Deposition reactors for solar grade silicon: a comparative thermal analysis of a Siemens reactor and a fluidized bed reactor

Polysilicon production costs contribute approximately to 25-33% of the overall cost of the solar panels and a similar fraction of the total energy invested in their fabrication. Understanding the energy losses and the behaviour of process temperature is an essential requirement as one moves forward to design and build large scale polysilicon manufacturing plants. In this paper we present thermal models for two processes for poly production, viz., the Siemens process using trichlorosilane (TCS) as precursor and the fluid bed process using silane (monosilane, MS).We validate the models with some experimental measurements on prototype laboratory reactors relating the temperature profiles to product quality. A model sensitivity analysis is also performed, and the efects of some key parameters such as reactor wall emissivity, gas distributor temperature, etc., on temperature distribution and product quality are examined. The information presented in this paper is useful for further understanding of the strengths and weaknesses of both deposition technologies, and will help in optimal temperature profiling of these systems aiming at lowering production costs without compromising the solar cell quality.

Understanding and improving the chemical vapor deposition process for solar grade silicon production

Esta Tesis Doctoral se centra en la investigación del proceso de producción de polisilicio para aplicaciones fotovoltaicas (FV) por la vía química; mediante procesos de depósito en fase vapor (CVD). El polisilicio para la industria FV recibe el nombre de silicio de grado solar (SoG Si). Por un lado, el proceso que domina hoy en día la producción de SoG Si está basado en la síntesis, destilación y descomposición de triclorosilano (TCS) en un reactor CVD -denominado reactor Siemens-. El material obtenido mediante este proceso es de muy alta pureza, pero a costa de un elevado consumo energético. Así, para alcanzar los dos principales objetivos de la industria FV basada en silicio, bajos costes de producción y bajo tiempo de retorno de la energía invertida en su fabricación, es esencial disminuir el consumo energético de los reactores Siemens. Por otro lado, una alternativa al proceso Siemens considera la descomposición de monosilano (MS) en un reactor de lecho fluidizado (FBR). Este proceso alternativo tiene un consumo energético mucho menor que el de un reactor Siemens, si bien la calidad del material resultante es también menor; pero ésta puede ser suficiente para la industria FV. A día de hoy los FBR deben aún abordar una serie de retos para que su menor consumo energético sea una ventaja suficiente comparada con otras desventajas de estos reactores. En resumen, la investigación desarrollada se centra en el proceso de depósito de polysilicio por CVD a partir de TCS -reactor Siemens-; pero también se investiga el proceso de producción de SoG Si en los FBR exponiendo las fortalezas y debilidades de esta alternativa. Para poder profundizar en el conocimiento del proceso CVD para la producción de polisilicio es clave el conocimiento de las reacciones químicas fundamentales y cómo éstas influencian la calidad del producto resultante, al mismo tiempo que comprender los fenómenos responsables del consumo energético. Por medio de un reactor Siemens de laboratorio en el que se llevan a cabo un elevado número de experimentos de depósito de polisilicio de forma satisfactoria se adquiere el conocimiento previamente descrito. Se pone de manifiesto la complejidad de los reactores CVD y de los problemas asociados a la pérdidas de calor de estos procesos. Se identifican las contribuciones a las pérdidas de calor de los reactores CVD, éstas pérdidas de calor son debidas principalmente a los fenómenos de radiación y, conducción y convección vía gases. En el caso de los reactores Siemens el fenómeno que contribuye en mayor medida al alto consumo energético son las pérdidas de calor por radiación, mientras que en los FBRs tanto la radiación como el calor transferido por transporte másico contribuyen de forma importante. Se desarrolla un modelo teórico integral para el cálculo de las pérdidas de calor en reactores Siemens. Este modelo está formado a su vez por un modelo para la evaluación de las pérdidas de calor por radiación y modelos para la evaluación de las pérdidas de calor por conducción y convección vía gases. Se ponen de manifiesto una serie de limitaciones del modelo de pérdidas de calor por radiación, y se desarrollan una serie de modificaciones que mejoran el modelo previo. El modelo integral se valida por medio un reactor Siemens de laboratorio, y una vez validado se presenta su extrapolación a la escala industrial. El proceso de conversión de TCS y MS a polisilicio se investiga mediante modelos de fluidodinámica computacional (CFD). Se desarrollan modelados CFD para un reactor Siemens de laboratorio y para un prototipo FBR. Los resultados obtenidos