RU2006984C1 - Process of rejection of semiconductor structures on semi-insulating backing by degree of display of effect of inverse control - Google Patents

Abstract

FIELD: semiconductor technology. SUBSTANCE: process of rejection of semiconductor structures includes formation of contacts to semi-insulating backing and active conductive layer, application of voltage to contact on backing with reference to contact on active layer, registration of change of conductance of active layer. For provision of capability of nondestructive testing and of increase of its graphicality change of conductance of active layer emerging as a result of application of voltage to contacts is determined by change of conductance of the whole structure by feeding SHF radiation to structure of conductance of the whole structure by feeding SHF radiation to structure and by measurement of reflected (absorbed) SHF power of structure. For these purposes contact to active layer is made pressure one and for contact with backing electrolyte transparent to SHF radiation is used. EFFECT: increased efficiency of nondestructive testing. 4 dwg

Description

 The invention relates to semiconductor technology and can be used for non-destructive input control of the properties of multilayer semiconductor structures designed for the manufacture of planar integrated circuits (ICs).

In the manufacture of ICs based on A ³ B ⁵ semiconductors, a serious obstacle to increasing the degree of integration and obtaining workable circuits is spurious coupling between adjacent devices, elements located on a semi-insulating substrate. The physical reason for this connection is the effect of controlling the properties of the device from the side of a semi-insulating substrate, which plays the role of a spurious shutter. This spurious effect is called the substrate control effect or the reverse control effect (EEC). The magnitude of this effect strongly depends on the content of impurities and defects in the semi-insulating substrate, on the degree of their compensation, and especially on the parameters of the film / substrate interface that are not controlled by the technology and the parameters of the buffer layer. Therefore, measuring the value of the EEC in the finished IC or test samples containing semiconductor devices can solve the problem of determining the suitability of structures for manufacturing IC.

 A known method of selecting multilayer semiconductor structures suitable for manufacturing ICs, including creating a field-effect transistor with a Schottky gate and controlling an ohmic contact to the substrate, applying a negative bias to it and registering the resulting change in the drain current of the transistor [1].

 The magnitude and kinetics of the change in the drain current at a given voltage V on the substrate, the threshold for the onset of this change, are the parameters of the EEC necessary to conclude the suitability of the controlled structure intended for the manufacture of ICs.

 The disadvantage of this method is the need to manufacture a transistor and a control contact, i.e., the need to perform a number of complex technological operations, such as photolithography, etching a multilayer structure, burning ohmic contacts, spraying the shutter.

 This method is time-consuming and leads to the destruction of part of the washer of a controlled multilayer structure and may not be accurate when the original washer is not uniform in area, and it is not profitable to produce test transistors over the entire area.

 The closest technical solution is a method for selecting multilayer semiconductor structures suitable for manufacturing ICs, including creating ohmic contacts to a semi-insulating substrate and a conductive layer of the structure to create a gate-free field effect transistor on it, applying voltage to the contact on the substrate relative to the contact on the active layer and registering the result of this change in the drain current of a gateless transistor [2].

 For the case of a gate-free transistor, it is proved that a change in the drain current is caused by a change in the conductivity of the active conductive layer (channel) of the structure. In this regard, in this method, the operation of detecting changes in the drain current of the transistor can be considered as the operation of detecting changes in the conductivity of the active layer of the structure.

 The disadvantage of this method is the need to perform a number of complex labor-intensive technological operations for the preparation of test transistors and control contacts: photolithography, etching of a multilayer structure, incineration of ohmic contacts. This method is destructive, since it requires the preparation of test structures, in addition, if the washer is not uniform in area, and it is unprofitable to produce test structures throughout the area, then the method becomes inaccurate.

 The purpose of the invention is the implementation of the possibility of non-destructive rejection of structures and increase expression.

