[go: up one dir, main page]

WO1995023429A1 - Providing a low resistance to integrated circuit devices - Google Patents

Providing a low resistance to integrated circuit devices Download PDF

Info

Publication number
WO1995023429A1
WO1995023429A1 PCT/US1994/011110 US9411110W WO9523429A1 WO 1995023429 A1 WO1995023429 A1 WO 1995023429A1 US 9411110 W US9411110 W US 9411110W WO 9523429 A1 WO9523429 A1 WO 9523429A1
Authority
WO
WIPO (PCT)
Prior art keywords
region
forming
over
conductive
insulation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/US1994/011110
Other languages
French (fr)
Inventor
Michael J. Grubisich
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
National Semiconductor Corp
Original Assignee
National Semiconductor Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by National Semiconductor Corp filed Critical National Semiconductor Corp
Priority to KR1019960704698A priority Critical patent/KR970701432A/en
Priority to EP94929971A priority patent/EP0748517A1/en
Publication of WO1995023429A1 publication Critical patent/WO1995023429A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L21/00—Processes or apparatus adapted for the manufacture or treatment of semiconductor or solid state devices or of parts thereof
    • H01L21/02—Manufacture or treatment of semiconductor devices or of parts thereof
    • H01L21/04—Manufacture or treatment of semiconductor devices or of parts thereof the devices having potential barriers, e.g. a PN junction, depletion layer or carrier concentration layer
    • H01L21/18—Manufacture or treatment of semiconductor devices or of parts thereof the devices having potential barriers, e.g. a PN junction, depletion layer or carrier concentration layer the devices having semiconductor bodies comprising elements of Group IV of the Periodic Table or AIIIBV compounds with or without impurities, e.g. doping materials
    • H01L21/28—Manufacture of electrodes on semiconductor bodies using processes or apparatus not provided for in groups H01L21/20 - H01L21/268
    • H01L21/283—Deposition of conductive or insulating materials for electrodes conducting electric current
    • H01L21/285—Deposition of conductive or insulating materials for electrodes conducting electric current from a gas or vapour, e.g. condensation
    • H01L21/28506—Deposition of conductive or insulating materials for electrodes conducting electric current from a gas or vapour, e.g. condensation of conductive layers
    • H01L21/28512—Deposition of conductive or insulating materials for electrodes conducting electric current from a gas or vapour, e.g. condensation of conductive layers on semiconductor bodies comprising elements of Group IV of the Periodic Table
    • H01L21/28518—Deposition of conductive or insulating materials for electrodes conducting electric current from a gas or vapour, e.g. condensation of conductive layers on semiconductor bodies comprising elements of Group IV of the Periodic Table the conductive layers comprising silicides
    • H—ELECTRICITY
    • H10—SEMICONDUCTOR DEVICES; ELECTRIC SOLID-STATE DEVICES NOT OTHERWISE PROVIDED FOR
    • H10D—INORGANIC ELECTRIC SEMICONDUCTOR DEVICES
    • H10D84/00—Integrated devices formed in or on semiconductor substrates that comprise only semiconducting layers, e.g. on Si wafers or on GaAs-on-Si wafers
    • H10D84/80—Integrated devices formed in or on semiconductor substrates that comprise only semiconducting layers, e.g. on Si wafers or on GaAs-on-Si wafers characterised by the integration of at least one component covered by groups H10D12/00 or H10D30/00, e.g. integration of IGFETs
    • H10D84/82—Integrated devices formed in or on semiconductor substrates that comprise only semiconducting layers, e.g. on Si wafers or on GaAs-on-Si wafers characterised by the integration of at least one component covered by groups H10D12/00 or H10D30/00, e.g. integration of IGFETs of only field-effect components
    • H10D84/83—Integrated devices formed in or on semiconductor substrates that comprise only semiconducting layers, e.g. on Si wafers or on GaAs-on-Si wafers characterised by the integration of at least one component covered by groups H10D12/00 or H10D30/00, e.g. integration of IGFETs of only field-effect components of only insulated-gate FETs [IGFET]
    • H10D84/85—Complementary IGFETs, e.g. CMOS

