No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Absence of enhanced stability in fully deuterated amorphous silicon thin-film transistors
Abstract
The stability of fully deuterated amorphous silicon a-Si: D thin-film transistors is compared with their hydrogenated equivalent a-Si: H in terms of gate bias stress. The amorphous silicon channel and silicon nitride gate insulator layers were deposited by radio-frequency plasma-enhanced chemical-vapor deposition. The use of SiD 4 rather than SiH 4 for the deposition of a-Si: D changes the physical properties of the plasma given the same conditions of rf power, pressure, and gas flow rates. Consequently, a higher gas pressure is required to produce a-Si: D at the same growth rate and with similar bulk properties as a-Si: H. It is shown that a-Si: H and a-Si: D deposited at the same growth rate have very similar structural properties. Therefore transistors deposited at the same growth rate may be more sensibly compared to determine the effect of replacing H with D in amorphous silicon without significantly changing the silicon continuous random network. Using this criterion for comparison, no detectable difference is observed between hydrogenated and deuterated transistors in terms of stability under the application of a gate bias. The experimental results rule out the possibility of a giant isotopic effect in amorphous silicon. Furthermore, this result supports the idea that the rate-limiting step for dangling-bond defect creation in amorphous silicon is the breaking of a weak Si–Si bond, rather than breaking of a Si–H bond. © 2005 American Institute of Physics.
... 8 Such an analysis has previously been used to describe a microscopic model for defect creation which explains the threshold voltage shift in a-Si:H TFTs under bias stress. [9][10][11] The concept of the thermalization energy starts with the assumption that a material contains a number of sites which may be converted into defects by some means (a gate bias, for example). Each site has a particular energy barrier to the conversion process, and so the material as a whole will have a distribution of energy barriers, DðEÞ. ...
Thin film transistors(TFTs) employing an amorphous indium gallium zinc oxide (a-IGZO) channel layer exhibit a positive shift in the threshold voltage under the application of positive gate bias stress (PBS). The time and temperature dependence of the threshold voltage shift was measured and analysed using the thermalization energy concept. The peak energy barrier to defect conversion is extracted to be 0.75 eV and the attempt-to-escape frequency is extracted to be 107 s−1. These values are in remarkable agreement with measurements in a-IGZO TFTs under negative gate bias illumination stress (NBIS) reported recently (Flewitt and Powell, J. Appl. Phys. 115, 134501 (2014)). This suggests that the same physical process is responsible for both PBS and NBIS, and supports the oxygen vacancy defect migration model that the authors have previously proposed.
... In order to evaluate a-Si:H-film stability grown on different gate dielectric surfaces such as SiNx and SiOx, we applied a thermalization-energy concept in the defect pool model that unifies the time and temperature dependence of Si dangling-bond-defect creation and removal in amorphoussilicon thin-film transistors. [19][20][21][22][23][24][25][26][27][28][29] A thermalization energy E th is defined by E th ¼ kT ln (t), where k is Boltzmann factor, T is temperature, is the attempt-to-escape frequency (¼10 10 Hz, same as in Ref. 23), and t is time. In case of (a) a-Si:H/SiNx, the transfer curves shift in parallel irrespective of the amount and polarity of gate bias stress. ...
In order to investigate the effects of interface and bulk properties of gate insulator on the threshold voltage (Vth) and the gate-bias induced instability of hydrogenated amorphous silicon thin-film transistors (a-Si:H TFTs), four kinds of TFTstructures were fabricated with SiNx and SiOx insulators stacked to make different combinations of the bulk and interface in the gate-dielectric layers. It was found that the Vth and the stability are independently controlled by tuning stoichiometry and thickness of the SiOx insertion layer between a-Si:H and SiNx. In TFTs with SiOx insertion layer of 50 nm thickness, on increasing oxygen/silicon (O/Si = x) ratio from 1.7 to 1.9, Vth increased from 0 V to 9 V. In these TFTs with a relatively thick SiOx insertion layer, positive Vth shift with negative bias stress was observed, confirmed to be due to defect creation in a-Si:H with the thermalization barrier energy of 0.83 eV. On reducing the thickness of the SiOx insertion layer down to approximately 1 nm, thin enough for hole injection through SiOx by tunneling effect, stable operation was obtained while keeping the high Vth value under negative stress bias. These results are consistently explained as follows: (1) the high value for Vth is caused by the dipole generated at the interface between a-Si:H and SiOx; and (2) two causes for Vth shift, charge injection to the gate insulator and defect creation in a-Si:H, are mutually related to each other through the “effective bias stress,” Vbseff = Vbs – ΔVfb (Vbs: applied bias stress and ΔVfb: flat band voltage shift due to the charge injection). It was experimentally confirmed that there should be an optimum thickness of SiOx insertion layer of approximately 1 nm with stable high Vth, where enhanced injection increases ΔVfb, reduces Vbseff to reduce defect creation, and totally minimizes Vth shift.
