2.1 Growth Mechanism of Atomic Layer Deposition
Atomic Layer Deposition (ALD) is a chemical vapor deposition technique with layer-by-layer growth mechanism and the features of low growth temperature, high uniformity, low vacuum demand, and thickness controlling in the nanometer range.
Such features are the results of the sequential and self-terminating chemical reactions of the precursors on substrate surface. For depositing the epitaxial ZnO thin film, the reacting temperature of the chemical reaction is ranged from room temperature to about 250 ºC, and the most suitable growing temperature is about 200 ºC. The precursor for providing zinc is the diethyl-zinc (DEZ), which is the colorless liquid contained in the bubbler and has vapor pressure about 15 torr at 25 ºC and boiling point at 118 ºC under 1 atm.
The sequential self-terminating gas-solid reactions for growing ZnO layers by ALD consist of the following four steps:
(a) The introducing and the self-terminating reaction of the first precursor A -- DEZ.
(b) Purging the residual precursor and the by-products, and then evacuating the reaction chamber.
(c) The introducing and self-terminating reaction of the second precursor B -- water (H2O).
(d) Purging the residual precursor and the by-products, and then evacuating the reaction chamber.
One reaction cycle consists of the steps 1 to 4, which would deposit amount of ZnO material to the surface.
13
Fig. 2-1 One cycle of ALD growth consists of the four steps: (a) introducing DEZ, (b) purging by-products, (c) introducing water (H2O), and (d) purging by-products.
The illustration of one ALD reaction cycle is sketched in Fig. 2-1. Each reaction cycle adds a given amount of ZnO onto the surface. To grow a ZnO layer, reaction cycles are repeated until the desired thickness of layer achieved. The ALD process starts when the substrate surface is in a stabilized state with the desired temperature and hydroxyl group on the substrate surface. The ALD growth mechanism bases on the self-terminating reactions, which means the dominating factor is the surface-state control while the other process parameters (such as the precursor flow rate and chamber pressure) have little or no influence.
14
2.2 Photoluminescence of ZnO
The wurtzite ZnO material has the wide direct band gap of about 3.37 eV and the high binding energy of the free-exciton (FX) about 60 meV. The properties of the emission transitions in ZnO attract variety of research attention. The main intrinsic excitonic emission transition is free exciton (FX), which is usually bound by the donors and/or acceptors to form donor-bound exciton (D0X) and/or acceptor-bound exciton (A0X), respectively. In certain conditions, the transitions of the two-electron satellites (TES) related to the D0X, LO-phonon replica of the main excitonic transitions, and the donor-acceptor-pair (DAP) transitions could be observed. Recently, the basal stacking fault (BSF) is found to form the type-II quantum well in the wurtzite structure and trap the excitons. Such excitons bound by the BSF is possibly influenced by intrinsic defects such donor and acceptor, which leads to a complex transition mechanism. [2] Through the Mg-doping technique, the energy band gap and band alignment could be engineered to fabricate the quantum wells (QWs) structure. By taking advantage of the designed micro-cavity to enhance the coupling of excitons and photons, the transition of excion-polaritons was observed. [3]
2.2.1 Free Exciton (FX)
The FX is a bound state of an electron and a hole, which attract to each other by the Coulomb interaction. It is taken as an electrically neutral quasi-particle. An FX could result from an excited electron from the valence band into the conduction band and leaves a positively charged hole. Such electron and hole pair is in the form of hydrogenic system, which provides a stabilizing energy state slightly less than the energy of
15
unbounded electron and hole. The binding energy is the difference of energy between the stabilizing energy of the hydrogenic system and the unbounded electron and hole.
The binding energy of the FX in ZnO is about 60 meV that prevents the FX from thermal ionization at room temperature while the thermal energy at room temperature is about 25 meV.
2.2.2 Bound Exciton Complex
The exciton as a quasi-particle is usually bound by the dopants or native defects, which results in the bound exciton complex and discrete electronic states in the band gap.
