Semiconductors and semimetals, volume 92

– High deposition rates due to a large silicon surface; different to the mens process where only toward the end of the process a large deposition... Furthermore, most ofthe side slabs of

SERIES EDITORS

EICKE R WEBER

Director

Fraunhofer-Institut

f€ur Solare Energiesysteme ISE

Vorsitzender, Fraunhofer-Allianz Energie Heidenhofstr 2, 79110

Freiburg, Germany

CHENNUPATI JAGADISHAustralian Laureate Fellow

and Distinguished Professor

525 B Street, Suite 1800, San Diego, CA 92101-4495, USA

125 London Wall, London, EC2Y 5AS, UK

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

First edition 2015

© 2015 Elsevier Inc All rights reserved

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing As new research and experience broaden our understanding, changes in research methods, professional practices,

or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-801021-1

ISSN: 0080-8784

For information on all Academic Press publications

visit our website at store.elsevier.com

Peter Dold

Fraunhofer CSP, Halle, Germany (ch1)

Hans Joachim M€ oller

Fraunhofer Technology Center for Semiconductor Materials, Freiberg, Germany (ch2) Thomas Walter

Faculty of Mechatronics and Medical Engineering, University of Applied Sciences Ulm, Ulm, Germany (ch3)

vii

The rapid transformation of our energy supply system to the more efficientuse of increasingly renewable energies is one of the biggest challenges andopportunities of the present century Harvesting solar energy by photovol-taics is considered to be a cornerstone technology for this truly global trans-formation process, and it is well on its way The speed of progress isillustrated by looking at some figures of the cumulative installed PV peakpower capacity In Part 1 of this series of “Advances of Photovoltaics,”published in 2012, the introduction mentioned 70 GWp installed at theend of 2011 As we write this preface of Part 4 in the spring of 2015, 1%

of the world electricity generation is now already supplied by PV, and inthe coming months the global PV installation figure will have tripledcompared with 2011! But this is just the beginning of the thousands of

GWpthat are likely to be installed in the decades to come

Key for this extraordinary development was the rapid decrease of PVprices and thus the cost of solar electricity This was fueled by a rapidtechnology development with soaring efficiencies at reduced productioncost, coupled with an effective market introduction policy, especially thewell-designed German feed-in tariff Today, we can harvest solar electricityeven in Germany—with insolation comparable to Alaska!—for about

10$ct/kWh, and in sun-rich areas for half of this amount, far below the cost,e.g., electricity obtained from Diesel generators

As already mentioned above, this book presents the fourth volume in theongoing series “Advances in Photovoltaics” within Semiconductors andSemimetals This series has been designed to provide a thorough overview

of the underlying physics, the important materials aspects, the prevailing andfuture solar cell design issues, production technologies, as well as energy sys-tem integration and characterization issues The present volume deals withthree important issues, of crystallizing silicon, the dominating PV material,the ways of how to transform it into wafers for solar cells, as well as the issue

of reliability of CIGS-based thin film solar cells and modules Following thetradition of this series, all chapters are written by world-leading experts intheir respective field

As we write this text, the German PV market is likely to collapse from a7.5 GWp/a market as recently as 2012 to a 1 GWp/a level in 2015, a marketsize that we last had in 2007 Fortunately, other markets in China, Japan, and

ix

the USA are now taking over by currently developing into 10 GWpper yearand more markets.

The solar PV revolution has started irreversibly, it is now fueled byeconomics in addition to the concern for reducing climate gas emissions,and it takes rapid foothold beyond Europe in Asia and the Americas, theother parts of our planet will follow in a few year’s time!

GERHARDP WILLEKE ANDEICKER WEBER

Fraunhofer ISE, Freiburg, Germany

Silicon Crystallization

Technologies

Peter Dold1

Fraunhofer CSP, Halle, Germany

1 Corresponding author: e-mail address: peter.dold@ise.fraunhofer.de

Contents

1.1 Polysilicon: The Base Material for over 90% of All Solar Cells 1

1.5 Different Poly for Different Crystallization Techniques 11

2.1 Material Properties, Material Utilization, and Chemical Reactivity 12

3.2 Directional Solidification: Growth of Multicrystalline Silicon 36

in the years 2007–2010, when companies could make billions of dollars ifthey were able to deliver polysilicon at all, was followed by the severe crush

in the years 2011–2012, when most of the newcomers marched into ruptcy and disappeared And, even some of the old ones had to fight heavily

bank-Semiconductors and Semimetals, Volume 92 # 2015 Elsevier Inc.

ISSN 0080-8784 All rights reserved 1

to survive During the golden years, spot market prices had reached highs of200–300 or even 400 US$/kg polysilicon, simply because the market wasswept and the order books of the cell and module manufacturers were full.The polysilicon industry was not prepared for such a fast ramp-up, invest-ment is high,1 and equipment could not readily be ordered The long-established companies either have an exclusive partnership with a specificequipment manufacturer, or they make the equipment in-house Produc-tion capacity could not easily be ramped up, but once the train was running,

it also could not be stopped so easily and could not be adjusted to the thenchanged market situation, partly because typical polysilicon projects takeseveral years from the financing phase all the way up to full production,and partly because the players did not want to believe that the siliconbonanza was over The huge shortage was followed by a tremendous oversupply with spot market prices as low as 14–16 US$/kg in 2013—which wasbelow the actual production costs Today, spot market prices leveled offaround 17–18 US$/kg and no significant changes are expected for the nearfuture

As a consequence, all (or at least as good as all) of the new and innovativeapproaches for polysilicon refinement, for upgrading metallurgical silicon(an excellent review was given byHeuer, 2013), or for alternative produc-tion methods (compare Bernreuter and Haugwitz, 2010) could not find amarket share and disappeared again The traditional Chemical Vapor Depo-sition (CVD)-based Siemens process (Fabry and Hesse, 2012), probably notthe most sophisticated technology for solar-grade-silicon production—butfor sure the most matured technique, was the match winner A good over-view of the market situation and an in-depth analysis of the trends are given

by Bernreuter every first or second year (Bernreuter, 2014)

Basically, two main routes might be distinguished for the refinement ofpolysilicon: (I) the chemical path: bringing silicon into the gas phase andpurifying it by distillation, followed by thermal pyrolysis of the gaseous spe-cies; and (II) the metallurgical path, where impurities are removed from sil-icon by mixing it with another metal or with a slag, then let the impuritiessegregate into the second phase, separate the different phases somehowmechanically, and clean the surface of the silicon crystallites by chemicaletching

1

Back in 2008, a polysilicon plant with a capacity of 10,000 t/a required an investment of at least 1 lion US $ Today, it might be something in the range of 400–600 M$, depending on the location.

bil-1.2 The Chemical Path

The Siemens process (or modified Siemens process, as many manufacturerslike to call their variation) allows to produce ultrapure polysilicon, withmetallic bulk impurity levels as low as a few tens of ppt (parts per trillion)

or an equivalent of 10–11N Electrically active elements (donors, acceptors)are in the ppt range and only carbon and oxygen show up in higher concen-trations, where lower single-digit parts per million levels are found Forsemiconductor applications, there is no alternative so far to the polysiliconproduced by the Siemens process

The Siemens process itself goes back to a patent in the late 1950s filed bythe German electronics company Siemens (Reuschel, 1963; Schweickert

et al., 1961), which stepped out of the polysilicon business long ago Itcan be described by the following process steps:

I Milling of the metallurgical silicon (purity: 98–99%) into millimeter/submillimeter particles