mediante simulación son comparados, en ambos casos, con resultados experimentales. Los modelos desarrollados se convierten en herramientas para la identificación de aquellos parámetros que tienen mayor influencia en los procesos CVD. En el caso del reactor Siemens, ambos modelos -el modelo integral y el modelado CFD permiten el estudio de los parámetros que afectan en mayor medida al elevado consumo energético, y mediante su análisis se sugieren modificaciones para este tipo de reactores que se traducirían en un menor número de kilovatios-hora consumidos por kilogramo de silicio producido. Para el caso del FBR, el modelado CFD permite analizar el efecto de una serie de parámetros sobre la distribución de temperaturas en el lecho fluidizado; y dicha distribución de temperaturas está directamente relacionada con los principales retos de este tipo de reactores. Por último, existen nuevos conceptos de depósito de polisilicio; éstos se aprovechan de la ventaja teórica de un mayor volumen depositado por unidad de tiempo -cuando una mayor superficie de depósito está disponible- con el objetivo de reducir la energía consumida por los reactores Siemens. Estos conceptos se exploran mediante cálculos teóricos y pruebas en el reactor Siemens de laboratorio. ABSTRACT This Doctoral Thesis comprises research on polysilicon production for photovoltaic (PV) applications through the chemical route: chemical vapor deposition (CVD) process. PV polysilicon is named solar grade silicon (SoG Si). On the one hand, the besetting CVD process for SoG Si production is based on the synthesis, distillation, and decomposition of thriclorosilane (TCS) in the so called Siemens reactor; high purity silicon is obtained at the expense of high energy consumption. Thus, lowering the energy consumption of the Siemens process is essential to achieve the two wider objectives for silicon-based PV technology: low production cost and low energy payback time. On the other hand, a valuable variation of this process considers the use of monosilane (MS) in a fluidized bed reactor (FBR); lower output material quality is obtained but it may fulfil the requirements for the PV industry. FBRs demand lower energy consumption than Siemens reactors but further research is necessary to address the actual challenges of these reactors. In short, this work is centered in polysilicon CVD process from TCS -Siemens reactor-; but it also offers insights on the strengths and weaknesses of the FBR for SoG Si production. In order to aid further development in polysilicon CVD is key the understanding of the fundamental reactions and how they influence the product quality, at the same time as to comprehend the phenomena responsible for the energy consumption. Experiments conducted in a laboratory Siemens reactor prove the satisfactory operation of the prototype reactor, and allow to acquire the knowledge that has been described. Complexity of the CVD reactors is stated and the heat loss problem associated with polysilicon CVD is addressed. All contributions to the energy consumption of Siemens reactors and FBRs are put forward; these phenomena are radiation and, conduction and convection via gases heat loss. In a Siemens reactor the major contributor to the energy consumption is radiation heat loss; in case of FBRs radiation and heat transfer due to mass transport are both important contributors. Theoretical models for radiation, conduction and convection heat loss in a Siemens reactor are developed; shaping a comprehensive theoretical model for heat loss in Siemens reactors. Limitations of the radiation heat loss model are put forward, and a novel contribution to the existing model is developed. The comprehensive model for heat loss is validated through a laboratory Siemens reactor, and results are scaled to industrial reactors. The process of conversion of TCS and MS gases to solid polysilicon is investigated by means of computational fluid-dynamics models. CFD models for a laboratory Siemens reactor and a FBR prototype are developed. Simulated results for both CVD prototypes are compared with experimental data. The developed models are used as a tool to investigate the parameters that more strongly influence both processes. For the Siemens reactors, both, the comprehensive theoretical model and the CFD model allow to identify the parameters responsible for the great power consumption, and thus, suggest some modifications that could decrease the ratio kilowatts-hour per kilogram of silicon produced. For the FBR, the CFD model allows to explore the effect of a number of parameters on the thermal distribution of the fluidized bed; that is the main actual challenge of these type of reactors. Finally, there exist new deposition surface concepts that take advantage of higher volume deposited per time unit -when higher deposition area is available- trying to reduce the high energy consumption of the Siemens reactors. These novel concepts are explored by means of theoretical calculations and tests in the laboratory Siemens prototype.

