JPH06216165A - Manufacture for semiconductor device - Google Patents

Abstract

PURPOSE:To manufacture a P-type semiconductor crystal of a II-VI compound with excellent reproducibility by alternately feeding a group II raw material and a group VI raw material and nitrogen raw material by a plurality of times, and performing nitrogen doping, when the crystal thin film of a II-VI compound semiconductor is formed on a substrate by epitaxial growing. CONSTITUTION:The growth of an undoped ZnSe layer, wherein nitrogen doping is not performed, and the irradiation of nitrogen plasma are alternately repeated until the total growing-layer film reaches 2mum. Thus, planar doping is realized. A substrate 3 is made of P<+> type GaAs (100). The control of the growing layer and the doping layer is performed by opening and closing shutters 4a-6a of cells 4-6 by using a sequencer. Thus, the doping efficiency of nitrogen is improved, and the P-type II-VI compound semiconductor crystal having the high effective acceptor concentration can be obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for forming a p-type crystal thin film by doping a nitrogen material when a II-VI group compound semiconductor crystal thin film is epitaxially grown on a substrate. Is.

[0002]

2. Description of the Related Art II-VI group compound semiconductors such as ZnSe and ZnS have been attracting attention as materials that can be applied to blue light emitting devices and have been researched. However, promising results have not been obtained until recently. I couldn't do it.

When manufacturing such a blue light emitting device, a pn junction structure is indispensable in order to realize highly efficient light emission, and these II-VI group compound semiconductor materials show an n-type conductivity type. This is because although the product can be obtained with good reproducibility, it was difficult to manufacture a product exhibiting the p-type conductivity type.

The reason is that the self-compensation effect is large in the crystal growth of II-VI group compound semiconductors. This is a remarkable effect with respect to a semiconductor crystal having a wide forbidden band, and is an effect that defects are easily generated and electrically active carriers existing in the sample are canceled.

On the other hand, in parallel with the study of the crystal growth technique of such a thin film, the study of additive impurities for producing p-type ZnSe is carried out, such as nitrogen, phosphorus, arsenic, oxygen and even alkali metals. Of the elements shown in the periodic table, almost all possible elements have been studied.

In particular, nitrogen has been considered as a promising impurity material for producing p-type ZnSe. In order to take in this as an effective impurity, it must be taken in as a nitrogen atom, but since this nitrogen forms a stable gas (N ₂ ) as is well known, it must be decomposed and taken into the sample. , This was difficult.

Therefore, in the MBE (Molecular Beam Epitaxy) method, it was reported that nitrogen doping was successful using a high frequency plasma generator. In this doping method, nitrogen is excited by plasma to be transported to the surface of the substrate, and N ₂ is efficiently decomposed into N on the surface of the substrate to be efficiently incorporated into the crystal as an atom. It is a method that can be manufactured simply by attaching the generator to a normal MBE crystal growth apparatus.

[0008]

However, the impurity concentration of p-type ZnSe produced by the nitrogen plasma doping method (uniform doping) in this epitaxial growth is still insufficient for practical device application. One of the reasons for this is that the activation rate of the nitrogen acceptor in the thin film (the ratio of the nitrogen element functioning as the acceptor in the doped nitrogen) is low.

An object of the present invention is to provide a method for producing a p-type semiconductor crystal of a II-VI group compound having a practically sufficient impurity concentration with good reproducibility.

[0010]

Therefore, in the manufacturing method of the present invention, when a II-VI group compound semiconductor crystal thin film is formed on a substrate by epitaxial growth, a group II source, a group VI source and a nitrogen source are used. Alternately supplied multiple times, or alternately supplied group II raw material, group VI raw material and nitrogen raw material multiple times,
Nitrogen doping is performed.

According to this manufacturing method, when the simultaneous supply of the group II raw material and the group VI raw material is switched to the supply of the nitrogen raw material, only one of the group II raw material and the group VI raw material is supplied to grow the growth surface. It is desirable to supply the above nitrogen raw material after coating with one of the raw material elements.

