DE2521909A1 - MEASURING METHOD AND MEASURING ARRANGEMENT - Google Patents

Description

The invention relates to a measuring arrangement and a method in particular for determining a property of a semiconductor, in which a measurable parameter that this property describes that varies with the depth of the semiconductor.

The measuring arrangement uses an electrolyte as "Schottky" ■ Rectifier or blocking contact on a semiconductor in order to obtain a profile of the (charge ^ carrier density in the semiconductor, and in any depth. This is done by means of an anodic dissolution reaction between the semiconductor and the electrolyte carried out in such a way that the semiconductor is resolved in a controlled manner, while simultaneously (or successively) Capacitance measurements can be carried out on the junction of the semiconductor which is polarized in the reverse direction. The capacity is related to the charge carrier diameter in a simple manner via a known equation (according to Schottky, 19 ^ 2),

41- (85

509849/0318

so that with the help of a simple analog electronic circuit the charge carrier density can be recorded over the depth.

Briefly, the measuring arrangement works as follows: For an n-type semiconductor, the anodic dissolution takes place with a fixed potential (which is determined by a "potentiostatic", d. H. the voltage constant controller is adjustable), at a rate that is determined by the Availability of minority carriers (holes) is determined, which are generated by irradiating the material. The anodic potential corresponds to a band curvature in the order of magnitude of 1 V in the depletion direction, whereby a simultaneous capacitance measurement by means of a phase-sensitive Procedure is possible. The integral of the dissolution current, together with a correction, gives the The width of the barrier layer corresponds to the depth of the profile representation. A control or regulating unit terminates the dissolution after a predetermined amount of material has been removed, or initiates a new run of a pen. The measuring arrangement can therefore be complete during the measurement left to itself after clamping the specimen and making initial adjustments.

In the case of a p-type semiconductor, small applied Large anodic currents are also available without irradiation. The resolution can either be constant Potential, or with a regulated constant current. To have a blocking state for the capacitance measurements set, a periodic switch to cathodic operation, i. H. to a predetermined one Potential made, with the restriction that only extremely small charge shifts in the cathodic direction are permitted are.

SÖ9Ö49 / Ö318

An automatic differentiation between p- or n-type material can be made by detecting the current or Capacitance values take place in anodic operation; this can be used to set the appropriate conditions for the Measurement of the junction capacitance can be used.

One advantage of the electrochemical profile measuring arrangement is that the depth is not limited. This distinguishes the arrangement from arrangements that have already been developed, in which an increasing reverse voltage is applied to a Schottky edge layer
must in order to extend the barrier layer width and thus the measurement depth as far as possible into the semiconductor. With these arrangements a limit is reached above which no further measurement is possible if a breakdown occurs in the reverse direction. In the case of gallium arsenide, for example, this limit is 3 μm or 0.1 μm,

16 18 / - "5 /

when the carrier density is 10 and 10 cm ^, respectively. For testing in greater depth must be in separate chemical Etching steps remove material so that the process is no longer continuous. The automatic electrochemical The measuring arrangement, on the other hand, carries out the etching and the measurement precisely and simultaneously (with the Schottky boundary layer is always biased well below the critical breakdown voltage), so a continuous Profile can be generated without depth restriction.

An additional property is given by the application for profile production or profiling of transition structures (p-n, n-p, p-n-p and n-p-n), where the material removal is an essential requirement, and for which the arrangements already developed are particularly unsuitable.

5098 ^ 9/0318

The electrochemical arrangement can also be used to distinguish between different semiconductors (e.g. in Profiling of heterostructures) can be used by the course of the band curvature potential, which is derived from the Capacitance measurement is derived, is shown over the depth. More accurate measurements of the band gap of Materials can be obtained by analyzing the photocurrent properties of the semiconductor-electrolyte interface as a function of the irradiation wavelength.

The principles have been applied to n- and ρ-GaAs and GaP with a 10 # (10 % weight by volume) KOH electrolyte, and there is no reason to believe that the method is not applicable to other materials is.

The measuring method and the measuring arrangement of the invention are particularly suitable for determining donor density profiles in semiconductor epitaxial layer substrate structures (n / n ⁺ ) for IMPATT and other microwave diodes, which has been greatly simplified with the aid of the electrochemical method according to the invention. In contrast to the processes already considered, the invention produces a continuous depth profile which encompasses the entire separation surface or interface area and the substrate itself at any depth. The invention thus provides information that is required not only for checking for specific circuit requirements, but also for optimizing layer growth processes.