Deposition-related sites K5/K12 in histone H4 are not required for nucleosome deposition in yeast

Histone H4 can be acetylated at N-terminal lysines K5, K8, K12, and K16, but newly synthesized H4 is diacetylated at K5/K12 in diverse organisms. This pattern is widely thought to be important for histone deposition onto replicating DNA. To investigate the importance of K5/K12 we have mutagenized these lysines in yeast and assayed for nucleosome assembly. Assaying was done in the absence of the histone H3 N terminus, which has functions redundant with those of H4 in histone deposition. Nucleosome assembly was assayed by three methods. Because nucleosome depletion may be lethal, we examined cell viability. We also analyzed nucleosome assembly in vivo and in vitro by examining plasmid superhelicity density in whole cells and supercoiling in yeast cell extracts. All three approaches demonstrate that mutagenizing K5 and K12 together does not prevent cell growth and histone deposition in vivo or in vitro. Therefore, K5/K12 cannot be required for nucleosome assembly in yeast. It is only when the first three sites of acetylation—K5, K8, and K12—are mutagenized simultaneously that lethality occurs and assembly is most strongly decreased both in vivo and in vitro. These data argue for the redundancy of sites K5, K8, and K12 in the deposition of yeast histone H4.

Neuronal overexpression of mutant amyloid precursor protein results in prominent deposition of cerebrovascular amyloid

Transgenic mice that overexpress mutant human amyloid precursor protein (APP) exhibit one hallmark of Alzheimer’s disease pathology, namely the extracellular deposition of amyloid plaques. Here, we describe significant deposition of amyloid β (Aβ) in the cerebral vasculature [cerebral amyloid angiopathy (CAA)] in aging APP23 mice that had striking similarities to that observed in human aging and Alzheimer’s disease. Amyloid deposition occurred preferentially in arterioles and capillaries and within individual vessels showed a wide heterogeneity (ranging from a thin ring of amyloid in the vessel wall to large plaque-like extrusions into the neuropil). CAA was associated with local neuron loss, synaptic abnormalities, microglial activation, and microhemorrhage. Although several factors may contribute to CAA in humans, the neuronal origin of transgenic APP, high levels of Aβ in cerebrospinal fluid, and regional localization of CAA in APP23 mice suggest transport and drainage pathways rather than local production or blood uptake of Aβ as a primary mechanism underlying cerebrovascular amyloid formation. APP23 mice on an App-null background developed a similar degree of both plaques and CAA, providing further evidence that a neuronal source of APP/Aβ is sufficient to induce cerebrovascular amyloid and associated neurodegeneration.

Apolipoprotein E is essential for amyloid deposition in the APPV717F transgenic mouse model of Alzheimer's disease

We quantified the amount of amyloid β-peptide (Aβ) immunoreactivity as well as amyloid deposits in a large cohort of transgenic mice overexpressing the V717F human amyloid precursor protein (APPV717F+/− TG mice) with no, one, or two mouse apolipoprotein E (Apoe) alleles at various ages. Remarkably, no amyloid deposits were found in any brain region of APPV717F+/− Apoe−/− TG mice as old as 22 mo of age, whereas age-matched APPV717F +/− Apoe+/− and Apoe+/+ TG mice display abundant amyloid deposition. The amount of Aβ immunoreactivity in the hippocampus was also markedly reduced in an Apoe gene dose-dependent manner (Apoe+/+ > Apoe+/− ≫ Apoe−/−), and no Aβ immunoreactivity was detected in the cerebral cortex of APPV717F+/− Apoe−/− TG mice at any of the time points examined. The absence of apolipoprotein E protein (apoE) dramatically reduced the amount of both Aβ1–40 and Aβ1–42 immunoreactive deposits as well as the resulting astrogliosis and microgliosis normally observed in APPV717F TG mice. ApoE immunoreactivity was detected in a subset of Aβ immunoreactive deposits and in virtually all thioflavine-S-fluorescent amyloid deposits. Because the absence of apoE alters neither the transcription or translation of the APPV717F transgene nor its processing to Aβ peptide(s), we postulate that apoE promotes both the deposition and fibrillization of Aβ, ultimately affecting clearance of protease-resistant Aβ/apoE aggregates. ApoE appears to play an essential role in amyloid deposition in brain, one of the neuropathological hallmarks of Alzheimer's disease.

