JP6116989B2 - Cu-doped p-type ZnO-based semiconductor crystal layer and manufacturing method thereof - Google Patents

Description

  The present invention relates to a Cu-doped p-type ZnO-based semiconductor crystal layer and a method for manufacturing the same. A II-VI group semiconductor containing Zn as a group II element and O as a group VI element is called a ZnO-based semiconductor. The amount of Cu dope is expressed as the atomic ratio (%) of the positive element.

  Zinc oxide (ZnO) is a direct transition type semiconductor having a band gap energy of 3.37 eV at room temperature, and the exciton binding energy is 60 meV, which is relatively large compared to other semiconductors. In addition, the raw material is inexpensive and has little influence on the environment and the human body. For this reason, realization of a light-emitting element excellent in environmental performance with high efficiency and low power consumption using ZnO is expected.

  The n-type conductivity of the ZnO-based semiconductor can be obtained without doping impurities. In order to increase the n-type conductivity, a donor impurity such as Ga can be doped. A ZnO-based oxide semiconductor is difficult to control the p-type conductivity because of a self-compensation effect due to strong ionicity. Therefore, in order to obtain a p-type ZnO-based oxide semiconductor having practical performance, molecular beam epitaxy (MBE), laser molecular beam epitaxy, metal organic chemical vapor deposition (MOCVD), remote plasma MOCVD, sputtering, etc. Development of p-type ZnO-based semiconductor layers has been promoted by various methods.

  To date, group VA elements such as N, P, As, and Sb, group IA elements such as Li, Na, and K, and group IB elements such as Cu, Ag, and Au have been studied as acceptor impurities. The carrier concentration is still low, and sufficient quality including reproducibility and stability has not been obtained. Techniques such as co-doping of two types of impurities have been considered. A technique for obtaining a p-type layer by co-doping an acceptor such as Cu and Ag together with a donor such as Ga (for example, see Patent Document 1), a technique for obtaining a p-type layer by co-doping nitrogen and a group IB element (for example, Patent Document 2) has been proposed. In addition, a technique for doping an acceptor such as N or Cu into at least one of superlattice layers having a stacked structure of MgZnO layers and ZnO layers has been proposed (for example, see Patent Document 3).

JP 2004-221132 A JP 2009-256142 A JP 2009-21540 A

One object of the embodiment is to provide a method of manufacturing a p-type ZnO-based semiconductor crystal layer in which a single Cu element is doped in a ZnO-based semiconductor.

According to one aspect of the present invention,
On the growth substrate, ZnO-based semiconductor layers and Cu layers are alternately grown, and a single Cu element is doped to form a ZnO-based semiconductor epitaxial layer doped with Cu in an atomic ratio of 5% or more. ,
Annealing the ZnO-based semiconductor epitaxial layer doped with 5% or more Cu to obtain p-type conductivity;
A method for producing a Cu-doped p-type ZnO-based semiconductor crystal layer,
The growth of the ZnO-based semiconductor layer is performed by molecular beam epitaxy having a period of supplying only Zn before and after a period of supplying Zn and O to grow a ZnO-based semiconductor layer,
The Cu layer is grown by molecular beam epitaxy for supplying Cu.
A method for producing a Cu-doped p-type ZnO-based semiconductor crystal layer is provided.

FIG. 1 is a schematic cross-sectional view showing an example of an MBE apparatus. 2A and 2B are a schematic cross-sectional view of a formed Cu-doped ZnO deposited film sample and a timing chart of shutters of Zn cells, O cells, and Cu cells in the growth of the deposited film of the sample. 3A, 3B, 3C, and 3D are RHEED images of the first, second, third, and fourth samples. 4A, 4B, 4C, and 4D are SIMS depth profiles of the first, second, third, and fourth samples. FIG. 5 is a graph showing a CV characteristic before annealing of the first sample and a depth profile of the impurity concentration. When, 6A, 6B (6B1, 6B2), 6C, and 6D are graphs showing CV characteristics and depth profiles of impurity concentrations after annealing of the first, second, third, and fourth samples. 7A, 7B, and 7C are a plan view schematically showing a six-membered ring structure, a perspective view showing a relative relationship of six-membered ring structures included in adjacent layers in the layered crystal, and two adjacent ones in the crystal. It is a diagram which shows the structure by which two Cu atoms were coordinated adjacently through the O atom at two group II sites of the layer. FIG. 8 is a cross-sectional view of a semiconductor device using a Cu-doped p-type ZnO layer.

