Where Vno(r – Rj) is the free atomic potential of the Rjth atom. The wave functions for the crystal orbitals may be expressed in term of a Bloch sum, which is given by
()()()1/21jikrRikrknreerNVφφ−−⋅⎛⎞=−⎜⎟⎝⎠
()1/21ikrkeurNV⋅⎛⎞=⎜⎟⎝⎠ (4.88)
Where uk(r) is the Bloch function. In Eq. (4.88), the atomic wave functions are being normalized (i.e., N represents the total number of atoms in the crystal). The factor (1/NV)1/2 is the normalization constant for the Bloch sum if overlapping of the atomic orbitals centered at different atomic sites is negligible. Thus, Eq. (4.88) is a good approximation for the crystal orbitals, provided that the energy levels of the atomic orbits are nondegenerate and overlapping between the orbital wave functions of the neighboring atoms is negligible. This condition can be expressed by
()()*njnirRrRdr φφ−−=∫ (4.89)
Note that in Eq.(4.89), δij = 0 if i ≠ j. Now, substituting Eq. (4.88) into Eq. (4.87), multiplying Eq. (4.87) by the conjugate wave functions, φn*(r - Ri), and integrating the expression over the entire space, one obtains
()()*3kkkErHrφφ=∫
()()()()22*312jiikRRninojnjijerRVrRrRdrNVmφφ⋅−⎧⎡⎤∇⎪⎛⎞=−−+−⎨⎢⎥⎜⎟⎝⎠⎪⎣⎦⎩Σ∫h
()()()()*jiikRRnjjnjijerRVrRrRdrφφ⋅−⎫⎪′+−−−⎬⎪⎭Σ∫ (4.90)
Using Eq. (4.89), Eq. (4.90) can be rewritten as follows:
()ijijikRknonnijREERe⋅=−−Σαβ (4.91)
Where Rij = Rj – Ri, and
22*12nonnonENVmφ∇=−+⎡⎤⎛⎞⎢⎥⎜⎟⎝⎠⎣⎦∫h (4.92)
()()2nnijrRVrRdrαφ′=−−−∫ (4.93)
()()()*nnijnjrRVrRrRdrβφφ′=−−−−∫ (4.94)
()()()jnojjVrRVrRVrR′−=−+− (4.95)
As shown in Figure 4.8, Vno(r – Rj) is the unperturbed atomic potential centered at Rj, and V′(r – Rj) is the perturbed crystal potential due to atoms other than the Rjth atom.
In general, the atomic orbital wave functions φn(r) falls off exponentially with the distance r, and hence overlapping of each atomic orbital wave function φn(r) is assumed to be negligibly small. Therefore, it is expected that the contribution to βn will come from a rather restricted range of r. Furthermore, it is also expected that βn will decrease rapidly with increasing distance between the neighboring atoms. Figure 4.8 illustrates the potential V′(r – Rj), which plays the role of the perturbing potential, is practically zero in the vicinity of Rj. The LCAO method may be applied to construct the energy band structures of the s-like states for a simple cubic lattice and a body-centered cubic lattice. This is discussed next.
