Thermal processing of semiconductor devices

Information

  • Patent Grant
  • 6444549
  • Patent Number
    6,444,549
  • Date Filed
    Tuesday, September 1, 1998
    26 years ago
  • Date Issued
    Tuesday, September 3, 2002
    22 years ago
Abstract
Upon fabrication of semiconductor devices, a semiconductor substrate is subjected to ion implantation with high energy. Subsequent annealing of the ion-implanted semiconductor substrate, when conducted by heating the substrate to a temperature of from 1,000° C. to 1,200° C. at a ramp-up rate of at least 200° C./sec, makes it possible to provide the resulting semiconductor devices with smaller leakage currents of reduced variations (σ/X). The present invention can therefore provide a process for the fabrication of semiconductor devices featuring both smaller leakage currents and reduced variations of the leakage currents even when ion implantation is conducted with high energy.
Description




BACKGROUND OF THE INVENTION




1. Field of the Invention




This invention relates to a fabrication process of semiconductor devices, and especially to a fabrication process of semiconductor devices suitable for use in bipolar transistors or biCMOS transistors.




2. Description of the Prior Art




With a view to reducing fabrication costs and shortening TAT (Turn Around Time; time from the acceptance of an order until its delivery), high-energy ion implantation (which means ion implantation with an acceleration voltage of several hundreds kev or higher) has been attempted as a substitute for the conventional epitaxial growth in recent years for the formation of buried collector layers for bipolar or biCMOS transistors.




Because of the need for a reduction in collector resistance to permit operation of a device at a higher speed, a dopant dose higher than that (<5×10


13


cm


−2


) employed to date for the formation of wells and the like is required upon high-energy ion implantation. However, use of a dose of about 1×10


14


cm


−2


or so is desired in view of the maintenance of an adequate balance between collector resistance and maximum collector-base breakdown voltage since a higher dose results in a lower maximum collector-base breakdown voltage.




All the same, formation of p+/n junctions by ion implantation at an injection dopant dose around 1×10


14


cm


−2


and subsequent annealing by RTA (rapid thermal annealing) or an electric furnace under conditions of 10° C./sec or lower in terms of ramp-up rate is accompanied by a problem in that very large leakage currents occur across the junctions under the influence of defects extended to the surface of the sample from the neighborhood of the projection range of the ion implantation.




As a solution to this problem, Mat. Res. Soc. Symp. Proc. 396, 739 (1996, Materials Research Society) discloses that the leakage current across a junction can be reduced by RTA which makes use of a ramp-up rate of 50° C./sec or higher. It also discloses that, as the above-mentioned defects begin to grow abruptly when the temperature of RTA rises to 800° C., the leakage current can be reduced when the RTA is conducted by heating a wafer at a ramp-up rate of 50° C./sec or higher in the temperature range higher than or equal to 800° C. It is also disclosed that the leakage currents across the junctions of individual chips in a wafer can be lowered and their variations can be reduced to about half when RTA to a temperature of 1,050° C. is conducted by controlling the ramp-up rate at 1° C./sec from room temperature to 600° C. and at 100° C./sec beyond 600° C. instead of performing the heat treatment to the heat treatment temperature of 1,050° C. by using a ramp-up rate of 100° C./sec or higher from room temperature.




A ramp-up rate as low as 100° C./sec or so, however, involves a potential problem in that the above-described defects propagated from the projection range of ion implantation may grow to an area to which a depletion layer may extend upon application of a reverse bias. As soon as the depletion layer reaches a depth where defects exist, an increased leakage current occurs across the junction. According to the above-described conventional art, variations of leakage currents at a wafer surface may reach as much as 50% or so.




Namely, the process of the above-described conventional art is very effective in reducing the average value of leakage currents across junctions in a wafer, but further improvements are desired for the reduction of variations.




SUMMARY OF THE INVENTION




With the foregoing problems in view, the present invention has as a primary object thereof the provision of a process for the fabrication of semiconductors featuring both smaller leakage currents and reduced variations of the leakage currents even when ion implantation is conducted with high energy.




In a first aspect of the present invention, there is thus provided a fabrication process of semiconductor devices, which comprises the following steps:




subjecting a semiconductor substrate to ion implantation with high energy; and




heating the resulting ion-implanted semiconductor substrate to a temperature of from 1,000° C. to 1,200° C. at a ramp-up rate equal to or higher than 200° C./sec.