  First, a crystal manufacturing apparatus used for growing a ZnO-based semiconductor layer or the like will be described. As a crystal manufacturing method, the sample described below was prepared using molecular beam epitaxy (MBE).

  FIG. 1 is a schematic cross-sectional view showing an example of an MBE apparatus. In the vacuum chamber 101, a Zn source gun 102, an O source gun 103, an Mg source gun 104, a Cu source gun 105, and another solid source gun 106 are provided.

Zn source gun 102, Mg source gun 104, Cu source gun 105, and other solid source gun 106 contain Zn (7N), Mg (6N), Cu (9N), and other solid solid sources, respectively. A Zn beam, an Mg beam, a Cu beam, and other molecular beams (for example, a Ga beam) can be emitted by heating the cells. The O source gun 103 includes an electrodeless discharge tube using a radio frequency (for example, 13.56 MHz), converts O ₂ gas (6N) into plasma, and emits an O radical beam. As the discharge tube material, alumina, high-purity quartz, boron nitride, or the like can be used. The discharge tube at the time of sample preparation was made of boron nitride (BN).

  A stage 107 including a substrate heater is disposed in the vacuum chamber 101 and holds the growth substrate 108. Each of the source guns 102 to 106 includes a cell shutter, and switches between a state of irradiating a beam on the growth substrate 108 and a state of not irradiating the growth substrate 108 by opening and closing the cell shutter. A ZnO-based compound semiconductor layer having a desired composition can be grown on the growth substrate 108 by supplying a desired beam at a desired timing.

By adding Mg to ZnO to form a Mg _x Zn _1-x O mixed crystal, the band gap can be expanded according to the Mg composition x. Since ZnO has a wurtzite structure (hexagonal crystal) and MgO has a rock salt structure (cubic crystal), the Mg composition x is limited. When maintaining the wurtzite structure, Mg _x Zn _1-x O (0 ≦ x ≦ 0.6) is set.

A film thickness meter 109 using a crystal resonator is provided in the vacuum chamber 101. From deposition rate measured by a film thickness meter 109, flux strength, such as Zn beam from a solid source (e.g. expressed as F _Zn) is obtained.

  A gun 110 for reflection high-energy electron diffraction (RHEED) and a screen 111 for projecting a RHEED image are attached to the vacuum chamber 101. From the RHEED image, the surface flatness and growth mode of the crystal layer formed on the substrate 108 can be evaluated.

  When the crystal is grown two-dimensionally and the surface is epitaxially grown (single crystal growth), the RHEED image shows a streak pattern. In the case of epitaxial growth (single crystal growth) where the crystal grows three-dimensionally and the surface is not flat, the RHEED image shows a spot pattern. In the case of polycrystalline growth, the RHEED image is a ring pattern. Therefore, from the RHEED image, it is possible to know whether the growth layer is a single crystal, a polycrystal, or a single crystal when the surface is flat or not flat.

Next, the VI / II flux ratio will be described. The flux intensity of Zn beam expressed as _{J Zn,} the flux intensity of the Mg beam expressed as _{J Mg,} representing the flux intensity of O radical beam and _{J O.} The beam of Zn or Mg, which is a group II material, contains Zn or Mg of atoms or clusters containing a plurality of atoms, and both atoms and clusters are effective for crystal growth. The O beam, which is a gas material, contains atomic radicals and neutral molecules. Here, the flux intensity of atomic radicals effective for crystal growth is considered.

The adhesion coefficient indicating the ease of adhesion of Zn to the crystal is denoted by k _Zn , the adhesion coefficient indicating the ease of adhesion of Mg is denoted by k _Mg, and the adhesion coefficient indicating the ease of adhesion of O is denoted by k 2 _O. Zn adhesion coefficient k _Zn and flux strength J _Zn product k _Zn J _Zn , Mg adhesion coefficient k _Mg and flux strength J _Mg product k _Mg J _Mg , and O adhesion coefficient k _O and flux strength J product k _O J _O of _O, respectively, Zn atoms adhering per unit time per unit area of the substrate, corresponding to the Mg atom, and the number of O atoms.

k _Zn J _Zn and _k Mg to the sum of _{J Mg} is the ratio of _{k O J O k O J O} / a _{(k Zn J Zn + k Mg} J Mg), defined as VI / II flux ratio. A case where the VI / II flux ratio is smaller than 1 is referred to as a group II rich condition (or a Zn rich condition when Mg is not included), and a case where the VI / II flux ratio is equal to 1 is referred to as a stoichiometric condition. A case where the II flux ratio is larger than 1 is called a VI group rich condition (or an O rich condition). In the crystal growth on the + c plane (Zn plane), if the substrate surface temperature is 850 ° C. or lower, the adhesion coefficients k _Zn , k _Mg , and k 2 _O can be regarded as 1, and the VI / II flux ratio is It can be expressed as J _O / (J _Zn + J _Mg ).

