JP7627071B2 - Mobility measurement device - Google Patents

Description

This relates to a method and measuring device for measuring the sheet resistance of a conductive substrate such as a semiconductor substrate or a conductive layer formed on the surface, the mobility of carriers such as electrons and holes in the conductive layer, and the carrier concentration in the conductive layer.

GaAs semiconductors and GaN semiconductors with AlGaN on the surface layer are fabricated on substrates on which an electrically conductive layer is formed by crystal growth using MOCVD or ion implantation. For this reason, it is necessary to measure the sheet resistance and carrier mobility of the electrically conductive layer of the substrate before fabricating the device. However, when a metal needle is directly contacted with the semiconductor surface of a GaAs semiconductor or a GaN semiconductor with AlGaN on the surface layer, a Schottky junction is formed between the metal needle and the semiconductor, making it impossible to conduct electricity. For this reason, metal is patterned by lithography and alloyed by high-temperature annealing to form an ohmic electrical contact that can conduct electricity. This is called the ohmic formation process, and it takes a lot of time to create a measurement sample. Such ohmic electrode creation is essential for evaluating electrical properties such as sheet resistance measurement, Hall measurement, and CV measurement, which are common for measuring the electrical properties of semiconductors.

The four-terminal probe method is used to measure sheet resistance. This method involves contacting four metal needles arranged in a straight line with a semiconductor, connecting a current source to the two outer metal needles, and measuring the voltage generated between the two inner metal needles, thereby measuring the sheet resistance. However, the four-terminal probe method cannot be applied to semiconductors such as GaAs and GaN, where the metal needles and the semiconductor form a junction that forms a Schottky diode, because it is not possible to pass a current through them.

As an alternative method, a device has been developed that measures sheet resistance by measuring the eddy currents generated in an electrically conductive layer by applying an AC magnetic field as impedance. However, when a conductive substrate such as silicon is used, the composite sheet resistance is measured, making accurate measurements impossible. In addition, commercially available sheet resistance measuring devices that use non-contact eddy currents have the problem that they cannot measure over a wide range because eddy currents are small on high-resistance substrates.

Furthermore, when measuring the Hall effect to measure carrier mobility, it is necessary to form an ohmic electrical contact. For this purpose, it is common to partially attach In solder to the sample surface and heat it to form an ohmic contact. This process is often performed by cutting out the sample, and is time-consuming.

In addition, to measure the carrier concentration, it is common to perform CV measurements (capacitance-voltage measurements). This involves applying a reverse voltage to a Schottky diode of a fixed area and measuring the capacitance value. Naturally, one side of the Schottky diode requires an ohmic electrical contact. For this reason, it is necessary to form an electrode pattern on the semiconductor surface that is separate from the Schottky electrode and create an ohmic contact by heat treatment.

JP 2014-29946 A (P2014-29946A) (Publication No.)

Catalog No.: B10-4500 Narica Co., Ltd. Small Piezoelectric Element A

A method is known in which electrical contact is obtained by forcibly passing electricity through a metal needle in contact with the semiconductor surface, and this method is known to be effective when a thin oxide film or dirt is attached to the semiconductor surface. However, in wide band gap semiconductors such as GaN semiconductors, the Schottky diode formed at the tip of the metal needle has an extremely high withstand voltage, so the Schottky junction cannot be destroyed by applying a voltage of approximately 100 V or less. If a Cockcroft-Waltron circuit or pulse transformer, known as a high-voltage generation method, is used, there is no limit to the generated voltage, but it is expensive and dangerous to handle because it uses high-voltage components. Furthermore, even if it is possible to destroy the Schottky junction, there is a problem that the measurement sample itself may be destroyed or altered.

In addition, if the current passing through is too large, the tip of the metal needle will melt. A method is needed to prevent melting of the metal needle by limiting the current passing through the metal needle and to prevent denaturation of the measurement sample.

Means for solving the problem

The means for forming an electrical contact according to the present invention is to apply a high voltage to a metal needle in contact with a semiconductor or the like using a piezoelectric element. Piezoelectric elements are widely used as ignition devices. This principle involves applying an external stress or striking the piezoelectric element to generate a high-voltage impulse. Since the piezoelectric element does not use electronic circuits, it is compact and yet can generate a voltage of about several kV. Since the amount of charge released is determined by the amount of polarization of the piezoelectric body, the amount of current, which is the time integral of the polarization change, is extremely small even at high voltages. For this reason, a high-voltage generator using a piezoelectric element can destroy the Schottky junction barrier without melting the electrode needle or causing denaturation of the measurement sample. This method for creating an ohmic electrical contact is very safe during measurements.