The goal is achieved in that in the known method for selecting multilayer semiconductor structures suitable for fabricating ICs, including creating contacts to a semi-insulating substrate and a conductive active layer of the structure, applying a voltage V to the contact on the substrate relative to the contact on the active layer and detecting the resulting conductivity change active layer (conductive channel of the structure), the contact to the substrate is carried out by applying a layer of liquid, for example, an electrolyte transparent to RF radiation, contact to the conductive active layer is performed by pressing the surface of the conductive layer structure, wherein the change in conductivity of the active layer structure determined by the change in the whole liquid contact conduction structure of a microwave. In order to increase its expressivity, the selection of structures suitable for the fabrication of ICs is carried out by the value of EOW by measuring the value ΔW _microwave ^V / Δ W _microwave ^F , where ΔW _microwave ^V is the change in microwave power transmitted through the structure (reflected from the structure) caused by the application voltage and to the substrate; W _microwave ^f - caused by the illumination of the structure from the side of the conducting layers with light with a quantum energy greater than the band gap of the semiconductor and flux f, leading to saturation of ΔW _microwave ^f .

10 МОм (R_подл = ρ·l/S = 10⁸ Ом˙см˙0,03 см/0,1 см² = 30 МОм). Более того, в электрической цепи жидкость - i-n-n⁺ - прижимной металлический зонд, как оказалось, самым высокоомным является не подложка, а n-i-переход, на котором и падает практически все напряжение. На переходе жидкость/подложка выделяется малая часть напряжения, а переход n⁺-металлический зонд смещается в прямом направлении и становится достаточно проводящим. Таким образом, основное напряжение выделяется на переходе полуизолирующая подложка/активная пленка (n-i или i-p). В результате этого толщина слоя обеднения обратно смещенного перехода (i-n или i-p) возрастает, слой обеднения распространяется в проводящую пленку и частично ее перекрывает, что приводит к уменьшению толщины проводящего канала пленки и, тем самым к изменению отраженного от пленки СВЧ-сигнала. Если величина СВЧ-сигнала прокалибрована заранее в единицах проводимости структуры, то можно получить непосредственно зависимость проводимости структуры от приложенного к n-i или i-p переходам напряжения V. Регистрируя эти изменения, получают величину и порог эффекта обратного управления без разрушения структуры (без стравливания n⁺-слоя, без вжигания оптических контактов и не касаясь n⁺ или n-слоев в центральной части структуры, идущей на изготовление ИС). При наличии "шунтирующих" n⁺-слоев изменения СВЧ-проводимости лежат в интервале Δσ^свч/σ_о ^свч = 10^-4-10^-1.The essence of the method is illustrated by the following. Consider the structure intended for the manufacture of ICs, a structure of the n ⁺ -ni type, where n ⁺ is the contact, high-alloy layer; n is the active layer; i is a semi-insulating substrate. According to the prototype method, in order to apply an offset to the junction, ohmic contacts are formed to the i-substrate and the n ⁺ layer. In this method, a liquid contact transparent to microwave radiation is formed on a substrate. As a liquid, an electrolyte, methyl alcohol, isopropyl alcohol, non-deionized water can be used. All these liquids do not form an ohmic contact to the semi-insulating substrate, moreover, when applying bias V to the liquid / substrate transition (V> 5 V), it can be seen that this is a rectifying contact forming a Schottky layer in a semiconductor. However, at small displacements (V <5 V), a current flows through the liquid / substrate transition through the transition, and the resistance of this transition together with the resistance of a thin liquid layer is much lower than the resistance of the semi-insulating substrate, the resistance of which is ≈0.1 for the substrate area through which the current flows cm ² is

10 MΩ (R _mean = ρ · l / S = 10 ⁸ Ω˙cm˙0.03 cm / 0.1 cm ² = 30 MΩ). Moreover, in the electric circuit, the liquid - inn ⁺ - clamping metal probe, as it turned out, the highest resistance is not the substrate, but the ni junction, at which almost all the voltage drops. A small part of the voltage is released at the liquid / substrate transition, and the n ⁺ metal probe transition is displaced in the forward direction and becomes sufficiently conductive. Thus, the main voltage is released at the junction of the semi-insulating substrate / active film (ni or ip). As a result, the thickness of the depletion layer of the reverse biased transition (in or ip) increases, the depletion layer propagates into the conductive film and partially overlaps it, which leads to a decrease in the thickness of the conductive channel of the film and, consequently, to a change in the microwave signal reflected from the film. If the value of the microwave signal is calibrated in advance in the units of the structure’s conductivity, then we can directly obtain the dependence of the structure’s conductivity on the voltage V applied to the ni or ip transitions. By recording these changes, we obtain the magnitude and threshold of the reverse control effect without destroying the structure (without etching the n ⁺ layer without burning optical contacts and without touching the n ⁺ or n-layers in the central part of the structure, which is going to fabricate the IC). In the presence of "shunting" n ⁺ layers, changes in the microwave conductivity lie in the range Δσ ^microwave / σ _about ^microwave = 10 ^-4 -10 ^-1 .

