Physical foundations of microelectronics. Physical foundations of microelectronics, lecture notes Designs and parameters of generators based on Gunn diodes

Sarapul Polytechnic Institute (branch)

State educational institution

higher professional education

"Izhevsk State Technical University"

Department of Cyprus

Course work

Discipline: Physical foundations of microelectronics.

On the topic: Dislocations. Burgers vector. Effect of dislocation on properties

construction materials.

Done: Checked:

student gr. 471 teachers

Volkov A.V. Ivannikov V.P.

Sarapul, 2010

Introduction........................................................ .................. 1

Types of dislocation................................................... ... ..2

Contour and Burgers vector...................................................2-3

Dislocation movement................................................... ...3-4

Dislocation density...................................................4

Force acting on a dislocation................................4-5

Dislocation energy................................................... ..5

Reproduction and accumulation of dislocations...................5-6

Frank dislocations and stacking faults................6

Dislocations and physical properties of crystals.....7

Dependence of strength on the presence of dislocation...7-8

Crystal growth......................................................... ..........8

Dislocations and electrical conductivity...................................8-9

Conclusion................................................. .....................10

List of references............................................. 11

Introduction

The dislocation theory appeared in the 50s. last century due to the fact that theoretical calculations of the strength of materials differed significantly from practical ones.

The theoretical shear strength of a crystal was first calculated by Frenkel, based on a simple model of two rows of atoms displaced by shear stress. The interplanar distance (distance between rows) is equal to A , and the distance between atoms in the sliding direction is equal to b . Under shear stress τ these rows of atoms are displaced relative to each other, ending up in equilibrium positions at points such as A , IN And WITH , D , where the shear stress required for a given shear configuration is zero. In intermediate positions, the shear stress has finite values, which periodically change in the volume of the lattice. Assume shear stress τ will be a function of offset X with period b :

(1.1)

For small offsets:

(1.2)

Using Hooke's law:

, (1.3)

where G is the shear modulus, and – shear deformation, find the proportionality coefficient To :

(1.4)

Substituting this value To in (1.1) we get:

(1.5)

Maximum value τ , corresponding to the voltage at which the lattice goes into an unstable state:

Can be accepted a ≈ b , then the shear stress

The theoretical shear stresses of various materials calculated in this way turned out to be significantly higher compared to practical values. So for copper

theoretical value

= 760 kgf/mm, and the practical value for real crystals = 100 kgf/mm.

Due to the strong discrepancy between theoretical and practical results, the presence of microscopic linear defects and dislocations in the crystal was assumed.

Dislocations are discontinuities in displacement between two parts of a crystal, one of which undergoes displacement and the other does not. Thus, deformation is represented by the sequential passage of dislocations along the slip plane, and not by simultaneous shear throughout the crystal.

Types of dislocations.

There are two main types of dislocations: edge and screw.

1. Edge dislocations.

The edge dislocation model can be represented by cutting a gap in a piece of an elastically solid body ABCD , ending along the line AB inside this piece (Fig. 1). The material on one side moves, creating a step CDEF . Line A B , corresponding to the end of the gap, is the boundary between the deformed and undeformed material, determines the points at which the dislocation line exits the surface of the body.

Fig.1 Fig.2

Figure 2 shows a visual model of an edge dislocation in a simple cubic lattice. The edge dislocation is caused by the presence of an extra half-plane A, perpendicular to the slip plane B (Fig. 2).

The extra half-plane can be above the slip plane (as in Fig. 2), then the dislocation is called positive, if the half-plane is below then negative.

2.Screw dislocations:

The screw dislocation model is similar to an edge dislocation, but the direction of the screw dislocation is parallel to the line AB, and a step ADEF is formed (Fig. 3).

Fig. 3 Screw dislocation model.

Burgers outline and vector:

To describe dislocations in crystals, the concept of a Burgers contour and vector is introduced. A contour drawn in a perfect lattice is a closed rectangle in which the last of the drawn vectors comes to the starting point in Fig. 4. The contour enclosing the dislocation has a discontinuity, and the vector that must be drawn in order for the contour to close is called the Burgers vector, and the drawn contour is called the Burgers contour. The Burgers vector determines the magnitude and directions of the rupture; it is usually equal to one interatomic distance and is constant along the entire length of the dislocation, regardless of whether its direction or location changes. In a perfect crystal, the Burgers vector is zero. In a crystal with an edge dislocation, it is parallel to the slip direction and corresponds to the slip vector in Fig. 5. In a crystal with a screw dislocation, it is perpendicular to the slip plane Fig. 6

Fig.4 Fig.5 Fig.6

In a crystal, dislocations are also possible that lie completely inside the crystal, and do not extend to its surface, as in those discussed above. Dislocations within a crystal can be interrupted at other dislocations, at grain boundaries and other interfaces. Therefore, dislocation loops or interconnected networks of dislocations are possible inside the crystal. Such a dislocation can be separated from the undeformed region by a dislocation line in the form of a ring or loop; in particular, it can be obtained by pressing a body into the crystal. Figure 7 shows the formation of a prismatic dislocation by indentation over area ABCD.

In this case, an edge and screw dislocation are formed, the Burgers vector, which is the vector sum of the components of the dislocation: (1.6)

At the point at which three dislocations join together, their Fig. 7 Burgers vectors are related by the relation:

(1.7)

Dislocation movement.

An important property of dislocations is their ability to move under the influence of mechanical stress. Let an elementary segment dl of a mixed dislocation with Burgers vector b move in the direction dz. The volume built on these three vectors:

dV = (dz×dl) b, (1.8)

is equivalent to the volume of material moving in the crystal when a dislocation moves. If V=0, the movement of the dislocation is not accompanied by mass transfer or change in the volume of the crystal. This is a conservative movement, or sliding. For edge and mixed dislocations for which the Burgers vector b is not parallel to the dislocation line dl, slip occurs in the plane defined by the vectors b and dl: expression (1.8) is equal to zero if dz lies in the same plane as the vectors b and dl. Obviously, the slip plane of an edge or mixed dislocation is the plane in which the dislocation and its Burgers vector lie. An edge dislocation is extremely mobile in its own slip plane. The movement of an edge dislocation can be represented as a sequential gradual movement of atoms adjacent along the entire length to the dislocation line, accompanied by a redistribution of bonds between these atoms. After each such event, the dislocation moves one interatomic distance. In this case, the stress causing the movement of dislocations is significantly less than the shear stress of the material. As a result of such movement, the dislocation can reach the surface of the crystal and disappear. Thus, the regions of the crystal separated by the slip plane, after the release of the dislocation, will be shifted by one interatomic distance (Fig. 8).

Ministry of Education of the Russian Federation

Oryol State Technical University

Department of Physics

ABSTRACT

on the topic: “The Gunn effect and its use in diodes operating in generator mode.”

Discipline: “Physical foundations of microelectronics”

Completed by a student of group 3–4 Senators D.G.

Supervisor:

Eagle. 2000

The Gunn effect and its use in diodes operating in generator mode.

To amplify and generate microwave oscillations, the anomalous dependence of the electron velocity on the electric field strength in some semiconductor compounds, primarily in gallium arsenide, can be used. In this case, the main role is played by processes occurring in the bulk of the semiconductor, and not in p - n-transition. Generation of microwave oscillations in homogeneous GaAs samples n-type at a constant electric field strength above a threshold value was first observed by J. Gunn in 1963 (therefore, such devices are called Gunn diodes). In Russian literature they are also called devices with volumetric instability or with intervalley electron transfer, since the active properties of diodes are determined by the transition of electrons from the “central” energy valley to the “side”, where they are characterized by a large effective mass and low mobility. In foreign literature, the last name corresponds to the term TED ( Transferred Electron Device).

In a weak field, the electron mobility is high and amounts to 6000–8500 cm 2 /(Vs). When the field strength is higher than 3.5 kV/cm, due to the transition of some electrons to the “side” valley, the average drift velocity of electrons decreases with increasing field. The highest value of the differential mobility modulus in the falling section is approximately three times lower than the mobility in weak fields. At field strengths above 15–20 kV/cm, the average electron velocity is almost independent of the field and is about 10 7 cm/s, so the ratio , and the velocity-field characteristic can be approximately approximated as shown in Fig. 1. The time to establish negative differential conductivity (NDC) is the sum of the heating time of the electron gas in the “central” valley (~10–12 s for GaAs), determined by the energy relaxation time constant and the intervalley transition time (~5–10–14 s).

One would expect that the presence of a falling section of the characteristic in the NDC region with a uniform distribution of the electric field along a uniformly doped GaAs sample would lead to the appearance of a falling section on the current-voltage characteristic of the diode, since the value of the convection current through the diode is defined as , where ; -cross-sectional area; – length of the sample between the contacts. In this section, the diode would have a negative active conductivity and could be used to generate and amplify oscillations similar to a tunnel diode. However, in practice, the implementation of such a regime in a sample of a semiconductor material with an NDC is difficult due to the instability of the field and space charge. As was shown in § 8.1, the fluctuation of the space charge in this case leads to an increase in the space charge according to the law

where is the dielectric relaxation constant; –concentration of electrons in the original n-GaAs. In a homogeneous sample to which a constant voltage is applied , a local increase in the electron concentration leads to the appearance of a negatively charged layer (Fig. 2), moving along the sample from the cathode to the anode.

Fig.1. Approximate dependence of electron drift velocity on electric field strength for GaAs.

Fig.2. To explain the process of formation of an accumulation layer in uniformly doped GaAs.

By cathode we mean a contact to the sample to which a negative potential is applied. The internal electric fields that arise in this case are superimposed on a constant field, increasing the field strength to the right of the layer and decreasing it to the left (Fig. 2, a). The speed of electrons to the right of the layer decreases, and to the left it increases. This leads to a further growth of the moving accumulation layer and to a corresponding redistribution of the field in the sample (Fig. 2, b). Typically, a space charge layer nucleates at the cathode, since near the cathode ohmic contact there is a region with an increased electron concentration and low electric field strength. Fluctuations that occur near the anode contact due to the movement of electrons towards the anode do not have time to develop.

