Bipolar transistor structure with a shallow isolation extension region providing reduced parasitic capacitance

This invention relates to bipolar transistor structures, and more particularly structures with reduced parasitic capacitance and methods of manufacture thereof.

FIG. 1 is a cross section of prior art Vertical Bipolar Transistor formed on a silicon semiconductor substrate. The transistor includes a doped emitter E, a doped collector C and a counterdoped compound base, in accordance with conventional bipolar device design. The collector C is formed between a pair of Shallow Trench Isolation (STI) (dielectric) regions in the surface of the silicon semiconductor substrate. The compound base comprises an intrinsic silicon (Si) or silicon-germanium (Si—Ge) base region and an extrinsic base region thereabove. The doped intrinsic base region, overlies the collector C and enclosed by the emitter area is in mechanical and electrical contact with the collector and emitter regions. The extrinsic base region overlies the outer surfaces of the intrinsic base region and portions of the STI regions. The emitter region is formed above the intrinsic base and is separated from the extrinsic base by dielectric regions. The Base-to-Collector Capacitance (Ccb) of the device is the sum of the components between (1) the base and the collector inside the active area (defined by the emitter opening) (2) the base and the collector outside the active area, but inside the STI edge (3) the base and the collector across the STI region.

Cut-off frequency (fT) and maximum oscillation frequency (fmax) are the most representative measures of operation speed for high-speed transistors. Hence, the design and optimization efforts for the high-speed transistors are mostly directed toward the maximization of these two parameters. One of the device parameters that influences the cut-off frequency (fT) and maximum oscillation frequency (fmax) is the Base-to-Collector capacitance (Ccb). The value of fT decreases with increasing Ccb as a result of increasing RC delay (charging time) associated with emitter and collector resistances and device transconductance. The impact of Ccb on fmax is even larger as fmax is more sensitive to RC delay associated with Ccb. Overall, the device component (resistance and capacitance) that has the largest impact on fT and fmax, or device operation speed, is Ccb. Therefore, the most effective way to improve device speed through parasitic component reduction is the minimization of extrinsic component of Ccb.

The extrinsic component, or parasitic component, of Ccb comprises more than half of the total Ccb for most conventional bipolar transistors. This parasitic capacitance results from the overlap between the collector and base (intrinsic and extrinsic) regions outside the active transistor area and extending over the shallow trench isolation (STI). The overlap between these regions can not be minimized by lithography due to limitation of overlay and alignment tolerances imposed by the requirement to minimize dimensions and increase the density of devices on the substrate. Moreover, the parasitic capacitance is further increased by the diffusion of the dopants from the base region to the collector region. Therefore, structural optimization of the device which would reduce the parasitic component is a key for the improvement of fT and fmax (i.e. the operation speed of the device).

U.S. Pat. No. 5,599,723 Sato Feb. 4, 1997 entitled “Method for Manufacturing Bipolar Transistor Having Reduced Base-Collector Parasitic Capacitance” teaches use of SiGe for the base, and that the parasitic capacitance formed between the collector epitaxial layer and the base electrode single crystal silicon film is reduced because the distance between them is set to about 1000 Å. In order to reduce the parasitic capacitance by the prior art technique, the intrinsic base must be thickened, and thus the cut-off frequency f T is lowered. A single crystal form of silicon formed by the selective epitaxial growth is used for the base electrode to reduce the parasitic capacitance between the base and the collector, particularly by forming the base of SiGe. The entire device including the collector regions is formed above the surface of the silicon semiconductor substrate. The approach to reducing the parasitic capacitance is to use selective epitaxy to grow the intrinsic base.

