DE4329837A1 - Method for producing a silicon semiconductor component - Google Patents

Abstract

A method is specified for producing a silicon semiconductor component with high sensitivity with respect to contamination by heavy metals, such as a Si CCD solid-state image sensor, in which a getter zone (7) is formed in an element-insulation region, such as found near an element region, a non-depleted n<+> or p<+> region and/or a region to be removed in a subsequent process from the surface of a semiconductor substrate (1) or to be insulated therefrom. This method makes it possible to produce semiconductor components with reduction of the contamination of the silicon substrate surfaces by heavy metals. The method also realises the production of silicon semiconductor components which have stable component properties and a reduced number of defects. <IMAGE>

Description

The invention relates to a method for producing a Silicon semiconductor device, more specifically a method for Manufacture components that have high sensitivity against contamination by heavy metals, like Si-CCD solid-state image sensors.

In the manufacture of Si semiconductor devices, the Verun cleaning of Si wafers by heavy metals and the like an important factor that affects the properties and properties reliability of the manufactured components deteriorated.

In particular, solid-state image sensors with charge coupling pelt devices (CCDs) that handle microsignal charges ben, very easily contaminated by heavy metals.

Such contamination by heavy metals  also for storage devices with very high integration density is an essential problem.

To cleanliness in plants for the production of Si half lead components and that of chemical materials improve, various ultra-clean techniques were featured beat. One is suggested in "Ultra LSI Ultra Clean Technology Workshop No. 7 Announcement ", Japanese Semicon ductor Basic Technology Research Committee, May 1990.

However, a more economical countermeasure is required Lich, since introducing the above new techniques into one large increase in installation costs.

Provided that a particular contaminant extent is unavoidable due to heavy metals Process for gettering the contamination in a subsequent process plays an important role in solving the contaminants problem.

So far, various gettering methods have been used as a measure Semiconductor components with high contamination sensitivity against heavy metals such as Si-CCD solid-state images sensors used. The known gettering methods are now exemplified.

The first method is illustrated in FIG. 3, wherein a layer 29 with microdefects distributed in volume is formed on a silicon substrate 30 , which is to be used as a getter sink.

The second method is illustrated by the Fig. Veran 4a and 4b. According to this method, a silicon substrate 18 is subjected to an exposure process on its rear side, as shown in FIG. 4a, before a process for the diffusion of phosphorus ions into a polysilicon layer 20 , which covers the silicon substrate 18 . Then, a high concentration phosphor diffusion layer 23 is formed on the back of the substrate 18 as shown in Fig. 4b.

Fig. 5, the third method is formed on the back of a Siliziumsub strats 18 in which a polysilicon layer 19 is illustrated.

The fourth method is shown in FIG. 6, in which a silicon substrate 18 is subjected to an ion implantation or a sandblasting treatment on its rear side to form a damaged layer 29 .

These methods produce an effect of improving the Gettering ability of silicon substrates. However, it was knows that heavy metals, when they are curbed, in a heating process are released, which at relatively low ger temperature is executed. It is also difficult to find one adequate gettering effect using the process for the diffusion of phosphorus ions into the back of a To achieve substrate because the diffusion process at low Temperature and running for a short time.

In addition, heavy metals can tend to build up in egg nem Si-Sio₂ border area or in areas of which mecha tensions, such as those arising from the structure of a manufactured component were produced to concentrate wear. It is also difficult to clean the surface due to heavy metals in the substrate or at the getter site reduce the back of the substrate.

Accordingly, front side gettering is examined in which a defect layer as a getter location in an area under egg Nem element area is formed in that ions higher  Energy can be implanted so that an adequate getter effect at lower temperature and shorter treatment time compared to previously achieved (Japanese Applied Physics Institute, Spring Conference 1991, 31a-X-8 to 11).

In particular, it was found that in the case of an ion implantation with boron or carbon is a heavy metal get tern should be achieved that does not lead to a re-delivery of Heavy metal leads.

However, this process requires new installations for Perform high energy ion implantation above 1 MeV.

When high energy ions are planted into a MOS transistor-formed substrate 18 , as shown in FIG. 7a, they pass through an element region where they cause a gettering region 28 to be formed under the element region, as in FIG Fig. 7b shown. As a result, a measurement of damage caused by the high-energy ions must take place for each component structure, and the damage must also heal. Accordingly, the process can hardly be regarded as an economical process.

The invention is directed to the above overcoming a difficulty. Tried for this purpose the invention to provide a getter method, which is in the Is able to effectively reduce heavy metals that are in an element area on the surface of a silicon sub strats, especially in a Si-SiO₂ border area on the Have accumulated silicon substrate surface without a be special installation to use at low cost aim.