... The recent study of this 'negative bias under illumination stress' (NBIS) by Chowdhury et al. is particularly interesting, 8 as the authors measure the effect of temperature on this process, and this permits an analysis based on the concept of the thermalization energy which has previously been applied to amorphous, microcrystalline and polycrystalline silicon TFTs. [11][12][13] In this work, such a thermalization energy analysis is applied to NBIS data of Chowdhury et al. 8 The parameters from this analysis are compared with previous studies on the stability of a-Si:H TFTs. It is clear that a very different microscopic mechanism is responsible for the threshold voltage shift in a-IGZO TFTs. ...
It has been previously observed that thin film transistors (TFTs) utilizing an amorphous indium gallium zinc oxide (a-IGZO) semiconducting channel suffer from a threshold voltage shift when subjected to a negative gate bias and light illumination simultaneously. In this work, a thermalization energy analysis has been applied to previously published data on negative bias under illumination stress (NBIS) in a-IGZO TFTs. A barrier to defect conversion of 0.65–0.75 eV is extracted, which is consistent with reported energies of oxygen vacancy migration. The attempt-to-escape frequency is extracted to be 106−107 s−1, which suggests a weak localization of carriers in band tail states over a 20–40 nm distance. Models for the NBIS mechanism based on charge trapping are reviewed and a defect pool model is proposed in which two distinct distributions of defect states exist in the a-IGZO band gap: these are associated with states that are formed as neutrally charged and 2+ charged oxygen vacancies at the time of film formation. In this model, threshold voltage shift is not due to a defect creation process, but to a change in the energy distribution of states in the band gap upon defect migration as this allows a state formed as a neutrally charged vacancy to be converted into one formed as a 2+ charged vacancy and vice versa. Carrier localization close to the defect migration site is necessary for the conversion process to take place, and such defect migration sites are associated with conduction and valence band tail states. Under negative gate bias stressing, the conduction band tail is depleted of carriers, but the bias is insufficient to accumulate holes in the valence band tail states, and so no threshold voltage shift results. It is only under illumination that the quasi Fermi level for holes is sufficiently lowered to allow occupation of valence band tail states. The resulting charge local- zation then allows a negative threshold voltage shift, but only under conditions of simultaneous negative gate bias and illumination, as observed experimentally as the NBIS effect.
Hydrogenated amorphous silicon has enabled the active matrix liquid crystal display to dominate the flat panel display market and indeed has supplanted the cathode-ray tube as the leading display technology. This is the triumph of manufacturability over performance. Hydrogenated amorphous silicon thin-film transistors have several fundamental performance limitations which are directly linked to the physics of the amorphous material. However, the amorphous structure coupled with the use of plasma-enhanced chemical vapor deposition (PECVD) allows devices to be manufactured with exceptional reproducibility and uniformity over very large-area display backplanes. Consequently, the material limitations are tolerated and engineering solutions found to mitigate their effects.
In this paper, hydrogenated amorphous silicon (a-Si:H) and microcrystalline silicon ( (mu ) c-Si:H) films are deposited by electron cyclotron resonance chemical vapor deposition with two separate silane gas inlets. One of the silane gases (S2) is introduced near the substrate region. Effects of S2 flow rate on film properties and solar cell performance are investigated in comparison to traditional plasma-enhanced chemical vapor deposition (PECVD). The results show that the introduction of S2 gas leads to: 1) significant reduction of higher order silane radicals participating film growth; 2) dense film structure with a low microstructure factor of 0.06; and 3) lower surface roughness of the interface between top a-Si:H and bottom (mu ) c-Si:H subcells of micromorph tandem cells, favoring bottom (mu ) c-Si:H deposition. Single-junction amorphous silicon solar cells show light-induced degradation (LID) of 7.8%, almost half of that observed in PECVD cells. Micromorph tandem solar cells show a 13.3% initial conversion efficiency and a 12.7% stabilized efficiency. Highly stabilized micromorph tandem solar cell with 4.7% LID can be achieved.