II Reaction between the fine silicon particles and gaseous HCl at atures around 300–350°C in a fluidized-bed reactor (FBR) The reac-tor might be heated from the outside, but the chemical reaction is alsostrongly exothermic Mainly copper is used as a catalyst The mainproduct is TCS (trichlorosilane, SiHCl3).

temper-III Fractional distillation of the TCS and the by-products, like metal rides, boron, and phosphorus components, and so on The result will

chlo-be ultrapure TCS

IV Pyrolytic decomposition of TCS in a bell-jar reactor (Fig 1) at increasedpressure (normally 6 bar) and temperatures of 1000–1150°C (Fig 2).High-purity polysilicon will be obtained (Fig 3)

Steps I–III are relatively straightforward, although the installation of thehardware reaches easily the size and complexity of a huge chemical plantfor typical production capacities of around 10,000 t/a Step IV is moredifficult:

– The high temperature required for the silicon deposition is rather energyintensive The silicon rods on which the deposition takes place aredirectly heated by an electrical current

– Deposition rates on these U-shaped rods are on the order of 0.5–1 mm/h(layer growth); beyond this rate, the rod morphology becomes unstableand so-called “popcorn” or “broccoli” growth takes place

– Only part of the TCS decomposes to silicon, and a significant part reactswith the HCl formed during the deposition to STC (silicon tetrachloride,

SiCl4) Decomposition of STC is too low at the typical rod temperatures

in the bell-jar; therefore, it has to be removed from the reactor and has to

be back-converted into TCS

In former times, back-conversion of STC to TCS was carried out mainly

in thermal STC converters (Paetzold et al., 2007; Sirtl et al., 1974), and theprocess is also referred as “hydrogenation.” At high temperature in a hotcarbon rod reactor (>1200 °C), STC reacts with hydrogen back to TCS(and other by-products), an another energy-intensive process step Nowadays,

Figure 1 Schematic drawing of a Siemens bell-jar reactor for polysilicon deposition from the gas phase The U-shaped silicon rods are heated up to a temperature of 1000–1150 °C by direct current The process gas enters and leaves the reactor chamber through the base plate By courtesy of Wacker Chemie AG.

Figure 2 Silicon deposition from TCS in a research reactor Left: beginning of the sition, right: after 30 h process time In particular, in the elbow area, current and tem- perature distribution might be nonuniform.

depo-“hydrochlorination” is more and more used (see, e.g.,http://www.gtat.com/products-and-services-trichlorosilane-and-silane-production-packages-HydrochlorinationTCS-Plant.htm), especially by the newcomers In thisprocess, hydrogen, metallurgical grade silicon, and STC are introduced into

an FBR At high pressure (20 bar and more) and temperature T>500 °C,TCS is formed

The Siemens process is a batch process The U-shaped rods in the bell-jarare heated with high current, starting with 6–8 mm starter rods (or slimrods) Today, most of the slim rods are prepared in so-called slim rod pullers

by the pedestal method: The top area of a cylindrical silicon rod of some4–600 in diameter is melted from above by an RF inductor with at leastone hole in center Through this hole, the slim rod is pulled, comparable

to a crucible-free Czochralski (Cz) approach In such a way, slim rods ofseveral meters are pulled, with pulling rates which might easily surpass half

a meter per hour

At the beginning of the deposition process, just a few tens of amperesare needed to keep the thin starter rods at deposition temperature

A certain challenge is to bring the starter rods to temperatures where theintrinsic carrier concentration of silicon becomes high enough that a decentcurrent can flow To bridge the gap from room temperature to the required300–400°C, where the rods become electrically conductive, variousmethods are in use: (I) preheating the starter rods with radiation lamps,(II) use of medium- or high-voltage power supplies (see, e.g., http://www.aegps.com/en/res/power-controllers/polysilicon-systems/), or (III)use of slightly predoped starter rods (Aulich and Schulze, 2009) The latterFigure 3 Polysilicon rods in an industrial multirod Siemens reactor The rod length might reach more than 3 m, at a maximum diameter of around 180 mm.

is not an option for electronic grade material, but quite an option for grade polysilicon At the end of the process cycle, when the rods have reachedtheir final size of 150–180 mm in diameter, several thousand amperes arerequired to keep them at the specific deposition temperature The wholecycle takes about 100 h, depending on the deposition rates and the final size.The maximum diameter is limited by the temperature gradient between therod surface (which has to stay around 1100°C and which cools down by radi-ation and by convection) and the hotter core of the rod, where the currentflows preferentially If the core or the elbow areas become too hot, there is arisk that the silicon is melting, which results in a strong decrease of the elec-trical resistivity, and finally a local shortcut and a burned-through rod.Some 10 years ago, with lower deposition rates, smaller reactors, andless-optimized processes, power consumption to produce 1 kg of siliconwas in the range of 150–200 kWh/kg (including STC conversion) Today,state-of-the-art reactors with some 48–72 rods (even 96 rod reactors are onthe market), and an annual capacity of some 400 t of silicon, high depositionrates, integrated hydrochlorination, and proper debottlenecking, the powerconsumption is as low as 50–70 kWh/kg Some manufacturers are claimingthat they can even reach values below 50 kWh/kg.

solar-As already mentioned in the beginning, the Siemens process is now verymatured, which also means that we cannot expect huge progress steps any-more, and further improvements will be rather incremental and less revolu-tionary A significant cost reduction is promised by the FBR technology

1.3 Fluidized Bed Reactor

In contrast to the batch-type Siemens process, the FBR operates in a tinuous mode Small seed particles (high-purity silicon with diameters ofsome tens of micrometers) are fed into a heated reactor, a strong gas flow(either TCS or silane, mixed with hydrogen) from the bottom part of thereactor keeps the particles floating (Fig 4) An excellent overview was given

con-inYdstie and Du (2011) Reaction with TCS (or silane as used in the case ofthe company REC) lets the silicon particles grow, until they reach a criticalmass and sink to the bottom area in the form of granules (or beads;Fig 5),where they can be harvested easily The technology has a certain charm andseveral advantages are obvious:

– Continuous operation—minimized downtime

– High deposition rates due to a large silicon surface; different to the mens process where only toward the end of the process a large deposition

Sie-Figure 4 Sketch of an FBR reactor: seeds entering the chamber from the top are tated by the strong gas stream and settle down once they have reached a certain weight At the bottom, the final granules are taken out of the process continuously.

levi-Figure 5 Solar-grade silicon: poly chunks (left-hand side) and granular material hand side).