When selecting standard structures of the same type, the change in microwave power caused by the application of voltage to the substrate, ΔW _microwave ^V and illumination of the structure by light from the side of the conducting layers, ΔW _microwave ^F. Since the measurements of these quantities are carried out at fixed signal amplification and liquid contact thickness, and the photoconductivity is almost the same for standard structures of the same type (since it is caused by the “collapse” of the substrate / film depletion layer), the value ΔW _microwave ^V / Δ W _microwave ^Ф is photoconductivity normalized value of the change in the conductivity of the structure, which characterizes the EEC in the structures. Measurement on 200 gallium arsenide structures of the n ⁺ -n-puff-i type, where n ⁺ , n are doped layers, puff. - a buffer layer, showed that the value of photoconductivity varies from structure to structure no more than 1.5 times, while the magnitude of the EEC varies more than 1000 times.

 The invention is illustrated by the following examples.

 Testing of the method was carried out on multilayer structures InP, GaAs, AlGaAs / GaAs, CdTe / GaAs.

Below is a description of the measurement procedure for two structures of type n ⁺ - n - buffer. - i. The structures contained an n ⁺ layer doped to a level of N _D ≈10 ¹⁹ cm ^-3 , an n-layer with N _D ≈2˙10 ¹⁷ cm ^-3 , a high-resistance buffer layer with a thickness of ≃ 1 μm and an i-semi-insulating substrate with a thickness of ≃ 300 μm . Each of the washers was cut into two parts. The first part was designed to measure the value of the EEC by the prototype method; therefore, ohmic contacts to the n ⁺ layer and i-insulating substrate were made on it by Ga: Au fusion. The upper n ⁺ layer between the contacts was etched, thereby obtaining the structure of a field effect transistor. In the table. Figures 1 and 2 show the results of changes in the conductivity of the structure σ caused by applying bias V to the substrate. The second part of the washer was measured by the proposed non-destructive method: the structure under study was placed in a holder that provided contact between the semi-insulating substrate and a thin layer of alcohol, in addition, there was a possibility in this area contact pass through the structure of microwave radiation. The conductivity of the structure was measured by the magnitude of the absorbed microwave structure. Measurements of the conductivity of the structure at microwave frequencies were performed on a setup designed to measure the conductivity of the structure (σ ^microwave ).

), легированный кремнием слой AlGaAs (d = 400

) и контактный n⁺-слой GaAs (d = 700

). Все из измеренных 10 структур были одинаково легированы и имели близкое поверхностное удельное сопротивление 300 Ом/□ , однако по величине ЭОУ они сильно различались. В табл. 3 и 4 приведены результаты измерений двух полярных структур, по предлагаемому способу и по способу, требующему изготовления беззатворного полевого транзистора (по способу-прототипу). Таким образом, изменения проводимости приведенных многослойных структур Δσ, установленные данным неразрушающим способом, хорошо согласуются с изменениями проводимости Δσ полевых транзисторов, изготовленных из данных структур. Величины Δσ при заданных V характеризуют величину ЭОУ, кроме того, зная Δσ, можно вычислить изменения тока стока Δ I_cтранзисторов с заданной геометрией при известном напряжении исток-сток V_ИС×(ΔI_c= a/lΔσ×V_ИС), где a; l - ширина и длина канала транзистора. Тем самым можно установить, насколько сильно будет подвержен ток транзистора, изготовленного на данной структуре, напряжению на подложке, т. е. установить величину ЭОУ.In addition, especially for verification, ^microwave measurements of σ were also performed on a setup combined with a holder for the structure, which made it possible to provide liquid contact to the substrate and pressure metal contact to the conductive layer of the structure. ^Microwave σ measurements were close to each other. All conductivity values are given in the table in units of mS˙. σ ^pt , Δσ ^pt - conductivity and conductivity changes measured in the structure of a gate-free field-effect transistor controlled from the substrate by voltage V. σ ^microwave , Δσ ^microwave - microwave conductivity of the structure and its changes caused by the application of voltage V to the substrate relative to the conductive layers; ΔW _microwave ^V , ΔW _microwave ^F - changes in the microwave power transmitted through the structure caused by the application of voltage V to the substrate and the illumination of the structure with light h ω> 1.9 eV and flux ≈3˙10 ^-5 W / cm ² . Testing of the method was also performed on heterostructures with modulated alloying containing two-dimensional electron gas (DEG). We measured standard structures intended for the production of HEMT transistors and containing a GaAs buffer layer (d = 0.6 μm) grown by molecular beam epitaxy and an AlGaAs spacer (d = 25