Specifically, the VI / II flux ratio can be calculated by the following procedure, for example. Take ZnO growth as an example. The Zn flux is measured as a _Zn deposition rate F _Zn (nm / s) at room temperature by a film thickness monitor using a crystal resonator. The unit of Zn flux is converted from F _Zn (nm / s) to J _Zn (atoms / cm ² s). O radical flux is calculated | required as follows. ZnO is grown by changing the Zn flux under constant O radical beam irradiation conditions (for example, O ₂ flow rate ₂ sccm / RF power 300 W), and the Zn flux dependence of the ZnO growth rate is experimentally determined. By fitting the result using an approximate expression of ZnO growth rate G _ZnO : G _ZnO = [(k _Zn J _Zn ) ^-1 + (k _O J _O ) ^-1 ] ^-1 , radical flux _{J O} is calculated. Thus from Zn flux _{J Zn} and O radical flux _{J O} obtained, it is possible to calculate the VI / II flux ratio.

  The inventors of the present invention have been conducting research on doping Cu as a p-type impurity (acceptor) in a ZnO-based semiconductor. When a Cu-doped ZnO crystal was grown by simultaneously supplying a Zn beam, a Cu beam, and an O beam, inconveniences occurred in crystallinity (see Japanese Patent Application No. 2012-251010, paragraphs 0028-0039). Cu oxidation was considered. Therefore, good crystallinity was obtained by setting the supply of the Cu beam at a different timing from the supply of the O beam (and Zn beam) (see Japanese Patent Application No. 2012-251010, paragraphs 0048-0083, etc.).

The present inventors considered obtaining p-type conductivity by using Cu as a single acceptor impurity. It is considered that the p-type conductivity of a ZnO-based semiconductor can be basically obtained by doping with a p-type impurity. However, a single crystal film in which ZnO layers and partial Cu layers (Cu layers of less than one atomic layer) are alternately grown does not exhibit p-type conductivity even if it exhibits a Cu concentration of about 10 ²⁰ atoms / cm ^3. It was.

  It was investigated what kind of phenomenon occurs when a higher concentration of Cu is doped. First, second, third, and fourth samples were prepared by changing the Cu concentration. First, the production of the base structure of each sample will be described.

FIG. 2A is a schematic cross-sectional view schematically showing the configuration of a sample. The Zn surface ZnO (0001) substrate 11 was subjected to thermal cleaning at 900 ° C. for 30 minutes, and then the substrate temperature was lowered to 300 ° C. Thereafter, at a growth temperature of 300 ° C., the Zn deposition rate F _Zn was set to 0.12 nm / s (Zn flux J _Zn = 7.9 × 10 ¹⁴ atoms / cm ² s), and the O radical beam irradiation conditions were set to RF power 300 W, A ZnO buffer layer 12 having a thickness of 30 nm was grown at an O ₂ flow rate of 2.0 sccm (O flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s). Then, annealing was performed at 900 ° C. for 10 minutes in order to improve the crystallinity and surface flatness of the buffer layer.

On the ZnO buffer layer, the growth temperature is 900 ° C., the Zn deposition rate F _Zn is 0.12 nm / s (Zn flux J _Zn = 7.9 × 10 ¹⁴ atoms / cm ² s), and the O radical beam irradiation condition is RF. An undoped ZnO layer 13 having a thickness of 100 nm was grown at a power of 300 W and an O ₂ flow rate of 2.0 sccm (O flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s). The obtained structure is the base structure of the sample, and is a configuration common to each sample. A Cu-doped ZnO layer was grown on the undoped ZnO layer 13.

  FIG. 2B is a timing chart of shutters of a Zn cell, an O cell, and a Cu cell in Cu-doped ZnO layer growth. Schematically speaking, a ZnO growth process in which the Zn cell shutter and the O cell shutter are opened and the Cu cell shutter is closed, and a Cu deposition process in which the Zn cell shutter and the O cell shutter are closed and the Cu cell shutter is opened are alternately repeated. .