Further, the group II raw material, the group VI raw material, the nitrogen raw material in this order, or the group VI raw material, the group II raw material,
It is desirable to supply the raw material to the substrate a plurality of times alternately in the order of the nitrogen raw material.

Further, it is desirable to adjust the spacing width formed by the simultaneous or staggered supply of the group II raw material and the group VI raw material to obtain an optimum acsector concentration.

Further, it is desirable that the nitrogen raw material is a raw material consisting only of nitrogen or a raw material containing nitrogen and other elements.

The above-mentioned II-VI group compound is ZnS.
e, ZnS, ZnTe, CdS, CdSe, CdTe,
Desirably, it is composed of a compound containing one or more selected from MgS, MgSe, and MgTe.

[0016]

EXAMPLES Examples of the present invention will be described below. This example is a method for producing a II-VI group semiconductor crystal thin film by combining molecular beam epitaxy and a nitrogen plasma method.
In particular, it is characterized in that the nitrogen doping is performed by using the planar doping method.

FIG. 1 is a diagram schematically showing an apparatus for performing molecular beam epitaxial growth and plasma doping. Reference numeral 1 denotes a growth chamber that is evacuated from a sluice valve 2 to and from an exhaust device. And a Se cell 5 for irradiating the same. Reference numeral 6 is a plasma cell for irradiating the crystal growth surface of the substrate 3 with nitrogen plasma. 7 is a jig for holding the substrate 3, 8 is an electron beam irradiation source for observation (RHEEDGUN) for observing the crystal growth surface of the substrate 3.
) And 9 are similar phosphor screens for observation (RHEED SCREE
N). Reference numerals 4a, 5a, and 6a are shutters of each cell.

The planar doping method used in this embodiment is a method of selectively doping on a specific crystal surface in the film thickness direction by alternately repeating crystal growth and deposition of dopant. .

In the following, by applying this planar doping method to nitrogen doping, attempts are made to improve the doping efficiency and at the same time the results are compared with the results of normal doping (uniform doping).

Further, the behavior of the nitrogen acceptor is observed in the doping on the Zn plane and the Se plane, and the mechanism of the doping by the nitrogen plasma method is examined.

First, in the plasma cell 6, the operating power 22
Nitrogen gas (purity 6N, moisture content 0.1ppm or less) during high frequency (13.56MHz) electrodeless discharge at 0W constant
Was introduced to generate radicals, the shutter 6a was released, and the radicals were ejected from the aperture (diameter 0.2 mm: not shown) toward the substrate 3 in the growth chamber 1.

At this time, the flow rate of nitrogen gas was controlled by a VLV (variable leak valve) so that the back pressure in the growth chamber 1 was kept constant at 1.5 × 10 ⁻⁶ Torr. This back pressure corresponds to a nitrogen gas flow rate of about 0.15 sccm. In addition, the substrate temperature is kept constant at 240 ° C.
And the molecular beam intensity ratio of Zn (J _Se / J _Zn ) was set to 8.

In the nitrogen plasma at this time, as shown in the spectroscopic result of FIG. 2, the second positive band (2nd positive), which is the light emission due to the transition of the N ₂ molecular state, is 300 to 400 nm.
, The first positive band (1st positive) is 550 to 7
It was observed near the wavelength of 00 nm.

Under this growth condition, the growth rate of the ZnSe layer is about 1.4 Å / s, and the growing surface becomes the Se stabilizing surface.

Using such growth conditions, growth of an undoped ZnSe layer (4 to 40 Å) without nitrogen doping and nitrogen plasma irradiation are alternately performed until the total thickness of the grown layer reaches 2 μm. By repeating, planar doping was realized. In addition,
The substrate 3 is p ⁺ type GaAs (100).

The growth layer and the doping layer are controlled by using a sequencer using the shutters 4a, 5a of the cells 4, 5, and 6.
6a was opened and closed. Here, the timing of doping in the crystal film growth process is
Went with seeds.