However, such an electric field distribution is unstable and, if there is inhomogeneity in the sample in the form of jumps in concentration, mobility or temperature, it can transform into the so-called strong field domain. The electric field strength is related to the electron concentration by the Poisson equation, which for the one-dimensional case has the form

(1)

An increase in the electric field in part of the sample will be accompanied by the appearance at the boundaries of this area of a space charge, negative on the cathode side and positive on the anode side (Fig. 3, a). In this case, the speed of electrons inside the region decreases in accordance with Fig. 1. Electrons from the cathode side will catch up with electrons inside this area, due to which the negative charge increases and an electron-rich layer is formed. Electrons from the anode side will move forward, due to which the positive charge increases and a depleted layer is formed in which. This leads to a further increase in the field in the fluctuation region as the charge moves towards the anode and to an increase in the extent of the dipole region of the space charge. If the voltage applied to the diode is maintained constant, then as the dipole domain grows, the field outside it will decrease (Fig. 3, b). The increase in the field in the domain will stop when its speed becomes equal to the speed of electrons outside the domain. It's obvious that . The electric field strength outside the domain (Fig. 3, c) will be below the threshold strength, which makes it impossible for the intervalley transition of electrons outside the domain and the formation of another domain until the disappearance of the one previously formed at the anode. After the formation of a stable high-field domain, the current through the diode remains constant during its movement from the cathode to the anode.

Fig.3. To explain the process of formation of a dipole domain.

After the domain disappears at the anode, the field strength in the sample increases, and when it reaches the value , the formation of a new domain begins. In this case, the current reaches a maximum value equal to (Fig. 4, c)

(2)

This mode of operation of a Gunn diode is called flight mode. In the transit mode, the current through the diode consists of pulses following with a period . The diode generates microwave oscillations with a flight frequency , determined mainly by the length of the sample and weakly dependent on the load (it was precisely these oscillations that Gunn observed when studying samples from GaAs and InP).

Electronic processes in a Gunn diode should be considered taking into account the Poisson equations, continuity and total current density, which for the one-dimensional case have the following form:

; (3)

. (4)

Fig.4. Equivalent circuit of a Gunn diode generator (a) and time dependences of voltage (b) and current through the Gunn diode in transit mode (c) and in modes with delay (d) and domain damping (e).

Instantaneous voltage across the diode. The total current does not depend on the coordinate and is a function of time. The diffusion coefficient is often considered to be independent of the electric field.

Depending on the parameters of the diode (the degree and profile of doping of the material, the length and cross-sectional area of the sample and its temperature), as well as on the supply voltage and load properties, the Gunn diode, as a microwave generator and amplifier, can operate in various modes: domain, limiting space charge accumulation (ONZ, in foreign literature LSA – Limited Space Charge Accumulation), hybrid, traveling waves of space charge, negative conductivity.

Domain operating modes.

Domain modes of operation of a Gunn diode are characterized by the presence of a formed dipole domain in the sample during a significant part of the oscillation period. The characteristics of a stationary dipole domain are discussed in detail in [?], where it is shown that from (1), (3) and (4) it follows that the speed of the domain and the maximum field strength in it are related equal area rule

. (5)

In accordance with (5), the areas shaded in Fig. 5, a and bounded by lines are the same. As can be seen from the figure, the maximum field strength in the domain significantly exceeds the field outside the domain and can reach tens of kV/cm.

Fig.5. To determine the parameters of the dipole domain.

Figure 5, b shows the dependence of the domain voltage on the electric field strength outside it, where is the length of the domain (Fig. 3, c). There, an “instrument line” of a diode with a length at a given voltage was built, taking into account the fact that the total voltage across the diode is . Intersection point A determines the voltage of the domain and the field strength outside it. It should be borne in mind that the domain occurs at constant voltage , however, it can also exist when, during the movement of the domain towards the anode, the voltage on the diode decreases to the value (dotted line in Fig. 5, b). If the voltage on the diode is further reduced so that it becomes less than the domain extinction voltage, the resulting domain will resolve. The damping voltage corresponds to the moment the “instrument straight line” touches the line in Fig. 5, b.

Thus, the voltage of domain disappearance turns out to be less than the threshold voltage of domain formation. As can be seen from Fig. 5, due to the sharp dependence of the excess voltage on the domain on the field strength outside the domain, the field outside the domain and the domain speed change little when the voltage on the diode changes. Excess voltage is absorbed mainly in the domain. Already at the domain speed is only slightly different from the saturation speed and can be approximately considered , and , therefore, the flight frequency, as a characteristic of a diode, is usually determined by the expression:

(6)

The domain length depends on the concentration of the donor impurity, as well as on the voltage on the diode and is 5–10 μm. A decrease in the impurity concentration leads to expansion of the domain due to an increase in the depletion layer. The formation of a domain occurs over a finite time and is associated with the establishment of negative differential conductivity and an increase in space charge. The time constant for the rise of the space charge in the mode of small perturbation is equal to the dielectric relaxation constant and is determined by the negative differential mobility and electron concentration. At the maximum value, while the ODP establishment time is less. Thus, the time of domain formation is determined to a large extent by the process of space charge redistribution. It depends on the initial field inhomogeneity, the doping level and the applied voltage.

Fig6. Gunn diode.

It is approximately believed that the Domain will have time to fully form in the following time:

where is expressed in . It makes sense to talk about domain modes only if the domain has time to form during the flight of electrons in the sample. Hence, the condition for the existence of a dipole domain is either .

The product of the electron concentration and the length of the sample is called critical and denote . This value is the boundary between domain modes of the Gunn diode and modes with a stable electric field distribution in a uniformly doped sample. When a strong field domain is not formed, the sample is called stable. Various domain modes are possible. The type criterion is valid, strictly speaking, only for structures in which the length of the active layer between the cathode and the anode is much less than the transverse dimensions: (Fig. 6, a), which corresponds to a one-dimensional problem and is typical for planar and mesastructures. Thin film structures (Fig. 6, b) have an epitaxial active layer of GaAs 1 length can be located between a high-resistance substrate 3 and insulating dielectric film 2 made, for example, from SiO 2. Ohmic anode and cathode contacts are manufactured using photolithography methods. The transverse size of a diode can be comparable to its length. In this case, the space charges formed during the formation of the domain create internal electric fields that have not only a longitudinal component, but also a transverse component (Fig. 6, c). This leads to a decrease in the field compared to a one-dimensional problem. When the thickness of the active film is small, when , the criterion for the absence of domain instability is replaced by the condition . For such structures, with a stable distribution of the electric field, it can be greater.

The domain formation time should not exceed a half-cycle of microwave oscillations. Therefore, there is a second condition for the existence of a moving domain, from which, taking into account (1), we obtain .

Depending on the ratio of the time of flight and the period of microwave oscillations, as well as on the values of the constant voltage and the amplitude of the high-frequency voltage, the following domain modes can be realized: flight-of-flight, mode with domain delay, mode with suppression (quenching) of the domain. Let us consider the processes occurring in these modes for the case of a Gunn diode operating on a load in the form of a parallel oscillating circuit with active resistance at the resonant frequency and the diode being powered by a voltage generator with low internal resistance (see Fig. 4a). In this case, the voltage on the diode changes according to a sinusoidal law. Generation is possible at .

At low load resistance, when , where – the resistance of the Gunn diode in weak fields, the amplitude of the high-frequency voltage is small and the instantaneous voltage on the diode exceeds the threshold value (see Fig. 4b, curve 1). Here, the previously considered transit mode takes place, when after the formation of the domain, the current through the diode remains constant and equal (see Fig. 9.39, c). When the domain disappears, the current increases to . For GaAs. The frequency of oscillations in the flight mode is equal to . Since the ratio is small, the efficiency The number of Gunn diode generators operating in transit mode is small and this mode usually has no practical application.

When the diode operates on a circuit with high resistance, when , the amplitude of the alternating voltage can be quite large, so that during some part of the period the instantaneous voltage on the diode becomes less than the threshold (corresponds to curve 2 in Fig. 4b). In this case they talk about mode with a delay in domain formation. A domain is formed when the voltage on the diode exceeds the threshold, i.e. at a moment in time (see Fig. 4, d). After the formation of the domain, the diode current decreases to and remains so during the time of flight of the domain. When the domain disappears on the anode at a moment in time, the voltage on the diode is less than the threshold and the diode represents an active resistance. The change in current is proportional to the voltage across the diode until the moment when the current reaches its maximum value and the voltage across the diode is equal to the threshold. The formation of a new domain begins, and the whole process repeats. The duration of the current pulse is equal to the delay time of the formation of a new domain. The domain formation time is considered small compared to and . Obviously, such a mode is possible if the time of flight is within the limits and the frequency of the generated oscillations is .

With an even greater amplitude of high-frequency voltage corresponding to the curve 3 in Fig. 4b, the minimum voltage on the diode may be less than the diode quenching voltage. In this case, mode with domain suppression(see Fig. 4, d). A domain is formed at a point in time and dissolves at a point in time when a new domain begins to form after the voltage exceeds a threshold value. Since the disappearance of a domain is not associated with its reaching the anode, the time of flight of electrons between the cathode and anode in the domain quenching mode can exceed the oscillation period: . Thus, in the damping mode. The upper limit of the generated frequencies is limited by the condition and can be .

Electronic efficiency generators based on Gunn diodes operating in domain modes can be determined by expanding the current function into a Fourier series (see Fig. 4) to find the amplitude of the first harmonic and the direct current component. Efficiency value depends on the relations , , , and at the optimal value does not exceed 6% for GaAs diodes in the domain delay mode. Electronic efficiency in the domain-quenching mode is less than in the domain-delaying mode.