Definitions

  • the present invention relates to fabricating integrated circuits, and more particularly to providing a low resistance to integrated circuit devices.
  • One technique for reducing the resistance in silicon circuits is the self-aligned silicide ("salicide") technique.
  • silicide a metal such as titanium is deposited after the formation of bulk transistors. The metal is reacted with the silicon that contacts the metal. The reaction produces a conductive metal silicide on the silicon surfaces. The unreacted metal is then stripped. The metal silicide has a lower resistivity than the silicon. Since practically all the current is carried by the metal silicide rather than the silicon, the resistance to the intrinsic device regions is reduced.
  • the silicidation reaction consumes silicon and thus reduces the depth of silicon regions such as extrinsic base regions of bipolar transistors and source and drain regions of field effect transistors.
  • the silicide must be sufficiently thick to obtain a desired low resistance.
  • the thicker the silicide the more silicon is consumed, and hence the greater the leakage currents between base and collector regions of bipolar transistors and from source and drain regions to the body regions of MOS transistors.
  • the increased leakage current, and in the extreme case shorting out of the base, source and drain regions due to total consumption of those regions reduce the integrated circuit yield.
  • the base, source and drain regions are made deeper. This, however, increases the integrated circuit size and capacitances and lowers ⁇ ource-to-drain punchthrough voltages.
  • the present invention allows in some embodiments obtaining a lower resistance to integrated circuit devices without a large consumption of the semiconductor material of the integrated circuit device regions.
  • the local interconnects are interconnects that are formed after the transistors but before formation of any interlevel dielectric in which contact openings are formed to contact the transistors. Thus local interconnects do not require a contact opening in order to contact a transistor region. However, because an interlevel dielectric has not been formed, a local interconnect may be unable to pass over a transistor region without contacting the region.
  • local conductive line means a local interconnect or a conductive line which, like a local interconnect, is formed after the transistors but before formation of any interlevel dielectric and which is used as an extension of a transistor region to allow contacting the transistor region over a field insulation. Contacting a transistor region over a field insulation allows making the transistor region smaller than a contact to the region, which in turn allows increasing the device packing density and the circuit speed.
  • a metal silicide is formed by the salicide technique described above, and then local conductive lines are formed in a separate sequence of steps to a desired thickness to achieve a desired low resistance.
  • the metal silicide is titanium silicide, and the conductive lines are formed from titanium nitride.
  • the local conductive lines are made sufficiently thick to have a desired low sheet resistance. Because much current is carried by the local conductive lines, a low resistance is obtained even with a thin metal silicide. Formation of the thin silicide does not require a large silicon consumption. Hence, the yield is increased, and the base, source and drain regions can be made shallower to decrease the circuit size and increase the circuit speed.
  • the emitter is defused from a doped semiconductor region such as a polysilicon emitter contact region.
  • the emitter resistance tends to be fairly high because of a high resistance of the polysilicon contact region.
  • the emitters are made narrow to reduce the base resistance.
  • emitter contact openings in the overlying dielectric are made larger than the emitter width in order to provide a low contact resistance.
  • the contact openings are formed at the ends of the emitter away from the transistor active region. Because the emitter is narrow and the contact openings are formed at the emitter ends, the emitter resistance is high. The high emitter resistance limits the possible emitter length and hence the emitter area and the emitter current.
  • high power transistors such as output transistors have to be provided with multiple emitters, which undesirably increases the transistor area.
  • the salicide technique has been used to silicide the polysilicon emitter contact region and thus to reduce the emitter resistance.
  • the silicide over the emitter contact region is formed at the same time as over the base region, and hence the silicide thickness over the emitter is limited by the base- collector leakage current as described above.
  • the emitter resistance is reduced further by providing a local conductive line that runs over the entire emitter contact region to the emitter contact openings.
  • the thickness of the local conductive lines is not limited by the base-collector leakage current considerations related to the salicide technique. Hence, the emitter resistance is reduced without creating the leakage and yield problems of the salicide technology.
  • FIG. 1 and 2 illustrate cross sections of a
  • Fig. 3 is a top view illustrating certain features of the BiCMOS structure.
  • Figs. 4-16 are cross-section illustrations of the
  • Fig. 17 is a top view illustrating certain features of the BiCMOS structure fabricated according to the present invention.
  • Fig. 1 illustrates the initial steps of the fabrication of a BiCMOS structure suitable for high frequency, high current drive applications.
  • the particular materials, layer thicknesses, temperatures, doping concentrations and doses and other particulars- are merely illustrative and do not limit the invention unless specifically noted otherwise.
  • the fabrication starts with P+ monocry ⁇ talline silicon substrate 110 having a resistivity of 24-36 ohm-cm.
  • P- layer 120 is grown epitaxially on substrate 110.
  • the thickness of layer 120 is 5 to 10 ⁇ m, and the doping concentration is lxio" atoms/cm 3 .
  • the fabrication starts with a Pr wafer, and epitaxial layer 120 is omitted.
  • the structure is denuded.
  • a silicon dioxide layer 130 is then grown thermally at 900°C to a thickness of 25 ⁇ A.
  • a photoresist mask (not shown) is formed and arsenic is implanted to form N+ buried layer 140 of the bipolar transistor.
  • the resulting net N type doping concentration is 2xl0 19 atoms/cm 3 .
  • the resist mask is stripped, and the buried layer is annealed at 1,100°C for 75 minutes. During the anneal, the thickness of oxide 130 increases, especially over the heavily doped buried layer.
  • a photoresist mask (not shown) is formed to expose channel stop region 150 laterally surrounding the buried layer 140. Boron is implanted into region 150 at an energy of 150 KeV. The dose is lxl0 M atoms/cm 2 , and the resulting net doping concentration is lxlO 17 atoms/cm 3 .
  • the doping concentration is chosen on the one hand sufficiently high to obtain a low leakage current and a high punchthrough voltage between the collector of the transistor being described and other collectors, if any, formed in the same integrated circuit. On the other hand, the doping concentration is sufficiently low to get a low capacitance between buried layer 140 and channel stop region 150.
  • channel stop region 150 is annealed in a non-oxidizing nitrogen- containing ambient at 950°C for 45 minutes.
  • N— layer 210 (Fig. 2) is grown epitaxially over the structure to a thickness of about 1.2 ⁇ .
  • the doping concentration in layer 210 is about 4.5xl0 14 atoms/cm 3 .
  • Buried layer 140 and channel stop region 150 extend some distance into layer 210 due to dopant diffusion.
  • Pad silicon dioxide layer 220 is formed over the layer 210 to a thickness of 45 ⁇ A.
  • Layer 220 is grown by thermal oxidation performed at 900°C for 20 minutes. Buried layer 140 and channel stop region 150 diffuse out further during the formation of layer 220.
  • N collector region 240 and N guard ring 250 are formed, and phosphorus is implanted to form N collector region 240 and N guard ring 250.
  • the implantation energy is 180 KeV
  • the dose is 3xl0 12 atoms/cm 2
  • the net doping concentration is 4.5xl0 16 atom/cm 3 .
  • Guard ring 250 (Figs. 2, 3) laterally surrounds the bipolar transistor active region.
  • Collector region 240 meets the guard ring on two sides as shown in Fig. 3. In some embodiments, collector region 240 meets the guard ring on one side only or does not meet the guard ring at all. Such a smaller collector region provides a lower collector-base capacitance but, on the other hand, a higher collector resistance.
  • the doping concentration stated above to be 4.5xl0 16 atoms/cm 3 , is generally optimized for a trade-off between a large current density and a small collector-base capacitance. Because the higher doped collector region 240 does not extend throughout the active region surrounded by guard ring 250, the extrinsic collector-base region has a reduced capacitance. The unity gain frequency Ft is high as a result.
  • guard ring 250 will meet buried layer 140. A portion of the guard ring will provide a sink region for contacting the collector region through the buried layer. Collector region 240 will meet buried layer 140 to provide a low collector resistance.
  • Silicon nitride layer 230 is deposited over the pad oxide to a thickness of 1350A.
  • a photoresist mask (not shown) is formed that exposes the location of a to-be-formed N-well 310 (Figs. 3, 4) of the PMOS transistor.
  • Nitride 230 is etched away at the location of well 310. The photoresist is then stripped, and phosphorus is implanted with nitride 230 as a mask to form well 310 in epitaxial layer 210.
  • the net doping concentration of well 310 is 7xl0 16 atoms/cm 3 .
  • the structure is then oxidized at 950°C for 180 minutes to increase the thickness of pad oxide 220 over well 310 to about 5,000A.
  • Silicon nitride 230 inhibits oxidation and oxidation-induced dopant updiffusion at the location of the bipolar and NMOS transistors.
  • photoresist mask 510 (Fig. 5) is formed over the bipolar transistor.
  • Nitride 230 is etched away over the to-be-formed P-well 520 (Figs. 3, 6). Resist 510 is stripped, and the bipolar transistor remains covered by nitride 230.
  • silicon nitride 230 inhibits oxidation over the bipolar transistor.
  • Layer 230 also inhibits oxidation-enhanced diffusion of buried layer 140. Because oxidation is inhibited, oxidation-induced stacking faults in the bipolar transistor are reduced, leading to a better yield.