The invention, like other processes already developed, is based on the formation of a Schottky edge layer

SÖ9849 / 0318

on the measured surface. A disadvantage of the methods already developed is that the maximum Depth below the surface, from which information z. B. by analyzing the relationships between capacity and Voltage or current and voltage can be derived is strictly limited. This limitation arises because the Measurements in the junction area of the semiconductor can be carried out in blocking mode, with the expansion this area is determined by the maximum potential that can be applied without a breakthrough. Since the Active area thickness of an IMPATT diode is necessarily greater than the junction width at breakdown, can provide an epitaxial layer structure for such a circuit arrangement without removing the surface material not be profiled as deep as the interface area. Because the maximum barrier width also decreases with increasing doping density, the profiling would go through the interface into the Substrate requiring many steps using the already developed one step-by-step process Material is removed by etching.

It is therefore the object of the invention to enable a continuous depth profile measurement, which is already the case developed method with vapor-deposited metal points or with a mercury probe is not possible.

The electrochemical process according to the invention is used as a barrier or edge layer or barrier layer material a concentrated electrolyte, which is also a means for the controlled dissolution of a surface of the Semiconductor forms such that a continuous depth profile is created. The depth profile characteristic can

-D-

can be determined by capacitance-voltage measurements on solid GaAs of the η-type with KOH as the electrolyte. In a method for electrochemical dissolution, the sample surface is irradiated through the electrolyte according to the invention, while the Schottky edge layer potential is kept at a sufficiently low value to avoid a breakdown. The irradiation produces minority carriers which are essential for the dissolution reaction, so that the dissolution rate is directly dependent on the irradiance. In order to obtain donor density _v * profiles in epitaxial layer structures, an arrangement was developed in which capacitance measurements are carried out continuously while the epitaxial layer is dissolved electrochemically.

The invention will now be explained in more detail with reference to the drawing. Show it:

1 shows the block diagram of an automatic measuring arrangement according to the invention;

2 graphically shows the relationship between the net donor density per cm and the depth of a profiled surface in , van for a GaAs sample;

3 shows a curve with non-uniform semiconductor doping at relatively high doping densities for a sample in an epitaxial growth device was established in the event of an error in the controller;

FIG. 4 shows an electrochemical cell of the arrangement according to FIG. 1;

5a graphically shows the relationship between the electrostatic potential and the depth of a sample for an irradiated sample according to FIG. 5b;

5Ö9843 / 0318

Pig. 6 shows a more detailed block diagram of the measuring arrangement according to FIGS

Fig. 7 is a block diagram for an improved measuring arrangement according to the invention.

1 shows an automatic measuring arrangement for the determination of donor density profiles in samples of semiconductor-epitaxial layer-substrate structures which are used, for example, in IMPATT and other microwave diodes. The sample forms part of an electrochemical Cell (also called element or chain), which is described in more detail below. The cell is called Part of the feedback path in a "potentiostatic" Circuit used to set a constant resolution potential that is independent of the irradiance allowed. A filtered 250 V quartz halogen lamp is used for irradiation used, wherein the wavelength range is below the absorption edge of the semiconductor. Of the Direct current through the cell indicates the rate of dissolution and a current integrator generates a signal according to Laraday's law of electrolysis proportional to the thickness of the material removed. The capacity of the semiconductor barrier layer at constant surface layer potential is determined by feeding in a 3 kHz signal with 50 mV rms value and measuring the Imaginary part of the alternating current in which a phase-sensitive Detector is used in a known manner. The detector signal is converted into a signal that is proportional to the logarithm of the donor density, and to the Y input of an X-Y recording device (hereinafter referred to as recorder called).

609849/0318

The simple Schottky relationship between the capacitance C and the net donor density (Np-N _A ) is given by the following equation:

Λα "

The Barrier and potential barrier height which is previously determined by a capacitance-voltage calibration at the chosen dissolution conditions for GaAs is typically 1.4 eV, and the area A is about 5 χ 1O ~ ² cm ^2.

The capacitance signal is used to determine the barrier layer · width W by means of a divider module, since the following applies: W = i

This gives a correction value that corresponds to the reduced correct value added by resolution / thickness

so that the effective depth is determined. This effective depth is shown on the X-axis of the X-Y axis applied.

No special preparation of the sample is required, since the area to be examined is defined by a sealing ring on the cell; Contacts to Back of the sample are simple hot-dip tinned probes that are easily secured with the help of a voltage pulse can be. After the start of the dissolution, the arrangement can work independently until a predetermined amount of the Semiconductor material is removed.

509849/0318

An example of a profile obtained for a Doppe1 epitaxial GaAs structure (n ~ / ⁿ / h ⁺ ) is shown by the continuous curve of FIG. It should be particularly emphasized that this profile could only be obtained extremely time-consuming and cumbersome with a conventional arrangement and step-by-step etching, taking into account that the maximum barrier widths before

■ I TJ ι Q -τ

the breakthrough at donor densities of 10 'and 10 cm are about 0.4 and 0.1 / µm, respectively. ^The continuous measurement

in the invention is particularly important when at relatively high doping densities an uneven doping is present, compare this as an example the curve shown in FIG. This curve represents a Illustrates the situation in which a control error occurred in the epitaxial growth device; the effects of this Errors can be seen from the exact information contained in the profile.