Case (a): Doping to Zn Termination Surface Only (Planar Doping) Case (b): Doping to Se Termination Surface Only (Planar Doping) Case (c): Uniform Doping

First, the case (a) will be described (see FIG. 3). As described above, the growth of the undoped ZnSe layer is performed under the Se stabilizing condition, and therefore it is necessary to obtain the Zn stabilizing surface before the nitrogen plasma doping. Therefore, the cell shutter 5a of Se is closed to stop the growth, and the surface of the substrate 3 is irradiated with only the Zn molecular beam for the time t ₀ . As a result, the RHEE of the phosphor screen 9 for observation
Observation of the D pattern confirms the formation of a Zn stabilized surface. Thereafter, with the Zn cell shutter 4a kept open, nitrogen plasma irradiation is performed for a time t ₁ .
After this nitrogen plasma irradiation, only the Zn molecular beam is continuously irradiated for a time t ₂ until the growth of the ZnSe layer is restarted.
After that, the Se cell shutter 5a is opened, and the growth of the ZnSe layer is restarted for a time t ₃ under the Se stabilizing condition. Thus, in the case (a), the irradiation is performed with the Zn molecular beam being irradiated.

On the other hand, in the planar doping on the Se final surface of the case (b) (see FIG. 4), time t ₀
Was fixed to 1 second which is the minimum setting time of the sequencer. The reason for this is that under this growth condition, Se is always
This is because the stabilizing surface is formed and Se irradiation prior to doping is unnecessary.

The nitrogen flow rate used for doping and the operating power of the radical beam source were constant throughout all samples. In addition, t ₁ (nitrogen plasma irradiation time) and t ₃ (ZnSe layer growth time) are set to be equal in order to equalize the total number of N ₂ molecules reaching the growth surface for all samples including the uniformly doped film. did.

As the effective acceptor concentration of the sample, a value obtained from CV (capacity-voltage) measurement was used. This is p-type Z
This is because it is difficult to measure holes because the technique for forming an ohmic electrode for nSe has not been established.

The sample for CV measurement was on the growth film side (p-type Z
An Au vapor-deposited film (160 μm in diameter) was used as the Schottky electrode on the nSe side, and In (for attaching the substrate during growth) on the back surface of the substrate GaAs was used as the ohmic electrode.
A Schottky diode structure was formed in the vertical direction with respect to the growth surface. The measurement frequency was 1 MHz.

FIG. 5 shows the undoped ZnS of [Case (a)] when planar doping is performed on the Zn final surface.
The change in effective acceptor concentration with respect to the e-layer thickness (hereinafter abbreviated as "spacing width") is shown. The horizontal axis is the spacing width (THICKNESS) and the vertical axis is the effective acceptor concentration (N
a-Nd).

First, it can be seen that the effective acceptor concentration increases as the width decreases in the range of 40 to 7 angstroms on the horizontal axis. Then, when the spacing width is 7 angstroms, the effective acceptor concentration has reached the 4.5 × 10 ^{1 7} cm ^-3, the value is the highest value. The effective acceptor concentration for uniform doping ZnSe film when was doped under the same conditions ^{5 × 10 1 6 ~1 × 10} 1 7 cm
It was ^-3 .

From this, it is understood that the planar acceptor concentration on the Zn plane increases the effective acceptor concentration. In this experiment, the conditions were set so that the total number of nitrogen molecules reaching the substrate surface was equal, so this result reflects the improvement in nitrogen addition efficiency and carrier activation rate. It can be said to be a thing.

The reason why the nitrogen addition efficiency was improved is considered to be that the following effects were exerted on the Zn plane. (1) Increasing the sticking coefficient of nitrogen molecules to the growth surface (2) Accelerating the dissociation of adsorbed nitrogen molecules (N ₂ ) to the atomic state (N) Furthermore, Si planar doping (EF
Schubert, JECunningham and WTTsang: Solid Stat
e Commnu. 63 (1987) 591.), Ga planar doping of ZnSe (JLMiguel, SMShibil, MCTamargo)
and BJSkrimme: Appl.Phys.Lett. 53 (1988) 2065.), (3) Suppression of generation of defects causing carrier compensation.