(right-surface is available, in the case of FBR it is provided right from thebeginning.2

– Significantly lower energy consumption, e.g., REC claims some 80–90%less energy consumption for their silane-based FBR process compared toTCS-based Siemens reactors (http://www.recsilicon.com/technology/rec-silicons-fluidized-bed-reactor-process/)

– The spherical silicon beads are ready to be shipped (and filled into thecrystallization crucible right away), and no crushing or mechanical han-dling is required

Of course, there are some obstacles to manage and one of the biggest is thepurity The moving particles in the reactor might touch the reactor wallwhere they might be contaminated, especially when steel-based/metal-based wall materials are in use Today, granular silicon is about two to threeorders higher in metals than high-class Siemens silicon Further, the swirlingand spinning in the reactor and the subsequent material handling producesome fines in the form of a black dust, which should be removed or washedoff; otherwise, the acceptance of the material suffers Finally, a major prob-lem is the melting of granular silicon in the subsequent crystallization pro-cess: it has a tendency for popping and splashing, and small silicon dropletsmight be found several centimeters away from the crucible Most likely, this

is related to process gases (hydrogen and/or chlorines) stored in the granules(Kajimoto et al., 1991) or it is related to stress at the interface seed shell Dur-ing crystal growth, evaporation of hydrogen might lead to a disturbed meltsurface during the Cz process Release of chlorine is affecting the crystalli-zation hardware, of course The popping problem might be reduced byproper charging of the crucible, blending the granular material with normalpolysilicon chunks, and avoiding that the granules are exposed to the freecrucible surface In the case of recharge processes, the splashing problem

is more difficult to overcome

Recently, quite some R&D activities are noticeable on FBR ogy For sure, it will not push the Siemens process out of the market, but

technol-it might gain a certain share of the poly market According to the 2014ITRPV report, today, granular silicon has a market share of some 15%(http://www.itrpv.net/Reports/Downloads/) It has still a significant costsavings potential, probably much more than the Siemens process Combined

2

Just 1 kg of granules provides a reactive surface of about two-and-a-half square meters, assuming an average diameter of 1 mm On the other hand, a full-size Siemens U-rod of 150 mm in diameter and a total length of 6.5 m possess a surface of about 3 m2at a weight of 280 kg.

with broken poly chunks from Siemens reactors, an improved crucible fillfactor is achieved, an improvement of 29.3% was reported (REC SiliconInc., 2013), the small granules fill perfectly the space between the largerchunks, and, the filling of crucibles with granules is fast.

1.4 The Metallurgical Path: UMG-Si

Over many years, photovoltaic industry (PV) used the leftovers from thesemiconductor industry, which was in most cases ultrapure poly-feedstock,cutoffs from Cz ingots, and so on The base material was in the range of 9–10

or even 11N purity Using it for multicrystalline ingots, there is hardly anydifference noticeable whether 6N or 8N or 10N polysilicon is used There-fore, the question seemed appropriate: Why not use silicon of purity justclean enough for cell processes and simplify the purification process accord-ingly? The metallurgical path seemed highly promising: easy to scale,low-energy consumption, low Capital Expenditures (CAPEX)—but stilldelivering a fully usable product Dozens and dozens of groups and compa-nies tried it worldwide (Bernreuter and Haugwitz, 2010), and only abouttwo survived on a scale somewhere between pilot and full production:Silicor Materials and Elkem (a subsidiary of China National Bluestar Group

Co Ltd.) The U.S.-based company Silicor Materials (former Calisolar) hadpurchased the UMG-process from the Canadian company 6N The 6N pro-cess (Nichol, 2011) is based on the alloying of silicon with aluminum3: Met-allurgical grade silicon of some 98–99% purity is mixed with aluminum, andthe hypoeutectic mixture becomes liquid in the range of 900–1000°C,depending on the silicon concentration The eutectic temperature itself is

577°C, with a silicon concentration of 12.2 at% Cooling down the eutectic mix, the excess silicon forms small crystallites or flakes, embedded inthe liquid Al–Si melt In silicon, all metals show small segregation coeffi-cients4and, consequently, are enriched in the melt, or better, are accumu-lated in the solid–liquid boundary layer The point with the accumulationwithin the boundary layer is a bit problematic: a proper separation of thesilicon crystallites from the melt is essential and a chemical etching step isrequired to dissolve the metals To get a good cost structure, the residual

hypo-3 Instead of aluminum, tin would also be an option, but aluminum can be separated from silicon more easily, either mechanically (e.g., centrifugation) or chemically etched off Basically, all materials used in former times for the liquid-phase epitaxy (LPE) of silicon could be used for alloying with silicon; restrictions result mainly from practical considerations like availability in large quantities and price.

4

The lower temperature of the Al–Si melt compared to pure liquid silicon reduces the segregation ficients even further (e.g., Morita and Yoshikawa, 2011 ).

coef-Al–Si melt—still slightly hypoeutectic—has to be sold, but there is a marketfor this kind of alloys The main trouble makers are, besides the properremoval of the aluminum, which might be trapped in inclusions, the elim-ination or reduction of boron and phosphorus Recently, plans for a 16,000 tplant in Island had been released (Kaes et al., 2014).

Core features of the ELKEM process are chemical etching and slag ment (Ceccaroli and Friestad, 2005; Heuer, 2013; Schei, 1998; Wang et al.,

treat-2014) A calcium-based slag is used, and during the cooling-down phase,most of the impurities are accumulated in the slag After solidification,the slag and the impurities are etched off and purified silicon is obtained.The process works very well for the metallic impurities, but again, boronand phosphorus are still present and the material is somewhat compensated.Boron and phosphorus had been the greatest bottleneck for all the dif-ferent UMG processes or better: their issue of failure Boron shows a segre-gation coefficient of k0¼0.8 (somewhat lower at reduced temperatures) andphosphorus 0.35 Removing boron and phosphorus simply by segregation isnot an option All the methods developed so far are either costly or compli-cated (or both):

– oxidizing the boron out (the Becancour/Timminco process): huge loss ofsilicon (Leblanc and Boisvert, 2008)

– removing it by slagging: expensive and risk of introducing other rities (Ceccaroli and Friestad, 2005; Schei, 1998; Wang et al., 2014).– gettering, forming a metal boride (e.g., TiB2): not efficient enough(Yoshikawa et al., 2005)

impu-– using low boron raw materials (SolSilc or SolSil process): helps cantly but requires a clean reduction process (Dosaj and Hunt, 1981;Geerligs et al., 2002)

signifi-Phosphorus might be reduced by vacuum treatment of the melt or by plasma(Alemany et al., 2002; Delannoy et al., 2002), but both approaches are costintensive Work-around solutions had been suggested using compensatedfeedstock (i.e., silicon-containing boron and phosphorus/adding boron orphosphorus during the solidification; Dethloff and Friestad, 2007) or addsome gallium (Forster et al., 2011; Kirscht et al., 2010) in order to compen-sate the accumulated phosphor toward the end of the block, but the point is,

so far, all UMG products are not reaching the purity of CVD-based Siemens

or FBR material Today, they are good with respect to metals, but boron inparticular is still an unsolved problem And even if the user is adding boronduring crystallization, and maybe much more than the remaining boronlevel in the UMG-feedstock had been, the product can be sold on the openmarket only with a certain discount

Today, UMG-Si suffers a hard time, but if the boron–phosphorus lem can be solved, it might be the path with the lowest cost structure, thelowest CAPEX, and the easiest to scale up or down, according to the marketrequirements.

prob-1.5 Different Poly for Different Crystallization Techniques1.5.1 Mono Growth, Single Batch Mode

In monocrystalline growth by the standard Cz method, the trend goes tohigh-efficiency cells Therefore, n-type cell structures will very likely gainmarket shares For these applications, high-quality wafers are essential andpolysilicon from CVD processes will be the standard A certain mixing withgranular material is possible, but only if it is low in metals and low withrespect to trapped gases During mono-crystal growth, the risk for structureloss is always given and ingot producers try to avoid any potential sourcewhich could jeopardize their yield Since high-quality material is available

in sufficient quantities right now, consumers favor 9N or 10N poly material.1.5.2 Feeding and Multipulling

Feeding and multipulling is used primarily for mono growth, although tain activities are visible in the multicrystalline sector (Mu¨ller et al., 2009),too Polysilicon for feeding processes has to show excellent transport prop-erties, with a minimum risk for clogging and low material abrasion Formono-ingot growth, the introduction of particles has to be avoided andaccumulation of impurities in the residual melt has to be minimized Melt-ing should be smooth and fast Theoretically, granular material would beperfect for feeding, and the spherical shape and the rather small size givethem perfect transport properties In practice, the high dust load, increasedmetal concentrations, and trapped process gases (Kajimoto et al., 1991)might cause problems Problems, the poly manufacturers still have to work

cer-on An alternative to granular material are crushed chunks: they are availablefrom so-called “size 0” on (smaller than 10 mm, often rather chip-like) andthe maximum size for feeding should not exceed some 10–20 mm; other-wise, the impact and the splashing when the solid silicon hits the melt mightbecome serious

