Method for evaluating electrical defect density of semiconductor layer, and semiconductor element

Abstract

One embodiment of the present invention provides a method for evaluating the electrical defect density of a semiconductor layer, which comprises: a step for measuring an electric current by applying a voltage to a semiconductor element 1 which comprises a GaN layer 12 that serves as a semiconductor layer; and a step for deriving the electrical defect density in the GaN layer 12 with use of the measured electric current value.

Claims

1. A method for evaluating an electrical defect density of a semiconductor layer, comprising: applying a voltage to a semiconductor layer comprising a substrate and a buffer layer and measuring a first electric current value at the substrate or both the first electric current and a second electric current value at an electrode contact with a surface of the semiconductor layer opposite to the substrate; and deriving the electrical defect density at a position with respect to an interface between the substrate and the buffer layer in the semiconductor layer with use of the first electric current value or a difference between the first electric current value and the second electric current value.

2. A semiconductor element, comprising: a semiconductor layer serving as a current path, wherein a maximum value of an electrical defect density obtained by using a charge amount immediately after current application and a charge amount in steady state of the semiconductor layer is 1.0×10.sup.19 cm.sup.−3 or less, wherein a planar density of electrical defects in the semiconductor layer obtained by using a net charge amount accumulated in the semiconductor layer by the current application increases with two gradients, in a region where the planar density of the electrical defects increases as a distance from a low potential side surface at the time of the current application to the semiconductor layer, wherein the two gradients are a first gradient of a planar density of electrical defects of the low potential side in the semiconductor layer and a second gradient of a planar density of electrical defects of a high potential side in the semiconductor layer, wherein the second gradient is larger than the first gradient.

3. The semiconductor element according to claim 2, wherein a bandgap of the semiconductor layer is 2.5 eV or more.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) FIG. 1 is a vertical sectional view of an example of a semiconductor element suitable for evaluation by the method for evaluating electrical defect density of a semiconductor layer according to this embodiment.

(2) FIG. 2A is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 10V.

(3) FIG. 2B is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 20V.

(4) FIG. 2C is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 30V.

(5) FIG. 2D is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 40V.

(6) FIG. 3A is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 50V.

(7) FIG. 3B is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 60V.

(8) FIG. 3C is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 70V.

(9) FIG. 3D is a graph showing temporal changes in the substrate current and the electrode current when the applied voltage is 80V.

(10) FIG. 4A is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 90V.

(11) FIG. 4B is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 100V.

(12) FIG. 4C is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 110V.

(13) FIG. 4D is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 120V.

(14) FIG. 5A is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 130V.

(15) FIG. 5B is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 140V.

(16) FIG. 5C is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 150V.

(17) FIG. 5D is a graph showing temporal changes in substrate current and the electrode current when the applied voltage is 160V.

(18) FIG. 6A is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 170V.

(19) FIG. 6B is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 180V.

(20) FIG. 6C is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 190V.

(21) FIG. 6D is a graph showing temporal changes in substrate current and electrode current when the applied voltage is 200V.

(22) FIG. 7 is a graph plotting Q.sub.cap and Q.sub.TOTAL for each applied voltage.

(23) FIG. 8 is a graph showing the relationship between the thickness z of the depletion layer and the applied voltage V.

(24) FIG. 9 is a graph showing the relationship between the electrical defect density ρ in the epitaxial layer in the steady state and the position z.

(25) FIG. 10 is a graph showing the relationship between the net charge amount Q.sub.net accumulated in the epitaxial layer and the applied voltage V.

(26) FIG. 11 is a graph showing the relationship between the planar density of electrical defects in the GaN layer and the position z.

DESCRIPTION OF EMBODIMENT

(27) (Semiconductor Element 1)

(28) FIG. 1 is a vertical cross-sectional view of a semiconductor element 1 which is an example of a semiconductor element suitable for evaluation by the method for evaluating electrical defect density of a semiconductor layer according to this embodiment.

(29) The semiconductor element 1 includes a substrate 10 made of a semiconductor such as Si, a nitride semiconductor layer 12 made of GaN or the like formed on the substrate 10 via a buffer layer 11, and an electrode 13 having a laminated structure made of a plurality of metals such as Ti/Al/Ni/Au. The nitride semiconductor layer 12 includes, e.g., an impurity-doped nitride semiconductor layer 12a made of a GaN film doped with impurities such as C (carbon) and an undoped nitride semiconductor layer 12b made of a GaN film or the like which is undoped with impurities, on the impurity-doped nitride semiconductor layer 12a.