Proline Accumulation in Maize (Zea mays L.) Primary Roots at Low Water Potentials. II. Metabolic Source of Increased Proline Deposition in the Elongation Zone1

The proline (Pro) concentration increases greatly in the growing region of maize (Zea mays L.) primary roots at low water potentials (ψw), largely as a result of an increased net rate of Pro deposition. Labeled glutamate (Glu), ornithine (Orn), or Pro was supplied specifically to the root tip of intact seedlings in solution culture at high and low ψw to assess the relative importance of Pro synthesis, catabolism, utilization, and transport in root-tip Pro deposition. Labeling with [3H]Glu indicated that Pro synthesis from Glu did not increase substantially at low ψw and accounted for only a small fraction of the Pro deposition. Labeling with [14C]Orn showed that Pro synthesis from Orn also could not be a substantial contributor to Pro deposition. Labeling with [3H]Pro indicated that neither Pro catabolism nor utilization in the root tip was decreased at low ψw. Pro catabolism occurred at least as rapidly as Pro synthesis from Glu. There was, however, an increase in Pro uptake at low ψw, which suggests increased Pro transport. Taken together, the data indicate that increased transport of Pro to the root tip serves as the source of low-ψw-induced Pro accumulation. The possible significance of Pro catabolism in sustaining root growth at low ψw is also discussed.

Callose Deposition Is Responsible for Apoplastic Semipermeability of the Endosperm Envelope of Muskmelon Seeds1

Semipermeable cell walls or apoplastic “membranes” have been hypothesized to be present in various plant tissues. Although often associated with suberized or lignified walls, the wall component that confers osmotic semipermeability is not known. In muskmelon (Cucumis melo L.) seeds, a thin, membranous endosperm completely encloses the embryo, creating a semipermeable apoplastic envelope. When dead muskmelon seeds are allowed to imbibe, solutes leaking from the embryo are retained within the envelope, resulting in osmotic water uptake and swelling called osmotic distention (OD). The endosperm envelope of muskmelon seeds stained with aniline blue, which is specific for callose (β-1,3-glucan). Outside of the aniline-blue-stained layer was a Sudan III- and IV-staining (lipid-containing) layer. In young developing seeds 25 d after anthesis (DAA) that did not exhibit OD, the lipid layer was already present but callose had not been deposited. At 35 DAA, callose was detected as distinct vesicles or globules in the endosperm envelope. A thick callose layer was evident at 40 DAA, coinciding with development of the capacity for OD. Removal of the outer lipid layer by brief chloroform treatment resulted in more rapid water uptake by both viable and nonviable (boiled) seeds, but did not affect semipermeability of the endosperm envelope. The aniline-blue-staining layer was digested by β-1,3-glucanase, and these envelopes lost OD. Thus, apoplastic semipermeability of the muskmelon endosperm envelope is dependent on the deposition of a thick callose-containing layer outside of the endosperm cell walls.