ONOZ mode.

Somewhat later, domain modes were proposed and implemented for Gunn diodes mode of limiting the accumulation of space charge. It exists at constant voltages on the diode, several times higher than the threshold value, and large voltage amplitudes at frequencies several times higher than the flight frequency. To implement the ONOS mode, diodes with a very uniform doping profile are required. The uniform distribution of the electric field and electron concentration along the length of the sample is ensured due to the high rate of change in voltage across the diode. If the time period during which the electric field intensity passes through the region of the NDC characteristic is much less than the domain formation time, then there is no noticeable redistribution of the field and space charge along the length of the diode. The speed of electrons throughout the sample “follows” the change in the electric field, and the current through the diode is determined by the dependence of the speed on the field (Fig. 7).

Thus, in the ONOS mode, the negative conductivity of the diode is used to convert the energy of the power source into the energy of microwave oscillations. In this mode, during part of the oscillation period, the voltage on the diode remains less than the threshold and the sample is in a state characterized by positive electron mobility, i.e., the space charge, which managed to form during the time when the electric field in the diode was above the threshold, is dissolved.

We will approximately write the condition for a weak increase in charge over time in the form , Where ; is the average value of the negative differential electron mobility in the region. Resorption of the space charge in time , will be effective if and where ; and – dielectric relaxation time constant and electron mobility in a weak field.

Counting , , we have . This inequality determines the range of values within which the ONZ mode is implemented.

The electronic efficiency of a Gunn diode generator in ONOS mode can be calculated from the current shape (Fig. 7). At The maximum efficiency is 17%.

Fig.7. Time dependence of the current on the Gunn diode in the ONOS mode.

In domain modes, the frequency of generated oscillations is approximately equal to the flight frequency. Therefore, the length of Gunn diodes operating in domain modes is related to the operating frequency range by the expression

where is expressed in GHz, and – in microns. In the ONOS mode, the length of the diode does not depend on the operating frequency and can be many times greater than the length of diodes operating at the same frequencies in domain modes. This allows you to significantly increase the power of generators in the ONO mode compared to generators operating in domain modes.

The considered processes in a Gunn diode in domain modes are essentially idealized, since they are realized at relatively low frequencies (1–3 GHz), where the oscillation period is significantly less than the domain formation time, and the diode length is much greater than the domain length at conventional doping levels . Most often, continuous-wave Gunn diodes are used at higher frequencies in so-called hybrid modes. Hybrid Modes The operation of Gunn diodes is intermediate between the ONOS and domain modes. It is typical for hybrid modes that the formation of a domain takes up most of the oscillation period. An incompletely formed domain resolves when the instantaneous voltage across the diode decreases to values below the threshold. The electric field strength outside the region of increasing space charge remains generally greater than the threshold. The processes occurring in the diode in the hybrid mode are analyzed using a computer using equations (1), (3) and (4). Hybrid modes occupy a wide range of values and are not as sensitive to circuit parameters as the ONOZ mode.

The ONOS mode and the hybrid operating modes of the Gunn diode are classified as “hard” self-excitation modes, which are characterized by the dependence of the negative electronic conductivity on the amplitude of the high-frequency voltage. Putting the generator into the hybrid mode (as well as into the ONOZ mode) is a complex task and is usually carried out by sequentially transitioning the diode from the transit mode to the hybrid mode.

Fig.8. Electronic efficiency of GaAs Gunn diode generators for various operating modes:

1–with domain formation delay

2–with domain suppression

Fig.9. Time dependence of voltage (a) and current (b) of a Gunn diode in the high-efficiency mode.

3-hybrid

Designs and parameters of generators based on Gunn diodes.

Figure 8 shows the values of the maximum electronic efficiency. GaAs Gunn diode in various operating modes. It can be seen that the values do not exceed 20%. Increase efficiency generators based on Gunn diodes is possible through the use of more complex oscillatory systems, which make it possible to provide the time dependences of the current and voltage on the diode, shown in Fig. 9. Expansion of functions and in the Fourier series at and gives electronic efficiency values for GaAs Gunn diodes of 25%. A fairly good approximation to the optimal curve is obtained by using the second voltage harmonic. Another way to increase efficiency consists of using materials with a high ratio in Gunn diodes. Thus, for indium phosphide it reaches 3.5, which increases the theoretical electronic efficiency of diodes to 40%.

It should be kept in mind that the electronic efficiency generators based on Gunn diodes decreases at high frequencies, when the oscillation period becomes commensurate with the establishment time of the NDC (this manifests itself already at frequencies of ~30 GHz). The inertia of the processes that determine the dependence of the average drift velocity of electrons on the field leads to a decrease in the antiphase component of the diode current. The limiting frequencies of Gunn diodes associated with this phenomenon are estimated at ~100 GHz for GaAs devices and 150–300 GHz for InP devices.

The output power of Gunn diodes is limited by electrical and thermal processes. The influence of the latter leads to the dependence of the maximum power on frequency in the form , where the constant is determined by the permissible overheating of the structure, the thermal characteristics of the material, and electronic efficiency. and diode capacity. Limitations on the electrical mode are due to the fact that at high output power the amplitude of oscillations turns out to be commensurate with the constant voltage on the diode: .

In domain modes therefore in accordance with we have:

where is the equivalent load resistance, recalculated to the diode terminals and equal to the module of the active negative resistance of the LPD.

The maximum electric field strength in the domain significantly exceeds the average field value in the diode, at the same time it should be less than the breakdown strength at which avalanche breakdown of the material occurs (for GaAs ). Usually the permissible value of the electric field is considered to be .

As with LPDs, at relatively low frequencies (in the centimeter wavelength range), the maximum output power of Gunn diodes is determined by thermal effects. In the millimeter range, the thickness of the active region of diodes operating in domain modes becomes small and electrical limitations prevail. In continuous mode in the three-centimeter range, a power of 1–2 W can be obtained from one diode with an efficiency of up to 14%; at frequencies 60–100 GHz – up to 100 Watt with an efficiency of a few percent. Gunn diode generators are characterized by significantly lower frequency noise than LPD generators.

The ONOZ mode is characterized by a much more uniform distribution of the electric field. In addition, the length of the diode operating in this mode can be significant. Therefore, the amplitude of the microwave voltage on the diode in the ONOS mode can be 1–2 orders of magnitude higher than the voltage in domain modes. Thus, the output power of Gunn diodes in the ONOS mode can be increased by several orders of magnitude compared to domain modes. For the ONOZ mode, thermal limitations come to the fore. Gunn diodes in ONOS mode operate most often in a pulsed mode with a high duty cycle and generate power up to several kilowatts in the centimeter wavelength range.

The frequency of generators based on Gunn diodes is determined mainly by the resonant frequency of the oscillatory system, taking into account the capacitive conductivity of the diode and can be tuned within a wide range by mechanical and electrical methods.

In a waveguide generator(Fig. 10, a) Gunn diode 1 installed between the wide walls of a rectangular waveguide at the end of a metal rod. Bias voltage is supplied through the inductor input 2 , which is made in the form of sections of quarter-wave coaxial lines and serves to prevent the penetration of microwave oscillations into the power source circuit. The low-Q resonator is formed by the diode mounting elements in the waveguide. The generator frequency is tuned using a varactor diode 3 , located at a half-wavelength distance and installed in the waveguide similar to a Gunn diode. Often the diodes are included in a reduced-height waveguide, which is connected to a standard-section output waveguide by a quarter-wave transformer.

Fig. 10. Design of generators based on Gunn diodes:

a-waveguide; b-microstrip; c–with frequency tuning by YIG sphere

In microstrip design(Fig. 10, b) diode 1 connected between the base and the strip conductor. A high-quality dielectric resonator is used to stabilize the frequency 4 in the form of a disk made of a dielectric with low losses and a high value (for example, barium titanate), located near an MPL strip conductor of width . Capacitor 5 serves to separate the power circuits and the microwave path. The supply voltage is supplied through the inductor circuit 2 , consisting of two quarter-wave segments of the MPL with different wave impedances, and the line with low resistance is open. The use of dielectric resonators with a positive temperature coefficient of frequency makes it possible to create oscillators with small frequency shifts when temperature changes (~40 kHz/°C).

Frequency tunable generators on Gunn diodes can be constructed using single crystals of yttrium iron garnet (Fig. 10, c). The frequency of the generator in this case changes due to the tuning of the resonant frequency of a high-quality resonator, which has the form of a YIG sphere of small diameter, when the magnetic field changes. Maximum tuning is achieved in unpackaged diodes that have minimal reactive parameters. The high-frequency diode circuit consists of a short turn enclosing the YIG sphere 6 . The connection of the diode circuit with the load circuit is carried out due to the mutual inductance provided by the YIG sphere and orthogonally located coupling turns. The electrical tuning range of such generators, widely used in automatic measuring devices, reaches an octave with an output power of 10–20 mW.

Fig. 11. Generalized equivalent circuit of a Gunn diode.

Amplifiers based on Gunn diodes.

The development of amplifiers based on Gunn diodes is of great interest, especially for the millimeter wavelength range, where the use of microwave transistors is limited. An important task when creating amplifiers based on Gunn diodes is to ensure the stability of their operation (diode stabilization) and, above all, to suppress small-signal domain-type oscillations. This can be achieved by limiting the diode parameter, loading the diode with an external circuit, choosing a diode doping profile, reducing the cross-section, or applying a dielectric film to the sample. As amplifiers, both planar and mesastructure diodes are used, which have negative conductivity at voltages above the threshold in a wide frequency range near the flight frequency and are used as regenerative reflective amplifiers with a circulator at the input, as well as more complex film structures that use the phenomenon of wave growth space charge in a material with NDP, often called thin film traveling wave amplifiers(UBV).