It has been widely observed that thin film transistors (TFTs) incorporating an hydrogenated amorphous silicon (a-Si:H) channel exhibit a progressive shift in their threshold voltage with time upon application of a gate bias. This is attributed to the creation of metastable defects in the a-Si:H which can be removed by annealing the device at elevated temperatures with no bias applied to the gate, causing the threshold voltage to return to its original value. In this work, the defect creation and removal process has been investigated using both fully hydrogenated and fully deuterated amorphous silicon (a-Si:D) TFTs. In both cases, material was deposited by rf plasma enhanced chemical vapour deposition over a range of gas pressures to cover the a-g transition. The variation in threshold voltage as a function of gate bias stressing time, and annealing time with no gate bias, was measured. Using the thermalisation energy concept, it has been possible to quantitatively determine the distribution of energies required for defect creation and removal as well as the associated attempt-to-escape frequencies. The defect creation and removal process in a-Si:H is then discussed in the light of these results.
Deposition of nanocrystalline silicon (nc-Si) on glass at very low temperatures by electron cyclotron resonance (ECR) plasma enhanced chemical vapour deposition (PECVD) was investigated. It was shown that nc-Si could be deposited from hydrogen diluted silane gas at a substrate temperature of 80° C with a crystalline fraction up to 80% and a lateral grain size of around 50 nm. This was achieved by growing the nc-Si in a low pressure regime which ensured that mono-silyl species were the dominant deposition precursor. Furthermore, a high flux of energetic hydrogen ions was required to induce crystallisation of the silicon material through a chemical annealing process.
The influence of radio frequency (rf) power and pressure on deposition rate and structural properties of hydrogenated amorphous silicon (a‐Si:H) thin films, prepared by rf glow discharge decomposition of silane, have been studied by phase modulated ellipsometry and Fourier transform infrared spectroscopy. It has been found two pressure regions separated by a threshold value around 20 Pa where the deposition rate increases suddenly. This behavior is more marked as rf power rises and reflects the transition between two rf discharges regimes (α and γ). The best quality films have been obtained at low pressure and at low rf power but with deposition rates below 0.2 nm/s. In the high pressure region, the enhancement of deposition rate as rf power increases first gives rise to a reduction of film density and an increase of content of hydrogen bonded in polyhydride form because of plasma polymerization reactions. Further rise of rf power leads to a decrease of polyhydride bonding and the material density remains unchanged, thus allowing the growth of a‐Si:H films at deposition rates above 1 nm/s without any important detriment of material quality. This overcoming of deposition rate limitation has been ascribed to the beneficial effects of ion bombardment on the a‐Si:H growing surface by enhancing the surface mobility of adsorbed reactive species and by eliminating hydrogen bonded in polyhydride configurations.
We study the magnitude of metastable light-induced changes in undoped hydrogenated amorphous silicon (the Staebler-Wronski effect) with electron-spin-resonance and photoconductivity measurements. The influence of the following parameters is investigated in a systematic way: sample thickness, impurity content, illumination time, light intensity, photon energy, and illumination and annealing temperatures. The experimental results can be explained quantitatively by a model based on the nonradiative recombination of photoexcited carriers as the defect-creating step. In the framework of this model, the Staebler-Wronski effect is an intrinsic, self-limiting bulk process, characterized by a strongly sublinear dependence on the total light exposure of a sample. The experimental results suggest that the metastable changes are caused by recombination-induced breaking of weak Si-Si bonds, rather than by trapping of excess carriers in already existing defects. Hydrogen could be involved in the microscopic mechanism as a stabilizing element. The main metastable defect created by prolonged illumination is the silicon dangling bond. An analysis of the annealing behavior shows that a broad distribution of metastable dangling bonds exists, characterized by a variation of the energy barrier separating the metastable state from the stable ground state between 0.9 and 1.3 eV.