  The open period of the Zn cell shutter is extended before and after the open period of the O cell shutter. That is, the open period of the Zn cell shutter includes the open period of the O cell shutter, and the surface of the ZnO single crystal film is covered with Zn before and after the Cu deposition step, thereby suppressing the direct reaction between the O radical and Cu. ing. By providing the ZnO growth step and the Cu deposition step separately, the open period of the O cell shutter and the open period of the Cu cell shutter are prevented from overlapping, and O radicals and Cu are not supplied onto the substrate at the same time. I have to. In the Cu adhesion step, Cu does not adhere to the entire surface and does not have to be an atomic layer.

  Returning to FIG. 2A, the alternate stack 15 of ZnO layers ML and Cu layers DL is formed on the undoped ZnO layer 13. The Cu layer DL is shown as one atomic layer or less. After the stacked structure 15 is formed, it is intended to diffuse Cu atoms into the ZnO layer ML by annealing. The Cu concentration is adjusted by adjusting the ratio between the thickness of the ZnO layer ML and the thickness (coverage) of the Cu layer DL.

  In the production process of the first sample, the growth substrate temperature was set to 300 ° C. The open period per O cell shutter is 16 seconds, and the open period of the Zn cell shutter is extended by 1 second before and after the open period of the O cell shutter. 18 seconds. 16 seconds in which both the Zn cell shutter and the O cell shutter are in the open state is the ZnO growth time per time. The open period per time of the Cu cell shutter was 180 seconds.

A 100-nm-thick Cu-doped ZnO epitaxial film was obtained by carrying out growth by repeating 30 steps of a process including one set of the ZnO growth step and the Cu deposition step. The Zn deposition rate F _Zn in the ZnO growth process is 0.15 nm / s (flux J _Zn = 9.9 × 10 ¹⁴ atoms / cm ² s), the O radical beam irradiation conditions are RF power 300 W, O ₂ flow rate 2. It was 0 sccm (flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s). The VI / II flux ratio is 0.82, which is a group II rich condition. The temperature of the Cu source was 870 ° C. The Cu deposition rate F _Cu was below the detection limit.

  In the production process of the second sample, the growth substrate temperature was set to 250 ° C. The open period per O cell shutter is 8 seconds, and the open period of the Zn cell shutter is extended by 1 second before and after the open period of the O cell shutter, that is, the open period per time of the Zn cell shutter is increased. 10 seconds. 8 seconds in which both the Zn cell shutter and the O cell shutter are open is the ZnO growth time per time. The opening period per time of the Cu cell shutter was set to 50 seconds.

A process in which the ZnO growth step and the Cu deposition step were set as one set was grown 60 times to obtain a 120 nm thick Cu-doped ZnO epitaxial film. The Zn deposition rate F _Zn in the ZnO growth step is 0.12 nm / s (flux J _Zn = 7.9 × 10 ¹⁴ atoms / cm ² s), the O radical beam irradiation conditions are RF power 300 W, O ₂ flow rate 2. It was 0 sccm (flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s). The VI / II flux ratio is 1.02, which is a slightly rich group VI condition. The temperature of the Cu source was 970 ° C. The Cu deposition rate F _Cu was 0.003 nm / s.

  In the third sample preparation process, the growth substrate temperature is 250 ° C., the opening period of each O cell shutter is 10 seconds, and the opening period of the Zn cell shutter is 1 second before and after the opening period of the O cell shutter. In other words, the Zn cell shutter was opened for 12 seconds. 10 seconds in which both the Zn cell shutter and the O cell shutter are open is the ZnO growth time per time. The opening period per time of the Cu cell shutter was set to 50 seconds.

The process which made the ZnO growth process and the Cu adhesion process one set was performed by repeating the growth for 60 sets to obtain a Cu-doped ZnO epifilm having a thickness of 145 nm. The Zn deposition rate F _Zn in the ZnO growth step is 0.13 nm / s (flux J _Zn = 8.6 × 10 ¹⁴ atoms / cm ² s), the O radical beam irradiation conditions are RF power 300 W, O ₂ flow rate 2. It was 0 sccm (flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s). The VI / II flux ratio is 0.94, which is a slightly rich group II condition. The temperature of the Cu source was 990 ° C. The Cu deposition rate F _Cu was 0.004 nm / s.

