Plasma Deposited Thin Films

J. Mort, F. Jansen

• Publisher: CRC Press
• Year: 2018
• Print ISBN: 9781315896717
• eBook ISBN: 9781351092715
• Topic: Tecnología e ingeniería
• Subtopic: Ciencias de los materiales

In Summary, the objective of this book is to present in one volume a review of the plasma deposition process and the present understanding of the most important and widely used plasma deposited thin film materials, devices and their applications.

Chapter 1

PLASMA DEPOSITION PROCESSES

Frank Jansen

TABLE OF CONTENTS

I.	Introduction
II.	Nature of the Deposition Process
III.	Operation and Equipment Considerations
IV.	Process Reliability and Diagnostics
V.	Scale-Up Issues and Economic Considerations
VI.	Conclusion
Acknowledgments
References

I. INTRODUCTION

The thin film deposition process, in which condensable radicals are created by the decomposition of precursor gases in nonequilibrium plasmas, is known by different names. In an early and extensive review of the process¹ the term “glow discharge deposition” was used. This term appropriately conveys the fact that gas decomposition takes place in a nonequilibrium plasma which is generally referred to as a “glow discharge.”² In subsequent reviews³, ⁴, ⁵ the process was referred to as “plasma deposition”. This designation nicely contrasts the process with plasma etching, which is a closely related technique, although it has the opposite effect of removing rather than depositing material. However, the term “plasma deposition” is also used, in the context of a thin film deposition process, to indicate a variety of sputtering processes,⁶ and the name could therefore be ambiguous. Recent reviews^7,8 use the terms “plasma-assisted” or “plasma-enhanced chemical vapor deposition” (PA or PECVD), which emphasize its similarity with chemical vapor deposition (CVD) processes.⁹ Both deposition processes use gaseous starting materials which are subsequently decomposed, by thermal energy for the CVD process and by electron impact for the PECVD process. Finally, the terms “glow discharge polymerization” and “plasma polymerization”¹⁰ also refer to “deposition by plasma decomposition” processes. However, these terms are used predominantly for film deposition from organic precursor materials. Given some justification for all of these designations, we have chosen, with some arbitrariness, to use the term “plasma deposition” in this chapter, notably excluding sputtering and equilibrium plasma processes.

A wide variety of materials have been deposited by the plasma process and referenced lists can be found in previous reviews.^1,4,5 Plasma deposition and etching processes have for many years been widely used for film patterning and deposition steps in the fabrication of microelectronic components.¹¹ The fact that insulating films for diffusion masks, interlayer dielectrics, and passivation layers can be deposited and be removed at low temperatures so that previous process steps are not affected, is extremely important for this industry. The plasma deposition of silicon nitride, particularly, has been extensively studied in certain reactor designs,¹², ¹³ has been reviewed in the scientific⁵ and trade literature,¹⁴ and will be further discussed in Chapter 5 of this book. This state of relative technical maturity has not been achieved for plasma deposition processes as applied to the field of macroelectronics. This term refers to the development and fabrication of devices, the dimensions of which are typically larger than those of a silicon wafer. Examples are photovoltaic modules, xerographic photoreceptors, large area displays, and other thin film transistor-based technologies.¹⁵ All these applications are further discussed in the last three chapters of this book. The field of macroelectronics has rapidly grown in importance since the recognition that the defect density in amorphous silicon could be reduced orders of magnitude by hydrogenation. It was realized¹⁶ in 1976 that, by appropriate fabrication techniques, amorphous silicon films could be deposited with such a small defect density (10¹⁵ to 10¹⁶ cm⁻³), that its electrical properties could be changed by substitutional doping. This recognition that plasma-deposited amorphous silicon behaves as an extrinsic semiconductor has spawned a new field of amorphous materials and device engineering with a degree of materials and device design flexibility which was previously unattainable.

The objective of this chapter is to provide the reader with a perspective on the unique features and the inherent limitations of the plasma deposition process for these macroelectronic applications. Because the nature of the plasma deposition process, particularly as it applies to amorphous silicon, will be treated in detail in Chapter 2 of this book, the emphasis here will focus on the practical aspects of plasma deposition. First, the unique characteristics of the plasma process as a deposition technique will be reviewed and compared to other physical vapor deposition processes. In a subsequent section, attention will be given to the practical requirements of the plasma reactor itself, as well as its most important peripheral hardware, such as pumps and gas-handling equipment. The point of view is sometimes taken that the plasma deposition process, due to the inherent complexity of the plasma chemistry and the multitude of phenomena in a low-pressure plasma, is basically uncontrollable. This might be true in the sense that it would be very difficult to control the pathway of certain reactions or to predict from basic principles the proper operating parameters for a certain reactor geometry. However, it is very clear that the characteristics of the end result of this complicated process, viz. the deposited film or device, are by and large reproducible, not only from run-to-run but also between quite dissimilar reactors in different laboratories. It is for this reason that in a later section, rather than emphasizing the complexity of the process, we briefly review the use of the diagnostic tools as they are commonly used to enhance the reproducibility and reliability of the process. Finally, because the process has considerable commercial potential for a variety of practical applications, some of the scale-up issues and the ecomonics of the process will be considered.