Two methods for characterizing semiconductor materials have already been described: The authors Baxandall, Colliver and Fray wrote the article "A device for the rapid determination of semiconductor impurity profiles" (J. Phys. E: Sei. Instrum. 1971. 4, p. 213-221), the author Copeland describes "A method for directly recording the inverse doping profile of semiconductor wafers" (IEEE Trans. Electron Devices I969 ED-I6 , pp. 445-449). The methods mentioned are particularly suitable for examining lightly doped layers with a limited thickness. In contrast, the method according to the invention is particularly suitable for samples with high doping densities, so that the method according to the invention is a supplement to these methods mentioned.

SÖ9849 / 0318

In the Copeland process, one is on the surface attached Schottky diode fed with a small constant high frequency current (some 100 / ttA at 5 MHz). The depth of the junction is changed by changing the DC voltage, but this is the only one Task of this DC voltage. The inverse doping profile n "(x) is obtained by monitoring the voltage across the Diode received at the fundamental frequency, which is proportional to the depth χ, and at the second harmonic, the is proportional to n ~.

In the arrangement according to Baxandall, Colliver and Pray there is a matrix of Schottky barrier diodes on the surface with the help of a small probe contact is made with each diode in turn. the The probe or the probe head is fed with a direct voltage and a small 100 kHz voltage, the 100 kHz current flowing in the diode is a measure of the small-signal capacitance C. Another small one Voltage of the frequency 1 kHz is superimposed and the resulting modulation depth of the 100 kHz current is a measure of dC / dV. The impurity density N and the depth are known functions of C and dG / dV, and the

Jv Jv

Order carries out a suitable analog calculation and results in a direct representation or recording of log N versus depth when the DC blocking voltage is changed.

The methods already developed allow the determination and automatic recording of the carrier density depending on the depth, but they are limited in terms of the maximum barrier width that is immediately before the barrier breakthrough of the transition is reached. In order to penetrate deeper into the material, these measurements must therefore with a series of separate chemical etching steps

S09849 / 0318

be bound. This process is very time consuming and is a potential source of error, with frequent Cracks in the profile occur if the etching steps are chosen too roughly. Since the maximum junction width is high doping densities is also very small, it is usually impossible to get a continuous profile through the interface between an epitaxial layer and a heavily doped substrate.

In contrast, the invention uses an electrochemical process that has a completely automatic, continuous profile production at any desired depth (including heavily doped areas) permitted, the usual metal Schottky boundary layer by a concentrated, liquid electrolyte is replaced, which is in contact with a precisely defined area of the semiconductor surface stands. This type of touch allows capacitance measurements as before and also sets Means for the anodic electrochemical dissolution of the semiconductor. The speed of the solution depends on on the availability of the minority carriers in the semiconductor, which is caused by irradiation by the electrolyte the semiconductor surface can be generated. Thus, continuous measurement is possible while the semiconductor is resolved in a controlled manner. The anodic dissolution current can be measured and integrated so that in each At the moment the total amount of material removed is known. A simple analog arrangement is permitted then the recording of the carrier density over the depth, without material limitation.

An arrangement based on the invention can be fully automated, so that after the assembly of a sample no further monitoring is necessary until a profile is made to the desired depth. In a reall *

609849/0318

based embodiment of this arrangement can Maximum or full sole depth gradually between 1.8 / um and 72 / um can be changed using a control unit is programmed in such a way that when the maximum value is reached either the arrangement is switched off, so that the dissolution ceases, or that the scribe automatically starts a new run. In the latter case, the depth range can of course be extended indefinitely will.

The arrangement can be used, with the aid of the profile, of a number of epitaxial layer substrate structures for n-type GaAs and -GaP, as well as structures with p-n semiconductor junctions to mark; the same applies to other materials.

Fig. 4 shows an electrolytic cell (or an electrolytic Element) with a flow or flux channel 1 between an electrolyte inlet tube 2 and one-exit pipe 5. The channel 1 is made in a pipe PTFE (polytetrafluoroethylene) made. An opening or orifice plate 4 in the channel 1 is to be checked Material sample 5 closed in such a way that one (front) surface of sample 5 is exposed to the electrolyte in channel 1. A seal against the electrolyte is arranged between the surface 6 and the channel 1, in the exemplary embodiment represented by a PVC mounting ring 7. Sample 5 is supported by gold alloy wire springs 8 and 9 held with respect to the mounting ring 7; the springs also make electrical contact with the sample 5 here. The spring contacts 8 and 9 are supported by a contact stamp or piston 10, which generates a sealing pressure between the surface 6 and the mounting ring 7.