4.6.1. The S-like States for a Simple Cubic Lattice
The LCAO method is first applied to the calculations of the energy band structure of the s-like states for a simple cubic lattice. In a simple cubic lattice, there are six nearest-neighbor atoms located at an equal distance, a, from any chosen atomic site. Therefore, the value of βn(a), given by Eq. (4.94), is the same for all six nearest-neighbor atoms. Since the perturbing potential V′(r) is negative, and the atomic wave functions are of the same sign in the region of overlapping, values for both αn and βn(a) are positive. Thus, the energy dispersion relation (E vs. k) for s-like states of a simple cubic lattice can be derived by substituting Rij = (a, 0, 0), (0, a, 0), (0, 0, a), (–a, 0, 0), (0, –a, 0), (0, 0, –a) into Eq. (4.91), and the result yields
()yyxxzzikaikaikaikaikaikakonnEEeeeeee−−−=−+++++−αβ
()()2coscoscosnoxyznnEakaka=+−−αβ (4.96)
Equation (4.96) shows the E-k relation for the s-like states of a simple cubic lattice. Figure 4.9a and b show the energy band diagrams plotted in the kx - direction and the kx–ky plane, respectively, as calculated from Eq. (4.96). The width of the energy band for this case is equal to 12 βn(a). It is of interest to note that the shape of the E-k plot is independent of the value of αn or βn used, but depends only on the geometry of the crystal lattice. Two limiting cases deserve special mention, namely, (i) near the top of the band and (ii) near the bottom of the band. First, in the case of
near the bottom of the band, the value of k is very small, and the cosine terms in Eq. (4.96) may be expanded for small ka [i.e., cos ka (1 – k≈2a2/2)]. If only the first-order term is retained, then the energy E is found to vary with k2 near the bottom of the band. This result is identical to the free-electron case. Under this condition, the E–k relation for the s-like states in a simple cubic lattice is reduced to
()()226knonnnEEaaka=−−+αββ (4.97)
From Eq. (4.97), the electron effective mass m* for small ka can be expressed as
()2*212222knaaEmk−=⎛⎞∂=⎜⎟⎜⎟∂⎝⎠βhh (4.98)
Which shows that the constant energy surface near the bottom of the band is parabolic (i.e.,), and the effective mass of electrons is a scalar quantity. Similarly, the E-k relation near the top of the band (i.e., k ≈ π/a) can be obtained by expanding cos(ka) in Eq. (4.96) at k22*/2kEk=h x = ky = kz = π/a. This is carried out by substituting kx = π/a-kx’, ky =π/a- ky’,and kz =π/a-kz’ into Eq.(4.96), (where kx’, ky’, kz’, are small wave vectors), which yields
22'*'2kkECm=+h (4.99)
Where C is a constant, and m* is given by
()2*22nmaa=−βh (4.100)
Equation (4.100) shows that the electron effective mass m* is negative near the top of the band. It is noted that the effective masses given by Eqs. (4.98) and (4.100) represent the curvatures of the bottom and top of the s-like energy band, respectively. The effective mass is an important physical parameter in that it measures the curvature of the (E–k) energy band diagram. It is noted that a positive m* means that the band is bending upward, and a negative m* implies that the band is bending downward. Moreover, an energy band with a large curvature corresponds to a small effective mass, and an energy band with a small curvature represents a large effective mass. The effective mass concept is important since the mobility of electrons in a band is inversely proportional to the effective mass of electrons. For example, by examining the curvature of the energy band diagram near the bottom of the conduction band one can obtain qualitative information concerning the effective mass and the mobility of electrons in the
conduction band. A detailed discussion of the effective masses for electrons (or holes) in the bottom (or top) of an energy band will be given in Section 4.8.
4.6.2. The S-like States for a Body-Centered Cubic Lattice
For a body-centered cubic (BCC) lattice, there are eight nearest-neighbor atoms for each chosen atomic site that are located at Rij = (±a/2, ±a/2, ±a/2). If one substitutes these values in Eq. (4.101), the E–k relation for the s-like states of the BCC crystal can be expressed as
()/2/2/28cos()cos()cos()knonnxyzEEkakaka=−−αβ (4.101)
In Eq. (4.101), values of k must be confined to the first Brillouin zone in order to have nondegenerate energy states. Using Eq. (4.101), the 2-D constant-energy contour plotted in the first quadrant of the kx–ky plane for the s-like states of a body-centered cubic lattice is shown in Figure 4.10. Although the constant energy surfaces are spherical near the zone center and zone boundaries, the constant-energy contours depart considerably from the spherical shape for other values of k. For small values of k near the zone center and for large values of k near the zone boundaries, the electron energy E is proportional to k2, and the effective mass of electrons can be derived from Eq. (4.101), which yields
()2*28nmaa=βh (4.102)
From Eq. (4.101), it can be shown that the total width of the allowed energy band for the s-like states in a body center cubic crystal lattice is equal to 16βn(a).
It is clear from the above examples that the tight-binding approximation is indeed applicable for calculating the energy states of the core electrons, such as the s-like states in the cubic crystals.
4.7. ENERGY BAND STRUCTURES FOR SOME SEMICONDUCTORS
Calculations of energy band structures for the elemental (Si, Ge) and III-V compound semiconductors (e.g., GaAs, InP, etc.) have been widely reported in the literature. As a result a great deal of information is available for the band structures of semiconductors from both the theoretical and experimental sources. In most cases theoretical calculations of the energy band structures for these semiconductor materials are guided by the experimental data from the optical absorption, photoluminescence, and photoemission experiments in which the fundamental absorption process is closely related to the density of states and the transitions from the initial to the final states of
the energy bands. The energy band structures for some elemental and III-V compound semiconductors calculated from the pseudopotential method are discussed in this section. In general, the exact calculations of the energy band structures for semiconductors are much more complex than those of the NFE approximation and the LCAO method described in this chapter. In fact, both of these approximations can only provide a qualitative description of the energy bands in a crystalline solid. For semiconductors, the two most commonly used methods for calculating the energy band structures are the pseudopotential and the orthogonalized plane wave methods. They are discussed briefly as follows.