U. S. Pat. No. 5,128,271 of Bronner et al. entitled “High Performance Vertical Bipolar Transistor Structure via Self-aligning Processing Techniques” describes a self-aligned, vertical bipolar transistor structure and a method of manufacturing such a structure with “reduced parasitic base collector capacitance” achieved by providing correct alignment. The Bronner et al. approach has similarities with the present approach to solution of the parasitic base collector capacitance problem. However, the approach of this invention has significant features not described in the Bronner et al. patent. The present invention decouples the primary STI formation from the self-aligned shallow isolation extension (or secondary shallow isolation) formation to reduce the parasitics. This major difference allows a robust manufacturing process and flexible device performance tailoring in many ways some of which are listed below:

(1) The present invention approach is more compatible with CMOS device fabrication for manufacturing BiCMOS technology, where the STI and the shallow isolation extension are formed independently.

(2) The present invention allows the use of different dielectric material(s), than that used to form the STI, to form the shallow isolation extension to further reduce the capacitance.

(3) The present invention utilizes the STI portion exposed to end-point of the RIE of the intrinsic base and as a self-aligning edge to extend the RIE into the collector pedestal.

(4) The present invention allows partial removal of the collector pedestal, which can be employed to tailor the collector pedestal shape within the STI.

(5) The present invention employs a raised extrinsic base and does not require the partial removal of the STI silicon oxide to form the extrinsic base.

In accordance with this invention, a structural modification is provided to reduce the parasitic component of Ccb in bipolar transistors with minimum adverse effect on other parameters. More specifically, the excess overlap region between the collector and the base is partially removed and filled with a dielectric prior to forming the extrinsic base region. The dielectric separates the collector and extrinsic base and acts as a barrier for dopant diffusion to reduce the parasitic component of Ccb. The parasitic component of Cbc in typical silicon based bipolar transistors is the result of the existence of depletion region of base-collector P—N junction formed at the extrinsic part of the devices. In accordance with this invention, the parasitic capacitance in such transistors is reduced by employing materials with reduced dielectric constants in the depletion region, since silicon has a high dielectric constant.

In accordance with this invention a structure and methods are provided whereby the silicon depletion region is partially or entirely replaced with a dielectric material with a low dielectric constant. One choice of dielectric is silicon oxide, dielectric constant of which is only 33% of silicon, which reduces the capacitance significantly.

In accordance with this invention, a bipolar vertical transistor is formed in a silicon semiconductor substrate which has an upper surface with STI regions formed therein composed of a dielectric material formed in the substrate having inner ends and top surfaces. A doped collector region is formed in the substrate between a pair of the STI regions. A counterdoped intrinsic base region is formed on the upper surface of the substrate between the pair of the STI regions with a margin between the intrinsic base region and the pair of STI regions, the intrinsic base region having edges. A doped emitter region is formed above the intrinsic base region spaced away from the edges. A shallow isolation extension region composed of a dielectric material is next to the edges of the intrinsic base region formed in the margin between the STI regions and the intrinsic base region. An extrinsic base region covers the shallow isolation extension region and extends partially over the intrinsic base region in mechanical and electrical contact therewith, whereby the shallow isolation extension region reduces the base-to-collector parasitic capacitance of the bipolar transistor. In accordance with another aspect of this invention, the shallow isolation extension region is either self-aligned with the emitter; or the shallow isolation extension region is non self-aligned with the emitter.

Preferably, the shallow isolation extension region is formed of an oxide material selected the group consisting of oxidized-doped-silicon and a deposited silicon oxide. Preferably the shallow isolation extension region is formed of a material having a lower dielectric constant than the STI regions, the shallow isolation extension region overlaps the inner ends and top surfaces of the STI regions. Preferably the shallow isolation extension region extends beneath the edges of the base region. Preferably the intrinsic base region is composed of a material selected from the group consisting of doped crystalline silicon and silicon germanium. Preferably the isolation extension region extends into the doped collector region thereby modifying the resistance of the collector region.

