Fig 6 - uploaded by Benedicte Le Tron
Content may be subject to copyright.
Schematic band diagram showing the base-collector parasitic barrier dimensions (its width 1W 3 ; and its height 1E 3 ); as well as the effect of the base-collector junction reverse bias on the parasitic barrier.  

Schematic band diagram showing the base-collector parasitic barrier dimensions (its width 1W 3 ; and its height 1E 3 ); as well as the effect of the base-collector junction reverse bias on the parasitic barrier.  

Source publication
Article
Full-text available
This paper describes a method for characterizing the bandgap narrowing and parasitic energy barrier in SiGe heterojunction bipolar transistors (HBTs), fabricated using a single-polysilicon self-aligned bipolar process. From a comprehensive study of the temperature dependence of the collector current, the bandgap narrowing in the base due to germani...

Contexts in source publication

Context 1
... simple way to highlight the existence of a parasitic energy barrier at the base-collector junction, is to reverse bias the base-collector junction, in order to widen the depletion layer and to reduce the barrier height, as illustrated schematically in Fig. 6. If we now plot the versus characteristic for a 3 V reverse bias, as shown in Fig. 7, we see that the slope increases to 67 meV, and becomes closer to the expected bandgap narrowing of 109 meV. In addition the intercept decreases to 1.5 and becomes closer to the expected value of unity. These results confirms the existence of a ...
Context 2
... Slotboom parasitic energy barrier model [13], [28] can be adapted to obtain an expression for the collector current density as a function of the barrier characteristics. If a barrier width and a barrier height are defined (Fig. 6), then the exp term in (6) can be replaced by a term For a large barrier can be approximated by In this case, the slope of a graph of versus gives and the intercept gives We can estimate the width and height of the parasitic energy barrier in transistor T2 (with 50-˚ A spacers) if we assume that the measurement on transistor T1 (with ...

Similar publications

Article
Full-text available
Mixed-phase hydrogenated silicon oxide (SiOx:H) is applied to thin-film hydrogenated amorphous silicon germanium (a-SiGe:H) solar cells serving as both p-doped and n-doped layers. The bandgap of p-SiOx:H is adjusted to achieve a highly-transparent window layer while also providing a strong electric field. Bandgap grading of n-SiOx:H is designed to...

Citations

... 29 High source doping induced band shrinking effects are captured using the OldSlotBoom bandgap narrowing model. 30 At A path = 1.70 × 10 15 cm −3 s −1 , B path = 1.6 × 10 7 V cm −1 for silicon, and A path = 2.27 × 10 15 cm −3 s −1 , B path = 1.55 × 10 7 V cm −1 for the Si 0.55 Ge 0.45 body, the simulated curve matches quite well with experimental results for V DS = 1.2 V, as shown in Figure 1E. It is also evident that a high-field-induced vertical onset is exhibited at mole fraction of 0.45. ...
Article
Full-text available
In this work, we propose and investigate a new pocket‐based Si0.55Ge0.45/Si gate normal tunnel FET design employing a gate over source with a single lateral pocket (GSLP) with and without a heterogeneous dielectric (HD) gate oxide. Miller capacitance is significantly reduced with the GSLP design, which is further improved by the HD gate oxide leading to full overshoot/undershoot suppression capability in transient response. Further, a steep switching with more than one order improvement in ON current is achieved when compared to state‐of‐the‐art line pocket designs. As a result, an ~98.8% and ~88% improved intrinsic delay (CGGVDD/ION) is achieved in comparison to dual line pocket and non‐pocketed designs, respectively. Additionally, an improved worst‐case trap‐charge tolerance, reduced pocket‐width‐induced fluctuations, and excellent immunity to gate misalignment‐induced fluctuations are achievable just by replacing the gate‐aligned parallel‐line pockets with a vertically aligned single‐lateral pocket (LP) improving the design reliability. A Ge/Si HD‐GSLP TFET design further offers stable ON currents with SS < 60 mV/dec over a feasible range of pocket widths between ~6 and 8 nm at VDS = VGS = 0.7 V.