Now a getter phenomenon is described as it is more important  Relation to the invention.

Although various silicon crystal defects for gettering Serving heavy metals, they will be even at a relative at a low temperature of 700 to 800 ° C (yes panese Applied Physics Institute, autumn conference 1992, 18p-ZH-5 announcement no. 1, page 314).

It was known that gettering methods, ion implantation use as gettering methods that are effective are able to re-release heavy metals prevent.

For example, very strong gettering ability was shown immediately after the Implantation of boron ions. By one according to the ion implant annealing treatment, Fe was soaked that it overlapped with the boron profile as it did directly after ion implantation showed (Y. Niki, S. Nadahara and M. Watabave: Proc. Int. Conf. Science and Tech. of defect Control in Semicond., Yokohama, 1989, Vol. 1, p. 329).

It has also been about getter phenomena of various elements reported. However, the gettering ability disappeared after ei ne treatment at a high temperature of about 900 ° C. was carried out. This may be the case because of the High temperature treatment implanted atoms directly put so that they penetrate into the crystal lattice.

It was also known that in the case of Koh the gettering ability was maintained by that carbon atoms hardly penetrate into the crystal lattice, even with high temperature treatment (Japanese Applied Physics Institute, Fall Conference 1992, 18p-ZH-11 Announcement No. 1, p. 313).  

In this case the getter location is a primary fault ordered structure, such as by implantation of coals atoms or through existing between lattice points Carbon atoms is caused. The gettering ability is approximately proportional to the dose amount of carbon.

The primary fault, which is a point defect delt is used to fix a heavy metal. After The primary fault increases in connection with the heavy metal not to a larger error structure. As a result, it is difficult that the primary fault becomes a transfer or a stacking error caused by oxidation grows.

It was also about using an element isolator empire or the like, not depleted n⁺ and p⁺ areas, or an area to be removed in a subsequent process reports as a getter area that is created by Implanta tion of carbon ions was formed. This is due to the Fact that by implanting carbon ions formed getter location is a point error that is not going to out of a big mistake that worsens of component properties would result, even if Heavy metals without the use of thermal diffusion be tert.

It was also confirmed that gettering by ion implant tation regardless of the acceleration voltage for the Ion takes place. Accordingly, ion implantation under Use conventional ion implantation devices only performed in an area near the substrate surface become. However, the non-depleted area can be used as a getter location be used.

If the area where gettering heavy metals he  follows, is oxidized, the heavy metals in which he witnessed oxide film drawn because stable fixation of heavy tallen occurs. Accordingly, it is possible to put in the oxide film drawn Si surface area in a subsequent Use process as getter location. In this procedure is it is necessary to prevent the pre-selected as getter location given area reaches an area in which components properties are adversely affected.

Because semiconductor devices have different sensitivity against defects, should be an appropriate span of up to 3σ or 4σ for an intended area at Implant ions for each component or dependent be selected from its structure.

The dose of implanted ions should also be taken into account be viewed. With a large dose of ions the damage density in the area with ion implantation corresponds higher. The amount of fixed heavy metals increases what leads to an increase in the possibility that a secondary mistake l grows. Accordingly, the dose amount implanted Io shows an upper limit. In fact, it has been shown that a secondary error with an ion dose amount of approximately 10¹⁶ / cm² occurs, causing deterioration of the getter assets (Japanese Applied Physics Institute, Fall conference 1992, 18p-ZH-10 announcement no. 1, p. 312).

It appears that the upper limit for the dose of ions is quite high is low when ions are in a range close to the Ele be implanted. However, it is when Io is carried out several times under the condition that the dose amount of the ions is divided into sub-dose amounts and a tempering process after each ion implantation leads, possible, a total ion dose amount of 10¹⁴ / cm² to use.

The invention has for its object a method for To manufacture a silicon semiconductor device, through which a getter location in an element isolation area is close an element area or the like, a non-impoverished one n⁺ or p⁺ range, one from the surface of a semi-lead tersubstrats in a subsequent process or area to be stripped is formed.