The means for solving the problems in the present invention will be described with reference to FIG.
The piezoelectric element is a ferroelectric ceramic such as PZT. Such a piezoelectric element 1 is sandwiched between an upper electrode 3 and a base electrode 2, and a striking rod 4 fixed to a spring 7 strikes the upper electrode momentarily when a stopper 5 is released. As a result, a voltage is generated between the upper electrode and the base electrode. This is a theoretical structure, and various mechanical structures for piezoelectric elements have been devised and put into practical use, and such commercially available piezoelectric elements can be used in the present invention.

The output of the piezoelectric element is connected to a 2X4 matrix switch 8. A matrix switch is a relay array installed at the junctions (intersections) of vertical and horizontal wiring, and has the function of shorting and opening the junctions. This allows the high voltage generated by the piezoelectric element 1 to be selectively distributed to any of the terminals A1 to A4.
The two-way switch 9 has metal needles 13, 14, 15, and 16 as poles, and is used to switch between the high voltage side and the measurement side of the 4X4 matrix switch 10. When S11 and S24 of the 2X4 matrix switch 8 are shorted and all two-way switches 9 are flipped to the high voltage generation side (upper side in the figure), a high voltage is applied to the metal needles 13 and 16 when a strike is applied to the piezoelectric element.

When the four metal needles 13, 14, 15, and 16 are brought into contact with the undoped AlGaN semiconductor layer 17, Schottky diodes represented by D1, D2, D3, and D4 are formed between the metal needles 13 to 16. Since back-to-back connected Schottky diodes are formed between the metal needles 13 and 16, one of them is reverse biased and cannot conduct electricity. A high voltage generated by the piezoelectric body is applied to this Schottky diode, but if the metal needle 13 is at a positive voltage with respect to the metal needle 16, the Schottky diode D1 under the metal needle 13 is forward biased and the Schottky diode D4 under the metal needle 16 is reverse biased.

If the current exceeds the reverse breakdown voltage of the Schottky diode D4, the Schottky barrier of D4 will be destroyed, but D1 may not be destroyed. This is because there is a difference in the energy consumption, i.e. heat generation, generated by the two Schottky diodes. Since the energy consumption is the product of the current and the voltage, the Schottky diode D4 with reverse bias generates more heat than the Schottky diode D1 with forward bias. As a result, the Schottky barrier of D4 is destroyed and electrical contact can be made with D4. The polarity of the generated high voltage pulse is determined by the polarization direction of the piezoelectric body. Therefore, in order to obtain an ohmic electrical contact, it is necessary to apply a high voltage pulse with the polarity reversed. For this reason, S21 and S14 of the 2X4 matrix switch 8 are shorted. This allows electrical contact to be made with the Schottky barrier of D1 formed by the metal needle 13. By selecting the shorted point of the 2X4 matrix switch 8 using this procedure, electrical contact can be made with any metal needle.

Figure 2 shows actual measurements of the current and voltage characteristics between the electrode needles before and after the application of a high voltage pulse and when the polarity is switched for a sample in which an undoped GaN epitaxial layer 18 and an undoped AlGaN epitaxial layer 17 are grown on a silicon substrate 19. Good ohmic properties are obtained by applying a high voltage pulse. No denaturation of the semiconductor substrate surface occurs.

The present invention shortens the sample preparation process, which previously required ohmic formation by patterning electrodes on the measurement sample or alloying by annealing. With the present invention, electrical contacts can be formed on the wafer and electrical characteristics can be measured using the metal needles. This enables rapid measurement of semiconductor electrical characteristics using methods such as sheet resistance measurement using a four-terminal probe, Hall measurement, and CV measurement.

Again, as a method for forming an ohmic contact by applying a high voltage pulse, a piezoelectric element is used instead of a high voltage generating circuit, which allows the current to be made extremely small, and therefore the electrode needle does not melt. In particular, even a Schottky junction with an extremely high breakdown voltage formed when a metal needle is brought into contact with a wide band gap semiconductor such as GaN can be easily broken down to form an ohmic electrode.
As a result of the experiment, it was confirmed that it is also possible to measure the sheet resistance of the two-dimensional electron gas layer at the heterojunction interface of the AlGaN and GaN layers. Electrical contact can be made to the p-type GaN semiconductor layer in a similar manner, although the Schottky diode polarity is forward-to-forward connected.