In subcritically doped diodes at the formation of a running domain is impossible even at voltages exceeding the threshold. As calculations show, subcritical diodes are characterized by a negative equivalent resistance at frequencies close to the flight frequency, at voltages exceeding the threshold. They can be used in reflective amplifiers. However, due to their low dynamic range and gain, they are of limited use.

Stable negative conductivity over a wide frequency range, reaching 40%, is realized in diodes with at short diode length (~8–15 µm) and voltages . At lower voltages, generation is observed, the breakdown of which with increasing voltage can be explained by a decrease in the NDC of the material with increasing temperature of the device.

A uniform distribution of the electric field along the length of the diode and stable amplification over a wide frequency band can be obtained due to non-uniform doping of the sample (Fig. 12, a). If there is a narrow lightly doped layer about 1 μm long near the cathode, then it limits the injection of electrons from the cathode and leads to a sharp increase in the electric field. Increasing the impurity concentration along the length of the sample towards the anode in the range from to makes it possible to achieve uniformity of the electric field. Processes in diodes with this profile are usually calculated on a computer.

Fig. 12. Doping profile (a) and field distribution (b) in a Gunn diode with a high-resistance cathode region.

The types of amplifiers considered are characterized by a wide dynamic range, an efficiency of 2–3%, and a noise figure of ~10 dB in the centimeter wavelength range.

Development of thin-film traveling wave amplifiers (Fig. 13) is underway, which provide unidirectional amplification over a wide frequency band and do not require the use of decoupling circulators. The amplifier is an epitaxial GaAs layer 2 thick (2–15 µm), grown on a high-resistivity substrate 1 . Ohmic cathode and anode contacts are located at a distance from each other and ensure electron drift along the film when a constant voltage is applied to them. Two contacts 3 in the form of a Schottky barrier with a width of 1–5 μm, they are used to input and output a microwave signal from the device. The input signal supplied between the cathode and the first Schottky contact excites a space charge wave in the electron flow, which changes in amplitude as it moves towards the anode with phase velocity .

Fig. 13. Diagram of a GaAs thin-film traveling wave amplifier with longitudinal drift

For the amplifier to operate, it is necessary to ensure uniformity of the film and uniformity of the electric field along the length of the device. The BW bias voltage lies in the GaAs NDC region, i.e., at . In this case, the space charge wave grows as it moves along the film. A stable, uniform distribution of the electric field is achieved in UWV by using films of small thickness and coating the GaAs film with a dielectric with a large value.

Application of the basic equations of electron motion for the one-dimensional case (1), (3), (4) and the small signal mode, when the constant components of the convection current, electric field strength and charge density are much greater than the amplitude of the variable components (), leads to the dispersion equation for the constant propagation, which has a solution in the form of two waves.

One of them is a direct wave propagating along the film from the cathode to the anode with phase velocity , and has an amplitude that varies according to the law:

where is the time of movement of electrons from the input of the device. When working in the ODP region, the direct wave also increases. The second wave is reverse, propagates from the anode to the cathode and attenuates in amplitude as . The diffusion coefficient for GaAs is , therefore the reverse wave quickly decays. From (9) the gain of the device is (dB)

(10)

Estimate by (10) at And gives a gain of the order of 0.3–3 dB/µm. It should be borne in mind that expression (10) is essentially qualitative. Direct use of it to calculate growing waves of a space charge can lead to errors due to the strong influence of boundary conditions for small film thickness, since the problem must be considered as two-dimensional. Electron diffusion must also be taken into account, limiting the frequency range over which amplification is possible. Calculations confirm the possibility of obtaining a gain of ~0.5–1 dB/μm in the UWV at frequencies of 10 GHz or more. Such devices can also be used as controlled phase shifters and microwave delay lines.

[L]. Berezin et al. Microwave electronic devices. – M. Higher School 1985.

Ministry of Education of the Russian Federation

Oryol State Technical University

Department of Physics ABSTRACT

on the topic: “The Gunn effect and its use in diodes operating in generator mode.”

Discipline: “Physical foundations of microelectronics”

Completed by a student of groups 3–4
Senatorov D.G.

Supervisor:

Eagle. 2000

The Gunn effect and its use in diodes operating in generator mode.

To amplify and generate microwave oscillations, the anomalous dependence of the electron velocity on the electric field strength in some semiconductor compounds, primarily in gallium arsenide, can be used. In this case, the main role is played by processes occurring in the bulk of the semiconductor, and not in the p-n junction. The generation of microwave oscillations in homogeneous n-type GaAs samples at a constant electric field strength above a threshold value was first observed by J. Gunn in 1963 (therefore, such devices are called Gunn diodes). In the domestic literature, they are also called devices with volumetric instability or with intervalley electron transfer, since the active properties of diodes are determined by the transition of electrons from the “central” energy valley to the “side”, where they are characterized by a large effective mass and low mobility. In foreign literature, the latter name corresponds to the term TED (Transferred Electron Device).

where is the dielectric relaxation constant; –electron concentration in the original n-GaAs. In a homogeneous sample to which a constant voltage is applied , a local increase in the electron concentration leads to the appearance of a negatively charged layer (Fig. 2), moving along the sample from the cathode to the anode.

Fig.1. Approximate dependence of electron drift velocity on electric field strength for GaAs.

Fig.2. To explain the process of formation of an accumulation layer in uniformly doped GaAs.

However, such an electric field distribution is unstable and, if there is inhomogeneity in the sample in the form of jumps in concentration, mobility or temperature, it can transform into a so-called strong field domain. The electric field strength is related to the electron concentration by the Poisson equation, which for the one-dimensional case has the form

(1)

Fig.3. To explain the process of formation of a dipole domain.

(2)

This mode of operation of a Gunn diode is called transit mode. In the transit mode, the current through the diode consists of pulses following with a period . The diode generates microwave oscillations with a flight frequency , determined mainly by the length of the sample and weakly dependent on the load (it was precisely these oscillations that Gunn observed when studying samples from GaAs and InP).

Electronic processes in a Gunn diode should be considered taking into account the Poisson equations, continuity and total current density, which for the one-dimensional case have the following form:

; (3)

. (4)

Fig.4. Equivalent circuit of a Gunn diode generator (a) and time dependences of voltage (b) and current through the Gunn diode in transit mode (c) and in modes with delay (d) and domain damping (e).

Domain operating modes.

. (5)

Fig.5. To determine the parameters of the dipole domain.

Figure 5, b shows the dependence of the domain voltage on the electric field strength outside it, where is the length of the domain (Fig. 3, c). There, an “instrument line” of a diode with a length at a given voltage was built, taking into account the fact that the total voltage across the diode is . The intersection point A determines the voltage of the domain and the field strength outside it. It should be borne in mind that the domain occurs at constant voltage , however, it can also exist when, during the movement of the domain towards the anode, the voltage on the diode decreases to the value (dotted line in Fig. 5, b). If the voltage on the diode is further reduced so that it becomes less than the domain extinction voltage, the resulting domain will resolve. The damping voltage corresponds to the moment the “instrument straight line” touches the line in Fig. 5, b.

(6)

Fig6. Gunn diode.

It is approximately believed that the Domain will have time to fully form in the following time:

The value of the product of the electron concentration and the length of the sample is called critical and is designated . This value is the boundary between domain modes of the Gunn diode and modes with a stable electric field distribution in a uniformly doped sample. When a strong field domain is not formed, the sample is called stable. Various domain modes are possible. The type criterion is valid, strictly speaking, only for structures in which the length of the active layer between the cathode and the anode is much less than the transverse dimensions: (Fig. 6, a), which corresponds to a one-dimensional problem and is typical for planar and mesastructures. For thin-film structures (Fig. 6, b), an epitaxial active GaAs layer 1 long can be located between a high-resistivity substrate 3 and an insulating dielectric film 2, made, for example, of SiO 2. Ohmic anode and cathode contacts are manufactured using photolithography methods. The transverse size of a diode can be comparable to its length. In this case, the space charges formed during the formation of the domain create internal electric fields that have not only a longitudinal component, but also a transverse component (Fig. 6, c). This leads to a decrease in the field compared to a one-dimensional problem. When the thickness of the active film is small, when , the criterion for the absence of domain instability is replaced by the condition . For such structures, with a stable distribution of the electric field, it can be greater.

When the diode operates on a circuit with high resistance, when , the amplitude of the alternating voltage can be quite large, so that during some part of the period the instantaneous voltage on the diode becomes less than the threshold (corresponds to curve 2 in Fig. 4b). In this case, we speak of a mode with a delay in domain formation. A domain is formed when the voltage on the diode exceeds the threshold, i.e. at a moment in time (see Fig. 4, d). After the formation of the domain, the diode current decreases to and remains so during the time of flight of the domain. When the domain disappears on the anode at a moment in time, the voltage on the diode is less than the threshold and the diode represents an active resistance. The change in current is proportional to the voltage across the diode until the moment when the current reaches its maximum value and the voltage across the diode is equal to the threshold. The formation of a new domain begins, and the whole process repeats. The duration of the current pulse is equal to the delay time of the formation of a new domain. The domain formation time is considered small compared to and . Obviously, such a mode is possible if the time of flight is within the limits and the frequency of the generated oscillations is .

With an even greater amplitude of the high-frequency voltage, corresponding to curve 3 in Fig. 4b, the minimum voltage on the diode may be less than the diode damping voltage. In this case, a mode with domain damping occurs (see Fig. 4d). A domain is formed at a point in time and dissolves at a point in time when a new domain begins to form after the voltage exceeds a threshold value. Since the disappearance of a domain is not associated with its reaching the anode, the time of flight of electrons between the cathode and anode in the domain quenching mode can exceed the oscillation period: . Thus, in the damping mode. The upper limit of the generated frequencies is limited by the condition and can be .

ONOZ mode.