This task is added by the subordinate Neten claims specified method solved. The invent Process according to the invention are characterized in that they one of the following procedures when manufacturing Si silicon semiconductor components by implanting carbon use fabric ions:

• 1. Implant carbon ions into an element iso lierbereich a Si semiconductor device;
• 2. Implant carbon ions into one under one Si-SiO₂ limit arranged area, performing a temper process, oxidizing the Si surface to a through the Implantation of carbon ion damaged area in to pull the oxide film formed on the Si surface Remove the oxide film and run the oxide again tion;
• 3. Implant carbon ions into an n⁺- or p⁺- Region of a Si semiconductor component; and
• 4. Perform the implantation of carbon ions according to the above procedures with a Be acceleration voltage that ensures that the target area ion implantation within 1000 Å (100 nm) of the Si-SiO₂ limit is removed.

The first step is to isolate the element rich, in which carbon ions are implanted, the getter place. Accordingly, contamination in the form of heavy  tallen in the element area in the element insulation area.

Generally, the element isolation area is used a local silicon oxidation (LOCOS) process forms. This process is where the implantation of Carbon ions run before oxidation, the Ver contamination by heavy metals from the implanted Carbon ions were saved at a subsequent Oxidation process drawn into the thick oxide film. The Ele The isolation area can be made using a different construction element isolation process that is based on the is used in the same way as the LOCOS process to produce a thick oxide film.

In the second approach, the Verun getstered Cleaning by heavy metals when tempering after implanting carbon ions. The contamination is fixed in such a way that it the distribution profile of the implanted carbon atoms overlaps. Because in an area where gettering from Heavy metals occurs, who is formed an oxide film which the scaled heavy metals are drawn into the oxide film. By forming a new oxide film after the oxide film with the heavy metals it is possible to have a Si surface with reduced contamination and an oxide film of good quality to maintain lity.

In the third procedure, the getter location is SiO₂ boundary area formed, whereby the ability verbes isert to effectively getter heavy metals, which in the Border area are concentrated. Use this procedure detects the phenomenon that heavy metals are distorted into one Area as in the crystal lattice of the Si-SiO₂ interface is present, so that it accumulates there meln.  

According to this method, the getter location is in the distorted Be richly trained, which improves getter efficiency becomes.

The fourth approach has the same effect as that first. When doing this on a CCD solid Image sensor with an Ele ment isolation area in addition to a general element isolation area such as a LOCOS area, ver it uses the channel stop area as a getter location so that a Contamination of the sewer area by the contamination is reduced, thereby enabling the charge to improve transmission efficiency and dark current to reduce. In this case, however, the concentra p-foreign substances in the sewer stop area are sufficiently high be so as not to impoverish it.

Other objects and aspects of the invention are possible from the following description of exemplary embodiments un ter reference to the accompanying drawings, in to whom:

Fig 1a to 1h are schematic cross sections showing a drive Ver for manufacturing a semiconductor device according to ei nem first embodiment of the invention veranschauli chen.

Figures 2a to 2i are schematic cross sections showing a drive Ver for manufacturing a semiconductor device according to ei nem second embodiment of the invention veranschauli chen. and

Figure 3 is. To 7b are schematic cross sections, veranschauli chen the herkömmli che method of gettering of heavy metals.

In the method according to the invention, independently of one another using the procedures above. However, in terms of efficiency, it is preferred to combine at least two of the approaches turn.

In Figs. 1a to 1h a method of manufacturing a semiconductor device according to a first illustrated embodiment of the invention, wherein the above-mentioned four approaches for an integrated Si-NOS component are used.

According to this method, an oxide film 2 and an SiN layer 3 are first formed on the surface of a p-Si substrate 1 , as shown in FIG. 1a.

A photoresist pattern 4 is then formed on the SiN layer 3 and has an opening which corresponds to an element isolation region. Using the photoresist pattern 4 as a mask, the SiN layer 3 is subjected to an etching process in order to be completely removed at the unprotected locations. Thereafter, an ion implant region 5 is formed by implanting boron ions and then carbon ions in the surface of the substrate 1 , as shown in Fig. 1b.

The photoresist pattern 4 is then removed. At 950 ° C temperature, nitrogen annealing and oxidizing processes are carried out to form a thick oxide film 6 in the element isolation region. These high-temperature processes diffuse boron ions deep into the Si substrate 1 , as a result of which a p ein region 8 is formed. On the other hand, the carbon ions contribute to gettering of heavy metals and they are subsequently drawn into the oxide film 6 , whereby a getteror layer 7 is formed, as shown in FIG. 1c.

The SiN layer 3 is then removed. As shown in Fig. 1d, carbon ions are then implanted again in the Si-SiO₂ interface identified by the reference numeral 9 . If processes for tempering and oxidizing are carried out, gettering takes place again, as a result of which heavy metals are drawn into the oxide film formed.