... Dengan memperhitungkan pengaruh rekombinasi, densitas arus kolektor (J C ) dapat dinyatakan dengan Persamaan 4,5 dan 6. D nB adalah konstanta difusi elektron pada basis untuk basis pendek (short base) dengan doping merata, [11] [12] , Besarnya densitas arus kolektor pada HBT SiGe dipengaruhi oleh energi bandgap total (ΔE g ) , sehingga current gain (β) dinyatakan Persamaan 6. [13]: ...
Article
Paper ini membahas pengaruh perubahan stripe emiter area (Ae) dan fraction mole (x) terhadap unjuk kerja HBT SiGe antara lain Resistensi parasitis RB dan RC, fT, fmaks, current gain (β) serta noise figure (Fn,), model dikembangkan dari HBT SiGe IBM generasi kedua dengan Ae 0,1810 m2. Saat Ae diturunkan menjadi Ae 0,1210 m2 dan Ae 0,0910 m2 dan fraction mole (x) dinaikkan menjadi dua kali (2) maka parameter RB, dan β mempunyai relasi positif sedangkan RC , fT, fmaks negatif terhadap perubahan tersebut. Model HBT SiGe dengan x: 0.1 dan Ae:0,1810 m2 mempunyai nilai Fn minimum terendah dibanding dengan Ae 0,1210 m2 dan 0.0910 m2 yaitu 0.57 dB , 0.64 dB, 0.69 dB. Jika nilai fraction mole (x) diturunkan 50% , menyebabkan kenaikkan Fn yang tidak linier yaitu 77%, 79% dan 89% dari nilai semula. Relasi noise figure (Fn) dengan stripe emiter area (Ae) dan fraction mole (x) diekspresikan dengan relasi berikut ; F k A x n e. 0  , jadi noise figure (Fn) dapat diperkecil dengan memperarea stripe emiter area (Ae) dan menaikkan fraction mole (x). Kata kunci : Noise Figure (Fn), stripe emiter area (Ae), fraction mole (x), SiGe HBT ABSTRAK Paper ini membahas pengaruh perubahan stripe emiter area (Ae) dan fraction mole (x) terhadap unjuk kerja HBT SiGe antara lain Resistensi parasitis RB dan RC, fT, fmaks, current gain (β) serta noise figure (Fn,), model dikembangkan dari HBT SiGe IBM generasi kedua dengan Ae 0,1810 m2. Saat Ae diturunkan menjadi Ae 0,1210 m2 dan Ae 0,0910 m2 dan fraction mole (x) dinaikkan menjadi dua kali (2) maka parameter RB, dan β mempunyai relasi positif sedangkan RC , fT, fmaks negatif terhadap perubahan tersebut. Model HBT SiGe dengan x: 0.1 dan Ae:0,1810 m2 mempunyai nilai Fn minimum terendah dibanding dengan Ae 0,1210 m2 dan 0.0910 m2 yaitu 0.57 dB , 0.64 dB, 0.69 dB. Jika nilai fraction mole (x) diturunkan 50% , menyebabkan kenaikkan Fn yang tidak linier yaitu 77%, 79% dan 89% dari nilai semula. Relasi noise figure (Fn) dengan stripe emiter area (Ae) dan fraction mole (x) diekspresikan dengan relasi berikut ; F k A x n e. 0  , jadi noise figure (Fn) dapat diperkecil dengan memperarea stripe emiter area (Ae) dan menaikkan fraction mole (x). Kata kunci : Noise Figure (Fn), stripe emiter area (Ae), fraction mole (x), SiGe HBT
... 4 corresponds to a BGN of roughly 30 mV. This is close to the value of 39 mV reported in [16] at a mean base doping concentration of 2 · 10 18 cm −3 in the absence of Ge. Consequently, bipolar transistors for precision bandgap reference applications shall be preferably characterized by models with the proposed extensions.Fig. ...
Article
Full-text available
A new compact model approach is suggested for the low-bias base charge of homo- and heterojunction bipolar transistors. It is shown that the junction-related (Early) components of this charge depend on an effective doping density. This concept leads to an accurate description with the conventional capacitance-charge model formulas but using different parameters directly extracted from direct-current measurements instead of the classical parameter values derived from alternating-current measurements. The temperature dependence of the low-bias base charge has also been revised, resulting in a new term added to the currently adopted temperature law. The applied novel analytic technique made it possible to determine the influence of bandgap narrowing (BGN) effects and Ge doping on the temperature behavior of SiGe transistors.
