Method and device for manufacturing semiconductor compound materials by means of vapour phase epitaxy
09856579 ยท 2018-01-02
Assignee
Inventors
Cpc classification
Y10T428/24612
GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
C30B25/00
CHEMISTRY; METALLURGY
C30B25/14
CHEMISTRY; METALLURGY
International classification
C30B25/00
CHEMISTRY; METALLURGY
C30B25/14
CHEMISTRY; METALLURGY
Abstract
A semiconductor compound material, preferably a III-N-bulk crystal or a III-N-layer, is manufactured in a reactor by means of hydride vapor phase epitaxy (HVPE), wherein in a mixture of carrier gases a flow profile represented by local mass flow rates is formed in the reactor. The mixture can carry one or more reaction gases towards a substrate. Thereby, a concentration of hydrogen important for the reaction and deposition of reaction gases is adjusted at the substrate surface independently from the flow profile simultaneously formed in the reactor.
Claims
1. A III-N bulk crystal as grown having a diameter of 5 cm (2 inches) or more, the III-N bulk crystal comprising a thickness layer homogeneity with a standard deviation of 7% or less relative to the layer thickness, wherein a surface of the crystal is mapped at multiple measurement points arranged across the surface at a pitch of 5 mm in each of two directions perpendicular to each other and an exclusion from the substrate edge of 4 mm is applied, and wherein in a rocking curve-mapping of an area parallel to the growth plane, and including 100 measurement points arranged in corresponding directions, the standard deviation of the full-width half-mean measured amounts to 3% or less.
2. The III-N bulk crystal according to claim 1, wherein in a micro-Raman-mapping of an area parallel to the growth plane, and including 100 measurement points arranged in corresponding directions, the standard deviation of the full-width half-mean being measured of the E2-phonon amounts to 3% or less.
3. A III-N-single crystal, obtained by separation of the III-N bulk crystal according to claim 1.
4. A III-N bulk crystal according to claim 1, further comprising constituents of a noble gas built in the crystal.
5. A III-N bulk crystal according to claim 1, further comprising constituents of a noble gas absorbed in the crystal.
6. A III-N bulk crystal according to claim 1, further comprising constituents of a chemically inert gas absorbed in the crystal.
7. A III-N bulk crystal according to claim 4, wherein the noble gas is helium or argon.
8. A III-N bulk crystal according to claim 5, wherein the noble gas is helium or argon.
9. The III-N bulk crystal according to claim 2, wherein the standard deviation of the full-width half-mean of the E2-phonon amounts to 2% or less.
10. A III-N bulk crystal as grown having a diameter of 5 cm (2 inches) or more, prepared by a hydride vapor phase epitaxy (HVPE) process in a vertical reactor, the process comprising the steps of: (a) introducing a mixture of carrier gases into the vertical reactor via concentrically arranged gas inlets to carry one or more reaction gases in a direction towards a substrate; (b) determining a volume flow of hydrogen as a first carrier gas in the mixture to set a predetermined value for the concentration of hydrogen at a point near a surface of the substrate; and (c) compensating for an influence exerted from step (b) on at least one of the concentrically arranged gas inlets in a laminar flow profile represented by a distribution of local mass flow rates in each of said concentrically arranged gas inlets in the reactor by determining corresponding volume flow proportions of a second and a third carrier gases in the mixture in at least another one of the concentrically arranged gas inlets.
11. A III-N-bulk crystal as grown having a diameter of 5 cm (2 inches) or more, prepared by a vapor phase epitaxy process in a reactor having a number of gas lines for introducing carrier and/or reaction gases, the process comprising: introducing three carrier gases into the reactor using the gas lines; wherein each gas line introduces at least one carrier gas, hydrogen as a first carrier gas and nitrogen as a second carrier gas are introduced in at least one of the gas lines simultaneously, at least one of the gas lines is a separation line in which at least a third carrier gas is introduced without introducing reaction gas, and which separates two reaction-gas-carrying gas lines from one another, and the third carrier gas has a specific gas weight different from that of hydrogen and nitrogen, wherein when the third carrier gas is introduced into the separation line, the third carrier gas is not introduced as a carrier gas into the two reaction-gas-carrying gas lines separated from each other by the separation line, and wherein proportions of the three carrier gasses in each of the gas lines are selected to maintain homogenous local mass flow rates to effect a laminar flow of the gasses in the reactor.
Description
(1) The invention will now be explained in more detail by means of embodiments with reference to the drawings. Therein:
(2)
(3)
(4)
(5)
(6) In the embodiments, an HVPE-arrangement is modified, such that at least one of the carrier gas lines is adjusted to carry a mixture of hydrogen, nitrogen and argon with respective volume stream percentages ranging from 0% to 100%.