1.5.3 Standard Multicrystalline Casting

The specifications for the polysilicon feedstock used for multicrystallinegrowth are lower and mainly driven by cost reduction A few particles or

a certain metal background are not affecting the quality of the ingot inthe same way as it would be in Cz growth One reason is that in any case,

the crucible and the crucible coating release a significant amount of rities during the crystallization process anyway (Schubert et al., 2013).Therefore, quite often a mix is used, composed of standard solar-grade poly-silicon mixed with second-grade poly (8N and lower) Furthermore, most ofthe side slabs of the ingot are recycled in order to minimize material losses.Most of the granular material is used for multicrystalline growth, where it isblended with poly chunks.

impu-1.5.4 Float Zone

Float-Zone (FZ) growth requires specific feed rods: crack-free, smooth face, minimum bending, high-purity, free of any oxide or nitride layers andwith uniform, microcrystalline morphology, to mention just the mostimportant characteristics FZ feed rods are produced in CVD reactors ded-icated to this purpose, and this requires special know-how with respect tothe control of the process gas composition and flow arrangements, as well

sur-as a uniform temperature distribution and a specific cool-down procedure(Freiheit et al., 2010) Only a very limited number of polysilicon producersare able to deposit feed rods for FZ applications; thus, the availability is lim-ited, production is low, and prices are high Alternatives will be discussed inSection 3.3

2 FUNDAMENTAL PARAMETERS FOR SILICON

The density of solid and liquid silicon differs by 10% Silicon shows asimilar density anomaly like water: at the phase transition to the solid, itexpands This property prevents the use of any kind of closed crucibles,and the sufficient space for volume expansion is always critical The densitychange might be used for the measurement of the solidification rate duringdirectional solidification, as we will see later on, but it bears a significant risk

for Cz and for vertical gradient freeze (VGF) growth In case of a power ure, the melt freezes from top to bottom and will unavoidably crack the cru-cible and will spill liquid silicon into the furnace chamber.

fail-The heat capacity for solid silicon is in the range of 0.7–1.0 J/g K andmight be described by a second-order polynomial fit (Gurvich et al.,

1990) Whether there is an anomaly around 560 K as described in Glazovand Pashinkin (2001) or not does not really affect crystallization since itwas only described for slow heating rates, not relevant for our consider-ations Quite significant is the high value for the latent heat of phase change.Values given in literature vary somewhat in the range of 40–50 kJ/mol (or3.3–4.2 kJ/cm3, seeTable 1), but in any case, it is extremely high Thus, alarge amount of energy is required for the melting process, which has to beremoved during crystallization As a matter of fact, more energy is requiredfor the melting itself than for the heating from room temperature to themelting point Heating needs approximately 0.33 kWh/kg silicon (assuming

an average heat capacity of 0.85 J/g K) and melting requires 0.5 kWh/kgsilicon (assuming 50.6 kJ/mol for the latent heat of phase change, according

toZulehner et al., 2012) Typical values for the crystallization by the Cz andthe VGF technique are summarized in Table 2

In the case of Cz growth, the heat is released by radiation mainly, but incase of VGF, it has to be extracted by heat conduction through the bottom of

Table 1 Specific Material Parameters of Silicon

Electrical conductivity

Liquid 1.33 10 6 Ω -1 m1( Brandes and Brook, 1992 )

the crucible, where the crucible made of sintered quartz ceramic acts as aninsulation barrier.

Discussing the crystallization of silicon for PV application, it is helpful tohave a look at the actual size and geometry of the ingots (Table 3) Today,standard wafer size is 156156 mm2

, either full square or pseudosquare incase of certain mono ingots Pseudosquare refers to the geometry with miss-ing corners: to do without the four corners (the missing triangles have thesize of approximately 101015 mm) reduces the active cell area by lessthan 1% but allows to reduce the ingot diameter from 222 mm down to

206 mm (also referred to 900 vs 800, even this is not exactly correct) Only

a few cell manufacturers are still using the 125125 mm2

of mono, it is due to the fact that the ingot is cylindrical but the wafer isrectangular; in the case of multi, it is due to the impurity-rich areas near

Table 2 Energy Balance for Heating, Melting, and Crystallization of Silicon

Heating and melting:

Czochralski

150 kg crucible load (Heating: 49.5 kWh; melting: 75 kWh)

total: 124.5 kWh Vertical gradient freeze

total: 207.5 kWh

total: 664.0 kWh Crystallization:

Czochralski

Diameter: 900; growth rate: 1 mm/min 5.5 kg/h 2.75 kWh

Vertical gradient freeze

the walls, the bottom, and the top For pseudosquare mono growth, the area

of the side slabs amounts to 27%, for full square even 37% Adding some 7 kgfor the top and tail part and some 3 kg for the residual melt, a 150 kg cruciblecharge results in 102 kg of bricks for wafering (pseudosquare) or 88 kg forwafering full square, respectively The material is not lost but will berecycled, apart from the residual melt, which is difficult to separate fromthe crucible Nevertheless, it is affecting the energy balance In the case

of VGF, the situation is slightly better, but still, about 2–2.5 cm from allthe edges have to be removed, which results in an optimistic scenario in

a material utilization of 73% (G4) and 77% (G6), respectively Part of theremoved side slabs will be recycled, but they are somewhat contaminatedwith iron, chromium, and copper

The cutoff size of the edge areas of VGF blocks are average values andmight vary somewhat from manufacturer to manufacturer In the case ofVGF, upscaling will improve the utilization factor somewhat, but the larger

Table 3 Geometry and Mass Balance for Czochralski (800Pseudosquare) and VGF (G6) Growth of Silicon

Czochralski

Ingot cross-section area (800) 333 cm2

melt volumes and the longer process times also increase the width of the face boundary layers with high metal contamination and low carrier lifetimes(“electrically dead zone”) The rather large loss of material was always astrong motivation for direct wafer casting technologies (until the final wafersare ready for the cell process, an additional 40–50% of the silicon from theready-to-cut bricks will get lost in the wire saw) However, as long as thedirect wafer technologies do not reach the same thickness as the wafers fromthe multiwire process, which is in moment between 150 and 180μm, there

sur-is not a real advantage from the viewpoint of material utilization In any case,the rather low material utilization factor for crystalline silicon wafer technol-ogy is a significant cost driver and it will be an important task for the future toimprove it

An important material property of liquid silicon is its high chemical tivity In contrast to solid silicon, which is protected by an oxide passivationlayer and thus is very easy to handle, liquid silicon is a highly aggressive sub-stance So far, no material is known, which is fully inert against silicon Even

reac-in the oxidized state as Si4+(e.g., as SiO2, SiC, or Si3N4), there is always aninteraction with the melt and a certain dissolution or formation of precip-itates can be observed In particular, in the case of SiO2, the reaction will notstop since the oxygen vapor pressure of SiO is rather high and it will evap-orate at the free melt surface Thus, the equilibrium always favors the furtherdissolution of the quartz crucible The dissolution rate for fused quartzglass in contact with liquid silicon was reported to be in the range of1.15105cm/min in the bulk of the melt and up to 8.4105cm/min

at the triple point melt–crucible–gas (Chaney and Varker, 1976)