Somewhat later than the domain modes, a mode of limiting the accumulation of space charge was proposed and implemented for Gunn diodes. It exists at constant voltages on the diode, several times higher than the threshold value, and large voltage amplitudes at frequencies several times higher than the flight frequency. To implement the ONOS mode, diodes with a very uniform doping profile are required. The uniform distribution of the electric field and electron concentration along the length of the sample is ensured due to the high rate of change in voltage across the diode. If the time period during which the electric field intensity passes through the region of the NDC characteristic is much less than the domain formation time, then there is no noticeable redistribution of the field and space charge along the length of the diode. The speed of electrons throughout the sample “follows” the change in the electric field, and the current through the diode is determined by the dependence of the speed on the field (Fig. 7).

Counting , , we have . This inequality determines the range of values within which the ONZ mode is implemented.

The electronic efficiency of a Gunn diode generator in ONOS mode can be calculated from the current shape (Fig. 7). At The maximum efficiency is 17%.

Fig.7. Time dependence of the current on the Gunn diode in the ONOS mode.

The considered processes in a Gunn diode in domain modes are essentially idealized, since they are realized at relatively low frequencies (1–3 GHz), where the oscillation period is significantly less than the domain formation time, and the diode length is much greater than the domain length at conventional doping levels . Most often, continuous-wave Gunn diodes are used at higher frequencies in so-called hybrid modes. Hybrid operating modes of Gunn diodes are intermediate between the ONOS and domain modes. It is typical for hybrid modes that the formation of a domain takes up most of the oscillation period. An incompletely formed domain resolves when the instantaneous voltage across the diode decreases to values below the threshold. The electric field strength outside the region of increasing space charge remains generally greater than the threshold. The processes occurring in the diode in the hybrid mode are analyzed using a computer using equations (1), (3) and (4). Hybrid modes occupy a wide range of values and are not as sensitive to circuit parameters as the ONOZ mode.

Fig.8. Electronic efficiency of GaAs Gunn diode generators for various operating modes:

1–with domain formation delay

2–with domain suppression

Fig.9. Time dependence of voltage (a) and current (b) of a Gunn diode in the high-efficiency mode.

3-hybrid

Designs and parameters of generators based on Gunn diodes.

In domain modes therefore in accordance with we have:

where is the equivalent load resistance, recalculated to the diode terminals and equal to the module of the active negative resistance of the LPD.

In a waveguide generator (Fig. 10, a), Gunn diode 1 is installed between the wide walls of a rectangular waveguide at the end of a metal rod. The bias voltage is supplied through choke input 2, which is made in the form of sections of quarter-wave coaxial lines and serves to prevent the penetration of microwave oscillations into the power source circuit. The low-Q resonator is formed by the diode mounting elements in the waveguide. The frequency of the generator is tuned using a varactor diode 3 located at a half-wave distance and installed in the waveguide similarly to a Gunn diode. Often the diodes are included in a reduced-height waveguide, which is connected to a standard-section output waveguide by a quarter-wave transformer.

Fig. 10. Design of generators based on Gunn diodes:

a-waveguide; b-microstrip; c–with frequency tuning by YIG sphere

In a microstrip design (Fig. 10, b), diode 1 is connected between the base and the strip conductor. To stabilize the frequency, a high-quality dielectric resonator 4 is used in the form of a disk made of a dielectric with low losses and a high value (for example, barium titanate), located near an MPL strip conductor of width . Capacitor 5 serves to separate the power circuits and the microwave path. The supply voltage is supplied through inductor circuit 2, consisting of two quarter-wave MPL sections with different wave impedances, and the line with low resistance is open. The use of dielectric resonators with a positive temperature coefficient of frequency makes it possible to create oscillators with small frequency shifts when temperature changes (~40 kHz/°C).

Frequency-tunable generators based on Gunn diodes can be constructed using yttrium iron garnet single crystals (Fig. 10, c). The frequency of the generator in this case changes due to the tuning of the resonant frequency of a high-quality resonator, which has the form of a YIG sphere of small diameter, when the magnetic field changes. Maximum tuning is achieved in unpackaged diodes that have minimal reactive parameters. The high-frequency circuit of the diode consists of a short turn enclosing the YIG-sphere 6. The connection of the diode circuit with the load circuit is carried out due to the mutual inductance provided by the YIG-sphere and orthogonally located coupling turns. The electrical tuning range of such generators, widely used in automatic measuring devices, reaches an octave with an output power of 10–20 mW.

Fig. 11. Generalized equivalent circuit of a Gunn diode.

Amplifiers based on Gunn diodes.

The development of amplifiers based on Gunn diodes is of great interest, especially for the millimeter wavelength range, where the use of microwave transistors is limited. An important task when creating amplifiers based on Gunn diodes is to ensure the stability of their operation (diode stabilization) and, above all, to suppress small-signal domain-type oscillations. This can be achieved by limiting the diode parameter, loading the diode with an external circuit, choosing a diode doping profile, reducing the cross-section, or applying a dielectric film to the sample. As amplifiers, both planar and mesastructure diodes are used, which have negative conductivity at voltages above the threshold in a wide frequency range near the flight frequency and are used as regenerative reflective amplifiers with a circulator at the input, as well as more complex film structures that use the phenomenon of wave growth space charge in a material with NDC, often called thin-film traveling wave amplifiers (TWA).

Fig. 12. Doping profile (a) and field distribution (b) in a Gunn diode with a high-resistance cathode region.

The types of amplifiers considered are characterized by a wide dynamic range, an efficiency of 2–3%, and a noise figure of ~10 dB in the centimeter wavelength range.

Development of thin-film traveling wave amplifiers (Fig. 13) is underway, which provide unidirectional amplification over a wide frequency band and do not require the use of decoupling circulators. The amplifier is an epitaxial layer of GaAs 2 thick (2–15 μm), grown on a high-resistivity substrate 1. Ohmic cathode and anode contacts are located at a distance from each other and ensure electron drift along the film when a constant voltage is applied to them. Two contacts 3 in the form of a Schottky barrier with a width of 1–5 μm are used to input and output the microwave signal from the device. The input signal supplied between the cathode and the first Schottky contact excites a space charge wave in the electron flow, which changes in amplitude as it moves towards the anode with phase velocity .

Fig. 13. Diagram of a GaAs thin-film traveling wave amplifier with longitudinal drift

One of them is a direct wave propagating along the film from the cathode to the anode with phase velocity , and has an amplitude that varies according to the law:

(10)

[L]. Berezin et al. Microwave electronic devices. – M. Higher School 1985.

Discussion of equations (1) with the aim of modifying them for the field of the EM vector potential, since the new equations will make it possible to consistently describe the processes of non-thermal action of electrodynamic fields in material media: the electric and magnetic polarization of the medium, the transfer of angular momentum of the EM impulse to it. The very initial relationships of the primary relationship between the components of the EM field and the field of the EM vector potential with...

Polarities of power supplies in Figure 3.4 and directions of currents for a p-n-p transistor. In the case of an n-p-n transistor, the voltage polarities and current directions are reversed. Figure 3.4 Physical processes in BT. This operating mode (NAR) is the main one and determines the purpose and name of the transistor elements. The emitter junction injects carriers into a narrow...

They are connected to secondary devices using thermoelectric wires, which, as it were, extend the thermoelectrodes. Secondary devices that work in conjunction with thermoelectric converters are magnetoelectric millivoltmeters and potentiometers. The operation of a magnetoelectric millivoltmeter is based on the interaction of a frame formed by a conductor through which current flows with...

Temperature control; Germanium and silicon flat diodes. Theoretical nutritional knowledge that is necessary for laboratory work: 1. Physical processes that occur as a result of contact of conductors with different types of conductivity. 2. Electronic-directory transition at the equal station. Energy diagram. 3. Injection and extraction of the charge. 4. Volt ampere characteristic (...

Send your good work in the knowledge base is simple. Use the form below

Students, graduate students, young scientists who use the knowledge base in their studies and work will be very grateful to you.

Posted on http://www.allbest.ru//

Introduction

The origin and development of microelectronics as a new scientific and technical direction that ensures the creation of complex radio-electronic equipment (REA) is directly related to the crisis situation that arose in the early 60s, when traditional methods of manufacturing REA from discrete elements by their sequential assembly could not provide the required reliability, efficiency, energy consumption, manufacturing time and acceptable dimensions of REA.

Despite the short period of its existence, the interconnection of microelectronics with other areas of science and technology has ensured unusually high rates of development of this industry and significantly reduced the time for the industrial implementation of new ideas. This was also facilitated by the emergence of peculiar feedback links between the development of integrated circuits, which are the basis for automation of production and management, and the use of these developments to automate the very process of design, production and testing of integrated circuits.

The development of microelectronics has made fundamental changes in the design principles of electronic devices and led to the use of complex integration, which consists of:

structural or circuit integration (i.e., integration of circuit functions within a single structural unit); with the degree of integration on the order of hundreds and thousands of components, existing methods of dividing systems into components, devices, subsystems and blocks, as well as forms of coordinating the development of components, devices and subsystems, become ineffective; at the same time, the center of gravity moves to the area of circuitry, which requires a radical restructuring of the methods for implementing electronic systems with the construction of equipment at the supermodular level;

1.The role of thin film technology in the production of integrated circuits

Integrated electronics is developing not as a new or separate field of technology, but by generalizing many technological techniques previously used in the production of discrete semiconductor devices and in the manufacture of topcoated film coatings. In accordance with this, two main directions have been identified in integrated electronics: semiconductor and thin-film.

The creation of an integrated circuit on a single monocrystalline semiconductor (so far only silicon) wafer is a natural development of the technological principles of creating semiconductor devices developed over the past decades, which, as is known, have proven themselves in operation.

The thin-film direction of integrated electronics is based on the sequential growth of films of various materials on a common base (substrate) with the simultaneous formation of micro parts (resistors, capacitors, contact pads, etc.) and in-circuit connections from these films.