Induced β-Carotene Synthesis Driven by Triacylglycerol Deposition in the Unicellular Alga Dunaliella bardawil1

Under stress conditions such as high light intensity or nutrient starvation, cells of the unicellular alga Dunaliella bardawil overproduce β-carotene, which is accumulated in the plastids in newly formed triacylglycerol droplets. We report here that the formation of these sequestering structures and β-carotene are interdependent. When the synthesis of triacylglycerol is blocked, the overproduction of β-carotene is also inhibited. During overproduction of β-carotene no up-regulation of phytoene synthase or phytoene desaturase is observed on the transcriptional or translational level, whereas at the same time acetyl-CoA carboxylase, the key regulatory enzyme of acyl lipid biosynthesis, is increased, at least in its enzymatic activity. We conclude that under normal conditions the carotenogenic pathway is not maximally active and may be appreciably stimulated in the presence of sequestering structures, creating a plastid-localized sink for the end product of the carotenoid biosynthetic pathway.

Matrilysin is expressed by lipid-laden macrophages at sites of potential rupture in atherosclerotic lesions and localizes to areas of versican deposition, a proteoglycan substrate for the enzyme.

Certain matrix metalloproteinases (MMP) are expressed within the fibrous areas surrounding acellular lipid cores of atherosclerotic plaques, suggesting that these proteinases degrade matrix proteins within these areas and weaken the structural integrity of the lesion. We report that matrilysin and macrophage metalloelastase, two broad-acting MMPs, were expressed in human atherosclerotic lesions in carotid endarterectomy samples (n = 18) but were not expressed in normal arteries (n = 7). In situ hybridization and immunohistochemistry revealed prominent expression of matrilysin in cells confined to the border between acellular lipid cores and overlying fibrous areas, a distribution distinct from other MMPs found in similar lesions. Metalloelastase was expressed in these same border areas. Matrilysin was present in lipid-laden macrophages, identified by staining with anti-CD-68 antibody. Furthermore, endarterectomy tissue in organ culture released matrilysin. Staining for versican demonstrated that this vascular proteoglycan was present at sites of matrilysin expression. Biochemical studies showed that matrilysin degraded versican much more efficiently than other MMPs present in atherosclerotic lesions. Our findings suggest that matrilysin, specifically expressed in atherosclerotic lesions, could cleave structural proteoglycans and other matrix components, potentially leading to separation of caps and shoulders from lipid cores.

An extended Herschel drop-out source in the center of AS1063: a normal dusty galaxy at z=6.1 or SZ substructures?

In the course of our 870 μm APEX/LABOCA follow-up of the Herschel Lensing Survey we have detected a source in AS1063 (RXC J2248.7-4431) that has no counterparts in any of the Herschel PACS/SPIRE bands, it is a Herschel “drop-out” with S_870/S_500 ≥ 0.5. The 870 μm emission is extended and centered on the brightest cluster galaxy, suggesting either a multiply imaged background source or substructure in the Sunyaev-Zel’dovich increment due to inhomogeneities in the hot cluster gas of this merging cluster. We discuss both interpretations with emphasis on the putative lensed source. Based on the observed properties and on our lens model we find that this source may be the first submillimeter galaxy (SMG) with a moderate far-infrared (FIR) luminosity (L_FIR < 10^12 L_⊙) detected so far at z > 4. In deep HST observations we identified a multiply imaged z ~ 6 source and measured its spectroscopic redshift to be z = 6.107 with VLT/FORS. This source may be associated with the putative SMG, but it is most likely offset spatially by 10−30 kpc and they may be interacting galaxies. With a FIR luminosity in the range [5−15] × 10^11 L_⊙ corresponding to a star formation rate in the range [80−260] M_⊙ yr^-1, this SMG would be more representative of the z > 4 dusty galaxies than the extreme starbursts detected so far. With a total magnification of ~25 it would open a unique window to the normal dusty galaxies at the end of the epoch of reionization.

Electrocatalytic Performance of SiO2-SWCNT Nanocomposites Prepared by Electroassisted Deposition

Composite materials made of porous SiO2 matrices filled with single-walled carbon nanotubes (SWCNTs) were deposited on electrodes by an electroassisted deposition method. The synthesized materials were characterized by several techniques, showing that porous silica prevents the aggregation of SWCNT on the electrodes, as could be observed by transmission electron microscopy and Raman spectroscopy. Different redox probes were employed to test their electrochemical sensing properties. The silica layer allows the permeation of the redox probes to the electrode surface and improves the electrochemical reversibility indicating an electrocatalytic effect by the incorporation of dispersed SWCNT into the silica films.