509849/0318

Directly opposite the measuring diaphragm 4 is a light channel 11, over which the surface 6 of the sample 5 can be irradiated with a number of light sources. The mounting ring 7 is arranged in such a way that it very precisely defines the size of the area 6 that extends over the light channel is irradiated.

The working or working cathode of the electrochemical The cell is represented by a graphically pure carbon electrode 12 at one end of the channel 1. At the other At the end of the channel 1 is a reference electrode IJ. These Reference electrode IjJ is a saturated calomel electrode with porous closure.

A measuring electrode 14 runs past the Möntagering 7 through the measuring orifice 4 and protrudes into channel 1. This measuring electrode consists of platinum wire and is used to measure capacitance used to make the influence of the series resistance of the electrolyte negligibly small.

The cathode 12 and the electrodes 1] 5 and 14 are in contact with the electrolyte in the channel 1. The electrolyte is a 10 ^ aqueous KOH solution that consists of a "Analyze grade" reactant is prepared with minimal exposure to air. The solution will be fed into the cell and pumped through it under pressure of "white spot" nitrogen minimize oxygen and carbon dioxide pollution. In operation, the flow rate is of the electrolyte 0.02 ml s ", with previously" white spot "nitrogen through the (not shown) Reservoir is blown through to remove contaminants from the electrolyte. It is very advantageous that

509849/0318

material is continuously removed from surface 6 of sample 5 by the action of the electrolyte in channel 1 will. Furthermore, the sample 5 is anodically polarized, so that upon irradiation of the surface β over the light channel 11 an electrostatic potential is built up. The electrostatic potential changes varies with the depth of the sample, which, as has already been said, is continuous due to the electrolytic action will be changed.

The relationship between the electrostatic potential and the depth for an irradiated sample is shown in FIG shown. As the theoretical model according to FIG. 5b shows, light is emitted via the electrolyte 15 onto the surface fed into the sample. The area β represents the boundary of a barrier layer 16. A quas! ^ Neutral diffusion area 16a separates the junction region 10 from the massive one Semiconductors 17.

Myamlin and Pleskov have in "their book entitled "Semiconductor Electrochemistry" (Plenum Press, I968) shown, how the model of Fig. 5b can be analyzed to obtain the To obtain relationship between the electrostatic potential and the depth of Fig. 5a.

The mode of operation of the described measuring arrangement can easily be explained with reference to FIG. In this Fig. 6 the electrolytic cell of FIG. 4 is denoted by the reference numeral 18. The semiconductor sample 5 is up of the cell 18 and has two point contacts 19a and 19b, (the cell 18 preferably has the one shown in FIG. 4, of course structure shown, where the point contacts 19a and 19b z. B. represented by springs 8 and 9 according to FIG. 4). A

509849/0318

"Potentiostat" generally drawn in dashed lines 21 consists of a differential amplifier 22, the output of which feeds a current source 23. The output signal from the power source is connected to the carbon working electrode 12, which is in contact with the electrolyte in the Cell 18 is standing. One input of the amplifier 22 is connected directly to the saturated calomel electrode 13, and the other input of the amplifier 22 is connected to a ramp generator 24, which is also equipped with a high impedance Voltmeter 25 and connected to the contact 19a is. The voltmeter 25 is connected to the Y input of an X-Y recorder 27 via a unit 26. The X input of the recorder 27 is provided by a unit 29 and a logarithmic micro-ammeter fed, which is between the back contact 19b and a ground line 30.

The potentiostat 21 is used to regulate the current through the sample 5 such that the potential between of the sample and the reference electrode I3 as well as possible corresponds to the potential of a reference voltage source. The generator 24 generates the reference voltage which increases linearly over time in the desired area. To prevent the influence of the ohmic voltage drop at the sample contact, contacts 19a and 19b represent separate current and voltage contact points, and the cell potential becomes independent of the grounded ramp generator 24 with the help a high-resistance voltmeter 25 measured. The logarithmic Micro-ammeter 28 in the circuit arrangement is used for the direct, semi-logarithmic representation of voltage-current curves on the X-Y recorder 27. A digital one

509849/0318

Integrator measures the time integral of the resolution current during the anodic removal of the epitaxial layers.

The measuring arrangement can generate pulse signals that automatically record the voltage-current curves in cause predetermined depth intervals of the removed material.

This experimental setup was successfully used for profile design used by GaAs and expitaxial GaP structures at depths up to 50 / µm. The measuring arrangement measures an effective mean value of the donor density over the entire barrier width. Since the barrier width of depends on the doping density and on the height of the potential wall, the depth of resolution is at the lowest recorded donor densities are limited. This is illustrated in Fig. 2, where an area of uncertainty is represented by the difference between the continuous course (with a depth scale correction equal to the total junction width is) and the dashed line (without correction) is shown.