In accordance with another aspect of this invention, a bipolar vertical transistor is formed in a silicon semiconductor substrate which has an upper surface and STI regions are formed therein composed of a dielectric material and having inner ends and top surfaces. A doped collector region is formed in the substrate between a pair of the STI regions. A counterdoped intrinsic base region is formed on the upper surface of the substrate between the pair of STI regions with a margin between the intrinsic base region and the pair of STI regions, the intrinsic base region having edges. The intrinsic base region is composed of a material selected from the group consisting of doped crystalline silicon and silicon germanium. A doped emitter region formed composed of doped polysilicon is formed above the intrinsic base region spaced away from the edges thereof. A shallow isolation extension region composed of a dielectric material is juxtaposed with the edges of the intrinsic base region formed in the margin between the STI regions and the intrinsic base region and overlapping the inner ends and top surfaces of the STI regions. An extrinsic base region covers the shallow isolation extension region and extends partially over the intrinsic base region in mechanical and electrical contact therewith. A silicon nitride cap is formed over the emitter region and silicon nitride sidewall spacers formed on sidewalls of the emitter region, whereby the shallow isolation extension region reduces the base-to-collector parasitic capacitance of the bipolar transistor.

Preferably, the shallow isolation extension region is self-aligned with the emitter, or alternatively the shallow isolation extension region is non self-aligned with the emitter. The shallow isolation extension region is formed of an oxide of the substrate. The shallow isolation extension region is formed of a material having a lower dielectric constant than the STI regions. The shallow isolation extension region extends beneath the edges of the base region.

In accordance with another aspect of the invention, a method of forming a bipolar vertical transistor is formed in a silicon semiconductor substrate having an upper surface includes forming STI regions composed of a dielectric material formed in the substrate having inner ends and top surfaces. The steps are performed including forming a doped collector in the substrate between a pair of the STI regions, forming a counterdoped intrinsic base region formed on the upper surface of the substrate between the pair of the STI regions with a margin between the intrinsic base region and the pair of STI regions, with the intrinsic base region having edges, and forming a doped emitter region above the intrinsic base region spaced away from the edges. The shallow isolation extension region is composed of a dielectric material juxtaposed with the edges of the intrinsic base region formed in the margin between the STI regions and the intrinsic base region. The method also includes forming an extrinsic base region covering the shallow isolation extension region and extending partially over the intrinsic base region in mechanical and electrical contact therewith, whereby the shallow isolation extension region reduces the base-to-collector parasitic capacitance of the bipolar transistor.

Preferably, the shallow isolation extension region is formed by a method selected from the group consisting of oxidation or RIE to form an isolation extension recess and then filling the isolation extension recess with a dielectric material. The shallow isolation extension region is self-aligned with the emitter or the shallow isolation extension region is non self-aligned with the emitter.

Preferably, the method includes forming of the shallow isolation extension region includes overlapping the inner ends and top surfaces of the STI regions, forming a thermal oxide layer upon the substrate, forming Shallow Trench Isolation (STI) regions in the substrate defining an active area, and a collector reachthrough area aside from the active area. Preferably, the method also includes forming a first silicon nitride layer over the active area, forming a stack of a first silicon oxide layer, a first undoped polysilicon layer and a second silicon oxide layer aside from the first silicon nitride layer, covering the collector reachtrhough area and partially covering the STI regions, stripping the first silicon nitride layer, forming an intrinsic base epitaxial layer over the active area, and forming a sacrificial emitter stack of a third silicon oxide layer, a second undoped polysilicon layer and a second silicon nitride layer centrally over a portion of the active area leaving exposed portions thereof on each side of the sacrificial emitter stack.

The foregoing and other aspects and advantages of this invention are explained and described below with reference to the accompanying drawings, in which:

FIG. 1 shows a cross section of prior art Vertical Bipolar Transistor formed on a silicon semiconductor substrate.

FIGS. 2 A- 2 AB are sectional views illustrating the process of manufacturing of a device in accordance with this invention employing a self-aligned technique to form the isolation extension.

FIG. 3 shows a device which is a variation of the device of FIG. 2 AB with the isolation extension extending below the epitaxial part of the extrinsic base.

FIG. 4 shows another variation of the device of FIG. 2 AB with the isolation extension formed by oxidation.