In this process, the temperature at which the heating is initiated at the ramp-up rate equal to or higher than 200° C./sec may preferably be from 600° C. to 800° C.




In a second aspect of the present invention, there is also provided a fabrication process of semiconductor devices, which comprises the following steps:




subjecting a semiconductor substrate to ion implantation with high energy;




heating the resulting ion-implanted semiconductor substrate to a temperature of from 1,000° C. to 1,200° C. at a ramp-up rate equal to or higher than 50° C./sec; and




subjecting the thus-heated semiconductor substrate to annealing at a temperature of from 1,000° C. to 1,200° C.




According to the first aspect of this invention, the heat treatment is conducted under hard conditions for the growth of defects occurred by the ion implantation so that, during the annealing step, the defects (dislocations) do not reach an area to which a depletion layer may extend. The process can therefore better reduce leakage currents and their variations than the conventional art. Further, the annealing at 1,200° C. and lower has made it possible to avoid the occurrence of slips (faults) in a wafer surface such as those occurring when the heat treatment is conducted at higher temperatures.




According to the second aspect of this invention, on the other hand, the ramp-up rate of 50° C./sec or higher is used, which includes ramp-up rates lower than those usable in the first aspect of the present invention. At such lower ramp-up rates, defects occurred as a result of the ion implantation may reach an area to which a depletion layer may extends. However, the subsequent heat treatment at a temperature of from 1,000° C. to 1,200° C. can eliminate the defects (dislocations) and restore the original crystalline structure, thereby making it possible to reduce leakage currents and their variations better than the conventional art.











BRIEF DESCRIPTION OF THE DRAWINGS




FIG.


1


A through

FIG. 1D

are cross-sectional views of a substrate-illustrating the formation of p/n diodes at various stages according to one example of fabrication processes of the present invention;





FIG. 2

is a diagrammatic representation of the ramp-up rate to an annealing temperature of 1,150° C. versus junction leakage current variation according to a first embodiment;





FIG. 3

is a diagrammatic representation of annealing temperature versus variations of junction leakage currents;





FIG. 4

is a diagrammatic representation of annealing temperature versus junction leakage currents;





FIG. 5

is a diagram showing leakage characteristics;





FIG. 6

is a diagrammatic representation of the heating commencing temperature in a second heat treatment step and variations of junction leakage current in a second embodiment;





FIG. 7

is a diagram comparing the variations of junction leakage currents in the first embodiment and those in the second embodiment;





FIG. 8

is a diagram illustrating a second (additional) heat treatment temperature and time in a third embodiment, which control at 20% the variations of junction leakage currents.











DETAILED DESCRIPTION OF THE INVENTION




[First Embodiment]




The first embodiment will hereinafter be described based on a specific fabrication process according to the first aspect of the present invention.




As is illustrated in

FIG. 1A

, a field oxide film


2


was first formed on a p-type single-crystal silicon substrate


1


, followed by the formation of a silicon oxide film


3


to a thickness of 7 mm. P (phosphorus) ions were next implanted at an acceleration voltage of 1 MeV and a dose of 1×10


14


cm


−2


. The resulting implanted substrate


1


was then annealed by RTA at a temperature of 1,150° C. by controlling the ramp-up temperature from room temperature at 100° C./sec (comparative example), 200° C./sec, 300° C./sec or 400° C./sec, whereby a buried n-type layer


4


was formed as shown in FIG.


1


B.




As is illustrated in

FIG. 1C

, ion implantation was then effected selectively at positions, where formation of p-type layers


5


was intended, at an acceleration voltage of 30 keV and a dose of 3×10


15


cm


−2


while using BF


2


as a dopant source. To establish a contact with the buried layer


4


, P (phosphorus) was ion-implanted selectively at positions, where formation of contact layers


6


to the buried n-type layer


4


was intended, at 70 keV and 5×10


15


cm


−2


and at 250 keV and 3×10


13


cm


−2


. The resulting substrate was subjected to heat treatment at 900° C. for several minutes in an electric furnace, so that the p-type layers


5


and the contact layers


6


were formed.




Subsequent to the formation of an interlayer film


7


, openings were formed through the interlayer film


7


and the silicon oxide layer


3


, and aluminum electrodes


8


and aluminum electrodes


9


were then formed in contact with the p-type layers


5


and the contact layers


6


, respectively. As a result, p/n diodes each of which had a cross-sectional structure depicted in

FIG. 1D

were formed at 53 locations in the wafer surface.