A correlation with melt stirring was reported by Hirata and Hoshikawa(1980) and a certain correlation to the boron concentration was found byAbe et al (1998), but the reported values were all in the same range Toget a better idea of the amount of quartz glass dissolved during the course

of the growth run, we might assume a process time of 50 h and an averagecrucible surface in contact to the melt of 2300 cm2(for a 2400crucible; at thebeginning, it will be around 5600 cm2but decreases continuously) The cru-cible wall would be reduced by about 0.35 mm on average, which correlates

to some 200–250 g of crucible material dissolved into the melt The sion rate of the quartz glass crucible is a fundamental issue for multipulling orfor continuous Cz processes, and the development of high corrosion-resistant crucible materials is essential In the case of multicrystalline growth,the crucible is protected by an Si3N4coating, which cannot be used for Czgrowth, of course Silicon nitride particles would result in structure loss

corro-With regard to metals, we might distinguish four classes (Table 4): siliconmight form (A) solid solutions, (B) eutectics, and/or (C) intermetallic com-ponents, or (D) shows a complete mixing in the liquid state, but as good as

no mixing in the solid Quite often, eutectics and intermetallic componentsare found in one phase diagram and sorting into the different classes is notalways a clear case However, it helps to understand the interactions andchemical reactions

Some of the silicides have rather high melting points, e.g., MoSi2(Tm¼2020 °C) or TaSi2 (Tm¼2040 °C) However, the tolerable levels

of these metals for solar applications are extremely low, and concentrations

in the ppt range affect the cell efficiency already heavily (Coletti et al., n.d.;Davis et al., 1980) Metals from class (D) are used for LPE and class (B) orclass (D) elements are candidates for the use in silicon refinement

Whereas the high reactivity in the liquid state makes it difficult to find theright crucible material, the low solubility in the solid helps quite significantlyfor purification Despite a few exceptions, most elements show small segre-gation coefficients (the segregation coefficient k0defines the ratio betweenthe concentration in the solid and the concentration in the liquid, under theassumption of thermodynamic equilibrium) and will not be incorporatedinto the crystal but will accumulate in the liquid boundary layer ahead ofthe solid–liquid interface (Table 5) One exception is boron (k0¼0.8).The large segregation coefficient of boron favors a uniform dopant distribu-tion for p-type ingots—but it is quite troublesome for silicon purification

A second exception is oxygen With a segregation coefficient around 1, allthe oxygen near the solid–liquid interface will be incorporated into the crys-tal To prevent this, the transport of oxygen toward the interface has to bereduced, which is possible by proper melt flow configurations The oxygen-rich melt should be moved away from the growing interface and should be

Table 4 Classification of Binary Silicon Phase Diagrams with Respect to the Formation of Solid Solutions, Silicides, or Eutectics

(B) Eutectics (low solubility in the solid) Al, Ag, Au, Bi, Pb (C) Intermetallic compounds/silicides Cu, Ta, Fe, Mg, Mo,

Ni, Ti, (D) Very limited solution in the solid, and complete

solubility in the liquid

Sn, In, Zn, Ga

transported toward the free surface, where the oxygen (in form of SiO) canevaporate and subsequently be removed from the growth chamber.2.2 Numerical Simulation

Today, numerical simulation is a standard tool for industrial crystallization

In most cases, it is an integral part for any hardware or hot-zone ment It helps to understand the heat fluxes (and losses), the material trans-port, and reveals which areas are crucial for the optimization of the energyconsumption The first attempts for computational simulation of crystalliza-tion processes go back to the 1970s (e.g.,Kobayashi, 1978) At that time, itwas still restricted to 2D axisymmetrical calculations based on finite differ-ences and nonstructured grids Now, modern software packages are running

develop-on PC systems and are able to handle transient processes, 3D flows, and some

of them even chemical reactions In particular, for the crystallization of icon, commercial codes are now tailored to specific growth technologies.Examples for software packages dedicated to silicon crystal growth are,e.g., CGSim (http://www.str-soft.com/products/CGSim/; Smirnov andKalaev, 2009), FEMAGSoft (https://www.femagsoft.com/; Collet et al.,

sil-2012), or CrysVUn (Kurz et al., 1999; http://www.iisb.fraunhofer.de/de/abteilungen/kristallzuechtung/crysmas.html) to mention just some ofthem, or of course ANSYS (http://www.ansys.com/) as a more general

Table 5 Segregation Coefficient k 0 for Various Elements in Silicon

software code for any kind of fluid dynamic problems Quite often, the user

is enabled to add and integrate user-based subroutines, e.g., in order to ulate external magnetic fields Therefore, numerical simulations became areliable and indispensable tool for any crystal grower Nevertheless, certainpoints have to be kept in mind when analyzing the results of numericalsimulations:

sim-– In the simulation, the heat transfer is always idealized In reality, it will bereduced due to small gaps, surface layers, cracks, etc., or it might beincreased by altered material properties, enhanced emissivities, etc.– Today, the material data are known much better than some 20 years ago.Still, they are often idealized or not available as a function of the temper-ature Furthermore, they might change over time

– Materials exposed to high temperatures and aggressive media will changetheir structure and their surface In particular, surface corrosion and sur-face coatings have a huge impact on the temperature Changes in theemissivity affect the radiative heat transfer, which has a T4 impact onthe heat flux

– The different length scales are difficult to handle We have to deal withmacroscopic features in the meter range, but at the same time, chemicalreactions and surface-related phase changes have to be resolved in themicrometer or even submicrometer range

– Certain features have a 3D or a time-dependent characteristic VGF isnonaxisymmetric by definition The large melt volumes result in largeGrashof and Reynolds numbers, indicating time-dependent 3D flowstructures

– For certain aspects like defect formation, structure loss, or grain tion, the physics behind is not fully understood yet and the physicalmodels are not always adequate

forma-As long as these limitations are kept in mind, numerical simulations are anextremely helpful tool Most software programs became rather user-friendlyand the profile of a typical operator is shifting from a highly specialized sci-entist toward an engineer with experimental background But in any case,the proper validation of numerical results by experimental data is absolutelycrucial

3 CRYSTALLIZATION TECHNOLOGIES

In the following chapter, the main technologies for silicon tion are described in detail: the Cz technique used for the majority of all

crystalliza-mono ingots, the directional solidification or VGF method used for crystalline ingot production, and finally the FZ technique, a method wellestablished for the crystallization of electronic grade ingots, whenever lowoxygen material is required, but not yet adapted to the PV market Also,

multi-FZ would provide many benefits, and there are certain bottlenecks whichprevented the cost-competitive introduction of FZ wafers for solar cellmanufacturing until now One serious problem is the availability of suitablefeedstock

Other crystallization techniques for silicon could not gain a significantmarket share so far For example, the electromagnetic casting had made sig-nificant progress; e.g., the Japanese company Sumco had shown impressivepictures of 7 m tall ingots (“taller than a giraffe”—as they claimed in their

SUMCO Annual Report, 2008), but the technique was considered not costcompetitive and production was stopped There had also been many activ-ities with respect to sheet growth (EFG - Edge-defined Film Fed Growth bySchott (Mackintosha et al., 2006), String Ribbon (van Glabbeek et al., 2008)

by Evergreen/Sovello, to mention just the most prominent ones), but so far,none of them had really been able to reach the cost structure and/or thequality of Cz and VGF We will therefore focus in the following on the pre-dominant and most promising PV silicon bulk crystallization technologies