... 41 shows the confinement of the emitter area and the preparation required to isolate the emitter from the base contact. Here, an over-etch is applied to ensure that epitaxial silicon is completely cleared from the surface (if the selective epitaxy method is used this step can be minimally applied or skipped altogether. ...
Thesis
p>This work provides a detailed study of device structures and fabrication routes required for the realisation of lateral SiGe heterojunction bipolar transistors (HBTs). After a comprehensive study of BJT and HBT device technologies our own designs of lateral SiGe HBT are introduced. The first design investigated is a device proposed in earlier work that utilises "SOI cavities" and confined selective epitaxial growth. Process and device characteristic simulations were performed for devices of modest lithographic mode and thick SOI suggested the device would be reach typical SiGe HBT performance levels with a cut off frequency (/r) of 15-22GHz, common emitter gain of 58-92 and maximum oscillation frequency (fmax) of 14GHz. The fabrication of the types of lateral HBT that been have explored require detailed understanding and control of selective epitaxial growth and etching techniques and each of these two important processes have been developed. Selective epitaxy growths on different crystallographic planes and seed window alignments have been carried out, and, as would be expected, selective epitaxial growth rate are seen to vary for each crystal plane. These different growth rates leads to different facet formations on lateral growth fronts that can inhibit lateral growth rates and thereby prevent the formation of suitable structures for HBT design. A mathematical model is developed that enables a prediction of these growth phenomenon. The epitaxial studies suggest that smooth vertical {100} sidewalls are required to enable suitable lateral growth two different approaches to achieve this are detailed. The first process uses conventional anisotropic plasma etching to establish followed by wet etching (KOH) for smoothing, the second technique is based on KOH etching alone. Both methods demonstrate smooth sidewalls suitable for epitaxy and the simplicity of the process that uses only KOH etching is therefore preferable. As with epitaxial growth, KOH etch rates are strongly dependent upon crystal plane a model is developed to allow an understanding of the different etch rates and facet formation. The epitaxy and etching development studies have allowed us to propose a new design of lateral SiGe HBT. Based on non-selective growth on SOI sidewalls. Simulations of this new design indicate high-speed performance with/r //max of 395/983GHz and 1.58ps ECL delay time on the 0.13mm lithographic node at collector currents of ~8mA and a common emitter gain around 600. Our results indicate great potential for such a device but highlight several fabrication challenges that requires further development before such devices can be realised.</p
... Les fréquences de transition sont passées en une dizaine d'années, typiquement de 20 GHz à 200 GHz. Par ailleurs, il faut mentionner que les éléments parasites associés à l'architecture du composant, essentiellement la résistance de base et la capacité collecteurbase [1,2], fortement liés à la technique d'intégration, limitent considérablement les performances électriques du HBT (fréquence de transition ft et temps de propagation). ...
Article
Full-text available
In this paper, we study bipolar transistors specified for analog and analog-digital circuits and working in the radiofrequency range . These devices are fully compatible with a deeply submicronic silicon CMOS technology. The ways to build such structures are twofold. On the first hand, silicon-germanium material replaces the silicon of the base ; on the other hand, a very thin thin oxide layer is introduced between the polysilicon and the underlying monocrystalline silicon of the emitter, acting as a tunneling barrier. This work, starting from a material point of view, is concluded by a fully numerical two-dimensional source code solving numerically the coupled equation of Poisson, ambipolar current continuity and Schroedinger equations.
... If the Si base is replaced by a SiGe base, the collector current expression (4) becomes [8] ...
Article
This paper describes a new way for measuring the apparent band gap narrowing in bipolar transistors with Si base or epitaxial Si0.80Ge0.20 base from the temperature dependence of the collector current JC(T). A graph of as a function of VBE is plotted based on the data extracted from the linear region of the Gummel plot as the temperature is kept as a constant. We have obtained two lines of at temperature, 300 and 77 K, respectively. From the intersection point of the lines we can calculate the values of band gap narrowing in Si and Si0.80Ge0.20 bases, which are 41 and 151 meV, respectively. The results are in good agreement with the measured values in reference.