The p/n diodes fabricated as described above were grounded on a back side


10


of the substrate. A voltage of −5 V was applied to the aluminum electrodes


8


, and leakage currents across the junctions were measured at the aluminum electrodes


9


. An investigation was made for a relationship between variations (relative standard deviation: σ/X) of the thus-obtained leakage currents across the junctions in the wafer surface and the ramp-up rate. The results are shown in FIG.


2


.




From the results, it has been found that upon formation of n-type buried layers, use of 200° C./sec or higher as a ramp-up rate to an annealing temperature of 1,150° C. can reduce the variations of leakage voltages across junctions to {fraction (1/10)} or even smaller compared with 100° C./sec.




Next, the annealing temperature (ultimate temperature) upon formation of the buried n-type layer


4


was changed to 700, 800, 900, 985, 1,015, 1,100 and 1,150° C. to investigate their characteristics.




As is illustrated in

FIG. 3

, it is appreciated that leakage currents have smaller variations when the heat treatment temperature (ultimate temperature) is 800° C. or lower or 1,100° C. or higher. From

FIG. 4

, on the other hand, it is also envisaged that leakage currents are small when the heat treatment temperature is 800° C. or lower or 1,000° C. or higher.




In view of these two results, 1,015° C. which is higher than 1,000° C. but exceeds 1,000° C. only a little seems to result in leakage voltages of large variations although the leakage voltages themselves are small. An investigation was therefore made for leakage characteristics of diodes fabricated at the heat treatment temperature of 1,015° C. Although most of them were found to have such characteristics as represented by curve A in

FIG. 5

, some devices were found to exhibit such characteristics as represented by curve B in the same diagram. This clearly explains the occurrence of large variations in terms of numerical values.




On the other hand, use of a heat treatment temperature higher than 800° C. but lower than 1,000° C. leads to such leakage characteristics as indicated by curve C in FIG.


5


. It is hence understood that such a heat treatment temperature is not preferred because it leads to large leakage voltages although their variations are apparently small.




Further, use of a heat treatment temperature of 800° C. or lower may look preferable at first view, because leakage currents are small and variations are small. Such a low heat treatment temperature is however not preferred from the practical viewpoint since, when exposed to high temperatures of 800° C. and higher in the subsequent step, defects (dislocations) occurred by the ion implantation grow and leakage currents increase.




[Second Embodiment]




In a similar manner as in the fabrication process described above as the first embodiment, p/n diodes having the cross-sectional structure shown in

FIG. 1D

were formed at 53 locations in a wafer surface. In the second embodiment, however, the annealing method employed upon formation of the buried n-type layer


4


in

FIG. 1B

was modified as will be described next.




Described specifically, a sample was introduced in an RTA apparatus which had been heated at 500° C. in advance, and was heated to a first annealing temperature of 600° C., 700° C., 800° C, 900° C. (comparative example) or 1,000 (comparative example) while controlling the first ramp-up rate to the first treatment temperature at 1° C./sec. After the sample was annealed for 1 minute at the first annealing temperature, the sample was heated to a second annealing temperature of 1,150° C. at a second ramp-up rate of 200, 300 or 400° C./sec, whereby a buried n-type layer


4


was formed.





FIG. 6

illustrates a relationship between the first annealing temperature, namely, the temperature at which heating was commenced at 200° C./sec and the variations (relative standard deviation) of leakage currents across junctions in a wafer surface.




From the results shown in

FIG. 6

, it is envisaged that the variations of junction leakage currents can be reduced when the 200° C./sec heating is commenced at a temperature of 800° C. or lower.





FIG. 7

makes a comparison in the variations (relative standard deviation) of junction leakage currents in a wafer surface between a case in which in accordance with this embodiment, a sample was introduced in an RTA apparatus heated at 500° C. in advance, was heated to a first annealing temperature of 600° C. at a ramp-up rate of 1° C./sec, and was then heated at a second ramp-up rate of 200, 300 or 400° C./sec to conduct heat treatment at a second heat treatment of 1,150° C. (the line representing the second embodiment in

FIG. 7

) and another case in which as in the first embodiment, heating of a sample was commenced at a ramp-up rate of 200-400° C./sec from room temperature and annealing was conducted at an annealing temperature of 1,150° C. (the line representing the first embodiment in FIG.


7


).




From the results of the above comparison, it is understood that the use of the annealing method according to this embodiment, at any ramp-up rate chosen here, makes it possible to reduce the variations of leakage currents across junctions in a wafer surface to about ½ of those obtained from the use of the annealing method in the first embodiment.