  • Thinner epitaxial layer 210 permits in some applications the P-well 520 to be shallower. More particularly, some applications require P-vell 520 to reach through epitaxial layer 210 to meet epitaxial layer 120 and thus to electrically contact the Pi- substrate 110 which in these applications is tied to the common ground of the NMOS transistors. In addition to contacting the common ground, P-well 520 diffusing through N— layer 210 allows isolating the N-well 310 from other N-wells (not shown) , if any, in the same integrated circuit. Since epitaxial layer 210 is thin, well 520 can be made shallower.
  • thinner epitaxial layer 210 advantageously reduces the size of active regions of parasitic transistors.
  • Nitride 230 over the bipolar transistor also reduces the diffusion of collector region 240, guard ring 250 and channel stop region 150. Reducing the diffusion of regions 240, 250, 150 allows reducing the bipolar transistor area and, therefore, the transistor capacitances including the base-collector capacitance. Reducing the lateral and downward diffusion of buried layer 140 reduces the collector-substrate capacitance.
  • a photoresist mask (not shown) is formed that exposes the to-be-formed well 520. The mask need not be perfectly aligned with the boundary between N-well 310 and P-well 520 because the thicker oxide 220 over N-well 310 masks the N-well during the P-well implant. Some embodiments do not use the oxide over N-well 310 as a mask, and instead rely on the photoresist to. mask the N-well.
  • Boron is implanted at an energy of 150 KeV and a dose of 3.3xl0 12 atoms/cm 2 to form P-well 520.
  • the photoresist is then stripped, and the wells are driven in at the temperature of 1,100°C for 50 minutes in an oxidizing ambient.
  • the thickness of oxide 220 increases over the MOS transistors to planarize the surface of wells 310, 520.
  • the oxidation is inhibited by nitride 230 as described above. Buried layer 140 and channel stop regions 150 increase during this step due to dopant diffusion. The dopant diffusion, however, is inhibited by layer 230 as described above.
  • Guard ring 250 and collector region 240 also increase and meet the buried layer to provide a well-defined guard ring and a well-defined collector region.
  • the active region of the bipolar transistor becomes completely enclosed by the guard ring and the buried layer within epitaxial layer 210.
  • epitaxial layer 210 is thin, the guard ring and the collector region need not be doped as heavily to meet buried layer 140 as would be required if epitaxial layer 210 were thicker.
  • the lighter doping of guard ring 250 and collector region 240 reduces the junction capacitances in the bipolar transistor and thus provides higher speed.
  • the implant that forms buried layer 140 of Fig. 1 forms also a buried layer 140.1 (Fig. 7) at the location of the to-be-formed PMOS transistor.
  • the implant that forms channel stop region 150 forms also a buried layer 150.1 at the location of the to-be-formed NMOS transistor.
  • N-well 310 and P-well 520 are created using shallow implants sufficient to cause the wells to meet the respective buried layers 140.1, 150.1.
  • the techniques of Figs. 1-6 allow high speed bipolar transistors to be integrated in an existing CMOS twin well process without using buried layers 140.1, 150.1.
  • the doping of regions 140, 150 can be optimized for the bipolar transistor to obtain a low leakage current, low capacitances and a low collector resistance, as the doping does not affect the MOS transistors.
  • the result is decoupling the MOS and bipolar transistor characteristics from each other without increasing the process complexity.
  • the remaining fabrication steps are similar for the embodiments of Figs. 6 and 7 and are illustrated only for the embodiment of Fig. 6 for simplicity.
  • Nitride 230 and oxide 220 are stripped.
  • Silicon dioxide layer 810 (Fig. 8) is formed to a thickness of 25 ⁇ A.
  • Layer 220 is formed by thermal oxidation performed at 900°C for 60 minutes.
  • Silicon nitride layer 820 is deposited to a thickness of l,85 ⁇ A.
  • a photoresist mask (not shown) is formed that exposes the regions in which field oxide 830.1, 830.2 is to be grown by the LOCOS process.
  • a composite nitride/oxide etch removes silicon nitride 820 and silicon dioxide 810 off the regions exposed by the mask.
  • the photoresist is then stripped, and another photoresist mask (not shown) is formed for a P+ boron implant that creates field implant regions 840, 840.1.
  • the boron is implanted at an energy of 30 KeV and a dose of 5.5xl0 13 atoms/cm 2 .
  • the resulting net doping concentration of regions 840, 840.1 is 2xlO ⁇ atoms/cm 3 .
  • the field implant region 840.1 that overlies channel stop region 150 meets the channel stop region to create a P+ ring around the bipolar transistor. This ring isolates the bipolar transistor from other transistors of the integrated circuit.
  • field oxide 830.1, 830.2 is thermally grown on the exposed silicon areas by the LOCOS process at 1,000°C for 160 minutes.
  • Oxide regions 830.1 overlie field implant regions 840, 840.1 and guard ring 250.
  • Field oxide regions 830.1 and field implant regions 840, 840.1 surround laterally each of the bipolar, PMOS, and NMOS transistors.
  • Field oxide region 830.2 extends between collector region 240 and the portion 250.1 of guard ring 250.
  • the guard ring portion 250.1 forms a sink region providing a low resistance path to buried layer 140 from the to-be- formed collector contact (not shown in Fig. 8) .
  • Field oxide region 830.2 separates the sink region 250.1 from the to-be-formed base region (not shown in Fig. 8) to reduce the base-collector capacitance.
  • Nitride 820 and oxide 810 are then stripped. Threshold-voltage-adjust implants are performed optionally into N-well 310 or P-well 520 or both.
  • Gate oxide layer 910 (Fig. 9) is thermally grown over the exposed silicon areas of the MOS and bipolar transistors at 900°C for 28 minutes to a thickness of 15 ⁇ A.
  • a protective intrinsic polysilicon layer 920 is deposited over the wafer to protect the gate oxide during subsequent processing steps.
  • a base implant mask 930 is formed from photoresist. Boron is implanted at an energy of 20 KeV and a dose of 5xl0 13 atoms/cm 2 to convert top portion 940 of epitaxial layer 210 in the bipolar transistor active region to the P conductivity type. Polysilicon layer 920 and gate oxide 910 combine to reduce the implant width, providing a shallower base with existing implanters. Because the base is formed after the gate oxide, the base region does not experience any oxidation or dopant redistribution during the gate oxide formation and in some embodiments does not see any significant diffusion/anneal cycles until the RTP. (rapid thermal processing) anneal which forms the emitter region (not shown in Fig. 9) . This RTP anneal is described below in connection with Fig. 14. Thus a shallow well-defined base region results. The shallow base region leads to a higher bipolar transistor speed and allows epitaxial layer 210 to be thinner.
  • RTP rapid thermal processing
  • collector region 240 is reduced by the base implant.
  • the enhanced doping of region 240 reduces the base width over region 240.
  • Field oxide region 830.1 and P+ regions 840.1, 150 laterally surround and isolate the bipolar transistor.
  • Substrate 110, and hence P+ regions 840.1, 150, are typically held at the lowest potential of the circuit during operation.
  • Guard ring 250 between base region 940 and field implant region 840.1 prevents the P dopant of region 840.1 from counterdoping the epitaxial layer at the guard ring location. Such counterdoping could cause field implant region 840.1 to meet, or come close to, the base region leading to a base-to- substrate short or high leakage current.
  • guard ring keeps the depletion layer between region 840.1 and the N type portion of epitaxial layer 210 farther from the base, thereby increasing the punchthrough voltage of the parasitic PNP transistor formed by the base region, the P+ region 840.1 and the portion of epitaxial layer 210 between the base region and the P+ region.
  • Guard ring 250 also impedes the formation of the inversion layer between the base region and the P+ region and thus further reduces the current leakage from the base region to substrate 110.
  • P+ regions 840.1, 150 are separated from buried layer 140 and guard ring 250 by portions of epitaxial layers 210, 120 which portions have a lower doping concentration than the two P+ regions, the buried layer and the guard ring.
  • the collector capacitance is reduced as a result, and the breakdown voltage between channel stop region 150 and buried layer 140 is increased.
  • the base region is separated from the guard ring and the buried layer by N— regions 950.1, 950.2 of epitaxial layer 210 which have a lower doping concentration than the base region and the guard ring.
  • the collector-base capacitance is reduced as a result.
  • Photoresist 930 is stripped, and a photoresist mask 1010 (Fig. 10) is formed to define the emitter.
  • Polysilicon 920 is etched through an opening in mask 1010 by a plasma etch. The etch stops on gate oxide 910.
  • resist 1010 is removed, and gate oxide 910 is etched with polysilicon 920 as a mask.
  • the resulting opening 1110 exposes region 940 at the location of the to-be-formed emitter.
  • polysilicon 920 protects the gate oxide over the wells of the MOS transistors.
  • the gate oxide etch is a wet buffered oxide HF etch designed to reduce any damage to, and removal of, the P doped region 940.
  • Intrinsic polysilicon is deposited by chemical vapor deposition (CVD) over the wafer at 630°C to a thickness of 3,25 ⁇ A to combine with the polysilicon 920 to form polysilicon layer 1210 (Fig. 12) . Even though layer 920 is removed at the emitter location, layer 1210 has a substantially uniform thickness because layer 920 is thin compared to layer 1210.
  • CVD chemical vapor deposition
  • a masked implant of arsenic at an energy of 100 KeV and a dose of 1.2xl0 16 atoms/cm 2 creates N+ emitter contact region 1220 and N+ gate regions 1230, 1240.
  • the net doping concentration of emitter contact region 1220 and gate regions 1230, 1240 is lxl0 20 atoms/cm 3 .
  • An optional P type masked implant forms resistors (not shown) in layer 1210.
  • the polysilicon is then annealed at 900°C for 15- inutes.
  • Polysilicon 1210 is then masked by a photoresist (not shown) and etched to define emitter contact region 1220 and gate regions 1230, 1240 (Fig. 13) .
  • Polysilicon resistors (not shown) , if any, are also defined at this step.
  • the etch stops on gate oxide 910 which protects epitaxial layer 210 and, in particular, the damage-sensitive base region 940.