(30) Although the materials and thicknesses of the buffer layer 11, the impurity-doped nitride semiconductor layer 12a, and the undoped nitride semiconductor layer 12b can be arbitrarily determined, in a method for evaluating an electrical defect density of a semiconductor described later, as an example, the nitride semiconductor layer 12, the impurity-doped nitride semiconductor layer 12a, and the undoped nitride semiconductor layer 12b are a GaN layer 12, the C—GaN layer 12a, and an undoped GaN layer 12b, respectively. The thicknesses of the buffer layer 11, the C—GaN layer 12a, and the undoped GaN layer 12b were 3.5 μm, 730 nm, and 570 nm, respectively. Also, the electrode 13 may have an arbitrary shape, but similarly, as an example, a circular electrode having a radius of 560 μm and an area of 1 mm.sup.2 was used.

(31) (Method for Evaluating Electrical Defect Density of Semiconductor)

(32) A voltage can be applied between the substrate 10 and the electrode 13 by a variable DC power source 14. The current (substrate current) flowing through the substrate 10 can be measured by the ammeter 15a and the current (electrode current) flowing through the electrode 13 can be measured by the ammeter 15b.

(33) FIGS. 2A to 2D, 3A to 3D, 4A to 4D, 5A to 5D, and 6A to 6D are graphs showing temporal changes in substrate current and electrode current when the applied voltage is 10 to 200V. The value of the substrate current is multiplied by −1.

(34) When the applied voltage is 90V or more, the electrode current flows backward, but this is presumably because the electrons emitted from the trap level of the nitride semiconductor layer 12 are excessive to the electrons emitted from the trap level of the buffer layer 11, so that a well is formed in the nitride semiconductor layer 12 and the potential of the nitride semiconductor layer 12 is increased so that electrons are supplied from the electrode 13.

(35) In the present embodiment, the density of electrical defects in the nitride semiconductor layer 12 made of, e.g., a GaN layer which is a part of the semiconductor layer is obtained. According to the present embodiment, since the density of electric charge emitted from a deep defect level can be obtained, it is also possible to obtain the electrical defect density of a wide bandgap semiconductor having a deep defect level. Hereinafter, two types of electrical defect density evaluation methods will be described using this semiconductor element 1. In the following description, charge and electric capacity of each part are defined per unit area.

(36) (First Method)

(37) The first method is a method of obtaining the density of the electrical defect of the semiconductor layer from the substrate current measured by the ammeter 15a.

(38) When the semiconductor element 1 is regarded as a capacitor, the capacitance C.sub.cap of the semiconductor element 1, which is an ideal capacitor in which all regions of the buffer layer 11 and the GaN layer 12 are depleted, is expressed by the following equation 1.

(39) $\left[\mathrm{Equation}1\right]$$\begin{array}{cc}\frac{1}{C_{\mathrm{cap}}}=\frac{d_G}{{\mathrm{.Math.}}_G{\mathrm{.Math.}}_0}+\frac{d_b}{{\mathrm{.Math.}}_b{\mathrm{.Math.}}_0}& \left(1\right)\end{array}$

(40) Here, ε.sub.0 is the dielectric constant of vacuum, ε.sub.G is the relative permittivity of the GaN layer 12, ε.sub.b is the relative permittivity of the buffer layer 11, d.sub.G is the thickness of the GaN layer 12, and d.sub.b is the thickness of the buffer layer 11. As described above, the method for evaluating the electrical defect density of the semiconductor layer according to the present embodiment can also be applied to a semiconductor element including a plurality of semiconductor layers having different dielectric constants.

(41) In the semiconductor element 1, ε.sub.G and ε.sub.b are 9.5 and 8.5 respectively, d.sub.G and d.sub.b are 1.3 μm and 3.5 μm, respectively, and the area of the electrode 13 corresponding to the area of the capacitor is 1 mm.sup.2. As a result, C.sub.cap is calculated to be 16 pF.

(42) Further, the charge Q.sub.total accumulated in the depletion layers formed in the buffer layer 11 and the GaN layer 12 is expressed by the following equation 2.
[Equation 2]
Q.sub.total=Q.sub.cap+Q.sub.dep  (2)

(43) Here, Q.sub.cap is the accumulated charge when the semiconductor element 1 is an ideal capacitor in which all the regions of the buffer layer 11 and the GaN layer 12 (hereinafter referred to as epitaxial layer), and Q.sub.dep is the charge released from the epitaxial layer.