Based on hydrogen evolution and secondary ion mass spectrometry measurements, the relation between incorporation and stability of hydrogen is discussed for undoped amorphous Si and Ge. In both materials, an increase of hydrogen concentration results in a decrease of H stability as monitored by an enhanced H diffusion and the appearance of void-related H surface desorption. Besides film microstructure, the hydrogen chemical potential defining the filling level of a film-specific hydrogen density of states is identified as an important parameter affecting H stability. For material with low H concentration, the H chemical potential moves with temperature and is enhanced by contact with material of high H concentration. While H surface desorption takes place for amorphous Ge at considerably lower temperature than for amorphous Si, less pronounced differences occur for H diffusion in the compact materials.
Light-induced changes in the photoconductivity and defect density have been compared in deuterated and hydrogenated amorphous silicon. There is no significant difference between the as-deposited properties of the two materials; however, the photoconductivity degrades more slowly but anneals more rapidly in material prepared using the heavier isotope. In hydrogenated material, the average photocarrier recombination cross-section of the defects increases with light soaking but the opposite occurs in deuterated material. These results are explained in terms of two kinds of recombination centre which are photocreated at different rates. The variation in the degradation and annealing behaviour of photoconductivity with deposition temperature implies subtle structural differences between the two materials which, for the first time, gives rise to the possibility of controlling photodegradation in amorphous silicon.
The first hydrogen dilution study of ammonia/silane plasmas tuned to an aminosilane plasma regime is reported. The use of formal statistical experimental design techniques have enabled us to determine the effects that hydrogen dilution level, rf power density and substrate growth temperature have on various film properties. Hydrogen dilution of the ammonia/silane plasma reduced the amino (NH2) content of the deposited nitride films while also driving the intrinsic film stress to compressive values. Increasing the substrate temperature also reduced the amino concentration, but drove the film stress tensile. It was therefore possible to control the film stress and produce nitride films that contained only NH and SiN bonds (with no detectable NH2 or SiH bonds) by choosing a suitable combination of H2 dilution level and growth temperature. Exodiffusion experiments on such ‘bond optimised’ films revealed only one hydrogen evolution peak at temperatures in excess of 900°C, with no ammonia exodiffusion detected.
We report experimental results that replacing hydrogen with deuterium during the final wafer sintering process greatly reduces hot electron degradation effects in metal oxide semiconductor transistors due to a new giant isotope effect. Transistor lifetime improvements by factors of 10–50 are observed. A plausible physical theory suggests that the benefits of deuterium use may be general and also applicable to other areas of semiconductor device processing and fabrication. © 1996 American Institute of Physics.
We have studied light-induced degradation in hydrogenated and deuterated amorphous silicon alloy solar cells. Replacing hydrogen with deuterium in the intrinsic layer of the cell improves stability against light exposure. Possible explanations for the improved stability are discussed. © 1997 American Institute of Physics.
To improve the bias-induced degradation in hydrogenated amorphous silicon thin film transistor, the hydrogen in the amorphous silicon film should be replaced by deuterium. The stability of deuterated amorphous silicon thin film transistors, i.e., the shifts of threshold voltage and subthreshold swing, is indeed improved compared to that of the hydrogenated ones. This result is consistent with the improvement of the light-induced degradation in deuterated amorphous silicon films and this improvement can be explained by the efficient coupling between the Si–D wagging mode and the amorphous silicon phonon mode. © 1999 American Institute of Physics.
We have studied light-induced photoconductivity degradation in an intrinsic hydrogenated and deuterated amorphous-silicon (a-Si) alloy. Deuterated a-Si turns out to be more stable under light exposure. A possible mechanism is proposed to explain this phenomenon. It is attributed to the highly efficient coupling between the localized Si–D wagging modes ( ∼ 510 cm−1) and the extended Si–Si lattice vibration modes ( ∼ 495 cm−1). The energy released from electron or hole capture at silicon dangling bonds causes localized vibrations of nearby Si–D bonds. The energy dissipates quickly to the background lattice and a higher recombination rate at local sites is needed in deuterated a-Si than in hydrogenated amorphous silicon to accumulate enough energy to break the nearby weak bonds. © 1997 American Institute of Physics.