FIGS. 5A-5D show sectional views of the processing steps of a device 10 C that is a modification of the process of FIGS. 2 A- 2 AB employing a non-self-aligned technique to form the isolation extension.

FIGS. 6A-6C show sectional views of the processing steps of a device that is a variation of the process of FIGS. 5A-5D with the isolation extension formed by oxidation.

FIGS. 2 A- 2 AB are sectional views illustrating the process of manufacturing of a device 10 in accordance with this invention.

FIG. 2A is a sectional view of the device 10 in an early stage in the manufacturing process. Device 10 includes a doped silicon semiconductor substrate 11 which has been coated with a thin thermal oxide layer TOX in accordance with a conventional process.

N-doped Collector Pedestal Formation

FIG. 2B shows the device 10 of FIG. 2A after a set of recesses have been formed in the surface of the substrate 11 . The recesses have been filled with shallow trench isolation (STI) regions 14 a dielectric material comprising silicon oxide. The STI regions 14 have been formed in the substrate 11 to define the silicon active area AA, the collector region C and the collector reachthrough region RT therebetween. The STI regions 14 have inner sidewalls, i.e. inner ends, juxtaposed with the central portion of the collector region C and top surfaces. The active area AA and the reachthrough region RT are located below the thin thermal oxide layer TOX which remains between the STI regions 14 . The STI regions 14 comprise any form of silicon oxide formed by a process selected from the group of processes including PECVD, RTCVD, OZONE TEOS, LPCVD. In the case of an NPN bipolar transistor, the active area AA which includes the collector region C of the transistor is doped with N type dopant by ion implantation or in-situ doped epitaxial growth, as will be well understood by those skilled in the art. The collector region C is located in the substrate, up to the surface thereof, as shown in FIG. 2 B.

FIG. 2C shows the device 10 of FIG. 2B after blanket deposition of a thin, protective, first silicon nitride layer SN 1 followed by formation of a first photoresist mask PR 1 with windows W 1 /W 1 ′ therethrough aside from the active area AA to prepare for patterning the thin, protective, first silicon nitride layer SN 1 over the active area AA in the silicon substrate 11 .

FIG. 2D shows the device 10 of FIG. 2C after etching of the portions of the thin, protective, first silicon nitride layer SN 1 aside from the active area AA, exposing portions of the surfaces aside of the STI regions 14 aside from the active area AA and the thermal oxide layer TOX above the reachthrough region RT. Then the first photoresist mask PR 2 was stripped away leaving the portion of thin, protective, first silicon nitride layer SN 1 above the active area AA and portions of the adjacent STI regions 14 intact.

FIG. 2E shows the device 10 of FIG. 2D after deposition of a stack of conformal, thin films of silicon oxide OX 1 , undoped polysilicon UP 1 and a second silicon oxide layer OX 2 , and a second patterning photoresist mask PR 2 is formed with a central window W 2 therethrough extending beyond the edges of the remainder of the underlying first silicon nitride layer SN 1 to form an opening to be used for epitaxial growth of an intrinsic base region.

FIG. 2F shows the device 10 of FIG. 2E after an RIE etching step removing the exposed silicon oxide OX 1 /undoped polysilicon UP 1 /silicon oxide OX 2 stack stopping on the first silicon nitride layer SN 1 . Then second photoresist mask PR 2 is stripped away from device 10 .

FIG. 2G shows the device 10 of FIG. 2F after stripping the first silicon nitride layer SN 1 with a wet etching process, i.e. hot phosphoric acid followed by the etching away of the thermal oxide layer TOX and the exposed second silicon oxide layer OX 2 with a wet etchant, i.e. HF acid.