(7)
(8) Prior to the growth procedure, the start or seed substrate 16 is provided. The GaN-substrate has for example a diameter of more than 5 cm at a (0001)-orientation, or at a slight inclination of the substrate surface against the exact (0001)-plane.
(9) The HVPE-arrangement 20 includes a horizontal glass reactor 21 according to one possible embodiment, further a multi-zone furnace 22 enclosing the reactor, gas inlets 23, 23 indicated by arrows and a pump and discharge system 24 also indicated by an arrow.
(10) The GaN-substrate 16 is introduced into the reactor 21 through the load-unload flange 25 and positioned on a substrate holder 26. Using the pump and discharge system 24, the reactor is brought to the desired process pressure, suitably in the range of 1.000 mbar, for example to a pressure of about 950 mbar.
(11) The multi-zone furnace has a first zone 22A, with which the growth temperature of the surface of the substrate is set, and a second zone 22B, with which the temperature in the area of a Ga-tray 28 is set. Via the gas inlets 23, 23 Ar, H.sub.2 and N.sub.2 as carrier gases are inlet into the reactor. The gas inlets 23, 23 each define one carrier gas line.
(12) For the purpose of an in situ generation of gallium chloride (GaCl), gallium (Ga) provided in the Ga-tray is heated to, e.g., 850 C. by adjusting a suitable temperature in the zone 22B of the multi-zone furnace 22. The gallium is then processed with hydrogen chloride (HCl), which is supplied from the gas supply 23 together with the Ar/H.sub.2/N.sub.2-carrier gas in a suitable gas mixture ratio and under a suitable flow rate.
(13) The gallium chloride (GaCl) generated in situ flows from the openings indicated in
(14) The chemical reaction of gallium chloride with ammonia (NH.sub.3) takes place in order to yield gallium nitride (GaN). Further, hydrogen chloride (HCl) as well as hydrogen (H.sub.2) are generated as side products, which leave the system together with the other carrier gases via the pump and discharge system 24.
(15) At the surface of the substrate 16 the hydrogen generated as stated above contributes to a local concentration of hydrogen in conjunction with that hydrogen, which has been supplied as a carrier gas. This may influence the local behaviour of the reaction sustainably, since it increases for example the surface mobility of the species. Consequently, the volume flow rate of hydrogen introduced through both gas lines according to this embodiment may be adjusted to an optimised value, which may be experimentally determined. Such determination may for example be carried out by investigations of the surface morphology as provided in Habel (see above). The determined value of the volume flow portion necessary to achieve a target concentration of hydrogen will generally have to be derived for each reactor separately.
(16) The local concentration of hydrogen is mainly determined by the supply of the carrier gas and only to a lesser degree by hydrogen as a side product of the reaction.
(17) The volume flow rate of each gas line can be different from next one. This also depends on the flow profile in the reactor 20. In case the flow profile differs, or if the mass flow rates of both gas lines are not adapted to each other, then the mixture ratios of the other two carrier gases Ar and N.sub.2 are adapted to each other for each of the carrier gas lines associated with the gas inlets 23, 23. The mixture ratios can particularly be different, such that the modification in the flow profile due to the adoption of the volume flow of H.sub.2 is just compensated.
(18) As can be seen from the temperature profile in the bottom part of
(19) If for example a (Ga, Al, or In)N, a (Ga, Al)N or a (Ga, In)N-layer is to be deposited instead of a GaN-layer, further trays for Al and/or In have to be provided in the HVPE-arrangement 20. The inflow of corresponding aluminium and/or indium chloride into the reactor is effected by supply of HCl in a suitable carrier gas of for example H.sub.2/N.sub.2, analogously to what is shown in
(20)
(21) After achieving the growth temperature, the III-N-crystal growth (i.e., of GaN) is started by connecting the supply of the group-III-source material (i.e., Ga). This means for example in the case of the GaN-bulk crystal growth, that hydrogen chloride gas is led over the gallium-source, wherein gallium chloride gas is yielded as in the first embodiment.
(22) Prior to the start of the III-N-crystal growth, a suitable composition of hydrogen, nitrogen and argon is adjusted in the carrier gas lines. This step will generally last until stable gas flows have been formed. The corresponding compositions in the carrier gas lines are chosen on the one hand, such that at the substrate surface a concentration of hydrogen is provided which may achieve an optimum crystal quality.
