1. Semiconductor Materials Revision 1.1 Introduction Our understanding of materials, in particular conductors and semiconductors, relies on the Band Theory of Solids, which means that solids contain energy bands, each being a continuum of energy levels, which may be occupied by an electron orbiting the nucleus of an atom. Of primary interest are the outermost conduction band and the valence band located a small energy gap above this. 1.2 Fermi Dirac Distribution and Fermi Energy Level The behaviour of electron and hole concentrations is complex but is described in simplified terms by the Fermi-Dirac Probability Function. This function indicates the probability that any particular energy state is occupied by an electron. The probability function applies to allowable energy levels in the case of energy bands. The Fermi Dirac function is given as: f(E) 1 1 e(E-EF )/kT where E is the energy of a particular state, T is temperature in degrees Kelvin and k = 1.38 x 10-23 J/K is Boltzmann’s Constant. The value of energy EF is known as the Fermi Energy, or Fermi Level, and by definition is given for a conducting metal as the highest energy level which is occupied by an electron at absolute zero temperature of 0K. The Fermi-Dirac function is shown plotted in Fig. 1.1. At absolute zero temperature of 0K, no electrons have any energy due to thermal agitation and consequently occupy all of the lowest energy levels available as shown for T=0K. All of the energy levels below the Fermi level are occupied and so the probability function has a value of 1 here. By the same token, all of the available energy levels above the Fermi level are unoccupied or empty and consequently the probability function has a value of 0 here. Therefore the function has the rectangular shape shown in Fig 1.1 for T=0K. When the temperature is raised above 0K, some electrons will gain thermal energy and will be able to occupy higher energy states greater than EF. These electrons will leave vacant energy levels behind below the Fermi level. Consequently, the curve spreads symmetrically on either side of the Fermi level as shown in Fig. 1.1 for T>0K. 1 Prob 1 T = 0K 1- f(E) 1/2 T > 0K f(E) 0 EF Fig. 1.1 E The Fermi-Dirac Probability Function This function can also be applied to semiconductor materials in order to calculate the concentration of charge carriers present in the energy bands. In the case of intrinsic Silicon the free electrons and holes are generated in pairs when electrons jump from states in the valence band to states in the conduction band. For each electron in the conduction band above the Fermi energy, there is a vacated state in the valence band an equal amount below the Fermi energy, which has the associated positive charge of the hole. Hence, the curve of the distribution function is symmetrical on either side of the Fermi level with f(EF) = ½. The concentration of free mobile electrons in the conduction band can be established from the Fermi Dirac function as: f(E) = 1 for (E-EF )/kT 1+ e EC ≤ E ≤ ∞ which is shown in blue in Fig. 1.1. In addition, the concentration of holes in the valence band can equally be established as: 1 - f(E) = 1 - 1 (E-EF )/kT 1+ e for 0 ≤ E ≤ EV which is shown in red in Fig. 1.1 and is the difference between unity and the probability function for electron occupancy. Hence for intrinsic Silicon, the Fermi level is located mid-way between the conduction and valence bands as can be seen in Fig. 1.2, which shows only the hole and electron free mobile charge carriers in the energy bands. 2 1.3 Intrinsic Silicon Intrinsic Silicon material is un-doped Silicon with a homogeneous crystalline structure. At absolute zero temperature 0K, all of the covalent bonds formed in the valence band are intact and there are no electrons present in the conduction band. At room temperature of 300K (27OC), atoms undergo thermal agitation and a significant number of electrons in the valence band gain sufficient energy to jump the energy gap to the conduction band. These electrons can move around easily in the conduction band as there are many vacant energy levels for them to transfer between, i.e. they become free mobile charge carriers. Electrons which have jumped to the conduction band leave behind empty energy states in the valence band. These empty energy states are associated with atoms which now have one more proton in the nucleus than the number of electrons closely bound to it in the atomic structure. This means that there is a net positive charge associated with the empty state known as a hole in the valence band for each atom in which a covalent bond has been broken and an electron freed. Electrons in the conduction band can move with ease