FIG. 2H shows the device 10 of FIG. 2G after epitaxial growth of an intrinsic base layer IB directly over the exposed surface of the monocrystalline active area AA combined with simultaneous formation of a first doped polycrystalline silicon layer DP 1 over the STI regions 14 and the remaining stack of the first undoped polysilicon UP 1 and the first silicon oxide layer OX 1 . The intrinsic base IB and the doped polysilicon layer DP 1 are deposited by Low Temperature Epitaxy (LTE) forming P-doped intrinsic base layer IB composed of P-doped crystalline semiconductor silicon (Si) or silicon-germanium (Si—Ge) intrinsic base region directly on the top surface of on the silicon substrate 11 directly above the active area AA which will comprise the collector region C of the vertical transistor. The LTE process that deposits the intrinsic base layer IB over the active area AA in the silicon substrate 11 also simultaneously forms a first P-doped polycrystalline silicon layer DP 1 .

FIG. 2I shows the device 10 of FIG. 2H after deposition of a stack of conformal layers comprising a third silicon oxide layer OX 3 /second undoped polysilicon UP 2 /second silicon nitride/SN 2 over the device 10 covering the intrinsic base IB and the first doped polysilicon layer DP 1 .

Next, a third photoresist mask PR 3 with windows W 3 therethrough aside from the location where an emitter pedestal EP is to be formed (as shown in FIG. 2J ) is patterned to form an emitter pedestal and a raised extrinsic base region in subsequent steps.

FIG. 2J shows the device 10 of FIG. 2I after an RIE etch through the windows W 3 in mask PR 3 of top layers of the stack added in FIG. 2I including the second silicon nitride layer SN 2 and the undoped polysilicon layer UP 2 . Next, the third photoresist mask PR 3 was stripped. The etch stops on the third silicon oxide layer OX 3 forming a raised extrinsic base region RER on either side of an emitter pedestal EP which is centered over active area AA.

FIG. 2K shows the device 10 of FIG. 2J after the formation of silicon nitride sidewall spacers SN 3 on the sides of the emitter pedestal EP and the stacks aside from the raised extrinsic base regions RER thereby covering exposed sidewalls of the second silicon nitride layer SN 2 and the undoped polysilicon UP 2 .

FIG. 2L shows the device 10 of FIG. 2K after the formation of silicon oxide spacers OX 4 to provide an extension window EW adapted to be used for the formation of the shallow isolation extension regions IE by the steps indicated by FIGS. 2M , 2 N and 2 O in a self-aligned manner with respect to the emitter pedestal EP. The spacers OX 4 are used as a hard mask while using RIE to etch the exposed doped silicon DP 1 in the extension window EW to form the shallow isolation extension regions IE and forming isolation extension recesses R 1 where the doped silicon DP 1 has been etched away as shown in FIG. 2 M. The width of spacers OX 4 defines and protects the perimeter area around emitter pedestal EP during the RIE to provide a contact area between the intrinsic base IB and the extrinsic base polysilicon DP 2 as shown in FIG. 2 P.

FIG. 2M shows the device 10 of FIG. 2L after the RIE etch through the exposed intrinsic base IB and doped polysilicon DP 1 over the STI and down into the collector region C in the active area AA using the silicon oxide spacers OX 4 as a hard mask to form isolation extension recesses R 1 around the emitter pedestal EP between the spacers OX 4 and the inner sidewalls of the shallow trench STI regions 14 adjacent to the emitter pedestal EP. The result is that the top central portion of the collector region C (which is formed from the substrate 11 ) have been shaped into an intrinsic base/collector pedestal located centrally between the isolation extension recesses R 1 . Moreover, the edges of the intrinsic base IB are aligned with the edges of the pedestal portion of the collector region C forming the top layer of the intrinsic base/collector pedestal. Later in the process, the shallow isolation extension regions IE will be formed between the sidewalls of the intrinsic base/collector pedestal by depositing silicon oxide into the isolation extension recesses R 1 as shown in FIG. 2 O. The width and depth of the isolation extension recesses R 1 formed in the collector region C are optimized to reduce the parasitic capacitance. In addition, the isolation extension recesses R 1 in the collector region C in the active area AA determines the final collector structure. Consequently, the width and depth of the isolation extension recesses R 1 can be optimized to tailor the collector resistance as will be well understood by those skilled in the art. The intrinsic base IB and the high portion of the collector region C are separated from the STI regions by the widths of the isolation extension recesses R 1 on either side thereof.