(23) On the other hand the respective portion of argon within the carrier gas lines ensures that a balancing of the respective momentum flows leads to a homogenous distribution of growth rates on the substrate. Therein, the effect of an additional use of argon is based on the fact that argon is heavier than nitrogen, whereas hydrogen is lighter as nitrogen. As a result the momentum flows can be balanced by varying the flow of argon with the proportion of hydrogen.
(24) In addition to the connection to the supply of group-III-source materials (herein Ga) corresponding source materials intended for doping may selectively be connected for supply.
(25) A modified embodiment of the AIXTRON HVPE-reactor will now be described: as a substrate 7 in this embodiment a GaN-substrate having a diameter between 50 and 60 mm may be chosen. The reactor is constructed such that not only in the separation line, which separates the ammonia and gallium chloride line from each other, but also in the peripheral wall purge line an arbitrary mixture of the three carrier gases hydrogen, argon and nitrogen can be adjusted. More specifically, the adjustment may concern each of the different volume flow rates, or the mass flow rates, respectively. The HVPE-process is carried out for example at a temperature of about 1.050 C. and at a pressure of 800 mbar with a V/III-mixture ratio of about 35.
(26) The carrier gas flow in the separation line (gas inlet 2) amounts to 50 Vol.-% of hydrogen, 30 Vol.-% of argon and of 20 Vol.-% of nitrogen. The carrier gas flow in the peripheral wall purge line (gas inlet 4) is composed as follows: 53 Vol.-% hydrogen and 47 Vol.-% nitrogen. The mean growth rate amounts to about 220 m/h and is determined by layer thickness mapping applying ellipsometry.
(27) The layer thickness distribution has in this experiment been determined by mapping the surface by means of an in situ-ellipsometer M2000 by J. A. Woollam Co., Lincoln, Nebr., U.S.A. The pitch of measurement points in each two directions perpendicular with respect to each other on the surface amounted to about 5 mm. Therein, from each spectrum the ellipsometric parameters Psi and Delta have been used to determine the layer thickness by means of model simulation (cf. H. G. Thomkins et al. in Spectroscopic Ellipsometry and Reflectrometry, Wiley, New York, 1989).
(28) The homogeneity of the distribution of growth rates further determined thereby is defined by the statistical standard deviation of the measured layer thicknesses. The standard deviation amounts to less than 10% according to the invention, in one embodiment even less than 7%with consideration of an edge exclusion area of 4 mm. Particularly note-worthy is the degree of a crystallinity yielded simultaneously with the layer thickness homogeneity, which will be detailed below.
(29) In the concentric, vertical construction as shown in
(30) In a reverse case, a third, lighter gas can be provided in the peripheral wall purge line as a carrier gas.
(31) The effect of argon may be explained as follows: by the larger mass of argon as compared with nitrogen, the momentum flow of an outer nitrogen dominated gas line can be compensated by adding argon to hydrogen in an inner gas line. Consequently, a constant homogeneous growth rate profile may be achieved.
(32) Basically, this would also be accomplished by varying only the ratio of N.sub.2 to H. However, this ratio is fixed (pinned), since the H.sub.2-concentration, or the ratio relative to N.sub.2 at the surface of the substrate, has to attain a predetermined value in order to obtain an optimum crystal quality.
(33) An exemplary sequence of steps according to the method of the invention is shown in
(34) On the right side of
(35) The simulations may also include the determination of the volume flow proportions of N.sub.2 and Ar. The basic principle is to achieve a flow profile as a target value. By adjusting the volume flow proportions of N.sub.2 and Ar (e.g., as compared with standard settings for these values) which may represent free parameters, a deviation from the target profile due to the calculated H.sub.2-volume flow proportion may be compensated. The dashed lines in
(36) The determination of the concentrations and flow profiles by experiment and measurement and the adjustment of the volume flow portions responsive thereto is basically encompassed by the invention.
(37) In a further embodiment of the invention, a III-N-substrate having a c-, a-, m- or r-plane as the growth plane is used as the substrate, and the III-N-bulk crystal is deposited on the selected growth plane.
(38) In a further embodiment of the invention a III-N-substrate having a growth plane is used, which is inclined relative to the c-, a-, m- or r-plane by 0.1-30, wherein the III-N-bulk crystal is deposited thereupon.
(39) In a still further embodiment of the invention, a doped III-N-substrate is used as the substrate. On the substrate a doped III-N-bulk crystal is deposited, wherein as dopants each one element is employed selected from the group comprising silicon, tellurium, magnesium and iron.