Diagram of the high voltage generation mechanism using a piezoelectric material and the four-terminal probe measurement configuration The initial current-voltage characteristics when two tungsten needles are in contact with the epitaxial surface on which an AlGaN/GaN heterojunction is formed, and the current-voltage characteristics after applying a high voltage using a piezoelectric material. Diagram of the Hall effect measurement device using a probe with metal needles placed at the vertices of a quadrilateral This is a diagram showing the configuration of a measurement device that performs CV measurements by forming a Schottky electrode on a GaN epitaxial layer and then contacting three metal needles around the electrode. This is a diagram showing the configuration of a measurement device for performing CV measurements between a Schottky junction made of an ionic liquid obtained by contacting an ionic liquid held in a cylindrical container with a semiconductor surface, and an Ohmic junction formed by arranging three metal needles around the junction. Measurement platform configuration common to 4-terminal sheet resistance measurement, Hall effect measurement, and CV measurement A diagram showing the configuration of a measurement method for measuring the Hall effect using two orthogonal metal needle arrays, although the ohmic electrode formation and switches are the same as those in the configuration shown in FIG.

A first embodiment of the present invention will be described with reference to FIG. 1. In this example, a commercially available piezoelectric element (Narica small piezoelectric element B10-4500) was used (Non-Patent Document 1). The generated high voltage is supplied to two horizontal lines of a 2X4 matrix switch 8, and the vertical output lines are output to four output terminals A1 to A4 by combining shorting and opening relays provided at the intersections, and are connected to a two-way switch 9. In this configuration, Omron's G4W relays with a dielectric strength of 10 KV or more are used for contacts S11 to S14 and S21 to S24 of the matrix switch 8 that short-circuit and open the intersections.

When electricity is passed through the metal needle, all two-way switches 9 are flipped to the upper side in the figure. The generated high-voltage waveform becomes a damped oscillation due to the vibration of the piezoelectric material and the inductance of the wiring. Therefore, since the first pulse is most effective in forming an ohmic contact, by switching the connection of the piezoelectric element connected to the electrode needle, the two diode barriers formed by point contact can be broken down alternately, realizing bidirectional electrical contact. The amount of current that flows here is due to the polarization change of the piezoelectric material and is therefore extremely small, so applying this high voltage multiple times does not cause the metal needle to melt or the measurement sample to deteriorate.

After the ohmic formation of each metal needle is completed, the two-way switch 9 is flipped downward in the figure to connect the metal needles 13 to 16 to the current source 11 and voltmeter 12 via the 4×4 matrix switch 10 connected to B1 to B4.
The two-way switch 9 is composed of four single-pole double-throw (SPDT) switches (SPDT1 to SPDT4), and uses a metal needle as an axis (single pole) to simultaneously switch between the high-voltage terminals A1 to A4 and the measurement terminals B1 to B4. The SPDT switches in the two-way switch are knife switches with a dielectric strength of 10 KV or more. With this configuration, there is no risk of accidentally connecting a high-voltage source to a measurement device and damaging it.

Relays for short-circuiting and opening are installed at the intersections in the 4x4 matrix switch 10. By shorting W11 and W42, the current source 11 is connected to the metal needles 13 and 16, and by shorting W23 and W34, the voltmeter 12 is connected to the metal needles 14 and 15.
From this measurement voltage and the set current value, the sheet resistance can be measured according to the principle explained in the normal four-terminal probe method. Regarding the principle of sheet resistance measurement, the sheet resistance R _S is often calculated from the current I and the measurement voltage V using the following formula:
R _S = (4.5324) V/I (1)

If the voltmeter used has a high input impedance, it is not necessarily necessary to form electrical contact with electrodes 2 and 3. However, it goes without saying that the measurement accuracy can be improved by forming electrical contact with electrodes 2 and 3 as described above.