Relatively recently, semiconductor (solid) and thin-film hybrid ICs were considered as competing directions in the development of integrated electronics. In recent years, it has become obvious that these two directions are not at all exclusive, but rather, on the contrary, mutually complement and enrich each other. Moreover, to this day, integrated circuits using any one type of technology have not been created (and, apparently, there is no need for this). Even monolithic silicon circuits, manufactured primarily using semiconductor technology, simultaneously use methods such as vacuum deposition of films of aluminum and other metals to produce in-circuit connections, i.e., methods on which thin film technology is based.

The great advantage of thin-film technology is its flexibility, expressed in the ability to select materials with optimal parameters and characteristics and to obtain, in fact, any required configuration and parameters of passive elements. In this case, the tolerances with which individual parameters of the elements are maintained can be increased to 1-2%. This advantage is especially effective in cases where the exact value of the ratings and the stability of the parameters of passive components are critical (for example, in the manufacture of linear circuits, resistive and RC circuits, some types of filters, phase-sensitive and selective circuits, generators, etc. .).

Due to the continuous development and improvement of both semiconductor and thin film technology, as well as the increasing complexity of ICs, which is reflected in an increase in the number of components and the complexity of their functions, it should be expected that in the near future there will be a process of integration of technological methods and techniques and most complex ICs will be manufactured using converged technology. In this case, it is possible to obtain such parameters and such reliability of the IC that cannot be achieved using each type of technology separately. For example, in the manufacture of a semiconductor IC, all elements (passive and active) are performed in one technological process, so the parameters of the elements are interrelated. The active elements are decisive, since usually the base-collector junction of the transistor is used as a capacitor, and the diffusion region resulting from creating the base of the transistor is used as a resistor. It is impossible to optimize the parameters of one element without simultaneously changing the characteristics of others. Given the characteristics of active elements, the ratings of passive elements can only be changed by changing their sizes.

When using combined technology, active elements are most often manufactured using planar technology in a silicon wafer, and passive elements are manufactured using thin-film technology on oxidized element-by-element (resistors and sometimes capacitors) - the surface of the same silicon wafer. However, the manufacturing processes of the active and passive parts of the IC are separated in time. Therefore, the characteristics of passive elements are largely independent and are determined by the choice of material, film thickness and geometry. Because the transistors of a hybrid IC are located inside the substrate, the size of such a circuit can be significantly reduced compared to hybrid ICs, which use discrete active elements that occupy a relatively large amount of space on the substrate.

Circuits made using combined technology have a number of undoubted advantages. For example, in this case it is possible to obtain resistors with a large value and a small temperature coefficient of resistance, having a very narrow width and a high surface resistance, in a small area. Controlling the deposition rate during the production of resistors allows them to be manufactured with very high precision. Resistors obtained by film deposition are not characterized by leakage currents through the substrate even at high temperatures, and the relatively high thermal conductivity of the substrate prevents the possibility of areas with elevated temperatures appearing in circuits.

Thin films, in addition to the production of ICs using epitaxial-planar technology, are widely used in the production of hybrid ICs, as well as in the manufacture of new types of microelectronic devices (charge-coupled devices, cryotron chargers based on the Josephson effect, chargers on cylindrical magnetic domains, etc.).

2. Thin-film metallization of semiconductor devices and integrated circuits

In the manufacture of semiconductor devices and ICs for producing ohmic contacts to silicon, interconnects and contact pads, as well as gate electrodes of MOS structures, aluminum films have become widespread, due to the following advantages of this metal:

low cost of Al and the possibility of using one metal for all metallization processes, which significantly simplifies and reduces the cost of the technology and prevents the occurrence of galvanic effects;

high electrical conductivity of Al films, close to the electrical conductivity of the bulk material; ease of evaporation of Al in vacuum from tungsten crucibles and electron beam evaporators;

high adhesion of A1 to silicon and its oxides; low-resistance contact of Al with silicon and n-type conductivity;

noticeable solubility of silicon in Al with the formation of a solid solution that almost does not reduce electrical conductivity;

the absence of chemical compounds in the Al--Si system;

chemical interaction of A1 with Si02, partially remaining on the contact pads; chemical resistance A1 in an oxidizing environment and radiation resistance;

ease of photolithographic operations to obtain the configuration of conductive tracks using etchants that do not react with silicon and silicon dioxide; good Al ductility and resistance to cyclic temperature changes.

The grain size of the deposited Al films depends significantly on the evaporation rate and temperature of the substrates. The larger the grain size and the more perfect the crystal structure of the film, the lower its resistivity, the less effect of electromigration and, as a result, current-carrying paths and ohmic contacts have a longer service life. The oriented growth of Al films on non-oxidized silicon surfaces in the (111) plane is observed at deposition rates of about 3 * 10-2 μm * s-1 and a substrate temperature of 200--250°C.

To obtain such high film deposition rates, electron beam evaporators are most often used. In this case, the degree of perfection of the crystalline structure of the films can change uncontrollably due to additional radiation heating of the substrates, the magnitude of which depends both on the power of the evaporator, and on the substrate material and the thickness of the deposited film. Uncontrolled changes in the structure of the film also arise due to the presence of charged particles in the molecular beam of evaporated Al vapor. The higher the cathode emission current and the higher the evaporation rate, the higher the concentration of charged particles.

One of the significant disadvantages of pure Al films is the transfer of matter as a result of electrodiffusion (the drift of material ions along a conductor, whether there is a potential difference at the ends of the latter). The speed of ion movement is a function of temperature and increases with temperature. In addition to electrodiffusion, diffusion of metal atoms is possible as a result of the temperature difference at the ends of the conductor. If Al is deposited on silicon oxide, this causes poor heat dissipation, the appearance of “hot” centers on the conductive paths and, as a result, significant temperature gradients. Electromigration of Al at current densities lower than for other metals leads to the appearance of voids in the film (Kirkendall effect).

Since electrodiffusion is an activation process, it significantly depends on the state of the grain boundary surface. Reducing the extent of boundaries by increasing grain sizes and selecting a protective coating material can significantly increase the activation energy and, as a consequence, the time between failures. A significant increase in the time between failures can be achieved by adding impurities of copper, magnesium, chromium, and aluminum oxide to aluminum.

After applying the film A1 and obtaining the required configuration of the current-carrying tracks, A1 is fused into silicon at a temperature of 500-550°C to obtain a low-resistance contact. Migration of excess silicon on the current paths adjacent to the contact substrates causes A1 peeling and IC failures. To prevent this, it is necessary to introduce about 2 wt. into it when A1 evaporates. % silicon. The addition of silicon to the contact pads from A1 reduces the migration of silicon from the shallow emitter layer (about 1 μm), which significantly increases the performance of the IC on bipolar transistors and prevents short-circuiting of the shallow emitter junctions in the IC. To prevent silicon migration into the A1 film, a titanium film can be used as an intermediate layer. The use of the method of creating ohmic contacts with a titanium sublayer in fast-acting ICs made it possible to increase the time between failures by 20 times. In addition to titanium, an underlayer of platinum or palladium can be used to form platinum silicide or palladium silicide.

Along with the previously listed advantages, aluminum metallization has a number of significant disadvantages, the most important of which are the following:

low activation energy of A1 atoms, causing electromigration at current densities of approximately 106 A/cm2 and elevated temperatures, resulting in the appearance of voids in the films;

the possibility of a short circuit through the dielectric in multi-level metallization systems due to the formation of sharp protrusions on the spit as a result of electromigration and recrystallization of A1;

danger of galvanic corrosion of Al when using other metals simultaneously; high diffusion rate of A1 along grain boundaries, which does not allow the use of devices with A1 metallization at temperatures above 500°C;

intense chemical interaction of A1 with silicon dioxide at a temperature of about 500°C;

low melting point in the eutectic of aluminum-silicon systems is about 577°C;

a large difference (6 times) between the thermal expansion coefficients A1 and 51;

softness of A1 and, therefore, low mechanical strength of films;

impossibility of connecting leads by soldering;

high threshold voltage in MOS structures due to the high work function.

Due to the listed disadvantages, aluminum metallization is not used in ICs and transistors with small emitter junctions, as well as in MIS ICs for ... creating gate electrodes. For this purpose, single-layer and multilayer systems made of various metals are used (including A1 for the top layer). The most suitable materials are tungsten and molybdenum. In particular, tungsten has almost the same TCR as silicon, good ohmic contact to silicon- and n-type conductivity, a small (2.5 times) difference from aluminum in electrical conductivity, the highest activation energy of all metals during self-diffusion, high temperature melting of eutectic with silicon, chemical inertness in air and in an aqueous solution of hydrofluoric acid, as well as high hardness, which eliminates the possibility of scratches on the film.

Due to the high temperature resistance of W, it can be used for multi-layer metallization, alternating layers of silicon dioxide with W. During heat treatment, no mounds are formed on the surface of the film and there is no danger of short circuiting between current-carrying paths in multi-layer metallization. In addition, W films (as well as Mo films) are a metallurgical barrier that prevents the formation of an intercrystalline structure of silicon and aluminum.

The disadvantage of W metallization is the difficulty of obtaining films (for which pyrolysis of tungsten hexofluoride is usually used) and etching them (in an alkaline solution of ferrocyanide). Both of these processes are complex and involve toxic substances. In addition, it is impossible to connect external leads directly to tungsten, so some other metal (Pt, Ni, Au, Cu, Al, etc.) is applied on top of it to the contact pads.