(40) According to the invention single crystalline sapphire, silicon carbide, gallium arsenide, lithium aluminate or silicon can further be used as the substrate. The III-N-bulk crystal is then deposited on this substrate. In a particularly preferred embodiment GaN-substrate is used as the substrate in free-standing form or in template-form, and a GaN-bulk crystal is then grown according to the invention.
(41) After the method according to the invention has been carried out, freestanding III-N-crystal substrates can be manufactured in a simple manner by separating one or more III-N-substrates from the III-N-bulk crystal. A particularly suited method for separation is represented by wire-sawing. Adjacently, further processing steps may follow, particularly lapping, polishing, thermal post processing and/or arbitrary final cleaning steps.
(42) According to the invention there is further provided a III-N-bulk crystal, which is available according to the method of the invention. Preferably, the crystal is manufactured according to the methods as provided in the appended claims or in the present embodiments. In conjunction with the bulk crystal also the single crystal obtained by separation due to wire- or inner hole-sawing inherits the excellent characteristics with concern to crystallinity and layer thickness homogeneity.
(43) A correspondingly manufactured III-N-crystal surprises by its unique crystallinity. This also holds true for the correspondingly separated, free-standing III-N-substrates. Particular emphasis has to be laid on the effect, that the crystallinity is obtained simultaneously and in combination with a high layer thickness homogeneity. As explained above with reference to the embodiments, the layer thickness homogeneity according to the invention amounts to less than 10%, in one embodiment even less than 7% in consideration of an edge exclusion area of 4 mm.
(44) The crystallinity of the crystal can be defined particularly by means of rocking-curve mapping and/or by micro-raman-mapping.
(45) Therein, the spatial distribution of absolute positions or the full-widths half-means of the X-ray diffraction curves corresponding to the diffraction at predetermined grid plane bands is recorded for example by means of X-ray diffraction in metrological manner. The homogeneity of the crystal quality (crystallinity) in the growth plane may in the case of rocking-curve mapping be examined by recording of -scans at distinct probe locations. The co-scan is recorded within a plane parallel to the growth plane. In the case of growth in [0001]-direction the reflection of the (0002)-grid planes can be used for the co-scans.
(46) The homogeneity of the crystal quality in the growth direction, however, may be determined by means of the standard deviation of the mean values of the full-width half-means of the (0002) co-scans of individual substrates obtained from corresponding III-N-bulk crystals.
(47) Alternatively, the homogeneity of the crystal quality in the growth direction can be determined by the rocking-curve-mapping, which is recorded within a plane including the growth direction. Regarding the growth in [0001]-direction, the reflection at an m-plane selected from {
(48) A second method for determining the homogeneity of the crystal quality relates to micro-Raman-mappings. For example the standard deviations of the frequency and the full-width half-mean of the E.sub.2-phonon of a scan on an area parallel to the growth plane, or along the growth direction, represent a measure for the homogeneity of the crystal quality in corresponding directions. The homogeneity of the crystal quality of the bulk crystal in growth direction is preferably determined by the standard deviation of the mean values of the full-width half-mean of the E.sub.2-phonon of the individual substrates, which have been obtained from the corresponding bulk crystal.
(49) The following results have been obtained for a crystal grown according to the invention:
(50) In a rocking-curve mapping of a III-N-bulk crystal according to the invention of an area parallel to the growth plane, the standard deviation of respectively measured full-width half-mean values amounts to 5% or less, in exceptional cases 3% or less. Within an area extending along the growth direction the full-width half-mean amounts to 10% or less, in exceptional cases 7.5%, and in particularly exceptional cases 5% or less.
(51) The standard deviation has been determined by performing at multiple measurement points, for example 100 points, in those areas intended to be measured each a measurement for the rocking curve-mapping. Of all measurements a mean value of the full-width half-means has been calculated and the standard deviation has been derived with respect to that mean value via a common statistical analysis.
(52) Alternatively, the crystallinity could be determined via a micro-Raman-mapping of the III-N-bulk crystal on a plane parallel to the growth plane and/or along growth direction. The standard deviation of the measured full-width half-means of the E.sub.2-phonon is then provided. In the first case (area parallel to the growth plane) the standard deviation amounts to 5% or less, in exceptional cases 3% or less, in particularly exceptional cases 2% or less. In the second case the standard deviation amounts to 10% or less, in exceptional cases 7.5% or less, and in particularly exceptional cases 5% or less. The measurement has been carried out analogously to that of the rocking-curve mapping over, e.g., 100 measurement points.
(53) The values provided with regard to crystallinity hold particularly true for the obtained GaN-crystals. It shall be repeatedly emphasized herein the combination of the excellent values for crystallinity with those for the layer thickness homogeneity, which may be obtained according to the invention.