The electrode needle that is in contact with the GaN epitaxial layer 18 and AlGaN layer 17 formed on the silicon substrate 19 forms a Schottky diode with the AlGaN layer, making it impossible to measure sheet resistance. In particular, it was previously thought that the four-terminal measurement method could not be used to measure the sheet resistance of the two-dimensional electron gas generated at the interface between the AlGaN layer and the GaN layer, because the Schottky diode of the metal needle has a high withstand voltage. However, because a voltage higher than the breakdown voltage of the diode can be applied to a piezoelectric element, it is possible to form an ohmic electrical contact in situ, making it possible to use the four-terminal measurement method.

The second embodiment of the present invention will be described with reference to FIG. 3. This is intended to measure the carrier mobility of a sample using the Hall effect. The measurement sample is placed on the N pole of a neodymium magnet 40, and metal needles 33, 34, 35, and 36 arranged at the vertices of a regular square with sides of 1 mm are in contact with the sample. The other configurations are the same as those of the first embodiment. The measurement sample is brought into contact with the metal needles 33, 34, 35, and 36, and electrical contacts are formed in the same manner as in the first embodiment. The measurement sample example shown in the figure has an undoped GaN epitaxial layer 38 and an undoped AlGaN layer 37 deposited in sequence on a silicon (111) substrate 39.

The two-way switch 9 is tilted upward in the drawing, and the 2×4 matrix switch 8 is set in the following order:
1) S11 and S24 are shorted and a high voltage generated by striking the piezoelectric element is applied.
2) S21 and S14 are shorted, and a high voltage generated by striking the piezoelectric element is applied.
3) S12 and S23 are shorted and a high voltage generated by striking the piezoelectric element is applied.
4) S22 and S13 are shorted and a high voltage generated by striking the piezoelectric element is applied.
This allows electrical contact to be made between all of the metal needles 33, 34, 35, and 36 and the measurement sample.

For the 4×4 matrix switch 10, the Hall electromotive force generated in the following connection order is measured.
1) W11 and W42 are shorted and a current is passed between metal needles 34 and 35 by current source 31. W23 and W34 are shorted and the Hall electromotive force V36 generated between metal needles 33 and 36 is measured by voltmeter 32.
Furthermore, in order to remove the offset voltage caused by the positional error of the metal needle, it is necessary to run the current in the reverse direction and average it, so measure again with the following connections.
2) W41 and W12 are shorted and a current is passed in the reverse direction between metal needles 34 and 35 by current source 31. W23 and W34 are shorted and the Hall electromotive force V63 generated between metal needles 33 and 36 is measured by voltmeter 32.
The difference between V63 and V36 is taken and multiplied by 1/2 to remove the offset voltage, thereby deriving an accurate Hall electromotive force.

Moreover, the measurement accuracy can be further improved by connecting a current source and a voltmeter in a direction perpendicular to the above and measuring the Hall electromotive force. The procedure is as follows.
3) W21 and W32 are shorted and a current is passed between metal needles 33 and 36 by current source 31. W13 and W44 are shorted and the Hall electromotive force V45 generated between metal needles 34 and 35 is measured by voltmeter 32.
Additionally, to eliminate any offset voltage, measure again with the following connections:
4) W31 and W22 are shorted and a current is passed between metal needles 36 and 33 by current source 31. W13 and W44 are shorted and the Hall electromotive force V54 generated between metal needles 34 and 35 is measured by voltmeter 32.
By averaging the Hall electromotive forces obtained in the above four steps, it becomes possible to measure the Hall effect with higher accuracy.

Furthermore, accuracy can be improved by changing the direction of the magnet. A commercially available neodymium magnet with a magnetic flux density of 0.4 T was used, but if the magnet is turned upside down and the measurement sample is placed on the south pole, the generated Hall electromotive force will be reversed. Using this, accuracy can be further improved by averaging.

The measured Hall electromotive force varies depending on the shape of the measurement sample. Also, since the magnetic flux density is that of the outermost surface, there is a slight difference from the magnetic flux density at the measurement position of the sample. Regarding the shape dependency of the Hall electromotive force, in terms of the relationship between the Hall electromotive force and the mobility of carriers, it is known that the Hall electromotive force is proportional to the magnetic flux density and the amount of current flowing, and is proportional to the mobility. Therefore, if a standard sample is prepared and a calibration coefficient is calculated based on the mobility measured in advance by a normal Hall effect measurement, the measurement sample can be multiplied by the coefficient to perform calculations.