Another effect that is important at low donor densities is the influence of the capacity of deep trap levels that may be present. Some of this may be ionized in the near surface area of the barrier layer so that the surface donor density is overestimated. For example, the donor density of the active layer of the sample according to FIG. 2 was given by the manufacturer as 7 χ 10 ^J cm.

When using an electrolyte to form the Schottky boundary layer medium, none of these restrictions apply. Accordingly, a capacitance analysis with modulation methods was developed in order to improve depth accuracy

6098 4 9/0318

to generate, combined with the additional option of displaying the surface and the Depth donor densities. The improved measuring arrangement is shown in FIG.

FIG. 7 shows the electrochemical cell according to FIG. 4 labeled 31. (Of course, the cell and the other components of FIG. 7 are not similar with the actual structure). The cell 31 contains the sample 32, an alternating current cathode 33, a direct current cathode 3 ^ and a saturated calomel electrode 35 The back of the sample 32 has probe contacts, which are represented by a contactor 36, which is conventional in Way springs like the springs 8 and 9 according to FIG.

As in the arrangement of FIG. 6, a potentiostat 37 is used, which has a single integrated operational amplifier, to provide a constant anodic dissolution potential to adjust. The desired potential is set in a built-in voltage reference source, and the potentiostat 37 regulates the cell current so that the same potential is applied to the saturated calomel reference electrode 35 occurs. The settling time of the potentiostat is greater than 100 ms, so that an external modulation of the cell potential for differential described below Capacitance measurements is possible. (The value of the anodic potential must be in a safety range in which no premature breakthrough occurs, so that preferential resolution in areas with material defects is avoided will). For GaAs, a value of -0.3 V with respect to the calomel electrode 35 is chosen, which corresponds to an overall band curvature or curvature of about 1.5 V minus one effective

609849/0318

fciven reverse voltage of 0.5 V.

So that the rest potential of the sample can be checked, the potentiostat contains 37 switch-controlled relays, via which the cathode can be switched off, so that the open-circuit voltage can be checked. Initially, a connection is made to the contactor 36 and the relays are brought into their rest position, in which the cathode is disconnected. The resolution is then started in which all connections between cell 31 and the potentiostat 37, and in ^ by removing a flap of the radiation source (not shown in FIG. 7). It is noteworthy that since only η-type material is radiation sensitive Has rest potential, the test of this rest potential in each case the determination of the Conductivity type of the sample allowed before its dissolution. It is recommended that the relays have an arrangement be controlled, which has protection against an interruption in the power supply.

A 3 kHz oscillator 38 feeds a signal of 50 mV RMS value in the sample 32. A variable resistor 39 is in series with the sample 32 and creates a current through the sample. For sensitivity calibration the resistor 39 is preset and has a resistance value (a few tens of ohms) compared to the lowest sample impedance at 3 kHz is negligible. To determine the junction capacitance, a 3 kHz phase-sensitive amplifier 40 the signal at resistor 39.

Expediently, fully encapsulated modules are used for both the oscillator and the amplifier 40.

609849/0318

applies, with an additional simple phase adjustment circuit with an integrated operational amplifier is available.

The phase and sensitivity of the phase sensitive amplifier 40 are adjusted with a standard test capacitor of size 10 riF to produce an output signal representative of the capacitance per unit area C / k of the sample 32. To do this, it is necessary to know the area A, which is irradiated and thus resolved during the analysis of the sample. By using a moving or moving microscope, it was found that this area can be reproduced very precisely in the measuring arrangement.

The area A can be surrounded by a ring of "excess" area A _E which is exposed to the electrolyte but not to the radiation. By carefully constructing the holder for the sample 32 (for example by using a mounting ring 7 according to FIG. 4), this excess area can be reduced and possibly made negligibly small. In any case, it can be assumed that before the dissolution begins, the "excess" capacitance attributable to the excess area is related to the measured total capacitance in the same way as the excess area A_ is related to the total area, provided that the doping is uniform in the transverse direction. In order to compensate for the influence of the excess area A ", a compensation device 41 is provided with which the output signal 42 of the phase-sensitive amplifier 40 can be shifted by a suitable fraction of the initial output signal, which is measured on the urine-irradiated sample before dissolution can..

509849/0318

The width of the barrier layer on the sample 32 is modulated by means of a 30 Hz Wieribrücken oscillator 43, which is located in the feedback path of the potentiostat 37. The Wien bridge oscillator 43 generates an output signal dV which is attenuated by the connection to the oscillator 38; it is set to 100 mV rms value. Thus, the DC output 42 of the 3 kHz phase sensitive amplifier 40 contains a 30 Hz component. Just like the component corresponding to the size C / A, this component dC / A must also be shifted in a similar manner in order to take into account a contribution from the excess surface area. This is brought about by adding a 30 Hz offset signal with a phase difference of 180 °, which is derived from the Wien bridge oscillator 43 via a phase shifter. The setting is carried out with the output signal of the compensation device 43, which is temporarily connected to a 30 Hz phase-sensitive amplifier 44 to be described in more detail. To compensate for the variables representing the variables C / A and dC / A, they are monitored on a four-digit digital voltmeter (not shown). The compensated signals 45 representing the quantities C / A and dC / A (the latter quantity remaining as a 30 Hz component) are fed into a plate 46. The divider 46 is an encapsulated analog divider and generates an output signal 47 which represents the junction width W (with a superimposed component dW _D ) A

509049/0318

The divider 46 is of course calibrated by a known one or a previously determined value for the relative dielectric constant £ of the semiconductor is used.