... The relative in¯uence of TED will be more pronounced for short annealing times and at low temperatures. Out-diusion of the intrinsic base pro®le has previously been investigated using process simulations [9], device simulations [10] and experiments [11,12]. The silicon interstitials have a high diusivity and the lateral range of the base pro®le broadening is large enough to aect the intrinsic device Solid-State Electronics 44 (2000) 1747±1752 [11]. ...
Article
The influence of transient enhanced boron out-diffusion from the intrinsic base, caused by excess silicon interstitials created during the extrinsic base implantation, has been investigated for a non-selective SiGe HBT process. Devices with different designs of the extrinsic base region were fabricated, where some designs allowed part of the epitaxial base to be implanted with a high boron dose, hereby increasing the number of silicon interstitials close to the intrinsic device. These devices showed a marked degradation of DC characteristics and HF performance. 2D-device simulations were used to investigate the sensitivity in DC and HF parameters to vertical base profile changes. Good agreement was obtained between measured and simulated DC and HF characteristics.
... Two existing methods were proposed by Slotboom W. [20] and Le Tron a/. [21]. where nib{SiGe) is the effective intrinsic carrier concentration in the SiGe base. ...
... Le Tron aZ. [21] haa proposed a way of measuring the total bandgap reduction due to germanium and heavy doping eSects, in a SiGe HBT. ...
... HBT is further complicated by the fact that both the germanium and doping proxies cannot be made truly abrupt in spatial composition which results in concentration "tails" that can inadvertently result in unwanted differences in bandgap across the neutral base. If the base doping extends beyond the Si-SiGe boundaries into the emitter and/or collector regions, parasitic potential barriers will form severely degrading current gain and transition frequency [18,20,21]. ...
Thesis
p>This thesis investigates advanced characterisation and modelling techniques for silicon-germanium heterojunction bipolar transistor (SiGe HBT's). Two characterisation techniques are proposed and evaluated to enable, as much as is possible, direct electrical extraction of physical device characteristics which are unique to SiGe HBT's, as opposed to standard silicon-only bipolar transistors. Principal objectives motivating the new characterisation techniques are the elimination of silicon control devices and the use of widely available measurement apparatus. The bandgap difference across the neutral base extraction method is further developed for extracting the parasitic potential barrier height in a SiGe HBT. The proposed method is shown to be able to extract the parasitic potential barrier heights at the emitter-base and collector base junctions of SiGe HBT's simultaneously in conjunction with numerical space-charge layer modelling using doping secondary-ion-mass spectroscopy (SIMS) profile data. It provides a more direct measure of parasitic barrier-related quantities and does not depend upon a detailed knowledge of the temperature dependence of SiGe density of states functions, carrier mobility, etc.. Numerical simulations and measurement results show that the method gives useful and representative values regarding the parasitic potential barrier height. Finally, the potential of a proposed novel lateral SiGe HBT structure for high performance rf/microwave applications is assessed. Various issues regarding the lateral structure, including base definition and base contact, are discussed. Compared to a state-of-the art vertical SiGe HBT structure, the lateral structure is found to have many advantages for low power rf/microwave applications. A realistic comparison is carried out by means of full 2-D numerical simulation. Numerical simulation results indicate that a lateral SiGe HBT structure can potentially out-perform a vertical structure in term of low power and high frequency performance.</p
... Ajjoys 7 Section 2.7 discusses advanced issues that affect device performance, such as parasitic energy barriers. Section 2.8 outlines details of a novel electrical method [48,73] that allows the bandgap narrowing in the base to be extracted from the temperature dependence of the collector current. This allows the presence and approximate magnitude of parasitic energy barriers to be determined. ...
... The Slotboom parasitic energy barrier model [14] can be adapted to obtain an expression for the collector current density as a function of the barrier dimensions [48]. For a barrier width AW and height AE*, the exponential term exp(qVBE/kT) ...