(44) Among them, Q.sub.cap can be obtained by integrating the substrate current I.sub.sub with the time (0 to 0+) until the charge accumulates in the capacitor as shown in the following equation 3, since the time until the charge accumulates in the capacitor is on the order of nanoseconds, the substrate current I.sub.sub which is enough for calculating Q.sub.cap cannot be measured under the usual measurement environment (the time resolution of the measuring apparatus is on the order of microseconds).
[Equation 3]
Q.sub.cap=ƒ.sub.0.sup.0+I.sub.subdt  (3)

(45) On the other hand, Q.sub.cap is expressed by the product of C.sub.cap and the applied voltage V as in the following equation 4. As described above, C.sub.cap can be obtained from Equation 1 and is 16 pF for the semiconductor element 1. Therefore, for example, Q.sub.cap is obtained as 1.12×10.sup.−9 C when the applied voltage V is 70V.
[Equation 4]
Q.sub.cap=C.sub.capV  (4)

(46) Q.sub.dep can be obtained by integrating the substrate current I.sub.sub with the time (0+ to ∞) after charge accumulation in the capacitor (steady state) as expressed by the following equation 5.
[Equation 5]
Q.sub.dep=ƒ.sub.0+.sup.∞I.sub.subdt  (5)

(47) From Equation 5, for example, Q.sub.dep when the applied voltage V of the semiconductor element 1 is 70V can be obtained as 1.73×10.sup.−10 C using the integrated value of the substrate current I.sub.sub up to 10 seconds. The substrate current I.sub.sub when the applied voltage V is 70V is shown in FIG. 3C.

(48) From Q.sub.cap obtained from Equation 4 and Q.sub.dep obtained from Equation 5, Q.sub.total is obtained using Equation 2.

(49) FIG. 7 is a graph obtained by plotting Q.sub.cap obtained from Equation 4 and Q.sub.total obtained from Equations 2, 4, and 5 for each applied voltage. The difference Q.sub.dep between Q.sub.total and Q.sub.cap at each applied voltage is caused by an increase in capacitance due to depletion of the epitaxial layer.

(50) The electric capacity C.sub.dep of the depletion layer in the steady state can be obtained using the following Equation 6.
[Equation 6]
Q.sub.total=C.sub.depV  (6)

(51) Also, when the thickness z of the depletion layer is larger than the thickness d.sub.b of the buffer layer 11, C.sub.dep is expressed by the following Equation 7.

(52) $\left[\mathrm{Equation}7\right]$$\begin{array}{cc}\frac{1}{C_{\mathrm{dep}}}=\frac{z-d_b}{{\mathrm{.Math.}}_G{\mathrm{.Math.}}_0}+\frac{d_b}{{\mathrm{.Math.}}_b{\mathrm{.Math.}}_0}& \left(7\right)\end{array}$

(53) From Equation 7, the thickness z of the depletion layer is calculated. For example, it is calculated that z is 4.07 μm when the applied voltage V of the semiconductor element 1 is 70V. In this case, since the thickness of the buffer layer 11 is 3.5 μm, the entire region of the buffer layer 11 is depleted, and the region of 0.57 μm thickness on the buffer layer 11 side of the GaN layer 12 is depleted.

(54) FIG. 8 is a graph showing the relationship between the applied voltage V and the thickness z of the depletion layer (at the position of the upper end of the depletion layer with reference to the interface between the substrate 10 and the buffer layer 11). When the applied voltage V is 5V, the entire region of the buffer layer 11 is depleted.

(55) When the relative permittivity ε.sub.G of the GaN layer 12 and the relative permittivity ε.sub.b of the buffer layer 11 can be regarded as being equal (ε.sub.Gε.sub.0=ε.sub.bε.sub.0=ε), the above equations can be simplified. In this case, from Equation 1 and Equation 4, C.sub.cap and Q.sub.cap are expressed as ε/d and εV/d, respectively. From Equations 2, 5, 6, and 7, C.sub.dep is expressed as ε/z. As a result, the following Equation 8 is established.

(56) $\left[\mathrm{Equation}9\right]$$\begin{array}{cc}{Q}_{\mathrm{dep}}=Q_{\mathrm{total}}-Q_{\mathrm{cap}}={\int }_{0+}^{\infty }{I}_{\mathrm{sub}}\mathrm{dt}=\mathrm{.Math.}\left(\frac{1}{z}-\frac{1}{d}\right)& \left(8\right)\end{array}$

(57) The following Equation 9 is a formula showing the relationship between the electrical defect density ρ in the epitaxial layer in the steady state and the applied voltage V. Here, x is a distance in the thickness direction with the interface between the substrate 10 and the buffer layer 11 as the origin, and ρ is a function of x.