  • Emitter contact region 1220 extends over the edges of the emitter opening in oxide 910.
  • Silicon dioxide 1310 is grown thermally at 900°C for 30 minutes over polysilicon 1210 to a thickness of
  • the NMOS transistor is then masked by a photoresist (not shown) , and a lightly doped drain (LDD) implant of boron is performed at an energy of 45 KeV and a dose of lxlO 13 atoms/cm 2 into N-well 310 and region 940 while gate region 1230 and emitter contact region 1220 mask, respectively, the PMOS channel region and the intrinsic base region.
  • LDD lightly doped drain
  • the resulting net doping concentration of the extrinsic base region is 4xl ⁇ 18 atoms/cm 3 .
  • This implant forms lightly doped source/drain regions 1320.1, 1320.2 of the PMOS transistor and reduces the extrinsic base resistance.
  • region 940 is not doped during this step.
  • the photoresist is stripped.
  • Another photoresist mask (not shown) is formed over the bipolar and PMOS transistors.
  • An LDD implant of arsenic is performed at an energy of 90 KeV and a dose of 4xl0 u atoms/cm 2 while gate region 1240 masks the NMOS channel region.
  • This implant forms lightly doped source/drain regions 1330.1, 1330.2 of the NMOS transistor.
  • sink region 250.1 is also doped during this implant.
  • a confor al layer of silicon dioxide is deposited by CVD at 650°C.
  • An optional "silicide exclusion" mask (not shown) is formed from photoresist over selected areas to protect the oxide and thus to prevent formation of a metal silicide on such areas in subsequent processing described below.
  • This CVD oxide is etched anisotropically to form spacers 1410 (Fig. 14) on the sidewalls of emitter contact region 1220 and gate regions 1230, 1240.
  • the gate oxide not covered by the spacers and by polysilicon regions 1220, 1230, 1240 is removed during this etch.
  • the remaining gate oxide over the bipolar transistor is completely covered by emitter contact region 1220 and spacers 1410.
  • the overetch required is uniform over the bipolar, PMOS and NMOS transistors because the gate oxide has a uniform thickness over the bipolar and MOS transistors, because the CVD silicon dioxide has a uniform thickness over the transistors except, perhaps, on the sidewalls of polysilicon portions 1220, 1230, and 1240, and because oxide 1310 is thin compared to the CVD oxide.
  • a masked N+ arsenic implant is performed into source/drain regions 1330.1, 1330.2 and sink region 250.1 at an energy of 30 KeV and a doping dose of 3xl0 15 atoms/cm 2 .
  • the resulting doping concentration in the top portion 250.2 of the sink region is lxlO 20 atoms/cm 3 .
  • a masked P+ boron implant is performed into source/drain regions 1320.1, 1320.2 and region 940 at an energy of 45 KeV and a dose of 3X10 15 atoms/cm 2 .
  • the resulting doping concentration in the extrinsic base region not covered by spacers 1410 is 5xl0 19 atoms/cm 3 .
  • RTP Rapid Thermal Processing
  • a refractory metal titanium in some embodiments, is deposited over the structure by chemical vapor deposition to a thickness of 60 ⁇ A.
  • the structure is heated and kept at 650°C for 30 minutes to react the titanium that contacts silicon surfaces with the silicon.
  • titanium silicide 1510 (Fig. 15) is formed on the exposed silicon surfaces including the surfaces of the extrinsic base region, emitter contact region 1220, sink region 250.1, and source/drain and gate regions of the MOS transistors.
  • the sheet resistance of layer 1510 is 3 ohms/square.
  • the titanium silicide reduces resistance to the intrinsic transistor regions and thus increases the circuit speed and reduces the power consumption.
  • Titanium nitride layer 1520 is deposited over, and in contact with, the entire silicide 1510 by chemical vapor deposition to a thickness of 300 ⁇ A. The deposition temperature is 650°C. Layer 1520 is patterned to form local conductive lines that contact titanium silicide 1510. As used herein, the term "local conductive lines” means conductive lines that are formed after the transistors but before formation of any interlevel dielectric in which contact openings are formed to contact the transistors. Layer 1520 has a sheet resistance of 1.5 ohms/square. A different sheet resistance can be obtained by changing the thickness of layer 1520. Conductive lines 1520 are used as interconnects and/or as extensions of the emitter, base, source/drain and gate regions. Such extensions allow forming contacts to these regions over field oxide 830.1, 830.2 rather than over those regions themselves.
  • TiN is etched off the gate regions.
  • a conductive line 1520 runs along the top surface of emitter contact region 1220, reducing the emitter resistance.
  • TiN lines run along, and/or contact, the top surfaces of gate region 1230 and/or gate region 1240.
  • the spacing between the TiN lines over base region 940 on the one hand and emitter contact region 1220 on the other hand is smaller than the minimal photolithography dimension obtainable by the fabrication equipment. The same is true for the spacing between lines 1520 and each of gate regions 1230, 1240. Spacers 1410 help isolate conductive lines 1520 from regions 1220, 1230, 1240.
  • lines 1520 are thick, a low resistance can be obtained with but a small overlap of lines 1520 over base region 940 and over the source and drain regions. The areas of the base, source and drain regions can therefore be reduced, providing a high packing density and a high speed.
  • interlevel dielectric 1610 is formed by TEOS deposition of silicon dioxide to a thickness of l,OO ⁇ A, TEOS deposition of boron and phosphorus doped silicon dioxide to a thickness of 7,OO ⁇ A, and a densification and reflow of the boron and phosphorus doped silicon dioxide.
  • Contact openings are etched in layer 1610. The contact openings terminate on conductive lines 1520. Tungsten is deposited by CVD and etched back to form tungsten plug contacts 1620.i in the contact openings. The plugs physically contact the titanium nitride lines 1520.
  • Conductive lines 1630 are formed from Al-Si-Cu or some other suitable material. The fabrication is completed using known processing technigues.
  • Fig. 17 is a top view of the resulting structure.
  • Emitter region 1420 is a non-walled emitter region.
  • Gate oxide 910 (Fig. 13) isolates emitter contact region 1220 from base region 940 outside emitter region 1420 and,, in particular, at the boundary between base region 940 and field oxide 830.1, eliminating yield and leakage problems described in U.S. patent application Serial No. 08/085,436 filed by M.J. Grubisich on June 30, 1993 and hereby incorporated herein by reference.
  • emitter 1420 is a walled emitter.
  • collector region 240 (Fig. 3) is in top view coextensive with emitter region 1420, providing a low base-collector capacitance.
  • Titanium nitride lines 1520 reduce the resistance allowing a low resistance to be obtained with thinner titanium silicide 1510 (Fig. 15) , especially since the titanium nitride has a lower resistivity than the titanium silicide. Because the titanium silicide is thinner, the consumption of silicon during the silicide formation is reduced, which lessens the chance of shorting out the junctions between the base and the collector and between the source/drain regions and the MOS wells 310, 520. Hence these junctions can be made shallower. The silicide outgrowth which results in stringers on spacers 1410 (Fig. 15) is also reduced, leading to a better yield.
  • the formation of stringers can be inhibited by lower silicidation temperatures, but the lower temperature also increases the silicide resistivity. Due to lines 1520, the stringer formation is inhibited by making titanium silicide thin. Hence, the silicidation temperature can be made higher to reduce the silicide resistivity.
  • Emitter contact opening 1620.7 and gate contact openings 1620.8 through 1620.11 are formed over field oxide 830.1. Titanium nitride line 1520 runs along emitter contact region 1220 to the contact opening 1620.7, reducing the emitter resistance and thus improving the VBE matching. As a result, the emitter current and the bipolar transistor speed can be increased without increasing the emitter area.
  • Figs. 1-8 Some embodiments combine the isolation techniques of Figs. 1-8 with other processes than those of Figs. 9-17.
  • the base is formed by diffusion from a P-doped polysilicon layer as described, for example, in the U.S. Patent No. 5,219,784 issued June 15, 1993 to A.G. Solheim and incorporated herein by reference.
  • the emitter is also formed as described in that patent.
  • Other fabrication techniques, including non-polysilicon emitter technigues, are used with the structure of Fig. 8.
  • isolation techniques of Figs. 1-8 are used in some embodiments with other silicidation/local conductive line techniques than those of Figs. 15, 16 or without any silicidation/local conductive line techniques.
  • base and emitter formation techniques of Figs. 9-14 are combined in some embodiments with other isolation techniques including trench isolation or junction isolation.
  • the techniques of Figs. 9-14 are used in some embodiments without any silicidation/local conductive line techniques or with silicidation/local conductive line techniques different from those of Figs. 15, 16.
  • silicidation/local conductive line techniques of Figs. 15, 16 are used in some embodiments with different isolation techniques than those of Figs. l-*8 and/or with different emitter/base formation techniques than those of Figs. 9-14.
  • Conductive lines 1520 are formed in some embodiments from other materials than titanium nitride, including other non-semiconductor materials. Lines 1520 are formed from titanium, tungsten or titanium tungsten in some embodiments. In some embodiments, the conductivity types are reversed to form PNP transistors in a BiCMOS circuit. Some embodiments include non-BiCMOS integrated circuits containing only bipolar or bipolar and MOS transistors. In some embodiments, an additional implant is performed into the collector region 240 to reduce the collector resistance. Other embodiments and variations are within the scope of the invention as defined by the following claims.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Condensed Matter Physics & Semiconductors (AREA)
  • General Physics & Mathematics (AREA)
  • Manufacturing & Machinery (AREA)
  • Computer Hardware Design (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Power Engineering (AREA)
  • Metal-Oxide And Bipolar Metal-Oxide Semiconductor Integrated Circuits (AREA)
  • Bipolar Transistors (AREA)
  • Electrodes Of Semiconductors (AREA)
  • Internal Circuitry In Semiconductor Integrated Circuit Devices (AREA)