Light-induced metastable defects in a-Si:H are proposed to be silicon dangling bonds accompanied by pairs of hydrogen atoms breaking a silicon bond, forming a complex with two Si-H bonds. This supports the model of Branz. These defects are the analog of the H2∗ defect in c-Si and their energy correlates with the bond-angle strain. Several features of the annealing are well described by this defect complex. © 1998 American Institute of Physics.
Scitation is the online home of leading journals and conference proceedings from AIP Publishing and AIP Member Societies
We have investigated the effect of depositing amorphous silicon (a-Si:H) under α and γ (or dusty) plasma conditions on hydrogen bonding and TFT performance. By infrared measurements three deposition regimes could be identified due to their distinctive absorption curves in the bending mode range of 800 to 900 cm'1. These corresponded to a films deposited below 300°C, γ films deposited below 300°C, and films deposited in either plasma condition at temperatures of 300°C and above. There was a correlation between the a-Si:H material regimes and TFT performance. When the deposition temperature was below 300°C the a-TFTs had a higher field effect mobility than γ-TFTs, but lower stability. For deposition temperatures of 300°C and above the films had more similar properties regardless of whether they were deposited under a or γ conditions. TFT mobilities were the same, but TFTs containing a-Si:H deposited under γ conditions were still more stable. These results show that the mobility and stability of TFTs are optimised for different growth conditions, and that the overall best conditions for TFT manufacture depends upon the specific application.
We present a thermalization-energy concept that unifies the time and temperature dependence of Si dangling-bond-defect creation and removal in amorphous-silicon thin-film transistors. There is a distribution of energy barriers for defect creation and removal, with the most probable energy barrier being 1.0 eV for defect creation and between 1.1 and 1.5 eV for defect removal, depending on how the defects were initially created. We suggest defect creation proceeds via Si-Si bond breaking, whereas defect removal proceeds by release of H from a SiHD complex. [S0163-1829 (98)00243-4].
We investigate the mechanism for Si dangling bond defect creation in amorphous silicon thin film transistors as a result of bias stress. We show that the rate of defect creation does not depend on the total hydrogen content or the type of hydrogen bonding in the amorphous silicon. However, the rate of defect creation does show a clear correlation with the Urbach energy and the intrinsic stress in the film. These important results support a localized model for defect creation, i.e., where a Si–Si bond breaks and a nearby H atom switches to stabilize the broken bond, as opposed to models involving the long-range diffusion of hydrogen. Our experimental results demonstrate the importance of optimizing the intrinsic stress in the films to obtain maximum stability and mobility. An important implication is that a deposition process where intrinsic stress can be independently controlled, such as an ion-energy controlled deposition should be beneficial, particularly for deposition temperatures below 300 °C. © 2000 American Institute of Physics. S0021-89790010001-5
We compared threshold voltage shifts in amorphous Si, microcrystalline Si and polycrystalline Si thin-film transistors TFTs in terms of a recently developed thermalization energy concept for a dangling-bond defect state creation in amorphous Si TFTs. The rate of the threshold voltage shift in microcrystalline Si TFTs was much lower than in amorphous Si TFTs, but the characteristic energy for the process, which we identified as the mean energy to break a SiSi bond, was virtually the same. This suggests that the same basic SiSi bond breaking process was responsible for the threshold voltage shift in both cases. The lower magnitude in microcrystalline Si TFTs was due to a much lower attempt frequency for the process. We interpreted the attempt frequency in amorphous and microcrystalline silicon in terms of the localization length of the electron wavefunction and the effect of stabilizing H atoms being located only at grain boundaries. 2001 Elsevier Science B.V. All rights reserved.