  In the fourth sample preparation process, the growth substrate temperature is 250 ° C., the open period of each O cell shutter is 8 seconds, and the open period of the Zn cell shutter is 1 second before and after the open period of the O cell shutter. In other words, each Zn cell shutter was opened for 10 seconds. 8 seconds in which both the Zn cell shutter and the O cell shutter are open is the ZnO growth time per time. The opening period per time of the Cu cell shutter was 60 seconds.

A process in which the ZnO growth step and the Cu adhesion step were set as one set was grown 60 times to obtain a Cu-doped ZnO epifilm having a thickness of 115 nm. The Zn deposition rate F _Zn in the ZnO growth step is 0.13 nm / s (flux J _Zn = 8.6 × 10 ¹⁴ atoms / cm ² s), the O radical beam irradiation conditions are RF power 300 W, O ₂ flow rate 2. It was 0 sccm (flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s). The VI / II flux ratio is 0.94, which is a slightly rich group II condition. The temperature of the Cu source was 1000 ° C. The Cu deposition rate F _Cu was 0.0045 nm / s.

  3A, 3B, 3C, and 3D show reflection high energy electron diffraction (RHEED) images from the [11-20] direction, which are obtained from Sample 1, Sample 2, Sample 3, and Sample 4. FIG. In the first sample, a streak pattern is observed, and it can be seen that the single crystal film is grown flat (two-dimensionally). As the Cu concentration increased from sample 2 to sample 4, it was confirmed that the single crystal grew epitaxially, although it gradually changed from the streak pattern to the spot pattern.

4A, 4B, 4C, 4D show the results obtained by secondary ion mass spectrometry (SIMS) of the grown Cu-doped ZnO sample (before annealing). The horizontal axis indicates the depth from the surface in the unit of nm, and the vertical axis indicates the Cu concentration in the unit of atoms / cm ³ (one logarithmic scale). The Cu concentration after performing the quantification and standardization treatment with the ZnO standard sample is approximately 7.0 × 10 ²⁰ atoms / cm ³ (1.7%, atomic ratio of positive element, the same applies hereinafter) in the sample 1. 2 is approximately 1.75 × 10 ²¹ atoms / cm ³ (4.3%), sample 3 is approximately 2.9 × 10 ²¹ atoms / cm ³ (7.0%), and sample 4 is approximately 4.4 × 10. ²¹ atoms / cm ³ (10.6%).

FIG. 5 is a graph showing a CV characteristic and a depth profile of the impurity concentration showing the relationship of the capacity to the applied voltage for the sample 1 before annealing. The graph showing the CV characteristics is shown in the upper row, and the graph (one logarithmic scale) showing the depth profile is shown in the lower row. The measurement was performed by an ECV method using an electrolytic solution as a Schottky electrode. The graph shows the results of analysis using a parallel model. The horizontal axis of the graph showing the CV characteristic represents the voltage in the unit “V”, and the vertical axis represents “1 / C ² ” in the unit “cm ⁴ / F ² ”. Both axes use a linear scale. The horizontal axis of the graph showing the depth profile represents the position in the depth (thickness) direction of the sample in the unit “nm”, and the vertical axis represents the impurity concentration in the unit “cm ⁻³ ”. The horizontal axis uses a linear scale, and the vertical axis uses a logarithmic scale.

Reference is made to the graph showing the CV characteristics of Sample 1 in the upper part of FIG. A clear upward curve (relationship where 1 / C ² increases as the voltage increases) is obtained. This indicates that Sample 1 has n-type conductivity. It is considered that Cu deposited on the ZnO layer functions as a donor and does not function as an acceptor. Note that the slope corresponds to the resistance value.

Reference is made to the graph showing the depth profile of Sample 1 in the lower part of FIG. The impurity concentration (donor concentration) _{N d} before annealing the sample of the sample 1 is ^{2.5 × 10 19 cm -3 ~3.5 ×} 10 approximately 20 ^{cm -3.}

  On the other hand, in the samples 2, 3 and 4 before annealing, the IV characteristics approached ohmic characteristics (that is, no Schottky junction was obtained), and the impurity concentration could not be calculated by CV measurement. . This is presumably because the impurity concentration was further increased as compared with Sample 1.

  Next, each sample was subjected to an annealing process (which can realize impurity activation), and it was examined whether or not p-type conversion was performed. In a sample in which Ga and Cu are co-doped, it has been found that there is a possibility of showing n-type in an as-grown state and showing p-type conductivity when annealed. Even in this sample doped only with Cu, there is a possibility of inversion from n-type to p-type.