Abstract

To obtain a low resistance to integrated circuit devices, a metal silicide (1510) is formed by a salicide technique. Then a conductive layer (1520) is deposited and patterned to form local conductive lines. The local conductive lines are sufficiently thick to obtain a desired low resistance even with a thin metal silicide (1510).

Description

PROVIDING A LOW RESISTANCE TO INTEGRATED CIRCUIT DEVICES
BACKGROUND OF THE INVENTION
The present invention relates to fabricating integrated circuits, and more particularly to providing a low resistance to integrated circuit devices. In integrated circuits, it is desirable to reduce the resistance to devices such as transistors and diodes because a lower resistance makes the integrated circuits faster and reduces the circuit power consumption. One technique for reducing the resistance in silicon circuits is the self-aligned silicide ("salicide") technique. In this technique, a metal such as titanium is deposited after the formation of bulk transistors. The metal is reacted with the silicon that contacts the metal. The reaction produces a conductive metal silicide on the silicon surfaces. The unreacted metal is then stripped. The metal silicide has a lower resistivity than the silicon. Since practically all the current is carried by the metal silicide rather than the silicon, the resistance to the intrinsic device regions is reduced.
One disadvantage of this technique is that the silicidation reaction consumes silicon and thus reduces the depth of silicon regions such as extrinsic base regions of bipolar transistors and source and drain regions of field effect transistors. The silicide must be sufficiently thick to obtain a desired low resistance. However, the thicker the silicide, the more silicon is consumed, and hence the greater the leakage currents between base and collector regions of bipolar transistors and from source and drain regions to the body regions of MOS transistors. The increased leakage current, and in the extreme case shorting out of the base, source and drain regions due to total consumption of those regions, reduce the integrated circuit yield. To increase the yield, the base, source and drain regions are made deeper. This, however, increases the integrated circuit size and capacitances and lowers εource-to-drain punchthrough voltages.
Thus, there is a need for a fabrication method that provides a low resistance to integrated circuit device regions but does not require a large consumption of the semiconductor material of regions such as base, source and drain regions.
SUMMARY OF THE INVENTION
The present invention allows in some embodiments obtaining a lower resistance to integrated circuit devices without a large consumption of the semiconductor material of the integrated circuit device regions.
This and other advantages are obtained in some embodiments by using local conductive lines such as local interconnects. The local interconnects are interconnects that are formed after the transistors but before formation of any interlevel dielectric in which contact openings are formed to contact the transistors. Thus local interconnects do not require a contact opening in order to contact a transistor region. However, because an interlevel dielectric has not been formed, a local interconnect may be unable to pass over a transistor region without contacting the region.
The term "local conductive line" as used herein means a local interconnect or a conductive line which, like a local interconnect, is formed after the transistors but before formation of any interlevel dielectric and which is used as an extension of a transistor region to allow contacting the transistor region over a field insulation. Contacting a transistor region over a field insulation allows making the transistor region smaller than a contact to the region, which in turn allows increasing the device packing density and the circuit speed.
In some embodiments of the present invention, a metal silicide is formed by the salicide technique described above, and then local conductive lines are formed in a separate sequence of steps to a desired thickness to achieve a desired low resistance. In some embodiments, the metal silicide is titanium silicide, and the conductive lines are formed from titanium nitride. The local conductive lines are made sufficiently thick to have a desired low sheet resistance. Because much current is carried by the local conductive lines, a low resistance is obtained even with a thin metal silicide. Formation of the thin silicide does not require a large silicon consumption. Hence, the yield is increased, and the base, source and drain regions can be made shallower to decrease the circuit size and increase the circuit speed.
An important goal in many integrated circuits is reducing the emitter resistance to increase the emitter current and improve the VBE matching. In some processes including so-called "polysilicon emitter" processes, the emitter is defused from a doped semiconductor region such as a polysilicon emitter contact region. In the resulting structures, the emitter resistance tends to be fairly high because of a high resistance of the polysilicon contact region. Moreover, in order to obtain a high frequency performance, the emitters are made narrow to reduce the base resistance. At the same time, emitter contact openings in the overlying dielectric are made larger than the emitter width in order to provide a low contact resistance. Hence, the contact openings are formed at the ends of the emitter away from the transistor active region. Because the emitter is narrow and the contact openings are formed at the emitter ends, the emitter resistance is high. The high emitter resistance limits the possible emitter length and hence the emitter area and the emitter current.
To provide a larger current, high power transistors such as output transistors have to be provided with multiple emitters, which undesirably increases the transistor area.
The salicide technique has been used to silicide the polysilicon emitter contact region and thus to reduce the emitter resistance. However, the silicide over the emitter contact region is formed at the same time as over the base region, and hence the silicide thickness over the emitter is limited by the base- collector leakage current as described above.
In some embodiments of the present invention, the emitter resistance is reduced further by providing a local conductive line that runs over the entire emitter contact region to the emitter contact openings. The thickness of the local conductive lines is not limited by the base-collector leakage current considerations related to the salicide technique. Hence, the emitter resistance is reduced without creating the leakage and yield problems of the salicide technology.
BRIEF DESCRIPTION OF THE DRAWINGS Figs. 1 and 2 illustrate cross sections of a
BiCMOS structure during fabrication according to the present invention.
Fig. 3 is a top view illustrating certain features of the BiCMOS structure. Figs. 4-16 are cross-section illustrations of the
BiCMOS structure during fabrication according to the present invention.
Fig. 17 is a top view illustrating certain features of the BiCMOS structure fabricated according to the present invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Fig. 1 illustrates the initial steps of the fabrication of a BiCMOS structure suitable for high frequency, high current drive applications. Below, the particular materials, layer thicknesses, temperatures, doping concentrations and doses and other particulars- are merely illustrative and do not limit the invention unless specifically noted otherwise.
The fabrication starts with P+ monocryεtalline silicon substrate 110 having a resistivity of 24-36 ohm-cm. P- layer 120 is grown epitaxially on substrate 110. In some embodiments, the thickness of layer 120 is 5 to 10 μm, and the doping concentration is lxio" atoms/cm3. In some embodiments, the fabrication starts with a Pr wafer, and epitaxial layer 120 is omitted.
The structure is denuded. A silicon dioxide layer 130 is then grown thermally at 900°C to a thickness of 25θA. A photoresist mask (not shown) is formed and arsenic is implanted to form N+ buried layer 140 of the bipolar transistor. The resulting net N type doping concentration is 2xl019 atoms/cm3.
The resist mask is stripped, and the buried layer is annealed at 1,100°C for 75 minutes. During the anneal, the thickness of oxide 130 increases, especially over the heavily doped buried layer.
A photoresist mask (not shown) is formed to expose channel stop region 150 laterally surrounding the buried layer 140. Boron is implanted into region 150 at an energy of 150 KeV. The dose is lxl0M atoms/cm2, and the resulting net doping concentration is lxlO17 atoms/cm3. The doping concentration is chosen on the one hand sufficiently high to obtain a low leakage current and a high punchthrough voltage between the collector of the transistor being described and other collectors, if any, formed in the same integrated circuit. On the other hand, the doping concentration is sufficiently low to get a low capacitance between buried layer 140 and channel stop region 150.
The resist is then stripped, and channel stop region 150 is annealed in a non-oxidizing nitrogen- containing ambient at 950°C for 45 minutes.
Oxide 130 is then stripped, and N— layer 210 (Fig. 2) is grown epitaxially over the structure to a thickness of about 1.2 μ . The doping concentration in layer 210 is about 4.5xl014 atoms/cm3. Buried layer 140 and channel stop region 150 extend some distance into layer 210 due to dopant diffusion.
Pad silicon dioxide layer 220 is formed over the layer 210 to a thickness of 45θA. Layer 220 is grown by thermal oxidation performed at 900°C for 20 minutes. Buried layer 140 and channel stop region 150 diffuse out further during the formation of layer 220.
Next a photoresist mask (not shown) is formed, and phosphorus is implanted to form N collector region 240 and N guard ring 250. The implantation energy is 180 KeV, the dose is 3xl012 atoms/cm2, and the net doping concentration is 4.5xl016 atom/cm3. Guard ring 250 (Figs. 2, 3) laterally surrounds the bipolar transistor active region. Collector region 240 meets the guard ring on two sides as shown in Fig. 3. In some embodiments, collector region 240 meets the guard ring on one side only or does not meet the guard ring at all. Such a smaller collector region provides a lower collector-base capacitance but, on the other hand, a higher collector resistance. The doping concentration, stated above to be 4.5xl016 atoms/cm3, is generally optimized for a trade-off between a large current density and a small collector-base capacitance. Because the higher doped collector region 240 does not extend throughout the active region surrounded by guard ring 250, the extrinsic collector-base region has a reduced capacitance. The unity gain frequency Ft is high as a result.
Due to dopant diffusion in subsequent processing steps, guard ring 250 will meet buried layer 140. A portion of the guard ring will provide a sink region for contacting the collector region through the buried layer. Collector region 240 will meet buried layer 140 to provide a low collector resistance.
Silicon nitride layer 230 is deposited over the pad oxide to a thickness of 1350A. A photoresist mask (not shown) is formed that exposes the location of a to-be-formed N-well 310 (Figs. 3, 4) of the PMOS transistor. Nitride 230 is etched away at the location of well 310. The photoresist is then stripped, and phosphorus is implanted with nitride 230 as a mask to form well 310 in epitaxial layer 210. The net doping concentration of well 310 is 7xl016 atoms/cm3.
The structure is then oxidized at 950°C for 180 minutes to increase the thickness of pad oxide 220 over well 310 to about 5,000A. Silicon nitride 230 inhibits oxidation and oxidation-induced dopant updiffusion at the location of the bipolar and NMOS transistors.