The oxide film containing the heavy metals is then removed. The oxide film formed on the element insulation region remains because it has a sufficient thickness, while the region arranged below the SiN layer 3 is exposed. A gate oxide film 10 is formed on the exposed area, as shown in FIG. 1e.

Thereafter, a polysilicon gate electrode 11 and an oxide film 12 covering the same are formed over the gate oxide film 10 . Using the polysilicon gate electrode 11 as a mask, As ions are then implanted in order to form an n + region 13 on the surface of the substrate 1 , as shown in FIG. 1f.

This is followed by implantation of carbon ions and annealing in order to form a getter location 14 in the n⁺ region 13 . As ions are diffused in to form the source and drain, as shown in FIG. 1g.

A smoothing layer 15 , contact holes and an Al wiring layer 16 and a surface passivation layer 17 are then produced, as shown in FIG. 1h.

In FIGS. 2a to 2i a method of manufacturing a semiconductor device according to a second Ausführungsbei of the invention illustrated game, which is used for producing a two-dimensional solid-state image sensor.

For producing peripheral areas of the two-dimensional Solid state image sensor uses this embodiment the same processes as the first.

According to this method, a first p-well layer 32- a, a second p-well layer 32- b, a vertical CCD channel region 5 , an Si oxide film 33- a and one are first on the surface of an n-Si substrate 31 CCD element isolation region corresponding channel stop region 34 is formed using well known manufacturing processes for solid state image sensors, as shown in Fig. 2a.

Thereafter, carbon ions, which is then subjected to a tempering process are implanted into the channel stop region 34, so that fixed 31 heavy metals present in a Getterort 36 in an area near the surface of the substrate who to, who was det gebil by the implantation of carbon ions, as shown in Fig. 2b shown.

The Si oxide layer 33- a is then removed. Thereafter, a gate oxide film 33- b is formed over the entire exposed surface of the resulting structure. A polysilicon transfer electrode 37 is formed on the vertical channel CCD region 35 , which in turn is oxidized to form an oxide film 38 . As already illustrated by Fig. 1d, carbon ions are then diffused again into the Si-SiO₂ boundary surface identified by the reference number 9 . Subsequently, carbon ions are implanted in the surface of an area in which a photodiode is to be formed, thereby forming a getter location 39 , as shown in FIG. 2c.

A photodiode 42 is then formed by implanting ions in the photodiode region. Oxidation is then carried out in such a way that the getter location 39 is drawn into an oxide film 40 formed by the oxidation, as shown in FIG. 2d. The reference number 41 denotes the heavy metals drawn into the oxide film 40 .

Then the gate oxide film 33- b and the oxide film 40 are removed , as shown in FIG. 2e. The get terort 39 and the heavy metals 41 are removed.

An oxide film 43 is then formed over the entire exposed surface of the resulting structure, as shown in FIG. 2f. Then boron and carbon ions are diffused into the Si-SiO₂ interface corresponding to the area of the photodiode 42 , thereby forming an ion implant layer 44 as shown in FIG. 2g.

If an annealing process is then carried out, the boron ions diffuse, whereby a p wodurch region 45 is formed on the surface of the photodiode 42 . On the other hand, the carbon ions contribute to gettering of heavy metals, so that a getter layer 46 is formed, as shown in FIG. 2h.

This is followed by the formation of a photo-shielding layer 47 and a surface passivation layer 48 in order to achieve a solid-state image sensor, as is shown in FIG. 2i.

Although this embodiment has been described as such ben that the above procedures in Com combination used, it can have sufficient effects even then  achieve if only one of the approaches is used.

As can be seen from the above description, he the invention enables semiconductor devices under Ver reduction of contamination of silicon substrate surface chen by heavy metals. The invention reali also the production of silicon semiconductor devices ten, the stable element properties and a reduced Show number of defects.

The implantation of carbon atoms, invented as a Appropriate means can be made using of a conventional ion implantation device the. Accordingly, it is not an expensive equipment for implantation of high energy ions required. Beyond that it is possible to achieve getters with high efficiency. in the The required standard regarding ver contamination by heavy metals can be loosened what it is possible to reduce installation costs.

In particular, if the invention is based on Si-CCD images sensors is used, image defects are considerably reduced become, which makes it possible to greatly increase the yield. A reduction in the dark current can also be achieved which ensures that the dynamic range is increased. Accordingly, semiconductor devices with a high degree of In density of integration can be obtained.

