 ℃ for 10 minutes in N2 ambient. It was found that both ITO and Ni/Au films could form
good ohmic contacts on p-GaN. Specific contact resistance was determined from plots of the measured total resistance versus the spacings between the TLM pads. The least square curve-fitting method was used to fit a straight line to the experimental data, as shown in Fig. 3-7. Although ITO and Ni/Au films both have ohmic contacts on p-GaN, Ni/Au still has the smaller specific contact resistance than all ITO films. To explain the low resistance of Ni/Au contacts, Ho et al. proposed that NiOx layer on the p-GaN epilayer surface causes a reduction of the schottky barrier height across the interface [40]. However, Maeda et al. suggested that oxygen reacts with hydrogen in the Mg-H bonds and, thus, reduces the hydrogen concentration and modifies the energy band structure in the interface region [41]. Moreover, another study suggested that ohmic contact characteristics result from the removal of surface contamination on GaN by Ni before or during layer reversal [42]. Figure 3-8 exhibits the specific contact resistances of ITO films on p-GaN which were evaporated with different oxygen flow rates and annealed at 400-600 .℃ It is obvious that the specific contact resistance had a minimum value at 500℃ with 15 sccm O2 flow rate and increased with the annealing temperature with 30 sccm O2 flow rate. This may be due to the variance of oxygen proportion in the ITO/p-GaN interface which will combine with gallium to form GaO. Moreover, ITO films that were evaporated with 30 sccm oxygen flow rate have a higher specific contact resistance than with 15 sccm. This may be due to the incorporation of oxygen. The relationship between specific contact resistance and ITO film thickness is not discussed because the thickness should not affect the interface characteristics. A particular data of all samples’ specific contact resistance is listed in table 3-2. The sheet resistance of p-GaN was also calculated simultaneously and is listed in table 3-3. The fluid values of p-GaN’s sheet resistance may be attributed to the instability of carrier concentration. However, the sheet resistance of p-GaN is larger than that of ITO film and result in a poor conductivity. That’s why we need an
additional transparent conducting layer upon p-GaN.
3.2.2 Current Spreading Behavior by Calculation
After the sheet resistances and specific contact resistances were known, we started the current spreading calculation. The calculated current distribution of the LED is shown in Fig. 3-9. As the results by calculation, we can conclude that thicker ITO film contributes to a uniform current spreading phenomenon. This instance is chiefly caused by the smaller sheet resistance of thicker film. It can be seen that current crowding can be alleviated via the use of a thicker ITO film. For simple expression, we define current spreading length Ls as the length where the current density has dropped to the 1/e value of the current density at the p-pad. Figure 3-10 shows the calculated current spreading length of different thicknesses of ITO films that were evaporated with 15 sccm oxygen flow rate. This figure indicates that no matter at what annealing temperature, current will spread farther by increasing ITO film’s thickness due to the lower sheet resistance. Figure 3-11 shows Ls of 300 nm ITO films which were evaporated with different oxygen flow rates. With a 30 sccm oxygen flow rate during evaporation, sample annealed at 600℃ has a longest Ls compared with other temperatures because of its largest specific contact resistance.
The large specific contact resistance makes the current tend to flow horizontally.
Same theory can also explain the shortest Ls at 500 with 15 s℃ ccm oxygen flow rate by having the smallest specific contact resistance. Ls of ITO samples are listed together in table 3-4.
3.3 Electrical and Optical Properties of Devices 3.3.1 Reverse Current-Voltage Characteristic of LEDs
Not only forward voltage was applied, reverse voltage was also used to check the
value of leakage current, as shown in Fig. 3-12. The leakage current is only several nA which is small enough as our toleration, and not affected by annealing temperature.
The fabricated LEDs’ leakage currents are generally small enough except two wafers.
Figure 3-13 points out the high leakage current of these devices with 180 nm ITO film evaporated with 15 sccm oxygen flow rate and annealed at 400℃. It may be caused by the bad epitaxy property of GaN wafer and will result in the output power’s degradation and a serious thermal effect.
3.3.2 Forward Current-Voltage Characteristic of LEDs
Figure 3-14 depicts the room-temperature forward current-voltage characteristics of the fabricated GaN-based LEDs. These samples were all annealed at 400 .℃ In Fig.
3-14(a), the 20-mA forward voltage measured from LEDs with 60 nm ITO, 180 nm ITO, 300 nm ITO, and Ni/Au on p-GaN was 3.45, 3.41, 3.32, and 3.14 V, respectively.
Apropos of LEDs with films which were evaporated with high oxygen flow rate, the forward voltage was 3.5, 3.31, 3.26, and 3.08 V, respectively, as shown in Fig. 3-14(b).
The forward voltage decreases with an increase of ITO film’s thickness. As the specific contact resistance doesn’t form an order, the forward voltage may be affected
The forward voltage decreases with an increase of ITO film’s thickness. As the specific contact resistance doesn’t form an order, the forward voltage may be affected