Next, photoresist mask 510 (Fig. 5) is formed over the bipolar transistor. Nitride 230 is etched away over the to-be-formed P-well 520 (Figs. 3, 6). Resist 510 is stripped, and the bipolar transistor remains covered by nitride 230. During the subsequent drive-in of wells 310, 520 (described below) , silicon nitride 230 inhibits oxidation over the bipolar transistor. Layer 230 also inhibits oxidation-enhanced diffusion of buried layer 140. Because oxidation is inhibited, oxidation-induced stacking faults in the bipolar transistor are reduced, leading to a better yield. Reducing oxidation of epitaxial layer 210 and the oxidation-enhanced upward diffusion of layer 140 allows using thinner epitaxial layer 210. Making epitaxial layer 210 thinner reduces the resistance of the vertical collector portions including the portion containing collector region 240 and that portion of ring 250 which provides the sink region. Higher speed is provided as a result. Of note, collector region 240 has a lighter doping than buried layer 140 and hence collector region 240 provides a large component of the collector resistance.
Thinner epitaxial layer 210 permits in some applications the P-well 520 to be shallower. More particularly, some applications require P-vell 520 to reach through epitaxial layer 210 to meet epitaxial layer 120 and thus to electrically contact the Pi- substrate 110 which in these applications is tied to the common ground of the NMOS transistors. In addition to contacting the common ground, P-well 520 diffusing through N— layer 210 allows isolating the N-well 310 from other N-wells (not shown) , if any, in the same integrated circuit. Since epitaxial layer 210 is thin, well 520 can be made shallower. Because shallower well 520 can be created with a lower doping dose, the lateral diffusion of the P-well dopant is reduced, allowing the MOS transistors to have a smaller area and, therefore, smaller capacitances and higher speed. In addition, thinner epitaxial layer 210 advantageously reduces the size of active regions of parasitic transistors.
Nitride 230 over the bipolar transistor also reduces the diffusion of collector region 240, guard ring 250 and channel stop region 150. Reducing the diffusion of regions 240, 250, 150 allows reducing the bipolar transistor area and, therefore, the transistor capacitances including the base-collector capacitance. Reducing the lateral and downward diffusion of buried layer 140 reduces the collector-substrate capacitance. A photoresist mask (not shown) is formed that exposes the to-be-formed well 520. The mask need not be perfectly aligned with the boundary between N-well 310 and P-well 520 because the thicker oxide 220 over N-well 310 masks the N-well during the P-well implant. Some embodiments do not use the oxide over N-well 310 as a mask, and instead rely on the photoresist to. mask the N-well.
Boron is implanted at an energy of 150 KeV and a dose of 3.3xl012 atoms/cm2 to form P-well 520. The photoresist is then stripped, and the wells are driven in at the temperature of 1,100°C for 50 minutes in an oxidizing ambient. During this step, the thickness of oxide 220 increases over the MOS transistors to planarize the surface of wells 310, 520. Over the bipolar transistor the oxidation is inhibited by nitride 230 as described above. Buried layer 140 and channel stop regions 150 increase during this step due to dopant diffusion. The dopant diffusion, however, is inhibited by layer 230 as described above. Guard ring 250 and collector region 240 also increase and meet the buried layer to provide a well-defined guard ring and a well-defined collector region. Thus the active region of the bipolar transistor becomes completely enclosed by the guard ring and the buried layer within epitaxial layer 210. Because epitaxial layer 210 is thin, the guard ring and the collector region need not be doped as heavily to meet buried layer 140 as would be required if epitaxial layer 210 were thicker. The lighter doping of guard ring 250 and collector region 240 reduces the junction capacitances in the bipolar transistor and thus provides higher speed. In an alternate embodiment, the implant that forms buried layer 140 of Fig. 1 forms also a buried layer 140.1 (Fig. 7) at the location of the to-be-formed PMOS transistor. The implant that forms channel stop region 150 (Fig. 1) forms also a buried layer 150.1 at the location of the to-be-formed NMOS transistor. N-well 310 and P-well 520 are created using shallow implants sufficient to cause the wells to meet the respective buried layers 140.1, 150.1. The techniques of Figs. 1-6 allow high speed bipolar transistors to be integrated in an existing CMOS twin well process without using buried layers 140.1, 150.1. Hence, in Fig. 6, the doping of regions 140, 150 can be optimized for the bipolar transistor to obtain a low leakage current, low capacitances and a low collector resistance, as the doping does not affect the MOS transistors. The result is decoupling the MOS and bipolar transistor characteristics from each other without increasing the process complexity. The remaining fabrication steps are similar for the embodiments of Figs. 6 and 7 and are illustrated only for the embodiment of Fig. 6 for simplicity.
Nitride 230 and oxide 220 are stripped. Silicon dioxide layer 810 (Fig. 8) is formed to a thickness of 25θA. Layer 220 is formed by thermal oxidation performed at 900°C for 60 minutes.
Silicon nitride layer 820 is deposited to a thickness of l,85θA. A photoresist mask (not shown) is formed that exposes the regions in which field oxide 830.1, 830.2 is to be grown by the LOCOS process. A composite nitride/oxide etch removes silicon nitride 820 and silicon dioxide 810 off the regions exposed by the mask.
The photoresist is then stripped, and another photoresist mask (not shown) is formed for a P+ boron implant that creates field implant regions 840, 840.1. The boron is implanted at an energy of 30 KeV and a dose of 5.5xl013 atoms/cm2. The resulting net doping concentration of regions 840, 840.1 is 2xlOπ atoms/cm3. The field implant region 840.1 that overlies channel stop region 150 meets the channel stop region to create a P+ ring around the bipolar transistor. This ring isolates the bipolar transistor from other transistors of the integrated circuit.
The photoresist is then stripped, and field oxide 830.1, 830.2 is thermally grown on the exposed silicon areas by the LOCOS process at 1,000°C for 160 minutes. Oxide regions 830.1 overlie field implant regions 840, 840.1 and guard ring 250. Field oxide regions 830.1 and field implant regions 840, 840.1 surround laterally each of the bipolar, PMOS, and NMOS transistors. Field oxide region 830.2 extends between collector region 240 and the portion 250.1 of guard ring 250. The guard ring portion 250.1 forms a sink region providing a low resistance path to buried layer 140 from the to-be- formed collector contact (not shown in Fig. 8) . Field oxide region 830.2 separates the sink region 250.1 from the to-be-formed base region (not shown in Fig. 8) to reduce the base-collector capacitance.
Using LOCOS isolation rather than a more stressful trench isolation reduces the stacking fault defect density thus improving the yield.
Nitride 820 and oxide 810 are then stripped. Threshold-voltage-adjust implants are performed optionally into N-well 310 or P-well 520 or both. Gate oxide layer 910 (Fig. 9) is thermally grown over the exposed silicon areas of the MOS and bipolar transistors at 900°C for 28 minutes to a thickness of 15θA. A protective intrinsic polysilicon layer 920 is deposited over the wafer to protect the gate oxide during subsequent processing steps.
A base implant mask 930 is formed from photoresist. Boron is implanted at an energy of 20 KeV and a dose of 5xl013 atoms/cm2 to convert top portion 940 of epitaxial layer 210 in the bipolar transistor active region to the P conductivity type. Polysilicon layer 920 and gate oxide 910 combine to reduce the implant width, providing a shallower base with existing implanters. Because the base is formed after the gate oxide, the base region does not experience any oxidation or dopant redistribution during the gate oxide formation and in some embodiments does not see any significant diffusion/anneal cycles until the RTP. (rapid thermal processing) anneal which forms the emitter region (not shown in Fig. 9) . This RTP anneal is described below in connection with Fig. 14. Thus a shallow well-defined base region results. The shallow base region leads to a higher bipolar transistor speed and allows epitaxial layer 210 to be thinner.
The width of collector region 240 is reduced by the base implant. The enhanced doping of region 240 reduces the base width over region 240.
Field oxide region 830.1 and P+ regions 840.1, 150 laterally surround and isolate the bipolar transistor. Substrate 110, and hence P+ regions 840.1, 150, are typically held at the lowest potential of the circuit during operation. Guard ring 250 between base region 940 and field implant region 840.1 prevents the P dopant of region 840.1 from counterdoping the epitaxial layer at the guard ring location. Such counterdoping could cause field implant region 840.1 to meet, or come close to, the base region leading to a base-to- substrate short or high leakage current. Further, the guard ring keeps the depletion layer between region 840.1 and the N type portion of epitaxial layer 210 farther from the base, thereby increasing the punchthrough voltage of the parasitic PNP transistor formed by the base region, the P+ region 840.1 and the portion of epitaxial layer 210 between the base region and the P+ region. Guard ring 250 also impedes the formation of the inversion layer between the base region and the P+ region and thus further reduces the current leakage from the base region to substrate 110. In some embodiments, as shown in Fig. 9, P+ regions 840.1, 150 are separated from buried layer 140 and guard ring 250 by portions of epitaxial layers 210, 120 which portions have a lower doping concentration than the two P+ regions, the buried layer and the guard ring. The collector capacitance is reduced as a result, and the breakdown voltage between channel stop region 150 and buried layer 140 is increased. The base region is separated from the guard ring and the buried layer by N— regions 950.1, 950.2 of epitaxial layer 210 which have a lower doping concentration than the base region and the guard ring. The collector-base capacitance is reduced as a result.
Photoresist 930 is stripped, and a photoresist mask 1010 (Fig. 10) is formed to define the emitter. Polysilicon 920 is etched through an opening in mask 1010 by a plasma etch. The etch stops on gate oxide 910.
As shown in Fig. 11, resist 1010 is removed, and gate oxide 910 is etched with polysilicon 920 as a mask. The resulting opening 1110 exposes region 940 at the location of the to-be-formed emitter. During the etch, polysilicon 920 protects the gate oxide over the wells of the MOS transistors. In some embodiments, the gate oxide etch is a wet buffered oxide HF etch designed to reduce any damage to, and removal of, the P doped region 940.
Intrinsic polysilicon is deposited by chemical vapor deposition (CVD) over the wafer at 630°C to a thickness of 3,25θA to combine with the polysilicon 920 to form polysilicon layer 1210 (Fig. 12) . Even though layer 920 is removed at the emitter location, layer 1210 has a substantially uniform thickness because layer 920 is thin compared to layer 1210.
A masked implant of arsenic at an energy of 100 KeV and a dose of 1.2xl016 atoms/cm2 creates N+ emitter contact region 1220 and N+ gate regions 1230, 1240. The net doping concentration of emitter contact region 1220 and gate regions 1230, 1240 is lxl020 atoms/cm3.