The output signal 47 is split in such a way that its component W _{0 is} fed into an adder to be described below, while its component dW _{D is processed} in an analog squarer 49. The squarer 49 consists of an encapsulated module, does not require any external adjustment and generates a 30 Hz signal that

is proportional to d (W _D ). This output signal 50 is converted into a DC voltage by the 30 Hz phase-sensitive amplifier 44. The amplifier 44 is to achieve a high sensitivity from an integrated operational amplifier with very low drift,

H.

low offset and a useful dynamic range of 10. This results in a carrier density measuring range of 10 (for example 10 ^ to 10 ° cm "* * ^) in the sample according to the following Equation:

As described below, the gain of the 30 Hz phase sensitive amplifier 44 can be adjusted by calibration can be set so that the output voltage 51

Represents £ / n. The output voltage 51 is converted into a logarithmic amplifier 52 fed, which is built into an encapsulated module. The logarithmic amplifier 52, taking into account the relative dielectric

S09849 / 0318

tricity constant of the sample calibrated in such a way that a Output voltage 53 is generated proportional to log η. This output voltage 53 is input to the Y-axis an X-Y recorder 54 is fed.

After describing the derivative of the differential Capacity in the measuring arrangement according to FIG. 7, the relationship with the depth will now be explained in more detail.

The anodic dissolution current is measured with a differential input ammeter 55, which is made up of four integrated operational amplifiers for this purpose and is _{calibrated to generate an output signal 56 which represents the dissolution rate dW R} / dt, ie the thickness of the material removed per unit of time. The calibration of the ammeter 55 *, which takes various parameters such as sample material and resolution area A into account, is carried out according to the equation

w _o «.

The output signal 56 of the ammeter 55 is fed into an integrator 57 which, in a known manner, generates an output signal 58 which represents the total thickness W- at each point in time, which has been removed by resolution. In the adder, this _{output signal 58 representing W R is} added to that component of the output signal 47 which represents the junction width W ₁ J, so that an output signal 59 is generated which represents the effective total depth χ. That

509349/0318

Output signal 59 is fed into the X input of the X-Y recorder 54 is fed so that the X-Y writer 54 represents the quantity log η over χ.

The input sensitivity of the integrator 57 1st Can be changed via a switch (not shown), so that different maximum depths ("full scale") are possible are. In the measuring arrangement mentioned, there are six possible values for the maximum depth from 1.8 / µm to 72 / µm. A control unit 6o indicates when the selected maximum depth has been reached and can adjust the arrangement to this Either switch off the point in time so that the dissolution is ended or another complete cycle by returning the pen to the beginning of the X-scale. In the latter case it will be natural the achievable total depth is extended indefinitely.

The output signal 58 of the integrator is expediently 57 changed step by step by an output stage fed by a stepper motor with a multi-turn potentiometer, so that long periods of integration are possible. These integration periods can last for several hours The integrator output signal switched to full scale about J56o times during a complete recording. at The pen of the X-Y recorder registers a single point for each step, with the pen registering one point via a control unit 61 by the integrator 57 controllable lowering device.

So that the depth of resolution W _R can be checked at any time during the recording, a (not shown)

509849 / G318

A switch is provided which removes the junction width correction Wp from the output signal 59.

The integrator 57 and the associated parts of the arrangement can be simply fed in with direct current test signals calibration, whereby a digital voltmeter is used to measure the output values. The calibration for measuring the barrier width parts used in the arrangement is also very simple, at least up to the exit of the divider 46, by simply putting on a fixed test capacitor and on DC measurements must be referred to. To calibrate the 30 Hz phase-sensitive amplifier 44, it is advisable to when accurate measurements of AC signal values (at 30 Hz) can be avoided. This can be done through a final setting phase gain of the amplifier ^ 4 is done by using an actual semiconductor device under test in below is used as described in more detail.