6A, 6B (6B1, 6B2), 6C, and 6D are graphs showing CV characteristics and depth profiles of impurity concentrations of samples 1, 2, 3, and 4 after annealing. Even if the sample 1 shown in FIG. 6A was annealed, the 1 / C ² graph was in the upper right and was n-type. That is, it is not made p-type. Sample 2 shown in FIGS. 6B1 and 6B2 exhibits a high resistance characteristic depending on the location, or a p-type characteristic. It does not show certain characteristics, suggesting an unstable state. The acceptor concentration in the p-type portion was 1.0 × 10 ^{19 to} 4.5 × 10 ¹⁹ (/ cm ³ ). Sample 3 shown in FIG. 6C was made p-type by annealing at 630 ° C. for 10 minutes. The CV characteristic in the upper stage shows clear right-sloping p-type conductivity. The depth profile in the lower part of FIG. 6C shows an acceptor concentration of about 0.9 × 10 ^{18 to} 4.5 × 10 ¹⁸ (/ cm ³ ). Sample 4 shown in FIG. 6D was also made p-type by annealing at 630 ° C. for 10 minutes. The CV characteristic in the upper stage shows clear right-sloping p-type conductivity. The depth profile in the lower part of FIG. 6D shows an acceptor concentration of about 0.3 × 10 ^{18 to} 6.5 × 10 ¹⁸ (/ cm ³ ).

  It was found that ZnO doped with 7.0% (positive element atomic ratio) of Cu (sample 3) and ZnO doped with 10.6% (positive elemental ratio) (sample 4) were converted to p-type by annealing. It was. ZnO doped with 4.3% (positive element ratio) of Cu (sample 2) shows p-type characteristics by annealing depending on the location, but there is a region that does not become p-type and shows stable characteristics. Has not reached.

  Simulations were performed with reference to these experimental results. A simulation was performed when Cu was doped into the crystal structure of ZnO and two Cu atoms entered the Zn site. The case where 2Zn atoms of 36Zn atoms are substituted with Cu was examined. At this time, the Cu concentration by the atomic ratio of the positive element is [Cu] / ([Zn] + [Cu]) = 2/36 = 5.6%.

  FIG. 7A is a plan view schematically showing a configuration of a six-membered ring. A ZnO crystal has a 6-membered ring structure in which Zn atoms and O atoms are alternately coordinated.

FIG. 7B schematically shows the structure of the ZnO crystal. Above the layer L _i of the Zn atoms and O atoms include a 6-membered ring structure bonded alternately, a layer L _{i + 1} include similar 6-membered ring structure are arranged. The position of the atoms in the layer is up and down. The average position is called an average layer. In the layer, Zn atoms are arranged above the average layer, and O atoms are arranged below the average layer. 6-membered ring of the upper layer L _{i + 1} has been positioned above the 6-membered ring of the lower layer L _i, are interchanged positions of Zn atoms and O atoms. And lower layer L _i O atom average lower than the upper Zn atoms and upper average layer of the layer L _{i + 1} from layer faces up and down to form a chemical bond between Zn and O atoms . The Zn atoms above the average layer of the upper layer L _{i + 1} form a chemical bond with the O atoms coordinated below the average layer of the upper layer L _{i + 2} . When Cu atoms are incorporated into ZnO crystals as added atoms (dopants), they occupy a certain Zn position in the host ZnO and form a chemical bond with the nearest O atom as Zn position substitution.

  In this case, the ZnO crystal is mainly divided into the following three different cases. In Zn at a certain Zn position in the host ZnO, (a) When the original Zn atom substituted by Cu atom stays in the vicinity, that is, stays in the crystal (this is called “interstitial atom”) There are two cases: (b) diffusion in the crystal, resulting in reaching the thin film surface and depositing from the surface. In addition, (c) Zn is already vacant at a certain Zn position in the host ZnO (this is generally referred to as “Zn vacancy”) and occupies the Zn vacancy position in the process of diffusion. There is a third case in which a chemical bond is formed with the adjacent O atom. Here, cases (b) and (c) are assumed.

  According to the analysis, there is no interaction between the Cu atoms diffusing in the ZnO crystal. As a result, when Cu is isolated from each other and substituted with Zn, the conductivity is estimated to be extremely low. This is because in this case, holes are trapped at the O atom position coordinating Cu along the c-axis, and the free movement along the electric field generated as a result of applying a voltage from the outside is limited. Because it is done.