An optional P type masked implant forms resistors (not shown) in layer 1210.
The polysilicon is then annealed at 900°C for 15- inutes.
Polysilicon 1210 is then masked by a photoresist (not shown) and etched to define emitter contact region 1220 and gate regions 1230, 1240 (Fig. 13) .
Polysilicon resistors (not shown) , if any, are also defined at this step. The etch stops on gate oxide 910 which protects epitaxial layer 210 and, in particular, the damage-sensitive base region 940. Emitter contact region 1220 extends over the edges of the emitter opening in oxide 910.
Silicon dioxide 1310 is grown thermally at 900°C for 30 minutes over polysilicon 1210 to a thickness of
The NMOS transistor is then masked by a photoresist (not shown) , and a lightly doped drain (LDD) implant of boron is performed at an energy of 45 KeV and a dose of lxlO13 atoms/cm2 into N-well 310 and region 940 while gate region 1230 and emitter contact region 1220 mask, respectively, the PMOS channel region and the intrinsic base region. The resulting net doping concentration of the extrinsic base region is 4xlθ18 atoms/cm3. This implant forms lightly doped source/drain regions 1320.1, 1320.2 of the PMOS transistor and reduces the extrinsic base resistance. In some embodiments, region 940 is not doped during this step.
The photoresist is stripped. Another photoresist mask (not shown) is formed over the bipolar and PMOS transistors. An LDD implant of arsenic is performed at an energy of 90 KeV and a dose of 4xl0u atoms/cm2 while gate region 1240 masks the NMOS channel region. This implant forms lightly doped source/drain regions 1330.1, 1330.2 of the NMOS transistor. In some embodiments, sink region 250.1 is also doped during this implant.
A confor al layer of silicon dioxide is deposited by CVD at 650°C. An optional "silicide exclusion" mask (not shown) is formed from photoresist over selected areas to protect the oxide and thus to prevent formation of a metal silicide on such areas in subsequent processing described below. This CVD oxide is etched anisotropically to form spacers 1410 (Fig. 14) on the sidewalls of emitter contact region 1220 and gate regions 1230, 1240. The gate oxide not covered by the spacers and by polysilicon regions 1220, 1230, 1240 is removed during this etch. The remaining gate oxide over the bipolar transistor is completely covered by emitter contact region 1220 and spacers 1410. The overetch required is uniform over the bipolar, PMOS and NMOS transistors because the gate oxide has a uniform thickness over the bipolar and MOS transistors, because the CVD silicon dioxide has a uniform thickness over the transistors except, perhaps, on the sidewalls of polysilicon portions 1220, 1230, and 1240, and because oxide 1310 is thin compared to the CVD oxide.
A masked N+ arsenic implant is performed into source/drain regions 1330.1, 1330.2 and sink region 250.1 at an energy of 30 KeV and a doping dose of 3xl015 atoms/cm2. The resulting doping concentration in the top portion 250.2 of the sink region is lxlO20 atoms/cm3. A masked P+ boron implant is performed into source/drain regions 1320.1, 1320.2 and region 940 at an energy of 45 KeV and a dose of 3X1015 atoms/cm2. The resulting doping concentration in the extrinsic base region not covered by spacers 1410 is 5xl019 atoms/cm3. These implants complete the doping of the source/drain regions of the NMOS and PMOS transistors and reduce the resistivity of the sink region and the extrinsic base region. Next, an RTP (Rapid Thermal Processing) anneal performed at 1030°C for 20 seconds causes the N-type dopant from emitter contact region 1220 to diffuse into region 940 and convert the top portion 1420 of the region to N conductivity type. This portion is the emitter region. The RTP anneal is the first major high temperature step after formation of base region 940 (Fig. 9) . The previous steps had but a small effect on region 940 compared with the RTP anneal. The base region is shallow and well-defined as a result. A refractory metal, titanium in some embodiments, is deposited over the structure by chemical vapor deposition to a thickness of 60θA. The structure is heated and kept at 650°C for 30 minutes to react the titanium that contacts silicon surfaces with the silicon. As a result, titanium silicide 1510 (Fig. 15) is formed on the exposed silicon surfaces including the surfaces of the extrinsic base region, emitter contact region 1220, sink region 250.1, and source/drain and gate regions of the MOS transistors. The sheet resistance of layer 1510 is 3 ohms/square. The titanium silicide reduces resistance to the intrinsic transistor regions and thus increases the circuit speed and reduces the power consumption.
Titanium nitride layer 1520 is deposited over, and in contact with, the entire silicide 1510 by chemical vapor deposition to a thickness of 300θA. The deposition temperature is 650°C. Layer 1520 is patterned to form local conductive lines that contact titanium silicide 1510. As used herein, the term "local conductive lines" means conductive lines that are formed after the transistors but before formation of any interlevel dielectric in which contact openings are formed to contact the transistors. Layer 1520 has a sheet resistance of 1.5 ohms/square. A different sheet resistance can be obtained by changing the thickness of layer 1520. Conductive lines 1520 are used as interconnects and/or as extensions of the emitter, base, source/drain and gate regions. Such extensions allow forming contacts to these regions over field oxide 830.1, 830.2 rather than over those regions themselves.
In the embodiment of Fig. 15, TiN is etched off the gate regions. A conductive line 1520 runs along the top surface of emitter contact region 1220, reducing the emitter resistance. In some embodiments, TiN lines run along, and/or contact, the top surfaces of gate region 1230 and/or gate region 1240. In some embodiments, the spacing between the TiN lines over base region 940 on the one hand and emitter contact region 1220 on the other hand is smaller than the minimal photolithography dimension obtainable by the fabrication equipment. The same is true for the spacing between lines 1520 and each of gate regions 1230, 1240. Spacers 1410 help isolate conductive lines 1520 from regions 1220, 1230, 1240. Because lines 1520 are thick, a low resistance can be obtained with but a small overlap of lines 1520 over base region 940 and over the source and drain regions. The areas of the base, source and drain regions can therefore be reduced, providing a high packing density and a high speed.
As shown in Fig. 16, interlevel dielectric 1610 is formed by TEOS deposition of silicon dioxide to a thickness of l,OOθA, TEOS deposition of boron and phosphorus doped silicon dioxide to a thickness of 7,OOθA, and a densification and reflow of the boron and phosphorus doped silicon dioxide. Contact openings are etched in layer 1610. The contact openings terminate on conductive lines 1520. Tungsten is deposited by CVD and etched back to form tungsten plug contacts 1620.i in the contact openings. The plugs physically contact the titanium nitride lines 1520. Collector contact
1620.2 physically contacts titanium nitride 1520 over- sink region 250.1. Conductive lines 1630 are formed from Al-Si-Cu or some other suitable material. The fabrication is completed using known processing technigues.
Fig. 17 is a top view of the resulting structure. Emitter region 1420 is a non-walled emitter region. Gate oxide 910 (Fig. 13) isolates emitter contact region 1220 from base region 940 outside emitter region 1420 and,, in particular, at the boundary between base region 940 and field oxide 830.1, eliminating yield and leakage problems described in U.S. patent application Serial No. 08/085,436 filed by M.J. Grubisich on June 30, 1993 and hereby incorporated herein by reference. In some embodiments, emitter 1420 is a walled emitter. In some embodiments, collector region 240 (Fig. 3) is in top view coextensive with emitter region 1420, providing a low base-collector capacitance. Titanium nitride lines 1520 reduce the resistance allowing a low resistance to be obtained with thinner titanium silicide 1510 (Fig. 15) , especially since the titanium nitride has a lower resistivity than the titanium silicide. Because the titanium silicide is thinner, the consumption of silicon during the silicide formation is reduced, which lessens the chance of shorting out the junctions between the base and the collector and between the source/drain regions and the MOS wells 310, 520. Hence these junctions can be made shallower. The silicide outgrowth which results in stringers on spacers 1410 (Fig. 15) is also reduced, leading to a better yield. Further, it has been found that the formation of stringers can be inhibited by lower silicidation temperatures, but the lower temperature also increases the silicide resistivity. Due to lines 1520, the stringer formation is inhibited by making titanium silicide thin. Hence, the silicidation temperature can be made higher to reduce the silicide resistivity.
Emitter contact opening 1620.7 and gate contact openings 1620.8 through 1620.11 are formed over field oxide 830.1. Titanium nitride line 1520 runs along emitter contact region 1220 to the contact opening 1620.7, reducing the emitter resistance and thus improving the VBE matching. As a result, the emitter current and the bipolar transistor speed can be increased without increasing the emitter area.
Some embodiments combine the isolation techniques of Figs. 1-8 with other processes than those of Figs. 9-17. For example, in some embodiments, after the structure of Fig. 8 is made, the base is formed by diffusion from a P-doped polysilicon layer as described, for example, in the U.S. Patent No. 5,219,784 issued June 15, 1993 to A.G. Solheim and incorporated herein by reference. The emitter is also formed as described in that patent. Other fabrication techniques, including non-polysilicon emitter technigues, are used with the structure of Fig. 8.
Further, the isolation techniques of Figs. 1-8 are used in some embodiments with other silicidation/local conductive line techniques than those of Figs. 15, 16 or without any silicidation/local conductive line techniques. Similarly, the base and emitter formation techniques of Figs. 9-14 are combined in some embodiments with other isolation techniques including trench isolation or junction isolation. The techniques of Figs. 9-14 are used in some embodiments without any silicidation/local conductive line techniques or with silicidation/local conductive line techniques different from those of Figs. 15, 16.
The silicidation/local conductive line techniques of Figs. 15, 16 are used in some embodiments with different isolation techniques than those of Figs. l-*8 and/or with different emitter/base formation techniques than those of Figs. 9-14.
While the invention has been illustrated by the embodiments described above, other embodiments and variations are within the scope of the invention. In particular, the invention is not limited by any particular materials or dimensions. For example, in some embodiments, other metal-containing materials are used instead of titanium to form the non-semiconductor conductive layer 1510. Conductive lines 1520 are formed in some embodiments from other materials than titanium nitride, including other non-semiconductor materials. Lines 1520 are formed from titanium, tungsten or titanium tungsten in some embodiments. In some embodiments, the conductivity types are reversed to form PNP transistors in a BiCMOS circuit. Some embodiments include non-BiCMOS integrated circuits containing only bipolar or bipolar and MOS transistors. In some embodiments, an additional implant is performed into the collector region 240 to reduce the collector resistance. Other embodiments and variations are within the scope of the invention as defined by the following claims.