The main difficulty of band calculations in a real crystal is that the only wave functions, which satisfy the boundary conditions imposed by the Bloch theorem in a simple manner are plane waves, but plane wave expressions do not converge readily in the interior of an atomic cell. The pseudopotential method is based on the concept of introducing the pseudopotential for a crystal that will lead to the same energy levels as the real crystal potential but do not have the same wave functions. The pseudopotential technique can greatly improve the convergence of the series of the plane waves that represent the pseudowave functions of electrons in a crystal.
ซึ่ง Vno (r-Rj) อยู่เป็นอะตอมอิสระของอะตอม Rjth ฟังก์ชันคลื่นสำหรับ orbitals คริสตัลอาจแสดงในเทอมของผลเม็ดเลือดขาว ซึ่งถูกกำหนดโดย()()() 1/21jikrRikrknreerNVφφ−−⋅⎛⎞ =−⎜⎟⎝⎠() 1/21ikrkeurNV⋅⎛⎞ =⎜⎟⎝⎠ (4.88)ที่ uk(r) เป็นการทำงานของเม็ดเลือดขาว มีอยู่ตามปกติใน Eq. (4.88), ฟังก์ชันคลื่นอะตอม (เช่น N แทนจำนวนของอะตอมในผลึก) ตัว (1/NV) 1/2 คือ ค่าคงฟื้นฟูสำหรับผลเม็ดเลือดขาวถ้าซ้อนทับกันของ orbitals อะตอมที่แปลกที่เว็บไซต์อะตอมแตกต่างกันเป็นระยะ ดังนั้น Eq. (4.88) เป็นการประมาณที่ดีสำหรับ orbitals คริสตัล ที่มีระดับพลังงานของวงโคจรอะตอม nondegenerate และทับซ้อนกันระหว่างฟังก์ชันคลื่นของวงโคจรของอะตอมใกล้เคียงเป็นระยะ เงื่อนไขนี้สามารถแสดงได้โดย()() * njnirRrRdr φφ−− =∫ (4.89)หมายเหตุว่า ใน Eq.(4.89), δij = 0 ถ้าฉัน≠ j ตอนนี้ แทนที่ Eq. (4.88) เป็น Eq. (4.87), คูณ (4.87) Eq. conjugate คลื่นฟังก์ชัน φn * (r - Ri), และรวมนิพจน์มากกว่าพื้นที่ทั้งหมด หนึ่งเหตุผล()() * 3kkkErHrφφ =∫()()()() 22 * 312jiikRRninojnjijerRVrRrRdrNVmφφ⋅−⎧⎡⎤∇⎪⎛⎞ =−− + −⎨⎢⎥⎜⎟⎝⎠⎪⎣⎦⎩Σ∫h()()()() * jiikRRnjjnjijerRVrRrRdrφφ⋅−⎫⎪′ + −−−⎬⎪⎭Σ∫ (4.90)ใช้ Eq. (4.89), Eq. (4.90) สามารถมีจิตเป็นดังนี้:ijijikRknonnijREERe⋅ () =−−Σαβ (4.91)ที่ Rij = Rj – Ri และ22 * 12nonnonENVmφ∇ =− + ⎡⎤⎛⎞⎢⎥⎜⎟⎝⎠⎣⎦∫h (4.92)2nnijrRVrRdrαφ′ ()() =−−−∫ (4.93)()()() * nnijnjrRVrRrRdrβφφ′ =−−−−∫ (4.94)jnojjVrRVrRVrR′− ()()() =− + − (4.95)ดังแสดงในรูปที่ 4.8, Vno (r-Rj) เป็นอะตอมหมายทันที Rj และ V′ (r-Rj) คริสตัล perturbed อาจเกิดขึ้นเนื่องจากอะตอมไม่ใช่อะตอม Rjthทั่วไป φn(r) ฟังก์ชันคลื่นโคจรอะตอมอยู่ปิดสร้างกับ r ระยะทาง และดังนั้น ซ้อนทับกันของแต่ละอะตอมของวงโคจรคลื่นฟังก์ชัน φn(r) จะถือขนาดเล็ก negligibly ดังนั้น คาดว่า สัดส่วนการ βn จะมาจาก r ค่อนข้างจำกัด นอกจากนี้ คาดว่า βn จะลดอย่างรวดเร็ว ด้วยการเพิ่มระยะห่างระหว่างอะตอมใกล้เคียง รูป 4.8 แสดง V′ เป็น (r-Rj), ซึ่งบทบาทศักยภาพ perturbing เป็นศูนย์จริงตั้ง Rj อาจใช้วิธี LCAO การสร้างโครงสร้างแถบพลังงานของอเมริกาเหมือน s โครงตาข่ายประกอบลูกบาศก์ง่าย ๆ และเนื้อหาแปลกลูกบาศก์โครงตาข่ายประกอบ นี้จะกล่าวถึงต่อไป4.6.1.อเมริกาเหมือน S สำหรับโครงตาข่ายประกอบลูกบาศก์อย่างก่อนมีใช้วิธี LCAO การคำนวณโครงสร้างแถบพลังงานของอเมริกาเหมือน s สำหรับโครงตาข่ายประกอบลูกบาศก์อย่าง ในเรื่องลูกบาศก์โครงตาข่ายประกอบ มีหกใกล้บ้านอะตอมมีระยะห่างเท่า ๆ a จากเว็บไซต์อะตอมใด ๆ ท่าน ดังนั้น βn(a) โดย Eq. (4.94), ค่าเป็นเหมือนกันทั้งหมดหกใกล้บ้านอะตอม เนื่องจาก V′(r) เป็น perturbing เป็นค่าลบ และฟังก์ชันคลื่นอะตอมเป็นเครื่องหมายเดียวกันในภูมิภาคของการซ้อนทับกัน ค่าสำหรับ αn และ βn(a) มีค่าเป็นบวก ดัง พลังงานกระจายตัวความสัมพันธ์ (E เทียบกับ k) สำหรับ s เหมือนอเมริกาของโครงตาข่ายประกอบลูกบาศก์อย่างสามารถได้มา โดยการแทนที่ Rij = (a, 0, 0), (0, a, 0), (0, 0 เป็น), (-a, 0, 0), (0, -a, 0), (0, 