The exciton bound by the neutral or charged donors and acceptors forms the neutral donor bound exciton (D0X), charged donor bound exciton (D+X), neutral acceptor bound exciton (A0X), and charged acceptor bound exciton (A-X). In high quality bulk ZnO, the D0X often dominates because of the unintentional impurities and/or shallow donor-like defects, and the A0X is usually observed in ZnO containing acceptors due to the intentionally p-type doping technique. Therefore, the transition of D0X is usually observed to dominate the PL spectra at low temperature, at which the thermal energy is able to ionize into FX. At low temperature under 10 K, the transition energy of the D0X is in the range of 3.360~3.368 eV and the transition energy of the A0X is in the range of 3.348~3.374 eV.
Another characteristic of the neutral D0X is the two electron satellites (TES). For the ZnO material, the transition energy of TES is in the range of 3.32~3.34 eV. The transition of TES results from an exciton bound to a neutral donor in excited state, and is usually observed in the samples with extreme crystalline quality, and has the intensity weaker than the D0X.
Donor-DFFHSWRU SDLUV '$3 FDQ EH FRQVLGHUHG DV ³SRO\FHQWULF´ ERXQG H[FLWRQV
16
When an electron on a donor has the wave function overlap with a hole on an acceptor, the transition of DAP would occur. The energy of the transition of the DAP is given by
LO and acceptor centers increases and their average distance necessarily decreases, which lead to the DAP blue-shift of the DAP transition energy EDAP. blend layer leads to the QW-like region act as type-II QWs with 147 and 37 meV negative band offsets of the conduction band minimum and the valance band maximum to those of the barriers, respectively. The sketch of the band alignment is plot in Fig. 2-2. This means that the BSF structure would be a potential barrier at the valence band and a potential well in the conduction band to capture the electrons, which attract the holes in wurtzite structure via Coulomb interaction to form the BSF confined exciton [5].
17
Fig. 2-2 The schematic plot of the band alignment of ZnO at WZ/ZB/WZ regions.
2.3 XRD Measurement
The X-ray measurements were conducted using a four-circle diffractometer at beamline BL13A of National Synchrotron Radiation Research Center (NSRRC) Taiwan with incident wavelength 1.02473 Å. Two pairs of slits located between the sample and a NaI scintillation detector were employed and yielded a typical resolution of better than 1u10-3 Å-1. The X-ray Diffraction measurements are the powerful technique to observe and determine the structural properties of the deposited crystalline film in the theory of diffraction.
2.3.1 X-ray Diffraction Theory
For the crystalline material, the periodical structural information can be determined by analyzing the reflected X-ray from the material. In Fig. 2-3, the incident X-rays are
18
reflected from the adjacent atomic planes with spacing of d in the crystalline material, and the reflected rays result in the constructive interference while match the condition of WKH%UDJJ¶VODZ nO 2d sin T , where the O is the wavelength of X-ray and T is the angle between incident X-rays and reflecting planes.
Fig. 2-3 Bragg diffraction condition in (a) real space and (b) reciprocal space.
In the elastic scattering process, the incident and emergent X-rays with wave vectors
k and kc, both of which have the magnitude of 2S /O , and the scattering vector q can be obtained by the equation: q { kck 2ksinT . In reciprocal space, every periodical plane with spacing d in real space is transformed as a point with a lattice spacing of
/d
2S , while q 2S /d WKH GLIIUDFWLRQ FRQGLWLRQ PDWFK WKH %UDJJ¶V /DZ 7KHUHIRUH
each set of periodical parallel planes in real space can be represented by the Miller indices
)
(hk l , and be expressed by corresponding lattice vector: ghkl , where the |ghkl | value is equal to 2S /dhkl . The ZnO crystal structure is hexagonal with lattice parameters
243 .
a 3 Å and c 5.203 Å (D E 90q,J 120 q). The reciprocal lattice vector
ghkl of a hexagonal lattice can be expressed by
19
Conventionally, for crystals with hexagonal and rhombohedral symmetry, crystallographic planes are denoted using the four indices based on a four-axis Miller-Bravias coordinate system, consisting of three basal plane axes (a1, a2, a3) at 120 q
angles to each other and the fourth axis c perpendicular to the basal plane. The Miller-Bravias indices (hk il) satisfy the conditions i (h k). In this thesis, 4-digit Miller-Bravias indices are used for materials with hexagonal and rhombohedral symmetries including ZnO and sapphire to distinguish them from those with cubic symmetry.