Claims

CLAIMSWhat is claimed is:
1. A method for fabricating an integrated circuit, said method comprising the steps of: forming a structure comprising one or more transistors each of which has three electrodes, the structure comprising for each of said transistors three semiconductor surfaces each of which is for electrically contacting a respective electrode of the respective transistor; after forming each said transistor but before forming any contact openings in any insulation formed after formation of all said transistors, forming a metal-containing material over said structure; reacting portions of the metal-containing material with semiconductor material of said semiconductor surfaces to form on all said surfaces a layer of a non-semiconductor conductive material; removing the metal-containing material that did not react with the semiconductor material; forming a conductive material M over and in contact with the entire non-semiconductor conductive material; and patterning the material M to form from the material M one or more conductive lines contacting the non-semiconductor conductive material.
2. The method of Claim 1 wherein the non- semiconductor conductive material comprises a metal silicide.
3. The method of Claim 1 wherein the material M comprises titanium nitride.
4. The method of Claim 1 wherein: at least one of said transistors is a bipolar transistor laterally surrounded by a field insulation, and one surface E of said surfaces is for contacting an emitter region of said bipolar transistor; the surface E extends over the field insulation; one or more of the one or more conductive lines extend over, and contact, the surface E; and said method further comprises the steps of: . forming a first insulation over the conductive lines; forming a contact opening in the first insulation over the field insulation to expose one of the conductive lines that extends over the surface E; and forming a conductive layer contacting in the contact opening the one of the conductive lines that extends over the surface E.
5. A method for forming an integrated circuit, said method comprising the steps of: forming a structure comprising monocrystalline silicon, the monocrystalline silicon comprising a base region of a bipolar transistor and a collector region of the bipolar transistor, the structure further comprising a field insulation laterally surrounding the bipolar transistor; forming a polysilicon emitter contact region over the monocrystalline silicon and the field insulation and diffusing a dopant from the polysilicon emitter contact region into the monocrystalline silicon to form an emitter region of the bipolar transistor; forming insulating spacers on sidewalls of the polysilicon emitter contact region, the spacers physically contacting the monocrystalline silicon to isolate the polysilicon emitter contact region from the base region; forming a metal-containing material over the monocrystalline silicon and the polysilicon emitter contact region; reacting the metal-containing material with portions of the monocrystalline silicon and of the polysilicon emitter contact region to form a metal silicide on selected surfaces of the monocrystalline silicon and polysilicon emitter contact region, at least one of the selected surfaces being a surface of the base region; removing unreacted metal-containing material; forming a conductive non-semiconductor material over and in contact with the entire metal silicide; patterning the conductive non-semiconductor material to form local conductive lines, a first one of the conductive lines overlying and running along the polysilicon emitter contact region to reduce an emitter resistance, and a second one of the conductive lines overlying the metal silicide on said surface of the base region, each of the first and second conductive lines overlying the field insulation; forming a first insulation over the conductive lines; forming contact openings in the first insulation, one of the openings terminating on the first conductive line over the field insulation and another one of the openings terminating on the second conductive line over the field insulation; and forming conductive contacts in the openings.
6. The method of Claim 5 wherein the step of forming a conductive non-semiconductor material comprises the step of forming titanium nitride by chemical vapor deposition.
7. The method of Claim 5 wherein: the monocrystalline silicon comprises a source region of a field effect transistor, a drain region of the field effect transistor, and-a channel region of the field effect transistor; the field insulation surrounds laterally the field effect transistor; the step of forming the polysilicon emitter region comprises the step of forming a polysilicon gate region over the channel region; the step of forming the insulating spacers on the sidewalls of the polysilicon emitter region comprises the step of forming insulating spacers on sidewalls of the gate region; the step of forming the metal-containing material over the monocrystalline silicon and the polysilicon emitter contact region comprises the step of forming a metal-containing material over the gate region; the reacting step comprises the step of forming a metal silicide on surfaces of the source, drain and gate regions; a third one of the conductive lines overlies the metal silicide on the surface of at least one of the source and drain regions and also overlies the field insulation; and at least one of the contact openings terminates on the third conductive line over the field insulation.
8. A method for forming an integrated circuit, said method comprising the steps of: forming a structure comprising monocrystalline silicon, the monocrystalline silicon comprising a source region of a field effect transistor, a drain region of the field effect transistor, and a channel region of the field effect transistor, the structure further comprising a field insulation laterally surrounding the field effect transistor; forming a polysilicon gate region over the channel region; forming insulating spacers on sidewalls of the gate region; forming a metal-containing material over the monocrystalline silicon and the gate region; reacting the metal-containing material with portions of the monocrystalline silicon and the gate region to form a metal silicide on surfaces of the source, drain and gate regions; removing unreacted metal-containing material; forming a conductive non-semiconductor material over and in contact with the entire metal silicide; patterning the conductive non-semiconductor material to form a local conductive line overlying the metal silicide on the surface of at least one of the source and drain regions and also overlying the field insulation; forming a first insulation over the conductive line; forming a contact opening in the first insulation, the opening terminating on the conductive line; and forming a conductive contact in the opening.
9. The method of Claim 8 wherein the step of forming a conductive non-semiconductor material comprises the step of forming titanium nitride by chemical vapor deposition.
SUBSTITUTESHEET{RULE26)
PCT/US1994/011110 1994-02-28 1994-09-30 Providing a low resistance to integrated circuit devices Ceased WO1995023429A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
KR1019960704698A KR970701432A (en) 1994-02-28 1994-09-30 PROVIDING A LOW RESISTANCE TO INTEGRATED CIRCUIT DEVICES
EP94929971A EP0748517A1 (en) 1994-02-28 1994-09-30 Providing a low resistance to integrated circuit devices

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US20253294A 1994-02-28 1994-02-28
US08/202,532 1994-02-28

Publications (1)

Publication Number Publication Date
WO1995023429A1 true WO1995023429A1 (en) 1995-08-31

Family

ID=22750280

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1994/011110 Ceased WO1995023429A1 (en) 1994-02-28 1994-09-30 Providing a low resistance to integrated circuit devices

Country Status (3)

Country Link
EP (1) EP0748517A1 (en)
KR (1) KR970701432A (en)
WO (1) WO1995023429A1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1987003425A1 (en) * 1985-11-29 1987-06-04 Plessey Overseas Limited Transistor having silicide contacts and method for producing same
EP0326218A1 (en) * 1988-01-29 1989-08-02 Koninklijke Philips Electronics N.V. Method of manufacturing a semiconductor device, in which a metal conductor track is provided on a surface of a semiconductor body
JPH02308569A (en) * 1989-05-23 1990-12-21 Fujitsu Ltd Semiconductor device and manufacture thereof
US5190893A (en) * 1991-04-01 1993-03-02 Motorola Inc. Process for fabricating a local interconnect structure in a semiconductor device

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1987003425A1 (en) * 1985-11-29 1987-06-04 Plessey Overseas Limited Transistor having silicide contacts and method for producing same
EP0326218A1 (en) * 1988-01-29 1989-08-02 Koninklijke Philips Electronics N.V. Method of manufacturing a semiconductor device, in which a metal conductor track is provided on a surface of a semiconductor body
JPH02308569A (en) * 1989-05-23 1990-12-21 Fujitsu Ltd Semiconductor device and manufacture thereof
US5190893A (en) * 1991-04-01 1993-03-02 Motorola Inc. Process for fabricating a local interconnect structure in a semiconductor device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
PATENT ABSTRACTS OF JAPAN vol. 15, no. 98 (E - 1042) 8 March 1991 (1991-03-08) *

Also Published As

Publication number Publication date
EP0748517A1 (en) 1996-12-18
KR970701432A (en) 1997-03-17

Similar Documents

Publication Publication Date Title
US5580798A (en) Method of fabricating bipolar transistor having a guard ring
US5171702A (en) Method for forming a thick base oxide in a BiCMOS process
US5387552A (en) Method of fabrication of PNP structure in a common substrate containing NPN or MOS structures
JP3232111B2 (en) Schottky barrier diode and Schottky barrier diode clamp transistor
KR100223098B1 (en) Bicmos device and manufacturing method thereof
EP0283135A1 (en) Fabrication of semiconductor structure
US4926233A (en) Merged trench bipolar-CMOS transistor fabrication process
US5107321A (en) Interconnect method for semiconductor devices
US5231042A (en) Formation of silicide contacts using a sidewall oxide process
US5389553A (en) Methods for fabrication of transistors
US20030080394A1 (en) Control of dopant diffusion from polysilicon emitters in bipolar integrated circuits
US7217609B2 (en) Semiconductor fabrication process, lateral PNP transistor, and integrated circuit
US5422290A (en) Method of fabricating BiCMOS structures
US5045483A (en) Self-aligned silicided base bipolar transistor and resistor and method of fabrication
EP0450503A2 (en) Semiconductor devices with borosilicate glass sidewall spacers and method of fabrication
US5348896A (en) Method for fabricating a BiCMOS device
US5010034A (en) CMOS and bipolar fabrication process using selective epitaxial growth scalable to below 0.5 micron
JP4956853B2 (en) Semiconductor device and manufacturing method thereof
EP0409041B1 (en) A method for forming a thick base oxide in a BiCMOS process
JP3190370B2 (en) BiCMOS device having closely spaced contacts and method of manufacturing the same
EP0748517A1 (en) Providing a low resistance to integrated circuit devices
WO2003043080A1 (en) Lateral pnp transistor device, integrated circuit, and fabrication process thereof
JPH08213491A (en) Method for manufacturing semiconductor integrated circuit
JPS63292666A (en) Manufacture of semiconductor device
JPH10199893A (en) Method for manufacturing integrated circuit of bipolar transistor

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): JP KR

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FR GB GR IE IT LU MC NL PT SE

DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
121 Ep: the epo has been informed by wipo that ep was designated in this application
WWE Wipo information: entry into national phase

Ref document number: 1994929971

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 1994929971

Country of ref document: EP

WWW Wipo information: withdrawn in national office

Ref document number: 1994929971

Country of ref document: EP