In the above description of the second embodiment, 1° C./sec was used as the first ramp-up rate. The first ramp-up rate is however not particularly important, but annealing at a temperature of from 600° C. to 800° C. as the first annealing temperature is important. It is to equalize the temperature of the whole substrate at this stage to prevent the variations of substrate temperatures from increasing further upon subsequently heating the substrate at a high ramp-up rate that the first heat treatment is conducted. Further, the use of 600° C. or higher as the first heat treatment temperature is needed because the equalizing effects of a substrate temperature by black body radiation are hardly available at a temperature lower than 600° C.




Incidentally, the retention time at the first heat temperature can be set as desired to equalize the substrate temperature.




[Third Embodiment]




The third embodiment will hereinafter be described based on a specific fabrication process according to the second aspect of the present invention.




In a similar manner as in the fabrication process described above as the first embodiment, p/n diodes having the cross-sectional structure shown in

FIG. 1D

were formed at 53 locations in a wafer surface. In the third embodiment, however, the annealing method employed upon formation of the buried n-type layer


4


in

FIG. 1B

was modified as will be described next.




Described specifically, as a first annealing step, a sample was introduced in an RTA apparatus which had been heated at 500° C. in advance, and was heated at a ramp-up rate of 1° C./sec to a heat treatment temperature of 600° C., at which the sample was annealed for 1 minute. The sample was then heated at a ramp-up rate of 50 or 100° C./sec to an annealing temperature of 1,015° C. or 1,100° C., at which the sample was subjected to heat temperature for 30 seconds.




Next, as a second annealing step, the sample was introduced in an electric furnace and was subjected to annealing for one of preset various annealing times at an annealing temperature in a range of from 1,015° C. to 1,200° C. with its furnace entering and leaving temperatures being set at 800° C., whereby a buried n-type layer


4


was formed.





FIG. 8

illustrates a relationship between the annealing temperature and the annealing time in the second annealing (additional annealing) conducted in an electric furnace so that the variations (relative standard deviation) of leakage currents across junctions in a wafer surface became 20%. When the annealing temperature and the ramp-up rate are set, for example, at 1,015° C. and 100° C./sec as RTA treatment conditions, respectively,

FIG. 8

indicates that the variations are smaller than 20% above the curve extending through the filled circles. To reduce the variations to 20% or smaller, it is therefore necessary to choose an additional annealing temperature and an additional annealing time on the curve or in an area above the curve.




It is envisaged from the results as shown in

FIG. 8

that the variations of leakage currents can be reduced by conducting the second annealing as described above. Further, it is appreciated that the time required for reducing the variations in the second annealing becomes longer when the annealing temperature is lowered and also that the time required for reducing the variations can be shortened by raising the heat treatment temperature. It is also understood that the variations can be reduced to 20% or smaller even when the ramp-up rate is decreased to 50° C./sec.




On the other hand, it was impossible to reduce the variations when the temperature of the second annealing was set below 1,000° C. Further, the second heat treatment, when conducted at a temperature higher than 1,200° C., has been found to develop a problem that crystal defects such as slips occur in the crystal.




As has been described above, the ramp-up rate can be decreased to 50° C./sec upon conducting the second annealing. On the other hand, the second annealing has also been found to be effective in reducing the variations even when a high ramp-up rate of 200° C./sec or higher is used. In this case, the second annealing can be conducted in the same RTA apparatus as that employed in the preceding heating step. Similar advantageous effects were brought about when the second annealing was conducted, for example, in a temperature range of from 1,000 to 1,200° C. subsequent to heating to 1,150° C.




As is readily appreciated from the foregoing, the second annealing step can bring about similar advantageous effects no matter whether it is conducted in an RTA apparatus or an electric furnace. In general, however, an RTA apparatus treats semiconductor substrates one after one, whereas an electric furnace can perform heat treatment of plural (about 100 or so) semiconductor substrates at once. Accordingly, an electric furnace can significantly shorten the total time required for the heat treatment of such many semiconductor substrates.




Namely, this embodiment, when the additional annealing is conducted separately, makes it possible to use a heating furnace having large processing capacity such as an electric furnace and hence to extremely shorten the fabrication time.




According to the fabrication process in each aspect of the present invention, defects which cause varied leakage currents across junctions are not allowed to grow to an area where a depletion layer extends even when ion implantation is conducted with high energy, thereby making it possible to fabricate semiconductors featuring both smaller leakage currents and reduced variations of the leakage currents.