For determining the crystalline structure properties, the four-circle diffractormeter is used which consists of four rotatable circles: T, 2T, F and I. The 2T circle is the detector axis controlling the magnitude of scattering vector q. The T, F, and I circles control the sample orientation. When the q vector coincides with the specific reciprocal lattice vector g, the Laue condition is satisfied. The I angle is equivalent to the azimuthal angle and the F angle is related to the polar angle of the crystal film. Different scan methods could perform the macro observation with different respect. The T/2T scan could observe the coherent length along the surface normal. The orientation of the deposited film could be observed through the T rocking curve scan, and the Iscan could determine the symmetry. The XRR method could confirm the roughness and thickness of the deposited layers.
2.3.2 Radial Scan
Radial scan is performed by driving the two rotatable circles: T and 2T , which is
20
shown in Fig. 2-4, to vary the q vector to scan the reciprocal space in the surface normal direction. The most commonly performed radial scan is the one along sample surface normal, which is often known as T 2T or Z 2T scan as shown Fig. 2-4. From the positions of diffraction peaks we can determine the corresponding interplanar spacing along the direction of q and the line width of the diffraction peak can yield the structural coherence length (grain size) and inhomogeneous strain along the same direction.
Fig. 2-4 The radial scan situation in the real space and reciprocal space.
2.3.3 Rocking Curve
As shown in Fig. 2-5 for a given incident x-ray direction, a detector is placed at the position of a diffraction spot with certain lattice vector, the scattered x-ray collected while the crystal is rotated by means of scanning the T DQJOHZKLFKLVDOVRFDOOHG³T rocking FXUYH´ 7KH ZLGWK RI D URFNLQJ FXUYH 'T is a direct measurement of the width of the diffraction spot in the reciprocal space. The 'T also presents the distribution of the sub-JUDLQV¶RULHQWDWLRQLQWKHILOP WKHZLGHGLVWULEXWLRQOHDGVWRWKHODUJHU 'T .
21
Fig. 2-5 T -rocking scan situations in the real space and reciprocal for ideal and non-ideal lattice structures.
2.3.4 Azimuthal Scan
Azimuthal scan means measuring the diffraction intensity as a function of azimuthal angle I by rotation the sample along an axis, which is usually parallel to surface normal or, in some cases, to a specific crystallographic axis. Figure 2-6 illustrates the scheme of the Azimuthal scan which can used to study the symmetry and crystal quality of the grown film and determine its relative orientation with substrate in epitaxy.
22
Fig. 2-6 Azimuthal angle scan situations in the real space and reciprocal for ideal and non-ideal lattice structures.
2.3.5 X-ray Reflectivity
X-ray reflectivity (XRR) is a surface-sensitive and non-destructive analytical technique to estimate the density, thickness and roughness of thin film structures by analyze the reflection of the X-rays from the surface and interfaces among layers. The setup of the XRR is sketched in Fig. 2-7(a). [6] Through scanning of the incident X-ray, the reflected X-rays result in interference that leads to the periodic interference stripes in Fig. 2-7(b). For the X-ray beam, the index of refraction n is defined as: reflection equation, the total external reflection of X-ray occurs at the angle of incidence smaller than the critical angle, Tc 2G , which depends on the electron density of the material. [17]
23
(a)
(b)
Fig. 2-7 X-ray reflectivity measurement situation of (a) the thin film structure and (b) the analysis of the obtained data.