FIG. 2N shows the device 10 of FIG. 2M after deposition of a conformal silicon oxide layer OX 5 to fill the isolation extension recesses (or trenches) R 1 to prepare for formation of the shallow isolation extension regions IE around the emitter pedestal EP.

FIG. 2O shows the device 10 of FIG. 2N after the oxide layer OX 5 has been planarized and etched or recessed down to the intrinsic base IB surface to form the shallow isolation extension regions IE.

FIG. 2 L through FIG. 2O show how the shallow isolation extension regions IE are formed independently from the formation of the primary shallow trench isolation STI. This decoupling of the formation of the shallow isolation extension regions IE allows compatibility of the bipolar transistor described herein with conventional CMOS technology to manufacture BiCMOS technology, where the primary shallow trench isolation STI is common to both bipolar and CMOS devices. This decoupling of the formation of the shallow isolation extension regions also allows control over parasitic base-to-collector capacitance reduction through control over the depth of the isolation extension recess R 1 and the option of using different materials with different dielectric properties than those of the primary shallow trench isolation STI material. For example, a different material with a lower dielectric constant can be used instead of the silicon oxide layer OX 5 to fill the isolation extension recesses R 1 to further reduce the parasitic capacitance. Such materials with lower dielectric constant include, but are not limited to, SiLK, fluorinated oxides, and porous oxides.

FIG. 2P shows the device 10 of FIG. 2O after deposition of doped polysilicon layer DP 2 followed by planarization and recess to form a raised extrinsic base DP 2 . FIG. 2P also shows the perimeter contact area between the intrinsic base IB and the extrinsic base doped polysilicon DP 2 which was defined and protected by the silicon oxide spacers OX 4 as shown in FIG. 2 M.

FIG. 2Q shows the device 10 of FIG. 2P after deposition of isolation oxide OX 6 followed by planarization and recessing of the isolation oxide OX 6 leaving the emitter pedestal EP and the lateral stack of layers SN 2 /UP 2 exposed.

FIG. 2R shows the device 10 of FIG. 2Q after an RIE etch of silicon nitride layer SN 2 and undoped polysilicon layer UP 2 thereby removing the emitter pedestal EP to make emitter opening EO between the isolation stacks IS which have been formed by the remaining stacks and the doped extrinsic base polysilicon DP 2 topped with the isolation oxide OX 6 . The silicon nitride spacers SN 3 are also removed with an RIE or a wet etch and the top surfaces of the third silicon oxide layer OX 3 are exposed.

FIG. 2S shows the device 10 of FIG. 2R after forming the final silicon nitride spacers SN 4 on the sidewalls of the isolation stacks IS.

FIG. 2T shows the device 10 of FIG. 2S after removal of the exposed portion of the third silicon oxide OX 3 at the bottom of the emitter opening EO and the field, thereby exposing the top surface of the intrinsic base IB.

FIG. 2U shows the device 10 of FIG. 2T after depositing a conformal layer of doped emitter polysilicon DP 3 followed by deposition of a conformal layer of silicon nitride to serve as a hard mask HM′. A fourth patterned photoresist mask PR 4 is formed over the layers DP 3 /HM to be used to form an emitter E as shown in FIG. 2 V. The mask PR 4 overlaps the isolation stacks IS.

FIG. 2V shows the device 10 of FIG. 2U after RIE etching of the silicon nitride layer HM′ to form the hard mask HM, followed by stripping mask PR 4 . This is then followed by RIE etching of doped emitter polysilicon DP 3 , doped polysilicon DP 1 (aside from silicon nitride spacers SN 4 ) and undoped polysilicon UP 1 to form the emitter E and to expose the collector reach-through region RT.

FIG. 2W shows device 10 of FIG. 2V after stripping the isolation silicon oxide OX 6 over the doped extrinsic base polysilicon DP 2 and the outer portions of silicon nitride spacers SN 4 and the outer portions of third silicon oxide layer OX 3 over doped polysilicon DP 1 .