(58) $\left[\mathrm{Equation}9\right]$$\begin{array}{cc}V=\frac{q}{\mathrm{.Math.}\left(z\right)}{\int }_0^z\rho \left(x\right).Math.\mathrm{xdx}& \left(9\right)\end{array}$

(59) In Equation 9 and Equations 10 and 11 to be described later, ε is the relative permittivity of the epitaxial layer composed of the buffer layer 11 and the GaN layer 12, and is expressed as a function of the thickness z of the depletion layer. Equation 7 can be expressed using this ε as the following Equation 10.

(60) $\left[\mathrm{Equation}10\right]$$\begin{array}{cc}\frac{1}{C_{\mathrm{dep}}}=\frac{z}{\mathrm{.Math.}\left(z\right)}& \left(10\right)\end{array}$

(61) The following Equation 11 is derived from the first-order differentiation of Equation 9 with z.

(62) $\left[\mathrm{Equation}11\right]$$\begin{array}{cc}\frac{\mathrm{dV}}{\mathrm{dz}}=\frac{q}{\mathrm{.Math.}\left(z\right)}\rho \left(z\right).Math.z& \left(11\right)\end{array}$

(63) Then, the following Equation 12 is obtained by modifying Equation 10.

(64) $\left[\mathrm{Equation}12\right]$$\begin{array}{cc}\rho \left(z\right)=\frac{\mathrm{.Math.}\left(z\right)}{\mathrm{qz}}.Math.\frac{\mathrm{dV}}{\mathrm{dz}}& \left(12\right)\end{array}$

(65) From Equation 12, it is possible to derive the electrical defect density ρ in the epitaxial layer at the position z (the distance from the interface between the substrate 10 and the buffer layer 11) with reference to the interface between the substrate 10 and the buffer layer 11.

(66) FIG. 9 is a graph showing the relationship between the electrical defect density ρ in the GaN layer 12 in the steady state obtained from Equation 12 and the position z.

(67) For example, in the case where the GaN layer 12 is used as a layer serving as a current path of a high electron mobility transistor (HEMT), the maximum value of the electrical defect density of the GaN layer 12, which is obtained by using the charge amount immediately after the current application and the charge amount in the steady state as described above, is preferably 1.0×10.sup.19 cm.sup.−3 or less, more preferably 2.0×10.sup.18 cm.sup.−3 or less. In the HEMT, an AlGaN layer is formed on the GaN layer 12, and a region having a depth of several nm from the interface with the AlGaN layer of the GaN layer 12 serves as a current path.

(68) (Second Method)

(69) The second method is a method of obtaining the density of the electrical defect of the semiconductor layer from the difference between the substrate current measured by the ammeter 15a and the electrode current measured by the ammeter 15b.

(70) The net charge amount Q.sub.net accumulated in the epitaxial layer of the semiconductor element 1 as a capacitor can be obtained by integrating the difference between the substrate current I.sub.sub and the electrode current I.sub.ele with time.
[Equation 13]
Q.sub.net=ƒ.sub.0.sup.∞(I.sub.sub−I.sub.ele)dt  (13)

(71) FIG. 10 is a graph showing the relationship between the net charge amount Q.sub.net accumulated in the epitaxial layer which is obtained by Equation 13 and the applied voltage V. The region where the net charge amount Q.sub.net decreases with the increase of the applied voltage V in FIG. 10 is the region where the electrode current flows backward.

(72) It can be assumed that the net charge amount Q.sub.net is equal to the planar density (surface density) of the defects in the region where the net charge amount Q.sub.net decreases with the increase of the applied voltage V in FIG. 10 (the region where the applied voltage V is 20V or less). Using the thickness z of the depletion layer at each voltage of the applied voltage V obtained by the first method, the relationship between the net charge amount Q.sub.net and the applied voltage V in FIG. 10 can be converted into the relationship between the planar density of the defects and the position z.

(73) FIG. 11 is a graph showing the relationship between the planar density of electrical defects in the GaN layer 12 and the position z.