  On the other hand, when two adjacent Zn atoms forming a chemical bond via an O atom in the same 6-membered ring are replaced with two Cu atoms, Cu is arranged along the c-axis for each Cu atom. The trapping of holes at the respective O atom positions is the same as in the case of the above-described isolated Zn-substituted Cu, but in addition to this, the repulsive force between the positive charges of the holes causes an electric field. From the disturbance, it is easily determined to limit the free movement of holes along the electric field. As a result, enhancement of p-type conductivity would not be expected.

  FIG. 7C shows a case where, in two adjacent layers, two Zn atoms that are different from each other are replaced by Cu atoms, and these two Cu atoms are adjacent via a first adjacent O atom. According to the analysis, when this Cu—O—Cu chemical bond is formed, an antibonded state is generated in the space region connecting the two Cu atoms. And the energy of the antibonding state exists in the band gap of ZnO which is a host material.

  In general, when two atoms form a chemical bond, the outermost electrons of the two atoms are divided into an electron that contributes only to the chemical bond and an electron that does not contribute to the chemical bond. Each of the bisected electronic states (both spatial distribution and corresponding energy distribution) is called a bonded state and an antibonded state. When the two Cus form a chemical bond, analyzing the result of theoretical calculation, the bond state is energetically stable, contributes to the chemical bond, and is not in a moving state, so it is located inside the valence band. I found out.

  On the other hand, from the results of theoretical calculations, it was found that the antibonded state has unique characteristics. The repulsive force between the electrons in the antibonded state is sufficiently small, and when the electrons in the bonded state absorb energy given from the outside and are excited thereby, they can receive the electrons. In this case, electrons are deficient in the bonded state, and holes, which are carriers having a positive charge, are generated. That is, when two Cu atoms in adjacent layers are adjacently coordinated via an O atom to form a chemical bond, an electron cloud EC connecting the two Cu atoms is generated, and p-type conductivity will be generated. This becomes clear by theoretical analysis. This coordination presupposes a model with a Cu concentration of 5.6% in terms of the atomic ratio of the positive element.

  In the tested samples 1, 2, 3, and 4, p-type conductivity was obtained in the sample 3 having a Cu concentration of 7% and the sample 4 having a Cu concentration of 10.6%, and the sample 2 having a Cu concentration of 4.3%. Although p-type conductivity is locally exhibited, there is a high-resistance region that does not exhibit p-type conductivity and is unstable. In ZnO doped only with Cu, it is considered that a Cu concentration of 5% or more is necessary to obtain a practical p-type conductivity. Also, from the viewpoint of crystal growth, although epitaxial growth has been obtained up to Sample 4 with a Cu concentration of 10.6%, it has shifted to three-dimensional growth, and if the concentration is further increased, polycrystallization (ring pattern) It is predicted that The polycrystalline Cu-doped ZnO film does not change to a single crystal even if it is subsequently annealed. Therefore, in order to obtain a single crystal p-type film, it is considered that the Cu concentration is preferably 15% or less, more preferably 10% or less.

  In order to obtain such a high-concentration Cu-doped ZnO crystal, it is preferable to form a Cu-doped ZnO crystal layer by the above-described alternate growth of the ZnO layer and the Cu layer, and to perform an annealing treatment. Since a p-type ZnO-based semiconductor layer can be realized, a semiconductor element such as a light emitting diode can be manufactured.

FIG. 8 is a schematic cross-sectional view of a ZnO based semiconductor light emitting device. On the Zn surface ZnO substrate 1, the growth temperature is 300 ° C., the Zn deposition rate F _Zn is 0.15 nm / s (flux J _Zn = 9.9 × 10 ¹⁴ atoms / cm ² s), and the O radical beam irradiation conditions are as follows. The ZnO buffer layer 2 having a thickness of 30 nm is grown at an RF power of 300 W and an O ₂ flow rate of 2.0 sccm (flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s). In order to improve the crystallinity and surface flatness of the buffer layer, annealing is performed at 900 ° C. for 10 minutes.

On the ZnO buffer layer 2, Zn, O and Ga are simultaneously supplied at a growth temperature of 900 ° C. to grow an n-type ZnO layer 3 having a thickness of 150 nm. Zn deposition rate F _Zn is 0.15 nm / s (flux J _Zn = 9.9 × 10 ¹⁴ atoms / cm ² s), O radical beam irradiation conditions are RF power 250 W, O ₂ flow rate 1.0 sccm (flux J _O = 4.0 × 10 ¹⁴ atoms / cm ² s) and the Ga cell temperature is 460 ° C. The Ga concentration of the n-type ZnO layer 3 is 1.5 × 10 ¹⁸ cm ⁻³ .