The test item or the sample should be uniform

be doped disc, with a doping density of about

17 -3
10 'cm, which is roughly in the middle of the logarithmic scale. The exact value does not have to be known in advance, since it is measured with the aid of the electrochemical measuring arrangement. First it is checked whether the dark current (ie without irradiation) is negligible with the usual resolution potential in order to ensure that there is no noticeable conductance component that would falsify the measurement. The 3 kHz phase-sensitive amplifier 40 has five switched capacitance measuring ranges and can therefore be used to measure the capacitance at a given reference potential. By increasing the reference potential linearly over time

509849/0318

over a range of about 0.5V and by absorbing this Potential above that of the 3> kHz phase-sensitive Amplifier 40 measured capacitance becomes a C-V representation obtained from which the carrier density can be determined by curve fitting. This procedure is used in the measuring arrangement used according to Fig. 6; it includes the adaptation of the curve obtained experimentally to a theoretical curve, those from the well-known Schottky boundary layer equation for the junction capacitance is derived.

The gain of the 30 Hz phase sensitive amplifier 44 is readjusted until the derived value of the carrier density η is reproduced by the X-Y recorder 54 will. At the same time, a phase correction takes place while the detector stage output signal of the amplifier 44 is on monitored by the oscilloscope. Of course, the above-mentioned excess area compensation must be carried out during this process applied to the signal values.

After the measuring arrangement has been calibrated in this way, the test specimen can be removed sufficiently wide to make the original one To confirm the assumption that the carrier density at least in the near surface area is uniform over the depth runs what for the successful application of the curve fitting procedure is required. If this is not the case, the calibration process must of course be carried out with another test item be repeated.

When the measuring arrangement is calibrated in this way, the analysis of n-type samples of the same semiconductor is extremely easy. The loading, the contact distribution and the area compensation corrections take only a few minutes,

509849/0318

and once a resolution has started, no further monitoring is required as the arrangement is automatic is controlled and either switches off or a new cycle begins when the maximum depth is reached. With GaAs dissolution rates of about 2 μm per hour were used, so that in the analysis of Thickness tests the measuring arrangement normally run over freight The material analysis over a similar depth range with the methods already developed by Copeland and Baxandall et al. Required an elaborate measuring process as well as a step-by-step chemical etching. These Procedures work with a test in depth by increasing the applied reverse voltage and increasing the barrier width is extended to the limit determined by the barrier breakdown, after which there is a new etching process necessary. The invention achieves a reliable and continuously running production, by using an electrolytic boundary layer medium for the first time. Although the principle of differential Capacitance analysis is the same as with Baxandall et al., The construction of the measuring arrangement differs somewhat, with it the full advantage of the electrochemical process indicated by the invention is used. The measuring arrangement in particular, has a large dynamic range in carrier density measurement and allows a single continuous one Representation over the entire density range that is expected in a structure of the sample without any Adjustment is required. This is essentially achieved by doing the analog operations on the 30 Hz signal rather than by signal detection at an earlier stage.

509849/0318

A clear advantage of the measuring arrangement according to FIG. 7 over the simpler measuring arrangement according to FIG. 6 is the better depth resolution or accuracy. For example In some cases, the simpler measuring arrangement can have wave profiles in the carrier density at an interface between the active and buffer layers of a sample are insufficiently accurate. The accuracy of representation the improved measuring arrangement according to Fig.7 is limited by the Debye length, which is smaller by a factor of 10 than that Is barrier widths that determine the resolution accuracy of the simpler measuring arrangement. This can in particular be important for sample testing in certain applications.

Other Aspects of Using the Invention are the type of irradiation and the irradiation area. One Uniform irradiation over the examined area is essential for a uniform dissolution rate. A multi-fiber light channel is currently being used, which gives a uniform irradiation sufficient for most purposes. Naturally occurs at the fiber ends a certain light scattering, which the resolution accuracy of sharp dividing surfaces with relatively large Limited depths. In this regard, some improvements can be achieved with microscopic optical methods.

Samples with an area of about 4 mm in diameter were examined. For practical reasons, this area cannot be reduced very much as there is a risk that the The ratio between the excess area and the resolved area becomes very large, so that errors in compensation

509 8 49/0318

Decrease the recording accuracy, especially if there is a separation area between zones * of high and low density. In some cases it may be advantageous to use a contact mask, but this measure would reduce the measuring arrangement complicate.

With the flux used, the influence of the irradiation on the carrier density measurement is obviously negligible. The external photopotential is taken into account by the potentiostat, although a small increase in the measured capacitance (typically less than 10 %) remains at a given anodic potential when irradiating. As a result, a decrease in impedance appears in the area very close to the surface in which the light is absorbed. In uniformly doped areas, however, no change in the carrier density derived from the modulation of the barrier layer width can be measured. At very high radiation densities, however, changes were found if the light source was not filtered very precisely.

A quartz-iodine-tungsten light source is used to analyze Ga-As used, with a narrow band filter at 550 nm. For GaP this light source is through a Replaced blue filter mercury arc radiation source.

With minor modifications, the measuring arrangement and the measuring method of the invention is also suitable for the analysis of ρ-type material, so that also structures with ρ-η transitions can be analyzed.