We present a microscopic model for metastable Si dangling-bond defect creation in hydrogenated amor-phous silicon, which is applicable to both light-induced defect creation in solar cells Staebler-Wronski effect and bias-stress-induced defect creation in thin-film transistors. Light or gate bias causes electron-hole pairs or electrons, respectively, to be localized on short, weak Si-Si bonds, which then break. A hydrogen atom, from a neighboring, doubly hydrogenated weak Si-Si bond SiHHSi moves to the T d site of the broken Si-Si bond. The other H atom from the SiHHSi is also located in the energetically favorable T d site. Overall, the reaction produces two SiHD defects. Each SiHD defect is an intimate Si dangling bond and Si-H bond, where the H atom is in the T d site, not the BC site. The distance between the dangling bond and the H atom in the T d site is in the range 4 –5 Å, in agreement with ESR data. The majority of silicon dangling bonds, both metastable and stable, exist as SiHD, with the H atom in the T d site. The microscopic process for defect creation is fairly well localized, requiring only short-range H motion, which proceeds via bond switching between neighboring T d sites. In contrast, the microscopic process for defect removal during thermal annealing involves reequili-bration of H in the a-Si:H network and is a global process involving a large fraction of H atoms. The rate-limiting step for this process is Si-H bond breaking from SiHHSi sites, which accounts for the maximum activation energy of 1.5 eV. We present a revised hydrogen density of states diagram, in line with this process.
A modified Schottky-contact gated-four-probe structure was applied to study the stability of the hydrogenated and deuterated amorphous silicon (a-Si:D) thin-film transistors under various bias conditions. It was found that after 10 V bias stress, the density of gap states generated in both the upper and lower part of the mobility gap of deuterated amorphous silicon is two to twenty times less than those of hydrogenated silicon. Besides, less density of states at the lower part of mobility gap of a-Si:D is generated after 20, 10, and 20 V bias stress. © 2003 American Vacuum Society.
Long exposure to light decreases the photoconductivity and dark conductivity of some samples of hydrogenated amorphous silicon (a‐Si : H). Annealing above ∼150 °C reverses the process. The effect occurs in the bulk of the films, and is associated with changes in density or occupation of deep gap states. High concentrations of P, B, or As quench the effect. Possible models involving hydrogen bond reorientation at a localized defect or electron‐charge transfer between defects are discussed. An example is shown where these conductivity changes do not affect the efficiency of an a‐Si : H solar cell.
The light‐induced decrease of the photoconductivity in deuterated amorphous silicon is a factor of 3 less even though the defect density increase is greater than in hydrogenated material having equivalent as‐deposited properties. Consequent changes in the average recombination cross section of the defects is illustrated. Since the differences in the light soaking behavior upon isotopic substitution has been found to disappear in films deposited at low temperatures, the changes are thought to arise from differences in the silicon network occurring during growth.
This article presents infrared absorption data of amorphous silicon alloys in which the hydrogen isotopes deuterium and tritium have been substituted for hydrogen. Silicon-deuterium and silicon-tritium vibration frequencies are related to silicon-hydrogen vibration frequencies by simple mass relationships. The silicon-deuterium wagging vibration is broadened and blueshifted due to strong coupling to the amorphous silicon network vibrations. (C) 1999 American Institute of Physics. [S0021-8979(99)08304-8].
The Raman spectra of hydrogenated and deuterated amorphous silicon films (a- Si:H , a- Si:D ) have been investigated. It is suggested that the asymmetrical broadening of the transverse-optical (TO) Raman peak of a- Si:D compared to the TO Raman peak of a- Si:H results from the coupling between the Si–D wagging mode and the Si–Si TO phonon mode rather than the structural difference. © 1999 American Institute of Physics.
The threshold voltage stability of fully deuterated (a- Si : D ) and hydrogenated amorphous silicon (a- Si : H ) thin-film transistors (TFTs) is compared. The difference in the kinetic energy of D <sup>+</sup> and H <sup>+</sup> ions upon impact with the growing surface during radio-frequency plasma-enhanced chemical vapor deposition leads to material having different physical properties for the same nominal deposition conditions. However, a- Si : D and a- Si : H grown at the same growth rate by adjusting the gas pressure have almost identical properties. By using the growth rate as a normalizing parameter for comparing a- Si : H and a- Si : D TFTs, it is shown that there is no difference in the stability of a- Si : D compared with a- Si : H TFTs. This study rules out the possibility of a giant isotopic effect in amorphous silicon TFTs, and supports the model for Si dangling bond defect creation in a- Si : H where the breaking of weak Si–Si bonds is the rate-limiting step.