On the n-type ZnO layer 3, the growth temperature is 900 ° C., the Zn deposition rate F _Zn is 0.03 nm / s (flux J _Zn = 2.0 × 10 ¹⁴ atoms / cm ² s), and the O radical beam irradiation condition is An undoped ZnO active layer 4 having a thickness of 15 nm is grown at an RF power of 300 W and an O ₂ flow rate of 2.0 sccm (flux J _O = 8.1 × 10 ¹⁴ atoms / cm ² s).

  The substrate temperature is lowered to 250 ° C., and a Cu-doped ZnO layer 5 is grown as a p-type layer on the undoped ZnO active layer 4. For example, the ZnO growth time in one ZnO growth step is 10 seconds, and the Cu supply period in one Cu deposition step is 50 seconds. The process which made the ZnO growth process and the Cu adhesion process one set is repeated 60 sets to grow a Cu-doped ZnO layer 5 having a thickness of 145 nm. The same process as the manufacturing process of the sample 3 may be sufficient. It was made p-type by annealing at 630 ° C. for 10 minutes in an oxygen atmosphere.

  Thereafter, the n-side electrode 6n is formed on the back surface of the ZnO substrate 1, the p-side electrode 6p is formed on the Cu-doped ZnO layer 5, and the bonding electrode 7 is formed on the p-side electrode 6p. The n-side electrode 6n is formed by laminating a 500 nm thick Au layer on a 10 nm thick Ti layer. The p-side electrode 6p is formed by laminating an Au layer having a thickness of 10 nm on a Ni layer having a size of 300 μm □ and a thickness of 1 nm, and the bonding electrode 7 is an Au layer having a size of 100 μm □ and a thickness of 500 nm. Form. In this manner, a ZnO-based semiconductor light emitting element is manufactured.

Note that the active layer may have a multiple quantum well structure. Although the case where Cu is doped into ZnO has been described, ZnO and Mg _x Zn _1-x O (0 ≦ x ≦ 0.6) can perform substantially the same crystal growth. Therefore, it may be applicable to the growth of Cu-doped Mg _x Zn _1-x O (0 ≦ x ≦ 0.6).

In addition, although O radical generated from the plasma of oxygen gas was used as the oxygen source, not only oxygen gas but also other gases having strong oxidizing power such as polar oxidants such as ozone, H ₂ O, and alcohol may be used. It is considered possible.

  Although the present invention has been described with reference to the embodiments, the present invention is not limited thereto. It will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

1 ZnO substrate, DL Cu layer,
2 ZnO buffer layer, 101 vacuum chamber,
3 n-type ZnO layer, 102 Zn source gun,
4 undoped ZnO active layer, 103 O source gun,
5 Cu-doped ZnO layer, 104 Mg source gun,
6n n side electrode, 105 Cu source gun,
6pp side electrode, 106Ga source gun,
7 Bonding electrodes, 107 stages,
11 ZnO substrate, 108 substrate,
12 ZnO buffer layer, 109 film thickness meter,
13 undoped ZnO layer, 110 RHEED gun,
15 Alternating stack: 111 screens.
ML ZnO layer,

Claims (2)

On the growth substrate, ZnO-based semiconductor layers and Cu layers are alternately grown, and a single Cu element is doped to form a ZnO-based semiconductor epitaxial layer doped with Cu in an atomic ratio of 5% or more. ,
Annealing the ZnO-based semiconductor epitaxial layer doped with 5% or more Cu to obtain p-type conductivity;
A method for producing a Cu-doped p-type ZnO-based semiconductor crystal layer ,
The growth of the ZnO-based semiconductor layer is performed by molecular beam epitaxy having a period of supplying only Zn before and after a period of supplying Zn and O to grow a ZnO-based semiconductor layer,
The Cu layer is grown by molecular beam epitaxy for supplying Cu.
A method for producing a Cu-doped p-type ZnO-based semiconductor crystal layer . 2. The method for producing a Cu-doped p-type ZnO-based semiconductor crystal layer according to claim 1 , wherein the Cu layer is grown by supplying Cu in an amount that provides a thickness of one atomic layer or less of Cu per one time.