The third embodiment of the present invention will be described with reference to FIG. 4. In this embodiment, a MIS (Metal Insulator Metal) junction or a Schottky junction is formed on a semiconductor sample, and a DC voltage is applied to the junction to measure the capacitance change. A metal electrode 48 that forms a capacitance part is formed on the measurement sample, and a metal needle 43 is in contact with the metal electrode 48, and other metal needles 41, 42, and 44 are in contact with the periphery of the metal electrode 48. The configuration of other parts is common to the first and second embodiments. The Schottky junction as an example of the measurement sample is formed by forming a metal electrode 48 on an N-type GaN epitaxial layer 45 and an undoped GaN layer 46 that are sequentially formed on a (111) silicon substrate 47. When measuring a MIS junction with an insulating film formed on the substrate, the insulating film can be destroyed and electrical contact can be made by applying a high voltage to the metal needles 41, 42, and 44 in the same manner as before. In this way, CV measurement can be performed on the metal electrode 48 formed on the insulating film in the same manner.

The ohmic formation procedure is performed by tilting the two-way switch 9 to the upper side in the figure, and setting the 2×4 matrix switch 8 in the following order.
5) S11 and S22 are shorted and a high voltage generated by striking the piezoelectric element is applied.
6) S21 and S12 are shorted and a high voltage generated by striking the piezoelectric element is applied.
7) S12 and S24 are shorted, and a high voltage generated by striking the piezoelectric element is applied.
8) S22 and S14 are shorted, and a high voltage generated by striking the piezoelectric element is applied.
By always leaving S13 and S23 open (not used), it is possible to prevent a high voltage from being applied to the metal electrode 48. In the case of measuring a MIS (Metal Insulator Metal) structure, the metal needle comes into contact with the insulating film, but if the insulating film is about 100 nm thick, the dielectric strength is usually 100 V or less, so that it can be easily broken down by the high voltage generated by the piezoelectric body, and an ohmic junction can be formed through the insulating film.

Matrix switch 10-2 represents a state in which W11, W21, W41, and W34 are always shorted in matrix switch 10. By joining electrode needles 41, 42, and 44 and connecting them to terminal E1, and shorting W34, it is possible to connect the upper electrode on the Schottky junction or MIS junction side to terminal E4.

E1 and E4 are connected to a capacitance meter 53 via a bias tee 52. The bias tee 52 connected to the E1 and E4 terminals is a combination of a variable DC voltage source 49 and an inductor 50 connected in parallel, and a DC blocking capacitor 51 connected in series, preventing the high frequency generated by the capacitance meter 53 from flowing into the variable DC voltage source 49 and preventing the DC voltage 49 from being applied to the capacitance meter. If the frequency at which the capacitance is measured is low, the inductor 50 and the capacitor 51 must be large. A bias tee is a module used to separate DC and RF, but capacitance meters 53 with this function built in are also available on the market. For example, if an Agilent E4980ALCR meter is used, a bias tee is not necessary.

A fourth embodiment of the present invention will be described with reference to FIG. 5. In this embodiment, a MIS junction or Schottky junction is formed on a semiconductor sample, and a DC voltage is applied to the junction to measure the capacitance change. In the fourth embodiment, CV measurements are performed using a conductive liquid instead of the metal electrodes used in the third embodiment. This eliminates the need to form electrodes.

The part that forms the capacitance part uses ionic liquid 59 filled in a container 60 with an open bottom. The container 60 is configured so that the bottom contacts the semiconductor substrate with an insulating O-ring around the periphery to prevent leakage of the liquid. The ionic liquid used is one that has been used as a gate electrode for carbon nanotubes and is composed of imidazolium cations and several types of inorganic anions. A platinum electrode 58 immersed in the ionic liquid serves as the upper electrode of the MIS junction or Schottky junction.
The material of the container 60 must be resistant to corrosion by the ionic liquid, but it is also possible to use quartz, Teflon, or platinum.

Metal needles 54, 55, and 57 are placed around a container 60 of ionic liquid, and the measurement configuration is the same as in Example 3. The ohmic electrode formation for electrode needles 54, 55, and 57, the two matrix switches, the two-way switch, the procedure for forming ohmic junctions using them, and the CV measurement system are also the same as in Example 3.