Optical, electronic, and structural properties of high‐quality a‐Si:H and a‐Si:D films are compared. While having essentially similar electrical properties, a‐Si:D shows higher stability of the photoconductivity (σ ph ) under illumination, and faster thermal recovery of the degraded σ ph . These differences may originate from differences in the Si—Si and Si—H/D bonding structures, indicated by Raman scattering and infrared measurements. In agreement, single‐junction solar cells with an intrinsic a‐Si:D layer show smaller light‐induced changes than a‐Si:H cells.
We investigate the scanning tunnelling microscopy-induced H and D atom desorption from Si(100)-(2 × 1):H(D). The desorption of both atoms shows the same energy threshold that corresponds well with the computed σ → σ∗ excitation energy of the SiH group. The H desorption yield, however, is much higher than the D yield. We ascribe this to the greater influence of quenching processes on the excited state of the SiD species. We use wavepacket dynamics to follow the motion of H and D atoms, and conclude that desorption occurs, for the most part, from the ‘hot’ ground state populated by the quenching process. Site-selective excitation-induced chemistry is found in the desorption of H from Si(100)-(3 × 1):H.
Hydrogen effusion results are discussed for hydrogenated amorphous silicon (a-Si:H) and related alloys as well as for crystalline silicon (c-Si). It is demonstrated that depending on the microstructure of the material, hydrogen effusion gives information on hydrogen diffusion or surface desorption. The results suggest for compact a-Si:H and for ion implanted c-Si a similar hydrogen diffusion process, which is a trap limited motion of atomic hydrogen. Hydrogen effusion from defect-free c-Si and from void-rich amorphous semiconductors is limited by surface desorption. Both hydrogen diffusion and desorption depend on the Fermi energy if hydrogen bonds to the host material are broken.
We present experimental and theoretical results on the STM-induced SiH bond-breaking on the Si(100)-(2 × 1):H surface. First, we examine the character of the STM-induced excitations. Using density functional theory we show that the strength of chemical bonds and their excitation energies can be decreased or increased depending on the strength and direction of the field. By shifting the excitation energy of an adsorbate below the tip, energy transfer away from this excited site can be suppressed, and localized excited state chemistry can take place. Our experiments show that SiH bonds can be broken when the STM electrons have an energy >6 eV, i.e. above the onset of the σ→σ∗ transition of SiH. The desorption yield is ∼2.4 × 10−6 H-atoms/electron and is independent of the current. We also find that D-atom desorption is much less efficient than H-atom desorption. Using the isotope effect and wavepacket dynamics simulations we deduce that a very fast quenching process, ∼1015 s−1, competes with desorption. Most of the desorbing atoms originate from the “hot” ground state produced by the quenching process. Most interestingly, excitation at energies below the electronic excitation threshold can still lead to H atom desorption, albeit with a much lower yield. The yield in this energy range is a strong function of the tunneling current. We propose that desorption is now the result of the multiple-vibration excitation of the SiH bond. Such excitation becomes possible because of the very high current densities in the STM, and the long SiH stretch vibrational lifetime. The most important aspect of this mechanism is that it allows single atom resolution in the bond-breaking process — the ultimate lithographic resolution.
We present a treatment of the defect-pool model, for the calculation of the density of electronic gap states in hydrogenated amorphous silicon, based on the equilibration of elemental chemical reactions involving the separate release and capture of hydrogen. We derive the corresponding hydrogen density of states, describing the distribution of hydrogen binding energies, and show that the two densities of states are completely consistent. Hydrogen can be captured into weak SiSi bonds, which can be occupied by one or two hydrogen atoms. These are the dominant chemical reactions controlling the defect density. The effective hydrogen correlation energy is variable, being negative for most sites but positive where most defects occur. We show that the electronic density of states reproduces the main features of our earlier defect-pool model, with more charged defects than neutral defects for intrinsic amorphous silicon. The electronic density of states and the corresponding hydrogen density of states are consistent with a wide range of experimental results, including hydrogenation-dehydrogenation and hydrogen diffusion.