A detailed discussion of the different ribbon and foil techniques is provided

byRodriguez et al (2011)

3.1 Pulling from the Melt: The Cz Technique

Initially, pulling a monocrystalline material from a melt goes back toCzochralski (1918) Although the initial intention was not the growth oflarge single crystals but the measurement of solidification velocities andlatent heat, it was soon realized that this method was perfectly suited forthe pulling of monocrystalline ingots There is no direct interaction ofthe growing crystal with the crucible material, and in situ observation ofthe success (or failure) of the growth process is easily carried out An excel-lent overview about the historical development of silicon pulling from themelt was given byZulehner (1999); unfortunately, to the knowledge of theauthor, it is only available as a conference proceeding paper

3.1.1 Standard Cz Growth

Since more than half a century, the Cz technique is the workhorse for thesemiconductor industry At the very beginning, there was a competition

between FZ and Cz The crucible-free growth and the lack of any graphite

or insulation material seemed to be in favor of the FZ method, but after thedevelopment of proper quartz glass crucibles and hot zones based on purifiedgraphite, the easy scale-up option for Cz and the easier handling and oper-ation of the crystallization process shifted the pendulum clearly toward Cz.The point that for most semiconductor devices a certain oxygen concentra-tion is beneficial for device manufacturing gave Cz additional credit andsoon the Cz technique had a market share for semiconductor ingots of morethan 90% Exceptions are the low-in-oxygen wafers for power electronics,which are still the domain of FZ silicon A schematic drawing of a Cz puller

is given in Fig 6, and state-of-art machines for the growth of up to 1200ingots are shown inFig 7

The basic features of the Cz technique might be summarized as follows: thefeedstock material (it might be broken chunks, or chips, or etched cutoffs, oreven granules) is loaded into a quartz glass crucible, which is sitting in a graphitesusceptor (Fig 8) The material is heated by a graphite-based resistance heater;

in most cases, it is a single, fence-shaped heater, sitting in a fixed positionand surrounding the crucible The crucible can be moved up and down, sim-ilar to the seed crystal The monocrystalline seeds of specific crystallographicorientation, which is machined from a dedicated Cz crystal, are clamped in

a seed holder and connected to a stainless steel rod or a wire Both the crucibleand the seed/growing crystal are rotating, normally in counter directions.The whole assembly is sitting inside a vacuum chamber During the pro-cess, a continuous argon flow is purging the puller, and the argon flow is in

Heater Heat shield

Quartz

Ingot

Figure 6 Sketch of the Czochralski method: The ingot is pulled upward and the crucible

is lifted according to the amount of solidified silicon in order to keep the melt level at a fixed position Crystal and crucible rotate in counter-direction For the increase of pro- ductivity, the radiation shield became an essential part in modern Cz puller.

Figure 7 Industrial-size Czochralski puller for the growth of 800 and 900mono crystals, with an ingot length of up to 2 m All the subsystems like vacuum unit, power supply, and dust filter are located at a lower level not visible in the image The total height of a puller might easily reach some 6 –8 m (PVA TePla puller EKZ-3500 at Fraunhofer CSP.)

Figure 8 Loaded Cz crucible The quartz glass crucible is sitting in a graphite support unit, surrounded by a fence-type heater For the picture, the outer insulation, as well as the water-cooled jacket, has been removed.

the range of 10–30 l/min—quite a significant contributor to the list of sumables The modern Cz puller, in particular the PV-related ones, operates

con-at a reduced pressure of approximcon-ately 5–50 mbar

For PV applications, onlyh100i-oriented ingots/wafers are in use, which

is quite convenient, since this is the easiest to grow direction

In order to pull a 1.5- to 2-m-long crystal, the pullers are rather tall, withall the crucible lifting devices, the pulling shafts, and the chamber itself; totalheight reaches easily some 6–8 m All the modern ones are equipped with agate valve between the upper chamber and the hot-zone area Not only doesthis allow multipulling (i.e., more than one ingot is pulled out of one cru-cible), but it also allows to remove a crystal in case of structure loss and pull asecond one from the remaining melt

An essential part of all modern Cz pullers is the radiation shield (or cone

or funnel;Fig 9) separating the growing crystal from the hot crucible walland the heater The first attempts with radiation shields go back to the 1980s(at Wacker;Zulehner and Huber, 1982); today, they are an inherent part ofany hot-zone design They might be manufactured from graphite, double-walled graphite, carbon re-enforced carbon (CFC), or even molybdenum,and the shape might be conical or rather straight—a variety of differentdesigns are in use Radiation shields have two clear benefits:

– Heat removal: critical for crystallization (and especially for achieving fastergrowth rates) is the removal of the latent heat during the solidificationprocess Using a radiation shield, the crystal faces a comparatively cold

Figure 9 Fully mounted hot zone, the radiation shield covers most of the crucible surface.

surface Without radiation shield, it would be exposed to the radiationfrom the heater and the crucible.

– Guidance for the argon flow: in order to keep the oxygen concentration low,the evaporating SiO from the melt surface has to be transported out of thehot zone as fast as possible The radiation shield guides the argon flowalong the crystal down to the melt surface, over the melt surface, upwardthe crucible wall, and then downward to the bottom area where the pip-ing for the exhaust system is located

The temperature distribution for an 800 process under an argon flow of

20 l/min is visualized by numerical simulations inFig 10 A certain lenge is the proper control of the gap between the radiation shield andthe melt surface: the smaller the gap, the better for heat removal and for crys-tallization But obviously, if it becomes too small, there is a high risk oftouching the melt, which would be the end of the process (and most likelythe end of the radiation shield) How precisely the gap might be controlled

chal-Figure 10 Numerical simulation of the Czochralski growth chamber Today, commercial software is readily available for the numerical simulation of basic process features like the temperature field, the gas flow, or the stress field Numerical simulation became a handy tool for the development and improvement of hot-zone designs (Simulation using CGSim software, Fraunhofer CSP.)

depends on the system used for growth control and for controlling the crystaldiameter Since PV-crystal pullers are rather low-budget machines, quiteoften, no fancy load cells or redundant measurements of weight and diam-eter are used, but simple optical control of the diameter using CCD cameras.Since the observation angle is rather steep with respect to the crystal axis(there is no option for any optical access to the hot zone from the side orany optical access through the crucible), accuracy is limited; e.g., an uncer-tainty of half a millimeter on the crystal radius (for an 800ingot) accumulatesover a crystal length of 200 cm to an incorrect calculation of the crystalweight of approximately 1.5 kg and consequently a misinterpretation ofthe position of the melt level.

it has to be kept in mind that solar-grade Cz growth is carried out underdifferent conditions than semiconductor ingot growth: first, with respect

Table 6 Summary of the Basic Parameters for Typical PV-Related Czochralski Ingot Growth

to the material preparation (no chemical etching; after crushing, thepoly-feedstock goes straight into the crucible) and second, with respect tothe growth environment It is obvious that higher defect densities are theprice for the envisaged cost savings.