0, –) Eq. (4.91), และผลผลิตผลลัพธ์yyxxzzikaikaikaikaikaikakonnEEeeeeee−−− () =− +++ −αβ2coscoscosnoxyznnEakaka ()() = + −−αβ (4.96)สมการ (4.96) แสดงความสัมพันธ์ E-k สำหรับอเมริกาเหมือน s ของโครงตาข่ายประกอบลูกบาศก์อย่าง รูป 4.9a และ b แสดงไดอะแกรมวงพลังงานที่พล็อตใน kx - ทิศทางและระนาบ kx-ky ตามลำดับ เป็นคำนวณจาก Eq. (4.96) ความกว้างของแถบพลังงานสำหรับกรณีนี้จะเท่ากับ 12 βn(a) มันจะสนใจสังเกตว่า รูปร่างของแปลง E-k ขึ้นอยู่กับค่าของ αn หรือ βn ที่ใช้ แต่ขึ้นอยู่กับรูปทรงเรขาคณิตของโครงตาข่ายประกอบคริสตัลเท่านั้น กรณีสองจำกัดสมควรพูดถึงพิเศษ ได้แก่, (i) ใกล้ด้านบนของวง และ (ii) ที่ด้านล่างของวง ครั้งแรก ในกรณีของnear the bottom of the band, the value of k is very small, and the cosine terms in Eq. (4.96) may be expanded for small ka [i.e., cos ka (1 – k≈2a2/2)]. If only the first-order term is retained, then the energy E is found to vary with k2 near the bottom of the band. This result is identical to the free-electron case. Under this condition, the E–k relation for the s-like states in a simple cubic lattice is reduced to()()226knonnnEEaaka=−−+αββ (4.97)From Eq. (4.97), the electron effective mass m* for small ka can be expressed as()2*212222knaaEmk−=⎛⎞∂=⎜⎟⎜⎟∂⎝⎠βhh (4.98)Which shows that the constant energy surface near the bottom of the band is parabolic (i.e.,), and the effective mass of electrons is a scalar quantity. Similarly, the E-k relation near the top of the band (i.e., k ≈ π/a) can be obtained by expanding cos(ka) in Eq. (4.96) at k22*/2kEk=h x = ky = kz = π/a. This is carried out by substituting kx = π/a-kx’, ky =π/a- ky’,and kz =π/a-kz’ into Eq.(4.96), (where kx’, ky’, kz’, are small wave vectors), which yields22'*'2kkECm=+h (4.99)Where C is a constant, and m* is given by()2*22nmaa=−βh (4.100)Equation (4.100) shows that the electron effective mass m* is negative near the top of the band. It is noted that the effective masses given by Eqs. (4.98) and (4.100) represent the curvatures of the bottom and top of the s-like energy band, respectively. The effective mass is an important physical parameter in that it measures the curvature of the (E–k) energy band diagram. It is noted that a positive m* means that the band is bending upward, and a negative m* implies that the band is bending downward. Moreover, an energy band with a large curvature corresponds to a small effective mass, and an energy band with a small curvature represents a large effective mass. The effective mass concept is important since the mobility of electrons in a band is inversely proportional to the effective mass of electrons. For example, by examining the curvature of the energy band diagram near the bottom of the conduction band one can obtain qualitative information concerning the effective mass and the mobility of electrons in theconduction band. A detailed discussion of the effective masses for electrons (or holes) in the bottom (or top) of an energy band will be given in Section 4.8.4.6.2. The S-like States for a Body-Centered Cubic LatticeFor a body-centered cubic (BCC) lattice, there are eight nearest-neighbor atoms for each chosen atomic site that are located at Rij = (±a/2, ±a/2, ±a/2). If one substitutes these values in Eq. (4.101), the E–k relation for the s-like states of the BCC crystal can be expressed