Amorphous-silicon thin-film transistors show a threshold voltage shift when subjected to prolonged bias stress. For transistors made with silicon oxide as the gate dielectric, the threshold shift induced under positive bias is due to the creation of dangling-bond states in the a-Si:H at low energy (${\mathit{D}}_{\mathit{e}}$ states). The threshold shift induced by negative bias stress is due to the creation of dangling-bond states at a higher energy (${\mathit{D}}_{\mathit{h}}$ states). In transistors made with silicon nitride as the gate dielectric, positive bias stress causes an increase in the density of ${\mathit{D}}_{\mathit{e}}$ states, but negative bias stress causes mainly a reduction in the density of ${\mathit{D}}_{\mathit{e}}$ states. Positive bias annealing of both oxide and nitride transistors leads to an increase in the density of ${\mathit{D}}_{\mathit{e}}$ states and a reduction in the density of ${\mathit{D}}_{\mathit{h}}$ states. Negative bias annealing leads to a reduction in the density of ${\mathit{D}}_{\mathit{e}}$ states and an increase in the density of ${\mathit{D}}_{\mathit{h}}$ states. The magnitude of each change depends on the initial Fermi-level position, which is the main difference between our oxide and nitride transistors. The results are explained by a defect-pool model for the dangling-bond states in a-Si:H. Dangling bonds are formed by a chemical equilibration process, resulting in the formation of dangling bonds in each of the possible charge states. This leads to a density of states in a-Si:H consisting of coexisting components formed as negatively charged dangling bonds (${\mathit{D}}_{\mathit{e}}$ states), positively charged dangling bonds (${\mathit{D}}_{\mathit{h}}$ states), and neutral dangling bonds (${\mathit{D}}_{0}$ states).
- S Ganguly
- A Yamasaki
- Matsuda
Ganguly, S. Yamasaki, and A. Matsuda, Philos. Mag. B 63, 281
1991.
- D Liu
- A Shih
- S D Chen
- S C Lee
D. Liu, A. Shih, S. D. Chen, and S. C. Lee, J. Vac. Sci. Technol. B 21,
677 2003.
- A Nevin
- H Yamagishi
- K Asaoka
- H Nishio
- Y Tawada
A. Nevin, H. Yamagishi, K. Asaoka, H. Nishio, and Y. Tawada, Appl.
Phys. Lett. 59, 3294 1991.
- J Beyer
- Non-Cryst
Beyer, J. Non-Cryst. Solids 198-200, 40 1996.
23 W. Beyer, Physica B 170, 105 1991.
- J Sugiyama
- S Yang
- Guha
Sugiyama, J. Yang, and S. Guha, Appl. Phys. Lett. 70, 378 1997.
- R Deane
- M Wehrspohn
- M Powell
- W Stutzmann
- C Jackson
- Tsai
Deane, R. Wehrspohn, and M. Powell, Phys. Rev. B 58, 12625 1998.
27 M. Stutzmann, W. Jackson, and C. Tsai, Phys. Rev. B 32, 23 1985.
- S C Shih
- C T Lee
- Chia
Shih, S. C. Lee, and C. T. Chia, Appl. Phys. Lett. 74, 3347 1999.
- J H Wei
- S C Lee
J. H. Wei and S. C. Lee, J. Appl. Phys. 85, 543 1999.
- I French
- S Deane
- D Murley
- J I Hewett
- M Gale
- Powell
I. French, S. Deane, D. Murley, Hewett, J. I. Gale, and M. Powell, Mater.
Res. Soc. Symp. Proc. 467, 875 1997.
- D Staebler
- C Wronski
D. Staebler and C. Wronski, J. Appl. Phys. 51, 3262 1980.
- C Powell
- A Van Berkel
- S Franklin
- W Deane
- Milne
Powell, C. van Berkel, A. Franklin, S. Deane, and W. Milne, Phys.
Rev. B 45, 4160 1992.
- M J Powell
- S C Deane
- R B Wehrspohn
M. J. Powell, S. C. Deane, and R. B. Wehrspohn, Phys. Rev. B 66, 155212
2002.
- R Wehrspohn
- S Deane
- I French
- I Gale
- J Hewett
- M Powell
- J Robertson
20 R. Wehrspohn, S. Deane, I. French, I. Gale, J. Hewett, M. Powell, and J.
Robertson, J. Appl. Phys. 87, 144 2000.