For an incident angle T , which is half of the scattering angle 2T in the reflectivity measurements, the X-ray momentum transfers along the surface normal could be presents as q 4OS sinT . Hence the period of interference fringes of the reflected X-ray beams is related to the thickness d of the film via 2S /d . And the roughness of the interfaces, which results in the dampling of reflevitity intensity, can be taken into account. [17] Fig.
24
2-8 shows the simulation results of reflectivity with different situations: ideal surface and film/substrate interface, surface roughness of 1 nm and ideal film/substrate interface, and ideal film surface with interfacial roughness of 1 nm. The simulated results reveal that the surface roughness of ZnO film has negative influence on the decay rate of the reflectivity curve, and the substrate roughness manily affects the amplitude of interference fringes.
Fig. 2-8 Simulated XRR curves of the ZnO films on sapphire substrate with different Rrms values: ideal interface and surface, film roughness of 1 nm with ideal substrate roughness, and ideal surface with substrate roughness of 1 nm.
2.4 Transmission Electron Microscope
The TEM is a very powerful method to analyze the structure characteristics of the
25
crystalline material in the atomic level, which make implementation of the direct observation of crystal defect possible.
2.4.1 TEM setup
A schematic presentation of the microscope is shown in Fig. 2-9. [7] The TEM instrument consists of an electron gun connecting to a high voltage (typically about 100-300 kV) accelerating electronic filed to emit electrons. By using condensor lenses (magnetic lens), the electron beam is focused to a spot of the order of 1 mm on the specimen. The image is magnified more than 106 times in the bright filed image mode (Fig. 2-9 (a)). In selected area diffraction mode (Fig. 2-9 (b)), the electron diffraction patterns are formed on the final image screen. In bright field imaging, the image of a thin sample is formed by the electrons, which pass the film without diffraction, the diffracted electrons being stopped by a diaphragm. In the corresponding dark-field-imaging mode, a diffracted beam is used for imaging.
26
Fig. 2-9 The measurement setup of TEM in (a) bright field imaging and (b) selected area diffraction modes.
2.4.2 Dark Field Image
Selected area electron diffraction (SAED), is a crystallographic experimental technique, which selects certain area of the specimen to obtain the diffraction pattern. The crystalline specimen is subjected to a parallel beam of high-energy electrons. Because the wavelength of high-energy electrons is close to the spacing of atoms in a crystal, the periodic atoms act as a diffraction grating to the electrons. As a result, the image on the screen of the TEM will be a series of diffraction spots (diffraction pattern), and each diffraction spot is corresponding to a satisfied diffraction condition of the crystal structure.
27
As a diffraction technique, SAED can be used to identify crystal structures and examine crystal defects. As shown in Fig. 2-10 below, the specimen holder is an objective aperture, which can be inserted into the beam path to block the electron beam except for the selected diffraction spot. [8] The diffraction spot, which is selected by the objective aperture can form the image on the screen is called dark-field (DF) images. Such an imaging mode is called the DF mode. In addition, the image formed by the unblocked spots including the direct beam and specific selected diffraction spots is called the bright filed (BF) image and the imaging mode is called the BF mode. Through these modes the TEM image can be formed by only the selected spot with a specific diffraction vector g.
Fig. 2-10 The measurement setup of dark filed imaging with the selected spots with specific diffraction vector g.
2.4.3 Basal Stacking Fault Analysis by TEM
The most common crystal structure of ZnO epitaxial film is wurtzite structure, which KDVWKHDWRPVVWDFNLQJVHTXHQFH«AaBbAaBbAaBb«DORQJWKH [0001 ] direction (c-axis direction), where Aa and Bb represent the planes with specific stacking positions of atom
28
layer and the capital letters represent the zinc atoms and the lowercase letters represent the oxygen atoms, the positions are shown in Fig. 2-11. [9] The stacking fault means that WKHLGHDOVWDFNLQJVHTXHQFHRIZXUW]LWHVWUXFWXUH «AaBbAaBb« KDVWKHIDXOWVVXFKDV
WKH VHTXHQFH «AaBbAaBbCcBbCc« ZKLFK LV WHUPHG DV WKH type-I stacking fault. In DGGLWLRQWKHVWDFNLQJVHTXHQFHRIWKH]LQFEOHQGVWUXFWXUHRI=Q2LV«AaBbCcAaBbCc«
along the [111 ] direction.