Here, it can be seen that the effective acceptor concentration is increased by one digit or more with respect to the change of the spacing width in the range of 40 to 7 angstroms, showing a very strong dependence. The reason for this is not currently understood, but is believed to be an essential feature of planar doping. The spatial extent of the carrier gas produced by the dopant may also be relevant. That is, in planar doping, since the dopant is two-dimensionally confined in a very narrow region of about one atomic layer, the quantum effect cannot be ignored, and carrier spreading may be affected. To be

However, when the spacing width was further narrowed to 4 angstroms, the effective acceptor concentration decreased, and the carrier concentration became almost equal to that in the case of uniform doping. The spacing width in this case is a thickness of 2 atomic layers or less, and it can be said that the planar doping condition at this time is similar to the uniform doping condition. Therefore, it is considered that the effect of increasing the effective acceptor concentration in the planar doping as described above has disappeared.

On the other hand, it was found that the sample showed a very low carrier concentration when the planar termination was performed on the Se termination surface. In this case, the spacing width is 4, 7,
Samples of 40 Å and 40 Å were prepared, both of which had a very low depletion layer capacitance (0.3
pF). Moreover, the dependency on the applied voltage is not observed, and the 1 / C ² -V characteristic (1 / C for the voltage change is
^The diffusion potential obtained from (Characteristic ² ) was 20 V or more, which was clearly an abnormal value.

From this, it is expected that the sample is almost depleted over the entire area (about 2 μm) in the film thickness direction. Further, when estimating the effective acceptor concentration of these samples than the value of the depletion layer capacitance at zero Baaisu, 10 ^{1 5} cm ^-3
It was found to be a value in the middle or below. As described above, in the case where the Se-stabilized surface is subjected to the perena doping, compared with the case where the uniform doping in the case (c) or the planar doping on the Zn stable surface in the case (a) is performed. It has been revealed that the effective acceptor concentration is reduced by about two digits or more.

Next, the optical characteristics will be described. It was also confirmed that the effect of the doping surface remarkably appeared on this optical property. The measurement temperature of the emission spectrum by PL (photoluminescence) measurement is 15K. Figure 6
2 shows the PL spectrum of a ZnSe film doped with nitrogen plasma.

Here, (a) is a sample on which the planar surface of Zn is terminated, and (b) is a sample on which the planar surface of Se is terminated by nitrogen plasma, and (c) is a sample by uniform doping. is there. First, when planar doping is performed on the Zn termination surface of (a), extremely strong DAP (donor acceptor pair) emission is dominant, and exciton-related emission is not observed. on the other hand,
In the case of planar doping on the Se termination surface of (b), sharp and strong acceptor-bound exciton emission (I ₁ line)
Has become dominant. Also, weak DAP emission appears, and the intensity is about 1/5 of the I ₁ line. The waveform on the right side of FIG. 6B is an enlarged waveform of the I ₁ line.

The difference between the spectra in FIGS. 6A and 6B is that the planar doping on the Zn termination surface is more efficient in acceptor than the planar doping on the Se termination surface. It can be said that it has been taken in. This result is in very good agreement with the result obtained by the CV measurement.

The change in the characteristics due to the difference in the doping surface becomes clearer by paying attention to the DAP emission. The spectrum of FIG. 7 is the DAP of FIG.
The light emitting area is enlarged.

Planar is formed on the Zn termination surface of FIG.
In the case of doping, extremely strong DAP emission appears, and its zero phonon line is located at 2.680 eV. On the other hand, when planar doping is performed on the Se termination surface of (b), the peak position of the zero phonon line of DAP emission is observed at 2.696 eV, which is on the higher side of 20 meV energy than in the case of Zn surface doping.

Of further note is the superposition of two different DAP emissions observed in (c) uniform doping. In this case, the peak position of the zero phonon line is 2.680 eV, which is the same as in the case of planar doping on the Zn termination surface.
However, a shoulder is clearly recognized on the high energy side, and it can be seen that the position thereof is almost the same as the case of doping on the Se termination surface.