Molecularly imprinted silica films prepared by electroassisted deposition for the selective detection of dopamine

Dopamine (DA) can be detected by electrochemical oxidation in conventional electrodes. However, the presence of other oxidizable species (interferents) usually present in physiological fluids at high concentrations (like ascorbic acid) makes very difficult its electrochemical detection. In the present work, glassy carbon electrodes have been modified with molecularly imprinted silica (MIS) films prepared by electroassisted deposition of sol–gel precursors. The production of MIS films was performed by adding the template molecule (DA) to the precursor sol. The molecular impression of silica was assessed showing a high coherency allowing a filtering capacity in the molecular scale. The MIS-modified electrodes present a high selectivity for the detection of DA in neutral or acidic solutions. The MIS-modified electrodes allow the amperometric determination of dopamine in solutions containing ascorbic acid with molar ratios lower than 1:50,000.

Electrochemical Behaviour of PSS-Functionalized Silica Films Prepared by Electroassisted Deposition of Sol–Gel Precursors

Porous, electrically insulating SiO2 layers containing polystyrene sulfonate (PSS) were deposited on glassy carbon electrodes by an electrochemically assisted deposition method. The obtained material was characterized by microscopic, spectroscopic and thermal techniques. Silica-PSS films modify the electrochemical response of the glassy carbon electrodes against selected redox probes. Positively charged species show reduced diffusivities across the SiO2-PSS pores, which resulted in a concentration ratio higher than 1 for these species. The opposite behaviour was found for negatively charged redox probes. These observations can be interpreted in terms of the different affinity of the GC/SiO2-PSS-modified electrode for the electroactive species, as a consequence of the negatively charged porous silica.

Simulation of the secondary electrons energy deposition produced by proton beams in PMMA: influence of the target electronic excitation description

We have studied the radial dependence of the energy deposition of the secondary electron generated by swift proton beams incident with energies T = 50 keV–5 MeV on poly(methylmethacrylate) (PMMA). Two different approaches have been used to model the electronic excitation spectrum of PMMA through its energy loss function (ELF), namely the extended-Drude ELF and the Mermin ELF. The singly differential cross section and the total cross section for ionization, as well as the average energy of the generated secondary electrons, show sizeable differences at T ⩽ 0.1 MeV when evaluated with these two ELF models. In order to know the radial distribution around the proton track of the energy deposited by the cascade of secondary electrons, a simulation has been performed that follows the motion of the electrons through the target taking into account both the inelastic interactions (via electronic ionizations and excitations as well as electron-phonon and electron trapping by polaron creation) and the elastic interactions. The radial distribution of the energy deposited by the secondary electrons around the proton track shows notable differences between the simulations performed with the extended-Drude ELF or the Mermin ELF, being the former more spread out (and, therefore, less peaked) than the latter. The highest intensity and sharpness of the deposited energy distributions takes place for proton beams incident with T ~ 0.1–1 MeV. We have also studied the influence in the radial distribution of deposited energy of using a full energy distribution of secondary electrons generated by proton impact or using a single value (namely, the average value of the distribution); our results show that differences between both simulations become important for proton energies larger than ~0.1 MeV. The results presented in this work have potential applications in materials science, as well as hadron therapy (due to the use of PMMA as a tissue phantom) in order to properly consider the generation of electrons by proton beams and their subsequent transport and energy deposition through the target in nanometric scales.

Deposition of Abraham Hill against Nathan Prince, Nov. 1, 1740

Abraham Hill (A.B. 1737) claimed that Prince had come to his College chamber "smoaking a pipe of Tobacco" the previous summer and asked numerous unusual questions. Hill also testified that Prince had accused fellow Tutor Daniel Rogers of being someone who "never did know what a scholar was" and Tutor Henry Flynt of having been "superannuated long ago." This deposition was attested by Justice of the Peace Trowbridge.