between the plentiful vacant states available here. This gives rise to an electron current. In the valence band, on the other hand, it is possible for an electron to move from a neighbouring atom to occupy the hole or vacant state in an atom. This is in effect the transfer of positive charge in the opposite direction. This means the current flowing in the valence band is composed essentially of positive charge carriers and is therefore a hole current. The total current then consists of both positive and negative charges and the technology is referred to as ‘bipolar’. In intrinsic or un-doped Silicon at room temperature, holes and free electrons are created in pairs. Some of the free electrons in the conduction band will collide with each other and with other atoms in the lattice and consequently lose some of their energy and return to the valence band. When this occurs they recombine with a vacant state here and neutralise the charge of a proton in reforming a covalent bond. There is always the same number of holes in the valence band as there is of free electrons in the conduction band. That is, they have the same concentrations in Intrinsic Silicon. Thermal equilibrium is reached when the rate of generation of free mobile electrons is equal to the rate of recombination. If the concentration of electrons is denoted n0 and that of holes as p0, then: 10 n0 =p0 =ni =1.5 x 10 cm-3 and the Law of Mass Action gives: 20 n0 =p0 =n2 i =2.25 x 10 3 E Intrinsic Silicon conduction band e- e- e- eE c conduction band edge energy gap E F Fermi Energy level Eg + h + h - e - e + h E v valence band edge h+ valence band 0 1/2 1 f(E) E n-type Silicon conduction band Ec Ed Donor level Fermi level EF EV e- h+ 0 valence band 1 1/2 f(E) E conduction band p-type Silicon EC EF Ea EV Fermi level Acceptor level - e h+ 0 Fig. 1.2 1/2 valence band 1 f(E) Mobile Carrier Concentrations in Intrinsic and Doped Silicon 4 1.4 n-Type Doped Silicon Intrinsic Silicon can be made n-type by doping the Silicon with a Group V material such as Phosphorous which has an atomic number of 15 and has 5 electrons in its outer sub-shell, one more than Silicon. The concentration of dopant atoms is usually between 1013 and 1018 cm-3, which is much lower than the density of the Silicon atoms themselves of 5 × 1022 cm-3. In n-type material the dopant atoms readily enter into covalent bonding with the neighbouring Silicon atoms but have an additional electron available that cannot form a covalent bond. The extra free electron originally occupies the donor energy level, Ed, in the dopant atom and this energy level is located just below the conduction band, EC, of the Silicon. At room temperature, practically all of the surplus electrons of the impurity atoms gain sufficient energy to leave the donor level and enter the conduction band of the Silicon lattice. The Law of Mass Action still applies to the n-type doped material and n0p0=ni2. However, the number of electrons in the conduction band in n-type Silicon due to doping far exceeds those generated by thermal agitation so that the electron concentration in the conduction band can be taken as equal to the doping concentration of n-type impurities present, i.e. n0 ≈ Nd . Moreover, the increased concentration of electrons in the conduction band also increases the rate of carrier recombination with the thermally generated holes in the valence band of the doped material. Hence, the concentration of holes in the valence band of ntype material falls and is much lower than is the case for un-doped material. The carrier concentrations in n-type material are given as: n0 p0 = n2i n0 ≈ Nd >> ni p0 = n2i << ni Nd majority carriers minority carriers When the Fermi-Dirac distribution is applied to doped, n-type material the electron concentration is higher in the conduction band so the distribution function, f(E), can be seen to shift upwards as can be seen in Fig. 1.2, where the electron concentration is shown in blue. Consequently, the function 1–f(E) becomes reduced in the valence band, as can be seen in red in Fig. 1.2. The result of this is that the Fermi level in n-type Silicon shifts upwards from its position in intrinsic Silicon and appears close to the donor energy level. 