FIG. 2X shows the device 10 of FIG. 2W after formation of silicon nitride spacers SN 5 on the sidewalls of the emitter stack E.

FIG. 2Y shows the device 10 of FIG. 2X after formation of a layer of a silicide SCD over exposed extrinsic base DP 2 and collector reachthrough region RT by a process well known to those skilled in the art of silicidation of polysilicon surfaces.

FIG. 2Z shows the device 10 of FIG. 2Y after depositing a thin conformal layer of silicon nitride layer SN 6 over the entire device.

FIG. 2 AA shows the device 10 of FIG. 2Z after deposition of a thick blanket layer of silicon oxide OX 7 over device 10 covering the conformal silicon nitride layer SN 6 . The silicon oxide layer OX 7 has been planarized to form contacts. Then a patterned layer of photoresist PR 5 was applied with openings to via holes VH 1 , VH 2 , VH 3 therethrough to form contacts to the Emitter E, the base silicide SCD over the extrinsic base DP 2 , and the collector region C via the reachthrough RT to the substrate 11 . The vias VH 1 -VH 3 are formed by RIE etching the silicon oxide layer OX 7 , followed by stripping the photoresist PR 5 .

FIG. 2 AB shows the device 10 of FIG. 2 AA after the conformal silicon nitride layer SN 6 and the silicon nitride hard mask HM above the emitter are RIE etched and the via metal contacts VI 1 , VI 2 and VI 3 are formed by deposition and planarization of a metal layer.

FIG. 3 shows a device 10 A which is a variation of device 10 of FIG. 2 AB. This structure is created the same way as the structure of FIG. 2 AB with the exception that the collector region C in the active area AA of silicon substrate 11 is further etched by wet or RIE chemistries to form undercuts UC of the intrinsic base IB. This can be implemented in the process flow shown in FIG. 2 A through FIG. 2 AB at the step shown in FIG. 2M by etching the collector region C in the active area AA to form an undercut UC that reaches underneath the intrinsic base region IB with an extra isotropic etch. Such an undercut UC can be accomplished by using isotropic RIE and/or wet etch chemistries that have high selectivity between the collector region C in the active area AA and the intrinsic base IB based on doping type, concentration and material type. For example, in the case of a heterojunction bipolar transistor with a silicon-germanium (SiGe) base and a silicon collector, a wet etch with ammonium hydroxide (HN 4 OH) can be used, which has a high etch selectivity of silicon to silicon-germanium. The width, depth and undercut of the isolation extension recesses R 1 into the collector region C in the active area AA are optimized to reduce the parasitic capacitance. In addition, the isolation extension recesses R 1 in the collector region C determines the final collector structure. Consequently, the width, depth, and undercut of the isolation extension recesses R 1 can be optimized to tailor the collector resistance, as will be well understood by those skilled in the art.

FIG. 4 shows a device 10 B which is another variation of the device 10 of FIG. 2 AB. This structure is created the same way as structure of FIG. 2 AB with the exception that the exposed portions of the intrinsic base IB and the doped polysilicon DP 1 over the shallow trench isolation STI in FIG. 2L are oxidized to form the isolation extension region IE. In this case, the process flow steps after FIG. 2L are skipped to the step in FIG. 2 P.

FIGS. 5A-5D show cross-sectional views of modified steps of the process shown in FIGS. 2 A- 2 AB to form the shallow isolation extension IE in a non-self aligned manner in making devices 10 and 10 A shown in FIG. 2 AB and FIG. 3 . The mask PR 6 is used to define and expose the region of the intrinsic base IB and the first doped polysilicon DP 1 over the primary shallow trench isolation to form the shallow isolation extension region IE prior to forming the emitter pedestal EP. The process steps in FIGS. 5A-5B replace the processing steps of FIGS. 2L-2O .