(74) In the example, the semiconductor element 1 having the structure shown in FIG. 1 was used as a sample of the electrical defect density evaluation method. The semiconductor element 1 has a substrate 10 made of Si having a (111) plane as a main surface, a GaN layer 12 formed on the substrate 10 via a buffer layer 11, an electrode 13 having a Ti/Al/Ni/Au laminated structure. The GaN layer 12 has a C—GaN layer 12a made of a GaN film doped with C (carbon) and an undoped GaN layer 12b composed of a GaN film not doped with impurities on the C—GaN layer 12a. The thicknesses of the buffer layer 11, the C—GaN layer 12a, and the undoped GaN layer 12b were 3.5 μm, 730 nm, and 570 nm, respectively. In addition, the electrode 13 was a circular electrode having a radius of 560 μm and an area of 1 mm.sup.2.

(75) The value of the intercept of FIG. 11, that is, the value of the planar density of the defects at the interface (z=3.5 μm) between the buffer layer 11 and the GaN layer 12 is 5.6×10.sup.10 cm.sup.−2. Assuming that the defects are uniformly distributed in the buffer layer 11 having a thickness of 3.5 μm, the electrical defect density in the buffer layer 11 is calculated as 5.6×10.sup.10 cm.sup.−2/3.5 μm=1.4×10.sup.16 cm.sup.−3. This electrical defect density is substantially equal to the electrical defect density obtained by the first method.

(76) The electrical defect density in the GaN layer 12 is obtained from the slope of the line in FIG. 11. That is, by dividing the minute change amount of the planar density in a certain region in the GaN layer 12 by the minute change amount of the distance, the electrical defect density in that region can be obtained.

(77) According to FIG. 11, in a region where the planar density of electrical defects increases (a region where the position z is about 3.5 to 3.9 μm) as a distance from the interface with the buffer layer 11 which is the low potential side surface at the time of current application, the planar density of electrical defects increases with two gradients.

(78) The electrical defect density in the two regions where the slopes of increase of the planar density of this defects differ from each other is calculated as 3.7×10.sup.14 cm.sup.−3 (in the region where the position z is approximately 3.5 to 3.8 μm) and 2.0×10.sup.15 cm.sup.−3 (in the region where the position z is approximately 3.8 to 3.9 μm). It is considered that 3.7×10.sup.14 cm.sup.−3 is equivalent to the electrical defect density obtained by the first method and indicates the electrical defect density in the region close to the buffer layer 11. Further, 2.0×10.sup.15 cm.sup.−3 is considered to indicate the electrical defect density of the interface between the C—GaN layer 12a and the undoped GaN layer 12b or the undoped GaN layer 12b.

(79) (Effect of Embodiment)

(80) According to the method for evaluating the electrical defect density of a semiconductor in the above embodiment, the density of the charge trapped in the deep defect level can be examined. Therefore, the method for evaluating the electrical defect density of a semiconductor in the above embodiment is particularly useful as a method for evaluating the electrical defect density of a wide bandgap semiconductor having a deep defect level, for example, a semiconductor layer having a bandgap of 2.5 eV or more.

(81) A highly reliable semiconductor device can be manufactured by using a semiconductor template including a semiconductor layer evaluated by the method for evaluating electrical defect density of a semiconductor in the above embodiment. In particular, it is useful for manufacturing a power device such as a power transistor using a wide bandgap semiconductor.

(82) As described above, the structure of the semiconductor element to be evaluated in the method for evaluating electrical defect density of a semiconductor layer according to the present invention is not limited to the structure of the semiconductor element 1. For example, the substrate is not particularly limited as long as it is an electrically conductive substrate, and even when an insulating substrate is used, a semiconductor layer to be evaluated may be formed on the insulating substrate via an electrically conductive layer. In this case, instead of the substrate current, the current of the conductive layer on the insulating substrate is measured. The layer structure and composition of the semiconductor layer are also not particularly limited. In addition, the electrode is not particularly limited as long as it is an ohmic electrode.

(83) Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the invention.

(84) In addition, the embodiment described above does not limit the invention according to the claims. It should also be noted that not all combinations of features described in the embodiments are indispensable to means for solving the problem of the invention.

INDUSTRIAL APPLICABILITY

(85) A method for evaluating the electrical defect density of a semiconductor layer applicable to a wide bandgap semiconductor having a deep defect level and a semiconductor element with a low electrical defect density which can be evaluated by the method are provided.

REFERENCE SIGNS LIST

(86) 1 Semiconductor element

(87) 10 Substrate

(88) 11 Buffer layer

(89) 12 GaN layer

(90) 12a C—GaN layer

(91) 12b Undoped GaN layer

(92) 13 Electrode