5 1.5 p-type Doped Silicon Intrinsic Silicon material can be made p-type by doping the Silicon with a Group III material such as Aluminium having an atomic number of 13 and has only 3 electrons in its outer sub-shell, one less than Silicon. The concentration of dopant atoms lies between 1013 and 1018 cm-3, much lower than the density of the Silicon atoms of 5 × 1022 cm-1. In p-type material the dopant atoms readily enter into covalent bonding with the neighbouring Silicon atoms but find themselves with one electron less than the number needed to form a complete set of covalent bonds in its outer sub-shell. The dopant atom has one vacant energy level in its outer shell, referred to as an acceptor level, Ea, and this energy level is located just above the valance band, EV, of the Silicon. At room temperature, electrons in the valence band of the Silicon easily gain sufficient energy to leave the valence band and occupy practically all of the dopant acceptor levels. Hence many covalent bonds in the Silicon lattice are broken and electrons move freely between vacant states in the valence band. Hence, positively charged holes are generated in the valence band and become mobile. The Law of Mass Action also applies to the p-type doped material and n0p0=ni2. The number of holes in the valence band in p-type Silicon due to doping far exceeds those generated thermally so that the hole concentration in the valence band can be taken as the concentration of p-type impurities present, i.e. p0 ≈ Na . The increased concentration of holes in the valence band also increases the rate of carrier recombination with the thermally generated electrons in the conduction band. Hence, the concentration of electrons in the conduction band of p-type material is much lower than is the case for un-doped material. The carrier concentrations for p-type material are given as: n0 p0 = n2i p0 ≈ Na >> ni n0 = n2i << ni Na majority carriers minority carriers When the Fermi-Dirac distribution is applied to p-type material the electron concentration is lower in the conduction band so the distribution function, f(E), shifts downwards as seen in blue in Fig. 1.2. Consequently, the function 1–f(E) reflects the increased concentration of holes in the valence band, as seen in red in Fig. 1.2. The result of this is that the Fermi level in p-type Silicon shifts downwards from its position in intrinsic Silicon and appears close to the acceptor energy level. 6 1.6 Drift Current When an electric field is applied across a semiconductor, by way of a battery or power supply for example, an electrostatic force is exerted on the charge carriers in the material. The carriers, as a result of the force they experience in the field, gain electrostatic energy. This gives rise to a gradient in the energy bands occupied by the carriers in the direction of the electric field as shown in Fig. 1.3 as well as a gradient of potential across the material. conduction band V eE qV C h+ semiconductor + E EV + - direction of electric field Fig. 1.3 valence band E direction of electric field - Effect of an Electric Field on Energy Bands Since electrons seek to occupy the lowest energy status available, they will tend to move from locations of higher electrostatic energy to locations of lower energy. This gives rise to a net movement of carriers through the material and hence sets up a flow of current. Electrons move in a direction opposite to the electric field while holes move in the direction of the field. When charge carriers drift in a uniform manner through a piece of homogeneous semiconductor material, there is an average sustained uniform flux of charge through the material. The average charge flow considered per unit area normal to the direction of flow is referred to as the Charge Flux Density as seen in Fig. 1.4 below. electric E field Jn drift nqμnE A charge flux density Jp drift pqμpE for electrons for holes J Fig. 1.4 Charge Flux Density in an Electric Field 7 Electrons and holes drift in opposite directions in any given electric field. However, since the charge on the electron is negative, both electrons and holes make positive contributions to conventional current specified as being positive in the direction of the electric field. When both flux components are integrated over the area, A, of the semiconductor, the Total Drift Current is obtained as: Idrift Jn drift A Jp drift A AEq(nμn pμp ) where: n and p are the free charge carrier concentrations, q is the magnitude of the charge on the carriers, μn , μp are the respective carrier mobilities and E is the electric field strength. 