as()/2/2/28cos()cos()cos()knonnxyzEEkakaka=−−αβ (4.101)In Eq. (4.101), values of k must be confined to the first Brillouin zone in order to have nondegenerate energy states. Using Eq. (4.101), the 2-D constant-energy contour plotted in the first quadrant of the kx–ky plane for the s-like states of a body-centered cubic lattice is shown in Figure 4.10. Although the constant energy surfaces are spherical near the zone center and zone boundaries, the constant-energy contours depart considerably from the spherical shape for other values of k. For small values of k near the zone center and for large values of k near the zone boundaries, the electron energy E is proportional to k2, and the effective mass of electrons can be derived from Eq. (4.101), which yields()2*28nmaa=βh (4.102)From Eq. (4.101), it can be shown that the total width of the allowed energy band for the s-like states in a body center cubic crystal lattice is equal to 16βn(a).It is clear from the above examples that the tight-binding approximation is indeed applicable for calculating the energy states of the core electrons, such as the s-like states in the cubic crystals.4.7. ENERGY BAND STRUCTURES FOR SOME SEMICONDUCTORSCalculations of energy band structures for the elemental (Si, Ge) and III-V compound semiconductors (e.g., GaAs, InP, etc.) have been widely reported in the literature. As a result a great deal of information is available for the band structures of semiconductors from both the theoretical and experimental sources. In most cases theoretical calculations of the energy band structures for these semiconductor materials are guided by the experimental data from the optical absorption, photoluminescence, and photoemission experiments in which the fundamental absorption process is closely related to the density of states and the transitions from the initial to the final states ofthe energy bands. The energy band structures for some elemental and III-V compound semiconductors calculated from the pseudopotential method are discussed in this section. In general, the exact calculations of the energy band structures for semiconductors are much more complex than those of the NFE approximation and the LCAO method described in this chapter. In fact, both of these approximations can only provide a qualitative description of the energy bands in a crystalline solid. For semiconductors, the two most commonly used methods for calculating the energy band structures are the pseudopotential and the orthogonalized plane wave methods. They are discussed briefly as follows.
The main difficulty of band calculations in a real crystal is that the only wave functions, which satisfy the boundary conditions imposed by the Bloch theorem in a simple manner are plane waves, but plane wave expressions do not converge readily in the interior of an atomic cell. The pseudopotential method is based on the concept of introducing the pseudopotential for a crystal that will lead to the same energy levels as the real crystal potential but do not have the same wave functions. The pseudopotential technique can greatly improve the convergence of the series of the plane waves that represent the pseudowave functions of electrons in a crystal.
การแปล กรุณารอสักครู่..