The advantage of using an ionic liquid in this embodiment is that it is possible to omit the formation of a metal electrode by using a liquid electrode directly on the measurement sample. This saves the effort required to prepare the sample. The area of the capacitance part needs to be known for CV measurement, but the area of the capacitance part can be determined by the size of the O-ring. One method of using mercury as such a liquid electrode. However, there are concerns about environmental pollution caused by mercury, and its use has been banned worldwide by the Minamata Convention. This problem can be solved by using an ionic liquid with a small environmental impact as in this embodiment. By combining it with the ohmic junction formation method of the present invention, there is no need to form a metal electrode, and rapid CV measurement is possible.

In Examples 1, 2, 3, and 4, the piezoelectric element, 2X4 matrix switch, two-way switch, 4X4 matrix switch, and other components can be made into a common configuration. Figure 6 shows the measurement platform configuration that is this common configuration. The aim is to make the device using the ohmic formation of the present invention multifunctional by making the sheet resistance measurement, Hall measurement, and CV measurement the same device configuration. The present invention makes sheet resistance measurement, Hall effect measurement, and CV measurement into a platform, and by replacing the metal needles with different arrangements and shapes and the base on which the semiconductor to be measured is placed, it is possible to measure multiple electrical characteristics with a single device.

The Van der Pauw method is a common method for determining carrier mobility and carrier concentration by Hall effect measurement, but it is difficult to apply to large-diameter substrates because it requires cutting the target sample into small pieces. Boundaryless Hall measurement does not depend on the size of the substrate, but the process of deriving the Hall coefficient is complicated and errors are likely to occur due to non-uniform current distribution in the sample through which current is passed. This embodiment provides a simple method for determining carrier mobility and carrier concentration by newly combining Example 1 and Example 2.

As shown in FIG. 7, a metal needle array 1 is provided with four metal needles arranged linearly and at equal intervals, and a similar metal needle array 2 is arranged perpendicular to the four metal needles, with the innermost two metal needles of metal needle array 1 and the innermost two metal needles of metal needle array 2 forming a regular quadrangle. The measurement substrate is an AlGaN layer epitaxially grown on a GaN layer. Two-dimensional electron gas is formed at the interface due to polarization. First, electrical contacts are formed to all eight metal needles in both rows according to the method described in Examples 1 and 2.

After connecting a current source to the two outermost metal needles of metal needle array 1, the voltage V _S of the two innermost metal needles of metal needle array 1 is measured, and a magnetic field is applied perpendicular to the semiconductor or conductive thin film, and the voltage V _H of the two innermost metal needles of metal needle array 2 is measured.

The Hall effect is a state in which the Lorenz force due to a magnetic field B and the force from the electric field due to the generated Hall electromotive force are in balance. Therefore, if the carrier velocity is V ₀ , the elementary charge is q, the Hall voltage is V _H , and the distance between the metal needles is L, then each force is expressed as follows:
Lorenz force: _qVOB ...(2)
Force due to Hall electric field: qV _H /L (3)
If we set (2) and (3) equal,
V _O B = V _H /L ・・・・・・・・・・・・(4)
In addition, the carrier velocity _V0 is the product of the carrier mobility μ and the electric field in the direction in which the current flows. Therefore, the carrier velocity _V0 can be expressed by the following formula from the detected sense voltage VS and the distance L between the electrode needles.
V ₀ =μV _S /L ・・・・・・・・・・・・(5)
From (4) and (5), L can be eliminated to obtain the carrier mobility μ.
This is very simple and becomes
μ=κVH/( _VSB )・・・・・・・・・・・・(4)
In this embodiment, the current uniformity is relatively high because the current source is far from the measurement point. However, a correction factor κ is introduced to take into account the spread of magnetic flux at the measurement point and the non-uniformity of the current distribution from the metal needle. κ can be determined by measuring a sample whose mobility is already known by another method.
In addition, since the sheet resistance R _S is obtained from the formula (1) in the first embodiment,
The sheet carrier concentration _{N_S} is calculated by the following formula:
N _S =1/(qR _S μ) (5)
In this way, the configuration of this embodiment makes it possible to obtain the mobility μ and sheet carrier density of the measurement sample.