That is, in uniform doping, D
As for the AP emission, it can be seen that the superposition of two different DAP emission is observed when the planar doping is performed on the Zn plane and the Se plane. The cause of the appearance of these two types of DAP emission is not clear at this time.

In understanding the mechanism of nitrogen plasma doping, adsorption and dissociation of N ₂ molecules on the ZnSe surface is extremely important. Regarding the surface adsorption mechanism of nitrogen plasma, the theoretical calculation by Nakao et al. Has already been reported (T. Nakao and T. Uenoyama; Extended Abstracts of I
nt.cof. of Solid State Devices and Materials, Tsuku
ba, 1992, PP.336).

According to this report, molecular N ₂ (A ³ Σ _u ⁺ ) in the lowest excited state can be adsorbed by Zn atoms on the ZnSe (100) surface and can cause dissociation. . On the other hand, it is said that N ₂ molecules cannot obtain sufficient energy for stabilization on the Se atom. Thus, the uptake of dopants by the terminal states of the surface can be explained by this model.

However, the characteristic seen in FIG. 5, that is, the spacing width dependence of the effective acceptor concentration in planar doping cannot be explained by this model. Through the planar doping experiment, the nitrogen flow rate was kept constant, and the molecular beam intensity of nitrogen was sufficiently higher than the molecular beam intensities J _Zn and J _Se of _Zn and _Se .

Here, when the model of Nakao et al. Is applied to the planar doping, the N irradiated on a specific surface is
It is considered that the number of attachable molecules among the _two molecules is uniquely determined by the terminal surface regardless of the irradiation time.
That is, the doping amount on each doping surface is constant, and the acceptor concentration in the total growth layer is inversely proportional to the spacing width.

However, when the planar doping is actually performed on the Zn plane, the spacing width is set to 40.
By changing from angstrom to 7 angstrom, the acceptor concentration increased more than one digit.
This result is considered to suggest that there is a more complicated process in the planar doping using nitrogen plasma.

Considering these theoretical results, Zn
In order to realize a high acceptor concentration in the nitrogen doping of Se, it is considered that doping under Zn-rich conditions is necessary. However, as is well known, in order to obtain a ZnSe film having good crystallinity in ordinary MBE growth, the growth must be performed under the Se-rich condition. Therefore, the conventional M
In BE growth, it seems difficult to further improve the nitrogen doping efficiency.

As a means for overcoming this current situation, growth using Se cells equipped with a high temperature cracker may be mentioned. Cammack et al. Reported that a high quality undoped ZnSe film was successfully prepared under Zn-rich conditions by providing a high temperature cracker on the top of the Se cell (DACammack, K. Shahzad and T. Marshall; Ap
pl.Phys. Lett.56 (1990) 845.). Therefore, if this Se cell with a high-temperature cracker is used, the nitrogen doping efficiency can be improved without impairing the crystallinity, and the growth of a ZnSe film of high quality and high acceptor concentration can be expected.

As described above, planar doping is applied to nitrogen plasma doping of ZnSe,
It has been revealed that the selective doping on the Zn-terminated surface improves the nitrogen doping efficiency. Effective acceptor concentration by planar doping of Zn terminating plane is reached the maximum of 4.5 × 10 ^{1 7} cm ^-3, the value that the uniform doping ⁽¹ × 10 1 ⁷
well above the cm ^-3 ).

Further, in the PL spectrum, it was found that the strong DAP emission is dominant in all the samples doped on the Zn termination surface. In contrast, weak DAP emission was observed with planar doping on the Se termination surface. It has been clarified that these DAP light emission appear at positions greatly different depending on the doping surface, and this result is considered to suggest that there are different impurity incorporation processes depending on the state of the termination surface.