Melting: The melting process is rather time consuming It takes some6–10 h, depending on the hot-zone design and the crucible charge Aftermelting, melt homogenization, and temperature stabilization, the seed crys-tal comes in contact with the melt

Necking: Introduction of the necking process in 1958 by Dash (1958,1959)was one of the great breakthroughs for silicon single-crystal growth

To keep a growing crystal free of dislocations is relatively easy, in particular

in the case of silicon As long as no major disturbances are imposed on thegrowth conditions, the value for the critical resolved shear stress is largeenough to avoid the generation of dislocations The challenge is, however,

to get a dislocation-free crystal Even starting with a dislocation-free seed, thethermal shock, when the seed crystal comes in contact with the melt, induces

a high concentration of crystal defects Silicon shows two distinct features:first, the velocity with which dislocations can travel through the crystal iscomparatively low and only effective near the melting point If the crystalhas reached some 1200°C, they are more or less immobile Second, siliconcan be crystallized with rather high growth rates, in particular for small diam-eters when the amount of latent heat to be removed is still small Since dis-location inh100i silicon grows outwards, the crystal has simply to be pulledfaster than the dislocations can move or multiply along the growth axis

Figure 11 The Czochralski process: top from left to right: the loaded crucible —melting

of the material —necking Bottom from left to right: shoulder growth—shortly after the transition to the body —body growth.

Impressive pictures of the dislocation elimination are provided byShimura(2007) For the necking during Cz silicon, typical parameters are as follows:– Pulling rate: >5 mm/min

– Diameter: <4 mm, preferably <3.5 mm, or even more preferably

<3 mm

– Length of neck: >50 mm

These values are guidelines to the best knowledge of the author No exactthreshold values have been measured or published so far and everyone has hisown set of parameters And once it works, it will not be changed anymore.For all the substances crystallized on an industrial scale, necking works bestfor silicon It is also applicable to other single elements (e.g., germanium ormetals) but fails for binary substances (e.g., GaAs or CdTe) In binary ormultinary crystals, dislocations travel too fast and the corresponding growthrates are too small

Shoulder: After necking, the growth velocity is reduced and the melttemperature is lowered in order to achieve an increase in the diameter.The shoulder length for an 800ingot is typically in the range of approximately50–100 mm

Body: The growth continues with constant diameter (and minimal iations of the crystal diameter), 800 for pseudosquare and 900 for full square.Standard growth rates are around 1 mm/min With sophisticated hot-zonedesigns, some 1.3 mm/min is possible Trends and new developments try tofurther increase it; a promising approach is the integration of an activecooling ring A 2-m-long ingot might be crystallized within 25–35 h,depending on the system But the whole cycle time amounts to 50–60 h,which explains the recent developments toward multipulling, continuouspulling, and so on

var-Cone: A slow reduction of the ingot diameter at the end of the run isessential to avoid thermal shock, which would immediately induce disloca-tions and slip planes Since they are preferentially arranged in 45° angles,they might move back some 200–250 mm If the diameter of the crystal

is controlled by a CCD camera only, the end cone is not visible and this parthas to be crystallized on preset parameters If a load cell is used, the cone iscontrolled actively

Cooling down: A quartz glass crucible survives only one heating cycle(Fig 12) Once it had been heated up, it cracks during the cooling-downprocess It might be refilled in the hot state using a feeder system, but duringthe cooling-down period, it will be damaged The remaining silicon (theso-called “pot scrap,” some 2–5 kg of residual silicon) sticks to the quartz

glass crucible, and due to the different thermal expansion coefficients, thequartz glass brakes In addition, a phase change of the quartz glass intocristobalite starts during the crystallization process, and these two substanceshave different thermal expansion coefficients Since the modern machinesare rather well insulated, the cooling down from over 1500°C to temper-atures low enough for cleaning and crucible replacement takes several hours,

a downtime of the puller which is unproductive of course

3.1.1.2 The Main Cost Drivers

Crucibles: As explained above, crucibles are for single use only Under thisconsideration, it would be attractive to use crucibles as cheap as possible

On the other hand, a crucible which does not survive the whole growth cyclemight destroy the entire hot zone Furthermore, crucibles are dissolved whilethey are in contact with liquid silicon Metallic impurities from the quartzglass will accumulate in the melt and finally will be incorporated in the ingot

In particular, Fe has to be mentioned, but Al, Ca Cr, Fe, Mn, Na, or K aretypical impurities, too In order to increase the lifetime of the crucible, which

is a crucial topic for all developments toward multipulling or continuousgrowth, multilayer designs are in use and/or doping the inner layer with bar-ium is another promising innovation (Wakita et al., 2013)

Figure 12 Pot scrap: 2 –5 kg of residual melt remains in the crucible and sticks to the quartz glass The crucible cracks during the cooling-down process and has to be dis- posed Recycling of the pot scrap is difficult; the separation of the silicon from the quartz requires mechanical and chemical process steps Further, impurities are accumulated in the pot scrap.

Hot zone: The graphite parts are consumables, too, and their life span isnormally in the range of 10 to maybe 50 cycles, depending on their work andheat load Since a full hot zone might easily cost some 20,000–40,000 US$depending on the graphite purity and manufacturing costs, the hot zonecontributes quite significantly to the operating cost Multipulling might help

to increase the lifetime of graphite parts, because they will see less heatingcycles, but at the end, there is not really a lot which can be done to lowerthe cost without affecting the purity and the quality of the final product.Coated graphite parts (e.g., SiC-coated) or CFC-enforced graphite felt isbeneficial for the purity and the life span of the corresponding part, but it

is significantly more expensive than the standard materials

Argon: During the whole growth cycle, a continuous argon flow is tial in order to remove the SiO Quartz glass is permanently dissolved byliquid silicon Fortunately, more than 99% of the oxygen evaporates asSiO, which condensates at cooler surfaces The argon flow helps to transportmost of the SiO to the outside of the growth chamber, where it might beoxidized in a controlled manner after the growth.5Further, the continuousremoval of SiO from the melt surface keeps the equilibrium on the silicon-rich side and supports the evaporation of the SiO Argon recovery systemsare available and they might be an option to reduce the Operational Expen-ditures (OPEX)

essen-Process time/productivity: A scale-up as we see it for directional tion, going from G4 to G5 to G6 and so on, is not possible for Cz growth.Increasing the diameter in such a way that four bricks are cut out of thegrown ingot instead of one brick would require a crystal diameter of

solidifica-445 mm, which is not cost effective In addition to the higher OPEXand CAPEX, the growth rate for such large ingots is reduced compared

to the one of an 800 standard ingot, since the removal of the latent heat ismore difficult Thus, an increase of the productivity has to be coupled with

an increase of the growth rate and/or a reduction of the downtime.Yield/structure loss: Structure loss is an important (or maybe the mostimportant) parameter for the calculation of the real cost of ownership andthe final dollar per wafer price One of the great features of the Cz method

is the fact that a loss of single crystallinity is detected immediately or it is evenanticipated by careful observation of the structure and pronunciation of thefour growth lines.Figure 13shows the growth facets during the crystalliza-tion of the shoulder If the facets are not all equally well pronounced or if one5

As a fine powder, SiO is a pyrophoric substance, which has to be handled with great care.

of them looks different compared to the other three, a structure loss is verylikely If a structure loss is unavoidable, a decision can be drawn whether it ischeaper to melt back the already grown part or to take it out and growanother ingot from the remaining melt.6 A problem with structure losses

is that quite often it is not clear why it happened Reasons might be:– Particles, dust, etc., have been brought in during the loading process oreven earlier, during the harvesting and crushing of the polysilicon In thiscase, backmelting would not help, of course The particle-contaminatedingot has to be removed

– Necking was not successful or the seed crystal was recycled too often.– Growth conditions are not appropriate (too fast, temperature fluctuationstoo high)

– Polysilicon was not pure enough, metallic impurities accumulated anddestabilized the growth interface (generation of morphologicalinstabilities)

– Particles are transported or introduced by the argon flow either from theargon itself or picked up by the argon

Figure 13 Growth facets (or growth lines), indicated by the blue (gray in the print sion) arrows: During the shoulder growth, the pronunciation of the four facets reveals relevant information about the defect-free character of the crystal (In the given image, the fourth facet is partially shadowed by the neck.)

ver-6

In the case of FZ growth, it will be seen immediately, too, but due to the growth setup, melting back is impossible.