Claims (8)

1. A method for producing a silicon semiconductor component with the following steps:

• - Implanting carbon ions in at least one area, which is either an element isolation region of a silicon substrate ( 1 ), an area present under an interface between the silicon substrate and an oxide film formed above this and / or an area formed in the silicon substrate with a high concentration of foreign substances; and
• - Heat treating the area ( 5 ; 9 ; 39 ; 46 ) with the implanted carbon ions to form a gettering location.

2. The method according to claim 1, characterized in that the step of implanting carbon ions is carried out at an accelerating voltage which causes the target region of the ion implantation within 100 nm starting from the interface between the silicon substrate ( 1 ) and an oxide film ( 2nd ) lies. 3. A method for producing a silicon semiconductor component with the following steps:

• - Forming a first oxide film ( 2 ) and a SiN layer ( 3 ) over a silicon substrate ( 1 ) of a first line type;
• - Removing a selected portion of the SiN layer, which portion lies over an element isolation region of the silicon substrate, implanting boron ions and carbon ions in the order mentioned in the element isolation region ( 5 ) and subsequent annealing;
• - Forming a second oxide film ( 6 ) with a large thickness over the element isolation region;
• - Removing the remaining portion of the SiN layer, implanting carbon ions in an area ( 9 ) which is formed under the interface between the first oxide film ( 2 ) and the silicon substrate ( 1 ), and subsequent Tem pern;
• - removing the second oxide film;
• - Forming a gate oxide film ( 10 ) over the entire exposed surface of the resulting structure after removing the second oxide film;
• - forming a gate electrode ( 11 ) over the gate oxide film;
• - Forming a region of the second conductivity type with a high impurity concentration above a predetermined portion of the silicon substrate using ion implantation; and
• - Implanting carbon ions in the area of the second conduction type with a high impurity concentration, and Tem pern to form a getter location ( 14 ) in the area with the implanted carbon ions.

4. The method according to claim 3, characterized in that in the step of forming the second oxide film ( 6 ) in the planted boron ions in the silicon substrate ( 1 ) are diffused, whereby an area ( 8 ) of the first line type with a high impurity concentration is formed while the implanted carbon ions together with the heavy metal ions are drawn into the second oxide film to form through a getteror layer. 5. The method according to claim 3, characterized in that in the step of implanting carbon ions in the loading area ( 9 ) under the interface between the first oxide film ( 2 ) and the silicon substrate ( 1 ) and the subsequent Tem perns heavy metal ions in the silicon substrate are present are drawn into the first oxide film to be gettered. 6. A method of manufacturing a semiconductor device, comprising the following steps:

• Forming a first well layer ( 32- a) and a second well layer ( 32- b) of a first conductivity type on the surface of a silicon substrate ( 31 ) of a second conductivity type, forming a channel stop region ( 34 ) in a predetermined section of the first well layer, Forming a vertical CCD region ( 35 ) in a predetermined section of the second well layer and forming a Si oxide film ( 33- a) over the entire exposed surface of the structure formed after the formation of the vertical channel CCD region, wherein all training operations are carried out using a general manufacturing process for a solid-state image sensor;
• - implanting carbon ions in the channel stop area, and annealing to form a getter location ( 36 );
• Removing the Si oxide film and then forming a gate oxide film ( 33- b) over the entire exposed surface of the structure resulting after the removal of the Si oxide film ( 33- a);
• - Forming a transfer electrode ( 37 ) over the vertical channel channel CCD area;
• Implanting carbon ions into a surface of an area of the first well layer where a photodiode ( 42 ) is to be formed to form a getter location ( 39 );
• - forming a photodiode ( 42 ) in said area of the first well layer using ion implantation, and forming a first oxide film ( 40 ) using an oxidation process;
• - removing the gate oxide film and the first oxide film;
• - Forming a second oxide film ( 43 ) over the entire exposed surface of the structure resulting after the removal of the gate oxide film and the first oxide film;
• - implanting carbon ions in an area ( 44 ) which is arranged below the interface between the second oxide film and the silicon substrate, and subsequent Tem pern; and
• - Form a photo-shielding layer ( 47 ) and an upper surface passivation layer ( 48 ) in each predetermined areas.

7. The method according to claim 6, characterized in that in the step of forming the first oxide film, the getter location together with heavy metals present in the silicon substrate ( 31 ) is drawn into the first oxide film. 8. The method according to claim 6, characterized in that in the step of implanting carbon ions in the loading area ( 44 ), which is arranged under the interface between the second oxide film ( 43 ) and the silicon substrate ( 31 ), and during the annealing the implanted Boron ions undergo diffusion to form a first conductivity type impurity region on the surface of the photodiode ( 42 ), while in the planted carbon ions form a getter layer.