In the planar doping of the above-mentioned embodiment, both the Zn molecular beam and the Se molecular beam are simultaneously irradiated to the substrate to form the undoped ZnSe layer, and the Zn molecular beam and the S molecular beam are added.
Although the doped layer is formed by simultaneous irradiation of one of the e molecular beam and nitrogen plasma, the same undoped ZnSe layer can be obtained by alternately supplying the Zn molecular beam, the Se molecular beam, and the nitrogen plasma in any order. And a doped layer is formed. That is, at this time, the material is supplied alternately in the order of the Zn molecular beam, the Se molecular beam, and the nitrogen plasma, and vice versa, in the order of the Se molecular beam, the Zn molecular beam, and the nitrogen plasma.

Further, in the embodiment, the simple substance of nitrogen is used as the doping raw material, but a compound containing nitrogen as an element can be applied to the planar doping. Further, the epitaxial growth method for II-VI group compounds is not limited to molecular beam epitaxy, and nitrogen doping is not limited to plasma doping. Further, when the terminal surface of the undoped ZnSe layer is coated with Zn, the Zn gas raw material or the metal material can be used. When the Se coating is performed, the Se gas raw material or the metal material is used. You can also do it. Furthermore, II-VI for forming a thin film by applying the present invention
Group compounds include ZnS, ZnT, in addition to ZnSe.
e, CdS, CdSe, CdTe, MgS, MgSe,
There are compounds containing one or more selected from MgTe. Furthermore, the substrate is not limited to GaAs.

[0059]

As described above, according to the present invention, a p-type II-VI group compound semiconductor crystal having a high effective acceptor concentration with improved nitrogen doping efficiency in epitaxy growth of a II-VI group compound semiconductor crystal is obtained. There is an advantage that it becomes possible to obtain.

[Brief description of drawings]

FIG. 1 is a schematic diagram of a molecular beam epitaxy plasma doping apparatus.

FIG. 2 is a diagram showing optical characteristics of nitrogen plasma.

FIG. 3 is an explanatory diagram of a planar doping sequence on a Zn termination surface.

FIG. 4 is an explanatory diagram of a planar doping sequence on the Se termination surface.

FIG. 5 is a characteristic diagram of the effective acceptor concentration with respect to the spacing width in planar doping on the Zn termination surface of ZnSe.

FIG. 6 ZnSe in nitrogen planar doping
FIG. 4 is a PL spectrum diagram of the film.

FIG. 7 is an enlarged view of a DAP region of the PL spectrum diagram of FIG.

[Explanation of symbols]

1: Growth chamber, 2: Gate valve, 3: Substrate, 4: Zn cell,
5: Se cell, 6: Plasma cell, 7: Substrate holding jig,
8: Observation electron beam irradiation source (RHEED GUN), 9: Observation phosphor screen (RHEED SCREEN).

Claims (6)

[Claims] 1. When forming a II-VI group compound semiconductor crystal thin film on a substrate by epitaxial growth, a group II raw material and
Alternately supplying Group VI raw material and nitrogen raw material multiple times, or II
A method for manufacturing a semiconductor device, which comprises supplying a group raw material, a group VI raw material, and a nitrogen raw material alternately a plurality of times to perform nitrogen doping. 2. When switching from the simultaneous supply of the group II source and the group VI source to the supply of a nitrogen source, only one of the group II source and the group VI source is supplied and the growth surface The method for manufacturing a semiconductor device according to claim 1, wherein the nitrogen source is supplied after coating with the element of the source. 3. The raw material is supplied to the substrate a plurality of times in the order of the group II raw material, the group VI raw material, the nitrogen raw material, or alternately in the order of the group VI raw material, the group II raw material, and the nitrogen raw material. The method for manufacturing a semiconductor device according to claim 1, wherein: 4. The optimum acsector concentration is obtained by adjusting a spacing width formed by supplying the group II raw material and the group VI raw material simultaneously or in a staggered manner. A method of manufacturing a semiconductor device according to item 1. 5. The method of manufacturing a semiconductor device according to claim 1, wherein the nitrogen source is a source consisting only of nitrogen or a source containing nitrogen and another element. 6. The above II-VI group compound is ZnSe,
ZnS, ZnTe, CdS, CdSe, CdTe, Mg
3. A compound containing one or more selected from S, MgSe, and MgTe.
The method for manufacturing a semiconductor device according to 3, 4, or 5.