In the manufacture of microwave ICs, special-purpose ICs, and also in hybrid technology, metallization is used, consisting of several layers of thin metals. In this case, usually the first (bottom) layer of metal must have high adhesion to both silicon and silicon dioxide and at the same time have low solubility and diffusion coefficients in these materials. These requirements are met by metals such as chromium, titanium, molybdenum, and platinum silicide. With two-layer metallization, the second (upper) layer of metal must have high electrical conductivity and ensure welding of wire leads to it. However, in some systems (such as Cr-Au, Ti-Au or Cr-Cu) contacts

During heat treatment, they lose mechanical strength as a result of the formation of intermetallic compounds at their boundaries. In addition, the overlying metal diffuses through the underlying layer into the silicon, which reduces the mechanical strength of the joint and changes the contact resistance. To eliminate this phenomenon, a third layer of metal is usually used, which is a barrier that prevents the interaction of the upper metallization layer with silicon. For example, in the triple system Tt-Pl-Au, which is used in the manufacture of beam terminals, the layer

Rice. 1. Scheme of the manufacturing process of two-level metallization in the A1-A1rOz-A1 system. microelectronics integrated thin film

a-- applying thick and thin layers of silicon oxide before metallization (the area of ohmic contact is shown); b - application of aluminum, forming the first level; c -- photoengraving of the first level of metal; d - anodizing of the first level of metallization with a photoresist mask; d - application of aluminum forming the second level; f - photoengraving of the second level of metallization.

Pt with a thickness of about 5X10-2 μm serves as a barrier against the diffusion of A1 into S1. In addition, for beam terminals in MIS ICs, Cr-Ag-Au, Cr-Ag-Pt, Pd-Ag-Au systems are used, in which a silver film plays the role of a barrier. For hybrid ICs and stripline microwave IC lines, the Cr-Cu and Cr-Cu-Cr systems are used.

An increase in the density of elements on a chip required the use of multi-level metallization. In Fig. Figure 1 shows the sequence of manufacturing two-level metallization in the A1-A1203-A1 system, which is used in charge-coupled devices.

A relatively new insulating material for multi-level metallization is polyimide, with which five-level metallization of LSIs on MOS transistors is obtained.

3. Factors affecting the properties of thin films

The growth of one substance on a substrate from another substance is a very complex process, depending on a large number of difficult-to-control parameters: the structure of the substrate, the state of its surface, temperature, the properties of the evaporated substance and the rate of its deposition, the material and design of the evaporator, the degree of vacuum, the composition of the residual environment and a number of others. In table Figure 1 shows the relationship between the properties of films and the conditions of their deposition.

Film properties

factors influencing these properties

Grain size

Substrate and film material. Substrate contamination.

Mobility of atoms of deposited material on the surface

substrates (substrate temperature, deposition rate).

Substrate surface structure (degree of roughness,

presence of crystals)

Crystal placement

Substrate structure "" (monocrystalline,

polycrystalline or amorphous). Substrate contamination

(violation of the film structure). Substrate temperature

(ensuring the necessary mobility of the atoms of the deposited

material)

Adhesion between film

Substrate and film material. Additional processes

(for example, the formation of an intermediate oxide layer

between the film and the substrate). Substrate contamination.

Mobility of atoms of deposited material

Pollution

Purity of the evaporated material. Evaporator material.

Substrate contamination. Degree of vacuum and composition

gases and deposition rate

Oxidation

The degree of chemical affinity of the deposited material to

oxygen. Absorption of water vapor by the substrate.

Substrate temperature. Degree of vacuum and composition

residual environment. The relationship between residual pressure

gases and deposition rate

Voltage

Film and substrate material. Substrate temperature.

Grain size, inclusions, crystallographic defects in

film. Annealing. Angle between molecular beam and substrate

Depending on the specific deposition conditions, films of the same substance may have the following main structural features: an amorphous structure, characterized by the absence of a crystal lattice; colloidal (fine-grained) structure, characterized by the presence of very small crystals (less than 10~2 µm); granular (coarse-grained) structure with large crystals (10-1 µm or more); monocrystalline structure, when the entire film is a continuous crystal lattice of atoms of a given material.

4.Substrates

The material used for the manufacture of substrates must have a homogeneous composition, a smooth surface (with a finishing grade of 12-14), have high electrical and mechanical strength, be chemically inert, have high heat resistance and thermal conductivity, thermal expansion coefficients of the substrate material and the deposited film should be close in value. It is quite clear that it is almost impossible to select materials for substrates that would equally satisfy all the listed requirements.

As substrates for hybrid ICs I use glass-ceramic, photositall, high-alumina and beryllium ceramics, glass, polycor, polyimide, as well as metals coated with a dielectric film.

Sitalls are glass-ceramic materials obtained by heat treatment (crystallization) of glass. Most glass ceramics were obtained in the systems Li2O-Al2O3-SiO2-TiO2 and RO-Al2O3-SiO2-TiO2 (CO type CaO, MgO, BaO).

Unlike most high-strength, refractory crystalline materials, glass-ceramic has good flexibility during formation. It can be pressed, drawn, rolled and centrifugally cast, and it can withstand sudden changes in temperature. It has low dielectric losses, its electrical strength is not inferior to the best types of vacuum ceramics, and its mechanical strength is 2-3 times stronger than glass. Sitall is non-porous, gas-tight and has insignificant gas evolution at high temperatures.

Since glass ceramics are multiphase in structure, when they are exposed to various chemical reagents used, for example, to clean the surface of the substrate from contaminants, deep selective etching of individual phases is possible, leading to the formation of a sharp and deep relief on the surface of the substrate. The presence of roughness on the substrate surface reduces the reproducibility of parameters and the reliability of thin-film resistors and capacitors. Therefore, to reduce the height and smooth the edges of micro-irregularities, sometimes a primer layer of a material with good dielectric and adhesive properties, as well as a uniform structure (for example, a layer of silicon monoxide several microns thick) is applied to the substrate.

Of the glasses, amorphous silicate glasses, alkali-free glass C48-3, borosilicate and quartz glass are used as substrates. Silicate glasses are obtained from a liquid melt of oxides by supercooling them, as a result of which the structure of the liquid is preserved, i.e., the characteristic amorphous state. Although glasses contain areas with a crystalline phase - crystallites, they are distributed randomly throughout the entire structure, occupy a small part of the volume and do not have a significant effect on the amorphous nature of the glass.

Quartz glass is a one-component silicate glass, consisting almost entirely of silicon and obtained by melting its natural varieties. It has a very low coefficient of thermal expansion, which determines its exceptionally high heat resistance. Compared to other glasses, quartz glass is inert to the action of most chemical reagents. Organic and mineral acids (with the exception of hydrofluoric and phosphoric acids) of any concentration, even at elevated temperatures, have almost no effect on quartz glass.

Ceramic substrates are of limited use due to their high porosity. The advantages of these substrates are high strength and thermal conductivity. For example, a BeO-based ceramic substrate has 200-250 times higher thermal conductivity than glass, so under intense thermal conditions it is advisable to use beryllium ceramics. In addition to beryllium ceramics, high-alumina (94% Al2Oz) ceramics, dense aluminum oxide, steatite ceramics, and glazed ceramics based on aluminum oxide are used. It should be noted that glazes are less than 100 microns thick and therefore do not provide a noticeable barrier between the film and the substrate at low power levels. The microroughness of untreated ceramics is hundreds of times greater than that of glass, reaching several thousand angstroms. They can be significantly reduced by polishing, but this significantly contaminates the ceramic surface.

The presence of contaminants on the substrate has a significant effect on both the adhesion and the electrical properties of films. Therefore, before deposition, it is necessary to thoroughly clean the substrates, as well as protect them from the possibility of the appearance of oil films that can arise as a result of migration of working fluid vapors from pumps. An effective cleaning method is ion bombardment of the substrate surface in a glow discharge plasma. For this purpose, special electrodes are usually provided in the working chamber of a vacuum installation, to which a voltage of several kilovolts is supplied from a low-power high-voltage source. Electrodes are most often made of aluminum because it has the lowest cathode sputtering rate among metals.

It should be borne in mind that even minor contamination can completely change the film growth conditions. If contaminants are located on the substrate in the form of small islands isolated from each other, then depending on which binding energy is greater: between the film material and the contaminant material or between the film material and the substrate, a film can form either on these islands or on a bare parts of the substrate.

Film adhesion depends to a very large extent on the presence of an oxide layer, which may arise during the deposition process between the film and the substrate. Such an oxide layer is formed, for example, during the deposition of iron and nichrome, which explains the good adhesion of these films. Films made of gold, which is not subject to oxidation, have poor adhesion, and therefore an intermediate sublayer of a material with high adhesion must be created between the gold and the substrate. It is desirable that the resulting oxide layer be concentrated between the film and the substrate. If the oxide is dispersed throughout the film or is located on its surface, then the properties of the film can change significantly. The formation of oxides is strongly influenced by the composition of residual gases in the working volume of the installation and, in particular, by the presence of water vapor.

5.Thin film resistors

The materials used in the manufacture of resistive films must provide the ability to obtain a wide range of time-stable resistors with a low temperature coefficient of resistance (TCR), have good adhesion, high corrosion resistance and resistance to prolonged exposure to elevated temperatures. When the material is deposited on the substrate, thin, clear lines of complex configuration should be formed with good repeatability of the pattern from sample to sample.

Resistive films most often have a fine-grained dispersed structure. The presence of dispersion r, the structure of the films allows, in a first approximation, to consider their electrical resistance as the total resistance of individual granules and barriers between them, in which the nature of the total resistance determines the magnitude and sign of TK.S. So, for example, if the resistance of the grains themselves is predominant, then the conductivity of the film is metallic in nature and the TCR will be positive. On the other hand, if the resistance is due to the passage of electrons through the gaps between the grains (which usually occurs with small film thicknesses), then the conductivity will be of a semiconductor nature and the TCR will accordingly be negative.

Monolithic IC manufacturing primarily uses high-impedance resistors. In order for resistors to be as small as possible, they must be manufactured to the same resolution and tolerance as other IC elements. This excludes the use of free metal masks to obtain the required configuration of resistors and allows it to be carried out only using photolithography.