509849/0318

Claims (18)

Claims ( \ .J measuring arrangement for measuring a predetermined semiconductor material parameter in such a way that a desired property of the semiconductor material can be determined, the value of the parameter in the semiconductor material being variable, marked by an electrolytic cell (51); a holder (7-10) for receiving the semiconductor material (5; 32) in cell (31) j a feed member (36) for feeding a measured electrical power into the semiconductor material (32) and into the cell (21); one
a mask for defining / surface area (6) a Schottky diode in the semiconductor material (32); an electrolyte for continuously etching the surface area (6) the Schottky diode in the semiconductor material (32); a capacitance measuring device (38-40) for measuring the capacitance of the semiconductor material (32) between the upper. surface area (6) and a rear side of the semiconductor material '; a display unit (54) for obtaining the measured Capacitance to the measured electrical power; and a light channel (11) for feeding light into the Schottky diode which is below its critical breakdown voltage is operated, whereby the surface area (6) is essentially continuously etched away by the electrolyte (FIGS. 4, 7). 609849/0318 2. Measuring arrangement according to claim 1, characterized by a potentiostat (37) for generating a predetermined one Etching potential, and by a current monitoring unit (55) to monitor the dissolution rate (Fig. 7). 3. Measuring arrangement according to claim 2, characterized by an integrator (57) for integrating the monitored current and for determining the thickness of the semiconductor (32) etched off on the surface area (6) (FIGS. 4, 7). 4. Measuring arrangement according to claim 3 »characterized by a X-Y recorder (54) connected to the capacitance measuring device (38-40) and is connected to the integrator (57) for generating a continuous graphic Representation of the relationship between the Net to Donatord ight on the surface of the Schottky diode and the thickness of the semiconductor etched off on the surface area (6) (Fig. 4,7). 5. Measuring arrangement according to claim 4, characterized by a sensing and re-actuation unit (60) for detecting the Maximum range and to operate the X-Y recorder again es (54) (Flg. 7). 6. Electrochemical line for determining donor density profiles in semiconductor samples with epitaxial structures, gekennzeichn e t by a container (1) for an electrolytic fluid; a first electrode (12) which protrudes into the container (1) in a first region (2); a second electrode (13) in a second area (3) of the container (1), the second area (3) from the first Area (2) is spatially separated; the semiconductor sample holder (7-10) in a third Area (4) of the container (1) between the first area (2) 509849/0318 and the second region (3) for defining a surface area (6) of the semiconductor in contact with the electrolytic fluid; and a light guide (11) for defining a path over transmit the light for irradiating at least part of the surface area (6) of the semiconductor (5) becomes (Fig. 4). 7. Cell according to claim 6, characterized in that the third region (4) in the container (1) between a Entry gate and an exit gate for the electrolyte Fluid is arranged such that the fluid on the surface area (6) of the semiconductor (5) flows past and the etched material is removed (Fig. 4). 8. Cell according to claim 7, characterized by an auxiliary electrode (14) in the third area (4) of the container (1) (Fig. 4). 9. Cell according to claim 8, characterized by the first electrode (12) made of carbon, the second electrode (13) with a saturated calomel solution, and the auxiliary electrode (14) made of platinum (Fig. 4). 10. Cell according to claim 6, characterized by the holder (7-10) with a plurality of electrical probe connections (8, 9) for contacting the half-lex sample (5) (Fig. 4). 11. Cell according to claim 10, characterized by the holder (7-10) with a sealing ring (7) made of plastic material to define the surface area (6) (Fig. 4). 509849/0318 12. Measurement method for measuring a predetermined semiconductor material parameter such that determines the change in a desired property of the material with depth where the value of the parameter changes with the material depth, marked by anodic etching on the material surface using a concentrated electrolyte solution by applying a PotentlaIsj Establish one below the critical breakdown voltage operated Schottky blocking contact through the Electrolytes on the material surface; and Measuring the semiconductor material parameter on the material surface, which is essentially continuously etched away. 13. A measuring method for measuring the parameter of a p-type material according to claim 12, characterized in that the applied potential to etch the semiconductor material the polarity is reversed briefly and periodically, during which a capacitance measurement is carried out. Ik. A measuring method for measuring the parameter of an η-type material according to claim 12, characterized in that the surface is removed in order to generate inority carriers on the surface and to enable etching. 15 «measuring method according to claim 12, characterized in that that the applied potential is modulated and the capacitance is measured differentially. 16. Measuring method according to claim 12, characterized in that that the potential of the applied electrical direct current 509849 / 03T8 power is kept essentially constant. 17. Measuring method according to claim 12, characterized in that the sample of the semiconductor material ρ- and n-type semiconductors and that the type of material on the etched surface is detected. 18. Measurement method for η-type material according to claim 14, characterized in that light niit predetermined! Frequency range is used. 509849/0318