Claims
  • 1. A fabrication process of semiconductor devices featuring both reduced leakage currents and reduced variations of leakage currents, which comprises the steps in sequence of:(A) subjecting a semiconductor substrate to ion implantation with high energy; (B) introducing the resulting ion-implanted semiconductor substrate into a first furnace preheated to a temperature of 500° C. and subjecting the substrate to annealing at a temperature selected from the group of temperatures consisting of 1015° C., 1100° C., and 1150° C.; and (C) introducing the resulting semiconductor substrate from step (B) into a second furnace, and heating the substrate in said second furnace to a temperature of from 1,000° C. to 1,200° C. at a ramp-up rate equal to or higher than 300° C./sec.
  • 2. A fabrication process according to claim 1, further comprising annealing said semiconductor substrate in said second furnace at said temperature of from 1,000° C. to 1,200° C. for thirty seconds.
  • 3. A fabrication process according to claim 2, wherein said second furnace comprises an electric furnace.
  • 4. A fabrication process of semiconductor devices featuring both reduced leakage currents and reduced variations of leakage currents, which comprises the following steps in sequence of:(A) subjecting a semiconductor substrate to ion implantation with high energy; (B) introducing the resulting ion-implanted semiconductor substrate into a first furnace preheated to a temperature of at least 500° C. and heating the substrate at a ramp-up rate of 1° C./sec to a temperature of 600° C.; (C) introducing the semiconductor substrate from step (B) into a second furnace with its entering and exiting temperatures set at 800° C., and heating the substrate to a temperature of from 1,000° C. to 1,200° C. at a ramp-up rate equal to 50° C./sec; and (D) subjecting the thus-heated semiconductor substrate to annealing at a temperature of from 1,000° C. to 1,200° C.
  • 5. A fabrication process according to claim 4, wherein said second furnace comprises an electric furnace.
Priority Claims (1)
Number Date Country Kind
9-248880 Sep 1997 JP
US Referenced Citations (20)
Number Name Date Kind
4432809 Chye et al. Feb 1984 A
4482393 Nishiyama et al. Nov 1984 A
4522657 Rohatgi et al. Jun 1985 A
4732866 Chruma et al. Mar 1988 A
4743569 Plumton et al. May 1988 A
4753895 Mayer et al. Jun 1988 A
4784964 Hulseweh et al. Nov 1988 A
4912053 Schrantz Mar 1990 A
5130261 Usuki et al. Jul 1992 A
5312764 Drowley et al. May 1994 A
5399506 Tsukamoto Mar 1995 A
5476799 Sakamoto et al. Dec 1995 A
5487127 Gronet et al. Jan 1996 A
5612253 Farahani et al. Mar 1997 A
5773337 Lee Jun 1998 A
5834353 Wu Nov 1998 A
5872047 Lee et al. Feb 1999 A
5940699 Sumi et al. Aug 1999 A
5950078 Maekawa et al. Sep 1999 A
5967794 Kodama Oct 1999 A
Foreign Referenced Citations (6)
Number Date Country
59-108315 Jun 1984 JP
4-249315 Sep 1992 JP
7-74180 Mar 1995 JP
8-107067 Apr 1996 JP
9-148570 Jun 1997 JP
10-55978 Feb 1998 JP
Non-Patent Literature Citations (8)
Entry
Shishiguchi et al., 1997 Symposium on VLSI Technology Digest of Technical Papers, pp. 89-90, 1997.*
Liu et al., Solid-State Electronics 38(8), 1995, pp. 1473-1477, 1995.*
Foad et al., IEEE 0-7803-4538-X/99, pp. 732-735, 1999.*
Downey et al., IEEE 0-7803-4538-X/99, pp. 55-58, 1999.*
Kang et al., IEEE Electron device Letters 21(1), 2000, pp. 9-11.*
Jang et al., IEEE TENCON 0-7803-5739-6/99, pp. 1121-1123, 1999.*
Hamada, K., Nishio N., and Saito, S., “Anomalous Leakage Current Reduction by Ramping Rate Control in MeV Implantation”, Mat. Res. Soc. Symp. vol. 396, 1996.*
“Anomalous Leakage Current Reduction by Ramping Rate Control in MEV Implnatation” Mat. Res. Soc. Symp. Prac. vol. 396; 1996 Materials Research Society; p. 739-743.