FIG. 5A shows the device 10 of FIG. 2H after deposition of silicon oxide layer OX 3 over the device 10 covering the intrinsic base IB and the first doped polysilicon layer DP 1 . Next a photoresist mask PR 6 with windows W 6 through it is patterned to form the shallow isolation extension region IE in subsequent steps.

FIG. 5B shows device 10 of FIG. 5A after an RIE etch through window W 6 in mask PR 6 through the silicon oxide layer OX 3 to form the extension window EW′. The etch stops on the region of the intrinsic base IB while exposing the right and left edges of the intrinsic base IB, and first doped polysilicon layer DP 1 to be RIE etched to form the isolation extension IE. Next the photoresist mask PR 6 is stripped. The width of the silicon oxide OX 3 portion inside the extension window EW′ defines and protects the emitter pedestal EP area and a perimeter area around the emitter pedestal EP to provide a contact area between the intrinsic base IB and the extrinsic base polysilicon DP 2 (shown in FIG. 2P ) during the RIE etch.

FIG. 5C shows the device 10 of FIG. 5B after the RIE etch through the exposed intrinsic base IB and doped polysilicon DP 1 over the STI and down into the collector region C in the active area AA using the silicon oxide OX 3 as a hard mask. This forms the isolation extension recesses R 1 ′ between the intrinsic base IB and the shallow trench isolations STI.

FIG. 5D shows the device 10 section of FIG. 5C after the silicon oxide OX 3 was removed with wet etch and a conformal silicon oxide layer OX 5 was deposited to fill the isolation extension recesses (or trenches) R 1 ′ to form the isolation extension regions IE′. The process flow then proceeds from the step in FIG. 2 I and the steps in FIGS. 2L-2O are skipped to create devices 10 and 10 A shown in FIGS. 2 AB and 3 .

FIGS. 6A-6C show cross-sectional views of a variation of the process shown in FIGS. 5A-5D to form the shallow isolation extension IE in a non-self aligned manner in making device 10 B shown in FIG. 4 using the same mask PR 6 as in FIG. 5 A. The mask PR 6 is used to define and expose the region of the intrinsic base IB and the first doped polysilicon DP 1 over the primary shallow trench isolation to form the shallow isolation extension region IE prior to forming the emitter pedestal EP. The process steps in FIGS. 6A-6C replace the processing steps of FIGS. 2L-2O . FIG. 6A shows the device 10 of FIG. 2H after deposition of silicon oxide layer OX 3 over the device 10 covering the intrinsic base IB and the first doped polysilicon layer DP 1 . Next a photoresist mask PR 6 with windows W 6 therethrough is patterned to form the shallow isolation extension region IE in subsequent steps.

FIG. 6B shows the device 10 of FIG. 6A after an RIE etch through the window W 6 in mask PR 6 through the silicon oxide layer OX 3 to form the extension window EW′. The etch stops on the region of the intrinsic base IB and first doped polysilicon layer DP 1 to be oxidized to form the isolation extension IE. Next the photoresist mask PR 6 was stripped. The width of the silicon oxide OX 3 portion inside the extension window EW′. defines and protects the emitter pedestal EP area and a perimeter area around the emitter pedestal EP to provide a contact area between the intrinsic base IB and the extrinsic base polysilicon DP 2 (shown in FIG. 2P ) during oxidation.

FIG. 6C shows the device 10 of FIG. 6B after the exposed portions of the intrinsic base IB and the doped polysilicon DP 1 over the shallow trench isolation STI in FIG. 6B are oxidized to form the isolation extension region IE′. The process flow then proceeds from the step in FIG. 2 I and the steps in FIGS. 2L-2O are skipped to create device 10 B shown in FIG. 4 .

While this invention has been described in terms of the above specific embodiment(s), those skilled in the art will recognize that the invention can be practiced with modifications within the spirit and scope of the appended claims, i.e. that changes can be made in form and detail, without departing from the spirit and scope of the invention. Accordingly all such changes come within the purview of the present invention and the invention encompasses the subject matter of the claims which follow.