) silicon-doped AlGaAs layer (d = 400

) and the contact n ⁺ GaAs layer (d = 700

) All of the 10 measured structures were equally doped and had a close surface resistivity of 300 Ohm / □, however, they differed greatly in the value of the EOS. In the table. 3 and 4 show the results of measurements of two polar structures, according to the proposed method and the method requiring the manufacture of a gateless field effect transistor (according to the prototype method). Thus, the changes in the conductivity of the reduced multilayer structures Δσ established by this non-destructive method are in good agreement with the changes in the conductivity Δσ of field effect transistors made from these structures. The values of Δσ for given V characterize the value of the EEC, in addition, knowing Δσ, we can calculate the changes in the drain current Δ I _{c of} transistors with a given geometry at a known source-drain voltage V _IC × (ΔI _c = a / lΔσ × V _IC ), where a ; l is the width and length of the transistor channel. Thus, it can be established how much the current of the transistor fabricated on this structure will be subject to voltage on the substrate, i.e., to establish the value of the EEC.

A comparison of the measurement results Δσ and the magnitude ΔW _microwave ^V / Δ W _microwave ^Φ allows us to state that the value Δ W _microwave ^V / ΔW _microwave ^Φ is a relative characteristic of the magnitude of the EEC standard multilayer structures. The measurements of the conduction relaxation time constant when applying bias to the substrate of a multilayer structure (n ⁺ -ni) and a transistor made of the structure testified to the identity of the relaxation time constants (τ = 1 s), which means that it is possible to measure the kinetics of the EOE in a non-destructive way.

 Compared with the prototype, the proposed technical solution allows for non-destructive express control of planar homo- and heterostructures designed for the production of IP. (56) 1. Khvelidze L.V., Khuchul N.P. Effect of control over a substrate in active elements of a GaAs-based IC. - Foreign electronic technology. 1987, N 9, p. 69-98.

 2. I. Itoh and all Stability of Performance and Interfacial Problems in GaAs MESFEI. IEEE Trans Electr. Dev. v. ED-27, N 6, 1980, pp. 1037-1044.

Claims (1)

METHOD rejection SEMICONDUCTOR STRUCTURES semi-insulating substrate by the degree of manifestation of the reverse control, comprising forming contacts to the substrate and the active layer structure, the application thereto varying in magnitude V ₀ voltage, detecting the active layer of conductivity depending on the applied voltage, assessment of symptoms reverse effect control, rejection of the structure, characterized in that, in order to implement the possibility of non-destructive rejection and increase expressivity Track to the substrate is carried by a transparent electrolyte microwave radiation, and contact to the active layer operate nip, affects the structure of the microwave radiation is measured reflected or absorbed by the structure of the radiation power, the change in conductivity of the active layer is determined by the change of microwave power ΔW

then, in the absence of voltage, the structure is illuminated with light with a quantum energy greater than the semiconductor band gap and with a stream at which the conductivity change reaches saturation, the measurement of the reflected or absorbed microwave power structure is repeated and the conductivity change is additionally determined by the change in the microwave power under the influence of lighting ΔW

, and the degree of manifestation of the reverse control effect is estimated with respect to ΔW

/ ΔW

.

Images