– Crucible corrosion results in a flaking off of quartz glass which might stick

to the growth interface

For sure, there are many other reasons not listed here, but it becomes clearthat the growth of single crystals by the Cz method does not allow manycompromises and shortcuts And, it requires skilled operators Figure 14shows an 800 ingot for the production of 156156 mm2 pseudosquarep-type wafers InFig 15, the ingot is seen after cutting top and tail, remov-ing the side slabs, and preparing wafers out of the brick

3.1.2 Actual Trends and Recent Developments

et al., 1998), rotating magnetic fields (Dold, 2003; Kakimoto, 2002), or eling ones (Virbulis et al., 2001) The benefits are obvious: with a properfield design, it is possible to reduce temperature fluctuations (Dold andBenz, 1995; Kanda et al., 1996), to minimize dopant striations (Kakimoto

trav-et al., 1995; Kim and Smtrav-etana, 1985), to stabilize the interface shape(Lu, 2007), and to lower the oxygen concentration (Gunjai andRamchandran, 2009) The latter aspect is actually a hot topic: is it possible

to reduce the oxygen level sufficiently to use MCz wafers instead of FZwafers? In particular for the prospering market of power electronic devices,this is of great interest In general, the large ingots with diameters of 300 mmand bigger are mostly grown with the support of magnetic fields in order toreduce convective flows and to keep the melt surface stable In particular,transverse fields generated by modern superconductive magnets are in use

In the case of solar applications, without doubt, it would also be ficial to have the oxygen levels reduced by up to half an order of magnitude,

bene-Figure 14 Czochralski mono ingot, diameter: 800, length: 600 mm.

especially for boron-doped wafers in order to counteract LID (light-induceddegradation due to formation of boron–oxygen complexes) On the otherhand, the installation of a magnet increases the equipment costs substantially.With the small margin ingot manufacturers are confronted with at themoment, magnetic fields are not really an option The improvement in qual-ity is not justifying the cost.

3.1.2.2 Active Cooling

As pointed out earlier, the main limitation for the growth rate in Czsilicon growth is the removal of the latent heat We will see later that,e.g., FZ ingots might be pulled twice as fast as Cz counterparts, simplybecause of the higher heat flow The first major breakthrough was theinsert of radiation shields in the 1980s Recently, in order to enhance theheat removal even more, active crystal cooling systems have been developed:the growing crystal is surrounded by a water-cooled ring or a water-cooledFigure 15 After the growth process, the ingot is squared and wafered The side slabs, the top, and the tail are recycled.

spiral (http://www.pvatepla.com/en/products/crystal-growing-systems/pva/cz—equipment/active-crystal-cooling) The axial position is some20–50 cm above the solid–liquid interface Since the optical access to themeniscus area is mandatory for diameter control, there is a limited degree

of freedom with respect to positioning of the water-cooling system ing to the manufacturer, the active cooling might increase the pulling veloc-ity by 20–40% Obviously, inserting a water-cooled system into the hot zoneand in particular within the proximity of liquid silicon bears a certain risk.Since it has to be machined from metal, any contact between the water-cooled device and the growing crystal has to be avoided Anyway, it is aninteresting approach and it will increase the productivity and thus will fur-ther reduce cost Furthermore, it shows that even for the rather matured Czpullers, there is still room for improvements

Accord-3.1.2.3 Multipulling, Feeding, and Continuous Growth

The Cz process is a batch process with high consumable costs Therefore,since the early days of Cz growth, there had been many ideas and many pat-ents to overcome these weaknesses by developing a semicontinuous or acontinuous process (Altekru¨ger, 1994; Lorenzini et al., 1977; Wang et al.,2003; Zulehner, 1982) The biggest problems when operating a continuoussystem are the reactivity of liquid silicon, the zero tolerance of the process forparticles, and the increase of complexity of the whole system For the pro-duction of electronic grade material for the semiconductor industry, thetechnical problems never had been solved satisfyingly With solar-grade sil-icon, the situation is now different Compared to the semiconductor mate-rial, the ingot specifications are somewhat more relaxed: whereas thesemiconductor industry tries to minimize the concentrations of vacanciesand interstitials and, in any case, dislocations are a complete no-go criteriafor solar applications, an ingot with a certain defect structure or even with

an etch pit density of 103cm2is still acceptable Nevertheless, even withthe development of new materials, better crucibles, and more advancedpullers, certain bottlenecks are still present:

– The crucible life span is a limiting factor

– “Feedable” material of high purity is required

– Accumulation of impurities in the melt has to be avoided and only quality feedstock should be used

high-The benefits of solving these problems are obvious: downtime is reduced,which increases the productivity, and in addition, cost for consumables islowered Actually, there are two different approaches: (I) batch feedingand (II) continuous feeding Figure 16 shows a feeder system for batch

feeding The feeder can be opened and recharged during ingot pulling, and it

is equipped with an independent gas and vacuum system

(I) Batch feeding: Granular silicon or small-size chips are fed into the cible before or in between growth periods, i.e., either before thegrowth starts, the crucible is topped up with additional silicon, orthe crucible is refilled after an ingot has been pulled, in which case,the time required to cool down and to remove the crystal out of thegate chamber is used to melt additional silicon The latter is illustrated

cru-inFig 17 Using batch feeding, the pulling process itself is not affected.The equipment manufacturer PVA TePla has batch feeding systems intheir portfolio, Kayex had offered one too, but in 2013, Kayex steppedout of the Cz business

Figure 16 Feeder system for batch feeding of small-size polysilicon The feeder is designed for topping-up the crucible after the initial melting of the first charge or for recharging the crucible after the first crystal has been grown and moved into the gate chamber for cooling down.

(II) Continuous Czochralski (CCz): In contrast to the batch feeding, tinuous feeding takes place during the growth process itself Silicon isfed into the crucible with the same rate as the ingot grows The amount

con-of melt remains constant; the crucible and the melt level stay at a fixedposition This bears a great advantage: the dopant concentration in themelt and in the ingot can be controlled perfectly Ingots with uniformaxial resistivity profiles can be grown, even for dopants with small seg-regation coefficients In particular for n-type material or for gallium-doped ingots, this is highly interesting Furthermore, since the meltvolume is kept relatively small, the energy to keep it at growth tem-perature is less than it is the case for a fully loaded 2400 crucible withsome 100–200 kg of liquid silicon in it Since the melting process takesplace parallel to the growth process, the process cycle time is reduced.However, feeding in solid silicon at the same time as a single crystal ispulled bears the risk that a solid particle is floating around and is touch-ing the growth interface, a problem that is somewhat reduced by usingdouble crucible arrangements (Bender and Smith, 2012; Swaminathan,2014a; Wang et al., 2003)

The U.S.-based company GT-AT has announced the introduction of a CCzsystem, after they had purchased a couple of years ago the American equip-ment manufacturer and ingot producer Confluence (Johnson and DeLuca,

2012) MEMC (Bender, 2013; Swaminathan, 2014b; now Sun Edison) isworking since several years on CCz systems, and their results with respect

to axial dopant uniformity are quite impressive (DeLuca and Delk, 2012).Furthermore, they purchased the company Solaicx a few years ago, a com-pany which was specialized on continuous pulling processes

Figure 17 Feeding process Feeding of small-size polysilicon chips into molten silicon for a multipulling process.