Of course, the above measurement method can be improved in accuracy by reversing the direction of the current source and averaging. The above description has been given with the metal needle array 1 as the axis, but the measurement accuracy can be further improved by alternating between the metal needle arrays 2 and 1 and performing measurements in perpendicular directions, or by reversing the direction of the applied magnetic field and performing measurements in the same manner. In this way, since the sheet resistance can be known in advance by the four-terminal method in this embodiment, the measurement of the carrier mobility and carrier concentration of the two-dimensional electron gas in the GaN/AlGaN heterojunction is simplified.
Needless to say, the measurement platform used in Example 5 can be applied twice to the metal needle array 1 and the metal needle array 2, and two voltmeters can be used.

1. Piezoelectric material 2. Lower electrode that holds the piezoelectric body 3. Upper electrode 4. Striking rod 5. Stopper 6. Insulating housing for piezoelectric element 7. Spring 8. 2x4 matrix switch 9. 2-way knife switch (2-way switch)
10. 4x4 matrix switch 11. Current source 12. Voltmeter 13-16. Metal needle 17. Undoped AlGaN epitaxial layer 18. Undoped GaN epitaxial layer (initial growth buffer layer includes AlGaN layer, etc.)
19. (111)-face silicon substrate or C-face sapphire substrate S11-S14, S21-S24. Relays A1-A4 that short-circuit and open at the intersections in the 2X4 matrix switch. Output terminals of the 2X4 matrix switch SPDT1-SPDT4. Single-pole double-throw switches B1-B4 that make up the 2-way switch. Switching side terminals C1-C4 to the measurement side of the 2-way switch. Single-pole terminals D1-D4 of the 2-way switch. Schottky diode R12 formed at the contact point between the metal needle and the semiconductor. Resistance R2 between metal needles 13 and 14. Resistance R3 between metal needles 14 and 15. Resistances 33-36 between metal needles 15 and 16. Metal needles arranged at the vertices of a regular square. 37. Undoped AlGaN epitaxial layer. 38. Undoped GaN epitaxial layer (initial growth buffer layer includes an AlGaN layer, etc.)
39. (111) face silicon substrate or C face sapphire substrate 40. Neodymium magnet (top face is N pole)
41. 42. 44. Metal needle 43. Metal needle in contact with Schottky electrode or MIS electrode 45. n-type GaN epitaxial layer 46. Undoped GaN epitaxial layer (initial growth buffer layer includes AlGaN layer, etc.)
47. (111) silicon substrate or C-plane sapphire substrate 48. Metal electrode (Schottky or MIS upper electrode)
49. Variable DC power supply 50. Inductor (for RF choke)
51. Capacitor (for DC cut)
52. Bias Tee 53. Capacitance Meter (LCR Meter)
54. 55. 57. Metal needle 56. Schottky or MIS upper electrode 58. Platinum electrode 59. Ionic liquid 60. Ionic liquid holding container 61. O-ring for sealing liquid 62. n-type GaN epilayer 63. Undoped GaN epilayer (initial growth buffer layer includes AlGaN layer, etc.)
64. (111) face silicon substrate or C face sapphire substrate 71 to 74. Metal needle array 1
75-78. Metal needle row 2
79. Current source 80. Sense voltage (VS) meter 81. Hall voltage (VH) meter 82. Magnet generating perpendicular magnetic field B 83. AlGaN epitaxial layer 84. GaN high resistance layer

Claims (1)

A method for forming an electrical contact between a metal needle and the semiconductor or the conductive thin film by contacting a metal needle connected to an electrode of a ceramic piezoelectric element with a semiconductor or a conductive thin film and applying a high voltage and a small current to the metal needle by applying stress or impact to the ceramic piezoelectric element, comprising the steps of:
A metal needle array 1 having four metal needles arranged linearly and at equal intervals;
a metal needle array 2 having four metal needles arranged in a straight line at equal intervals and disposed perpendicular to the metal needle array 1;
a current source connected between the two outermost metal needles of the metal needle array 1;
a voltage Vs measuring unit for measuring a voltage Vs between the two innermost metal needles of the metal needle array 1;
a voltage _VH measuring unit that applies a magnetic field B perpendicularly to a semiconductor or conductive thin film and measures a voltage _VH between the two innermost metal needles of the metal needle array 2;
Equipped with
The metal needle array 1 and the metal needle array 2 form electrical contacts with the semiconductor or the conductive thin film,
The mobility measuring method is characterized in that the two innermost metal needles of said metal needle array 1 and the two innermost metal needles of said metal needle array 2 form a regular quadrilateral.