Fig. 2-11 Three kinds of planes with different atom stacking arrangements in ZnO material marked as A-, B-, and C- planes.
The BSF can be classified into four types: type-I, type-II, type-III, and Extrinsic SF.
The scheme of the four types of stacking sequences are shown in Fig. 2-12. [10]
(a) Type-I stacking fault (I1):
The type-I BSF is shown in Fig. 2-12(a). It is commonly expected to have the lowest formation energy. For the type-I SF, two stacking sequences are FRQVLGHUHG «AaBbCcBbCc« DQG «AaBbAaCcAaCc« ERWK W\SHV RI 6) VWDFNLQJ
sequences have the same energy.
(b) Type-II stacking fault (I2):
29
7KH%6)KDVWKHVWDFNLQJVHTXHQFHRI«AaBbAaBbCcAaCcAaCc«ZKLFKKDVWKH
second lowest formation energy.
(c) Type-III stacking fault:
These are intrinsic BSFs, in which one of the Aa or Bb layers occupied by the Cc SRVLWLRQ «$D%E$D&F$D%E«
(d) Extrinsic stacking fault:
These SFs have the additional Cc layer inserted in the midst of the normal stacking VHTXHQFH «$D%E&F$D%E«
Fig. 2-12 Four types of stacking faults in ZnO: (a) type I, (b) type II, (c) type III, and (d) extrinsic. The arrows indicate the position of the stacking faults and the black and white circles denote zinc and oxygen atoms, respectively.
Depending on the type of error in stacking sequence or equivalently the displacement vector R defines the relative displacement between the unfaulted lattices above and below the fault. In wurtzite crystal structure the displacement vectors associated with the I1, I2
and Extrinsic type of BSFs are 2203 6
1 , 1100
3
1 , and 0001
2
1 , respectively; [11-13]
but the type-III BSF has no displacement vector. From the TEM image of a crystal, the intensity of the electron beam diffraction could be described as following:
30
In this formula, the part of the phase factor is as the following:
R g
e
2Si& & .According to this formula, a stacking fault will be out of contrast if the dot product of its displacement vector R with the diffraction vector g used for imaging equals to 2Sn, where n = 0, r1, r«[14, 15] Consequently, all three types of stacking faults are visible in image with g = (10 11) and out of contrast as g = (0002). In the case of g = (10 10) , only intrinsic stacking faults of types I1 and I2 are in contrast. Table 1 shows the visibility of the stacking faults with different g vectors.
(0002)
Type-I Invisible Visible Visible
Type-II Invisible Visible Visible
Extrinsic Invisible Invisible Visible
Table. 1 The visibility of the BSFs in TEM dark-field imaging with different selected g vectors.
2.5 Atomic force microscopy
Atomic force microscopy (AFM) is a very high-resolution type of scanning probe microscopy with demonstrated resolution on the order of fractions of a nanometer. The
31
brief setup of the AFM is shown in Fig. 2-13. The AFM consists of a cantilever with a sharp tip (the radius of curvature is on the order of nanometers), which is used to scan the specimen surface. When the tip is brought into the vicinity of the specimen surface, the forces between the tip and the surface deflects the cantilever according to Hooke's law.
Depending on the scanning modes, the forces that are measured in AFM include mechanical contact force, van der Waals force, electrostatic force, magnetic force, etc.
Depending on the scanning modes, the forces that are measured in AFM include mechanical contact force, van der Waals force, electrostatic force, magnetic force, etc.