When manufacturing micro-power monolithic ICs using combined technology, it becomes necessary to place high-resistance resistors with a resistance of up to several megaohms on a relatively small area of the crystal, which can only be achieved if the resistor material has Rs (10--20) kOhm/ c. The process of manufacturing resistors must be combined with the main technological process of manufacturing the entire silicon IC using planar or epitaxial-planar technology. For example, resistive films should not be sensitive to the presence of silicon nitride, phosphorus, borosilicate glass and other materials used in the production of monolithic ICs on the silicon wafer. They must withstand the relatively high temperature (500-550°C) that occurs during the IC sealing process, and in some cases must not change their properties under the influence of an oxidizing environment. Monolithic ICs mainly use nichrome and tanta to make resistors.

In the manufacture of hybrid ICs, a much wider range of thin-film resistor materials is used.

As low-resistance films with Rs from 10 to 300 Ohm. films of chromium, nichrome and t-tal are used. The production of chromium films with reproducible electrophysical properties is somewhat complicated by its ability to form compounds (especially oxide ones) when interacting with residual gases during evaporation and deposition. Resistors based on chromium-nickel alloy (20% Cr and 80% Ni) have significantly more stable characteristics. Tantalum films, due to the presence of various structural modifications, have a very wide range of surface resistances (from several Ohm/s for a-tantalum to several MOhm/s for low-density tantalum ) Tantalum nitride is also used as a highly stable resistive material.

A significant expansion of resistor ratings is achieved by using metal-ceramic films and films of silicides of some metals. In these systems, chromium is most often used as a metal, and oxides, borides, nitrides and silicides of transition metals, as well as oxides of some metalloids, are used as a dielectric. Films made of chromium disilicide, as well as films made of an alloy of silicon, chromium and nickel, have Rs up to 5 kOhm/s; for films based on systems chromium --- silicon monoxide Rs, depending on the chromium content, can vary from units to hundreds of Ohms/s.

6.Thin film capacitors

Thin-film capacitors, despite the apparent simplicity of the three-layer structure, are the most complex and labor-intensive compared to other film passive elements.

Unlike resistors, pads and switching, in the manufacture of which it is sufficient to deposit one or two layers (sublayer and layer), the manufacture of thin film capacitors requires the deposition of at least three layers: the bottom plate, the dielectric film and the top plate (the use of more plates complicates the manufacturing process of capacitors and increases their cost).

The material used for the manufacture of dielectric films must have good adhesion to the metal used for the capacitor plates, be dense and not subject to mechanical destruction when exposed to temperature cycles, have a high breakdown voltage and low dielectric losses, have a high dielectric constant, and not decompose during the process of evaporation and deposition and have minimal hygroscopicity.

The most common materials used as dielectrics in film capacitors are silicon monoxide (Si0) and germanium monoxide (GeO). In recent years, aluminosilicate, borosilicate and antimonidogermanium glasses have been used for this purpose.

The most promising dielectrics are composite glassy compounds, since they have the ability to change the electrophysical, physicochemical and thermodynamic properties over a wide range by selecting the composition of the glass and implementing the features of the aggregative state of glassy systems in thin-film metal-dielectric-metal structures.

7. Films of tantalum and its compounds

In recent years, films of tantalum and its compounds have become increasingly widespread in the manufacture of film elements of integrated circuits. The choice of tantalum as the starting material is largely explained by the fact that, depending on the conditions for obtaining tallalum films, they can have a different structure and, accordingly, change both their resistivity and its temperature coefficient within wide limits.

In terms of crystal structure and electrical properties, b-tantalum films are closest to the bulk sample; they have a coarse-crystalline body-centered structure and have a relatively low resistivity (20-40 μOhm-cm). Unlike k-tantalum, p-tantalum, which has a tetragonal fine-crystalline structure and a resistivity of 160-200 km Ohm * cm, is not found in massive samples. This metastable modification of tantalum is characteristic only of thin films.

The production of films of b - and c - tantalum is usually carried out by cathode sputtering at a voltage of 4--5 kV and a current density of 0.1--1 mA/cm2. If you reduce the voltage and do not increase the argon pressure, the discharge current will decrease, which will lead to a significant decrease in the deposition rate. This produces films of low density, having a highly porous structure with pore sizes of (4--7)-10-3 µm, consisting of a larger number of k- or p-tantalum grains with crystal sizes of (3--5) * 10-2 µm. The high porosity of the films and the appearance of the metal-dielectric mixture system cause an anomalous increase in resistivity (about 200 times compared to b-tantalum) and a change in its temperature coefficient. If nitrogen is added to argon in an amount significantly exceeding the background of residual gases, tantalum nitride films can be obtained having two stable states Ta2N and TaN with different crystal structures and electrical properties.

The presence of several modifications of tantalum (b- and b-tantalum, low-density tantalum) and its nitride makes it possible to choose a variety of topological solutions when designing the passive part of microcircuits.

Pure b-tantalum, due to high mechanical stresses in the film and poor adhesion to the substrate, has not found widespread use in the manufacture of RC elements of microcircuits; b-tantalum is used for the manufacture of the lower plates of capacitors and partially for the production of resistors. Tantalum nitride and low density tantalum are used to make resistors. The practical value of low-density tantalum lies in the ability to obtain highly stable thin-film resistors (from 10 kOhm to several megaohms) that are small in size and have a simple configuration. Thin-film capacitors can be made much more easily from low-density tantalum, since in this case the top electrode, as well as the bottom, can be obtained by sputtering tantalum, whereas when using tantalum of normal density, attempts to obtain the top electrode in this way often resulted in damage to the dielectric layer. In addition, low-density tantalum makes it possible to produce RC circuits with distributed parameters and an adjustable resistor value, which can be used as the upper electrode of a capacitor.

Tantalum pentoxide (Ta2O5), obtained by electrolytic or plasma anodization, has low dielectric losses and can be used both as a dielectric for a capacitor and as an insulator or protective layer for a resistor. In addition, anodizing can be used to accurately adjust the values of capacitors and resistors. The use of ion etching, as well as the solubility of tantalum nitride, pure tantalum and its oxides in various etchants, makes it possible to use a variety of methods to obtain the required configuration of microcircuits.

Thus, based on tantalum, it is possible to ensure the group production of passive elements (resistors, capacitors, connecting conductors and contact pads) with both concentrated and distributed parameters, which in their complexity are not inferior to elements made on the basis of other materials, but at the same time have significantly greater accuracy, stability and reliability. The versatility of tantalum and the lack of need to use other materials indicate that the vast majority of passive IC elements can be manufactured based on “tantalum technology”.

Conclusion

The current stage of development of integrated electronics is characterized by tendencies to further increase operating frequencies and reduce switching times, increase reliability, and reduce costs for materials and the IC manufacturing process.

Reducing the cost of integrated circuits requires the development of qualitatively new principles for their manufacture using processes based on similar physical and chemical phenomena, which, on the one hand, is a prerequisite for the subsequent integration of homogeneous technological operations of the production cycle and, on the other hand, opens up fundamental the ability to control all operations from a computer. The need for qualitative changes in technology and technical re-equipment of the industry is also dictated by the transition to the next stage of development of microelectronics - functional electronics, which is based on optical, magnetic, surface and plasma phenomena, phase transitions, electron-phonon interactions, effects of accumulation and charge transfer, etc. .

The criterion for the “progressiveness” of the technological process, along with the improvement of the parameters and characteristics of the product itself, is high economic efficiency, determined by a number of private, interrelated criteria that ensure the possibility of building sets of fully automated, high-performance equipment with a long service life.

The most important particular criteria are:

universality, i.e. the ability to carry out the entire (or the overwhelming number of operations) of the production cycle using the same technological methods;

continuity, which is a prerequisite for subsequent integration (combination) of a number of technological operations of the production cycle, combined with the possibility of using simultaneous group processing of a significant number of products or semi-finished products;

high speed of all main operations of the technological process or the possibility of their intensification, for example, as a result of exposure to electric and magnetic fields, laser radiation, etc.;

reproducibility of parameters at each operation and a high percentage of yield of both semi-finished and suitable products;

manufacturability of the design of a product or semi-finished product that meets the requirements of automated production (possibility of automated loading, basing, installation, assembly, etc.), which should be reflected in the simplicity of the form, as well as limited tolerances for overall and basic dimensions;

formalization, i.e. the possibility of drawing up (based on analytical dependencies of product parameters on technological process parameters) a mathematical description (algorithm) of each technological operation and subsequent control of the entire technological process using a computer;

adaptability (vitality) of the process, i.e. the ability to exist for a long time in conditions of the continuous emergence and development of new competitive processes and the ability to quickly rebuild equipment for the manufacture of new types of products without significant capital costs.

Most of the listed criteria are satisfied by processes that use electronic and ionic phenomena occurring in vacuum and rarefied gases, with the help of which it is possible to produce:

ion sputtering of metals, alloys, dielectrics and semiconductors in order to obtain films of various thicknesses and compositions, interconnections, capacitive structures, interlayer insulation, interlayer wiring;

ion etching of metals, alloys, semiconductors and dielectrics in order to remove individual localized areas when obtaining an IC configuration;

plasma anodizing to obtain oxide films;

polymerization of organic films in areas irradiated with electrons to obtain organic insulating layers;

cleaning and polishing the surface of substrates;

growing single crystals;

evaporation of materials (including refractory ones) and recrystallization of films;

micro-milling of films;

micro-welding and micro-soldering to connect IC leads, as well as sealing housings;

non-contact methods for monitoring IC parameters.

The commonality of the physical and chemical phenomena on which the listed processes are based shows the fundamental possibility of their subsequent integration in order to create a new technological base for high-performance automated production of integrated circuits and functional electronics devices.

Posted on Allbest.ru

...

Physical foundations of microelectronics. Physical foundations of microelectronics, lecture notes Designs and parameters of generators based on Gunn diodes

Department of Physics

ABSTRACT

Eagle. 2000

Send your good work in the knowledge base is simple. Use the form below

Similar documents