Thesis
p>A background doping concentration (10<sup>20</sup>cm<sup>-3</sup>) of C has been introduced into the base of SiGe HBTs with the aim of studying the effects of C on TED of B from the base. An electrical method is used to extract the bandgap narrowing in the base of SiGe and SiGe:C HBTs through measurements of the temperature dependence of I<sub>c</sub> at different C/B reverse biases. The method is very sensitive to small amounts of dopant out-diffusion from the base and hence is ideal for determining the effect of C on TED. Extracted BGN values of 11meV and 173meV were obtained for the SiGe and SiGe:C HBTs respectively, for a C/B reverse bias of 0V. Increasing the C/B reverse bias to IV increased the extracted BGN of the SiGe HBT to 145meV, but left the SiGe:C value unchanged. This demonstrates that no parasitic energy barrier exists in the SiGe:C HBT and that TED has been suppressed. The effect of carbon position and concentration has been studied by introducing a peak C concentration of 10<sup>20</sup>cm <sup>-3</sup> in the collector and 1.1 x 10<sup>19</sup>cm <sup>-3</sup> or 1.5 x 10<sup>19</sup>cm <sup>-3</sup> C in the base. From these measurements it has been shown that TED is only suppressed in the device with 1.5 x 10<sup>19</sup>cm <sup>-3</sup> is needed to suppress TED and that the C needs to be co-located with the B profile. The effects of carbon on the electrical properties of polycrystalline Si and SiGe films have been investigated. The resistivity, Hall mobility ( μ<sub>H</sub> ) and effective carrier concentration ( N<sub>EFF</sub> ) of n- and p-type polySi<sub>1-y</sub>C<sub>y</sub> and polySi<sub>0.82-y</sub>Ge<sub>0.18</sub>C<sub>y</sub> layers have been measured for carbon contents between 0% and 8%. For the n-type polySi<sub>1-y</sub>C<sub>y</sub> and polySi<sub>0.82-y</sub>Ge<sub>0.18</sub>C<sub>y </sub>layers, the addition of small amounts of C (≤ 0.9%) was found to severely increase the resistivity of the layers, caused by a drop in N<sub>EFF</sub> AND μ<sub>H</sub> . In contrast, for the p-type polySi<sub>1-y</sub>C<sub>y</sub> and polySi<sub>0.82-y</sub>Ge<sub>0.18</sub>C<sub>y</sub> layers, the effect of C on the resistivity was much less dramatic for C concentrations up to 7.8%. Measurements of the grain boundary energy barriers for the n-type polySi<sub>1-y</sub>C<sub>y</sub> and polySi<sub>0.82-y</sub>Ge<sub>0.18</sub>C<sub>y</sub> layers, extracted from the temperature dependence of the resistivity, showed that there was a square law dependence on carbon content.</p
... In this letter, an electrical method [6], [7] is applied to characterize parasitic energy barriers in SiGe HBT's, with and without a background C concentration 10 cm This method is extremely sensitive to small amounts of boron out-diffusion from the base and hence allows accurate determination of the presence of parasitic energy barriers. It is shown that this relatively low background carbon concentration is completely effective in suppressing the transient ehanced boron diffusion in SiGe HBT's. ...
... The collector current and base sheet resistance were measured as a function of temperature in the range of 200 to 400 K. The measured collector current is normalized against a saturation current density , given by (1) where A graph of as a function of reciprocal temperature gives a straight line with the slope equal to the total bandgap narrowing (BGN) in the base (due to heavy doping effects and the presence of Ge) [6], [7]. ...
... The slope of the graph is used to characterize the parasitic energy barriers, since a lower slope is observed when a barrier is present [7]. The existence of barriers is confirmed by varying the collector-base reverse bias. ...
Article
Full-text available
An electrical method is applied to SiGe and SiGe:C heterojunction bipolar transistors (HBTs) to extract the bandgap narrowing in the base layer and to characterize the presence of parasitic energy barriers in the conduction band, arising from boron transient enhanced out-diffusion from the SiGe layer. It is shown that a background carbon concentration within the base (/spl ap/10/sup 20/ cm/sup -3/) eliminates parasitic energy barriers at the collector/base junction, and hence shows that transient enhanced diffusion of boron from the base has been completely suppressed.