1.7 Diffusion Current If the concentration of free charge carriers at a particular location in a semiconductor is higher than at other surrounding locations, then electrostatic forces cause carriers to migrate away from the regions of higher concentration towards regions of lower concentration. This means that the concentration at the higher point falls while in the surrounding regions it rises, thereby establishing a concentration gradient as it does so. This process is known as diffusion and gives rise to a diffusion current. Fig. 1.5 shows the corresponding change in the concentration gradient with distance from the source and its variation with time. Note that the movement of carriers is always in the direction of decreasing concentration. n n t = t 2 > t1 t = t1 Carrier flow Carrier flow x Fig. 1.5 x Change in Concentration Gradient with Diffusion 8 The diffusion of carriers will continue as long as the concentration gradient is maintained and will cease if the gradient eventually becomes zero. The rate at which diffusion takes place determines the associated charge flux density which depends on: electrons holes dn dx dp dx the carrier concentration gradients, q q the charge on the carriers, Dn Dp the diffusion coefficients. The charge flux density for diffusion is then: Jndiff -q .Dn . - dn dx Jpdiff q .Dp . - dp dp -qDp dx dx qDn dn dx for electrons for holes The negative signs in front of the concentration gradients above account for the fact that the carriers move in the direction of decreasing concentration. Note that if the concentration gradients for both holes and electrons are in the same direction, then both types of carrier will be travelling in the same direction and the charge flows will tend to neutralize each other. If, on the other hand, the gradients are in opposite directions, then the flows of charge of opposite polarity will combine and add together. For a piece of homogeneous semiconductor of uniform crosssectional area, A, the Total Diffusion Current due to both electrons and holes in the positive x direction is given as: Idiff Jn diff A Jp diff A Aq(D n dn dp Dp ) dx dx The polarity of the gradients will determine the direction of flow of individual carriers and hence the direction of the overall diffusion current. 9 1.8 Carrier Injection, Diffusion and Concentration Gradient Consider a length of homogeneous p-type semiconductor of uniform cross-sectional area, A, as shown in Fig. 1.6. Under equilibrium conditions, the concentration of majority carriers is determined by the doping concentration, p0 = Na, and the concentration of minority carriers governed by the Law of Mass Action, n0 = ni2/Na. If now, additional or ‘excess’ minority charge carriers are injected into the semiconductor from an external source, this equilibrium is disturbed as the excess minority carriers injected begin to recombine with the majority carriers present within the material. This gives rise to changes in the charge concentration within the material. The rate of change of carrier concentration depends on the rate of injection of excess carriers and the rate of recombination in the volume. For minority carrier electrons: Total Carrier Concentration: n Equilibrium Carrier Concentration: n0 Excess Carrier Concentration: n’ Excess Carrier Concentration is: excess minority carriers recombine with majority carriers inside the volume of the semiconductor n’ = n – n0 A Jn (x) uniform cross sectional area charge flux density becomes a function of distance into the material excess minority carriers injected from an external source Fig 1.6 Excess Charge Carriers Injected into a Semiconductor 10 If there is no electric field present across the semiconductor and the material is homogeneous or uniform, then the injected minority carriers will move into the material in the x direction by diffusion. As minority carriers diffuse into the doped semiconductor some will recombine with the majority carriers in the material. If the supply of excess minority carriers into the material can be maintained by some external means and the majority carriers which recombine can also be replaced, then this will support the recombination process and eventually a new steady-state condition will become established and the concentration of charge within the volume will reach a stable value. The excess minority concentration profile with distance x into the semiconductor is then described as: Lx n'(x) n'(x 0)e n where n’(x = 0) is the quantity of excess minority carriers injected into the material at x = 0. The excess injected minority carriers will penetrate a distance Ln on average into the material before recombining. This distance is equivalent to the depth of penetration which would take place if the concentration gradient at the boundary face of the material persisted into it. The diffusion length is related to the average carrier lifetime through the diffusion coefficient so that for electrons: Dn n L n 2 The profile of the minority carrier concentration across the material into which excess carriers are injected therefore has an exponential form as shown in Fig. 1.7. n (x) n(x = 0) excess carriers n’(x = 0) thermal n carriers 0 0 Ln distance x Fig. 1.7 Profile of Minority Carrier Concentration across Semiconductor 11