II. NATURE OF THE DEPOSITION PROCESS

At the heart of any physical vapor deposition process is the creation of a flux of condensable species. In the case of evaporation and sputtering processes, these species originate from a spatially constrained source such as a solid or an intermediate liquid. The workpiece, or substrate, is required to be in direct line of-sight with this source in order to be covered with a thin film of the source material. Plasma and CVD processes are significantly different from other vapor deposition processes in that the depositing species are created from precursor gases and vapors which, without energy input, are not in themselves condensable. These deposition processes are therefore not in direct line-of-sight in the sense that the precursor materials surround the substrate, irrespective of its shape or its position in the deposition chamber. This, in principle, enables the substrates to be densely packed during the deposition process. In the CVD process, substrates are heated to a high temperature to cause the precursor gases to decompose. In a plasma deposition process, electrical energy is supplied to the gas through the mediation of electrons which are accelerated in an applied electric field. A nonequilibrium plasma, in which the electrons have a much higher effective temperature than the ions and uncharged species, is created through the action of the electric field. The extent of the electric field defines approximately the deposition area. The substrate is commonly close to an electrode or functions itself as an electrode.

Different methods can be used to couple the electrical energy into the plasma. High frequency electric power can be coupled into the discharge through the capacitance of a dielectric wall of a vacuum chamber by the use of external electrodes or a coil. Where a coil is used, the method is often referred to as “inductive coupling”, although the coupling in most cases is not through the action of the magnetic field.¹⁷ Because the deposit pattern on the walls matches the coil, it is evident that the power is, even though nonuniformly, capacitively coupled through the wall. In any case, this external electrode geometry is primarily restricted to tubular reactor shapes of relatively small diameter. A more common electrode configuration is the direct capacitive coupling arrangement, which consists of two opposing electrodes located inside the vacuum chamber. Direct capacitive coupling allows a degree of flexibility in reactor design and control over the process which would be impossible to obtain with external electrodes or coils. The inherent applicability of the plasma deposition process to the uniform deposition over large areas is basically a result of the relative ease by which uniform electric fields can be created over large areas with internal electrodes.

The use of gaseous precursor materials has significant a priori advantages over the use of solid or liquid sources. These latter sources are not only spatially but also compositionally constrained. This is a serious limitation for the demanding materials requirements of macroelectronic devices, which often involve doped and alloyed multilayers. The optimization of device characteristics may even require the grading of interfaces to reduce internal fields, or the creation of dopant or alloying profiles in the film. Controlled doping, alloying, and multilayering are very hard to accomplish with solid sources of material of fixed composition, but this flexibility is easily obtained by the variable flow control of gaseous precursors. Therefore, although it is in principle possible to obtain equivalent¹⁸ or, or some instances, unique¹⁹ materials by hydrogenation of the vapor flux derived from solid source material, the flexibility, mechanical simplicity, and control over the process in practice favors the plasma deposition process over other methods, such as reactive sputtering, for the fabrication of macroelectronic devices.

The most important phenomena that take place in the plasma excitation process, in which condensable species are created from the precursor gases, can be understood by significantly simplifying an inherently complicated process. Precursor gases are always molecular compounds, and a wide variety of species are created by the electron impact dissociation, excitation, and ionization of the gas molecules. The electrons are accelerated in the electric field applied between two electrodes, and their velocity distribution is determined by frequent collisions with gas molecules. This distribution can, to a good approximation, be assumed to be Maxwellian, with a mean speed which is equivalent to an energy of a few electronvolts. As molecular dissociation energies are usually significantly smaller than atomic or molecular ionization energies, the density of neutral radicals is much higher than the density of ionic species. For a gas at a pressure of 1 Torr, the density of gas molecules is 10¹⁶ cm⁻³ With a mean electron energy of 1 eV, a molecular dissociation energy of 5 eV, and an ionization energy of 10 eV, the density of dissociated molecules can be estimated to be 10¹⁴ cm⁻³, whereas the density of ions and electrons is about 10¹² cm⁻³. This estimate for the ion density is an upper limit because it does not take into account that the ionic recombination lifetime is generally much smaller than the lifetime of neutral species due to the long range coulombic attraction force. The important points are that the combined density of condensable species, namely neutral radicals and ions, is small relative to the gas density, and the density of condensable neutrals is much higher than that of charged particles.

The film is therefore predominantly formed by the bonding of neutral radicals to the surface of the growing film. However, there is not necessarily a simple relationship between the weighted chemical composition of the condensable radicals and the chemical composition of the film,²⁰ because surface chemical reactions²¹ can take place between adsorbed species. The latter possibility will be discussed in detail for the case of amorphous silicon in Chapter 2. Ions do not contribute measurably to the deposition process^22,23 although the ionic bombardment during film growth can have a significant effect on the physical properties of the film. The fact that the species which contribute predominantly to film growth are electrically neutral and have to find their way to the surface of the growing film by diffusion has important practical consequences. The diffusing radicals have a relatively low density and the plasma deposition process is therefore inherently materials inefficient. The deposition on a flat surface is isotropic in nature, i.e., the vapor at any point of a flat surface originates from a hemisphere with a radius of about the mean free path of the radicals. Step coverage of surface features depends on the size of the surface feature relative to the radical mean free path²⁴ and is subject to electric field distortion. Furthermore, columnar structures and nodular film defects, resulting from self-shadowing or macroscopic shadowing during film growth,²⁵ are prevalent defect patterns in plasma deposited film...

Table of contents

• Cover
• Title Page
• Copyright Page
• Table of Contents
• Chapter 1 Plasma Deposition Processes
• Chapter 2 Plasma-Deposited Films: Kinetics of Formation, Composition, and Microstructure
• Chapter 3 Electronic Properties of Plasma-Deposited Semiconductor Films
• Chapter 4 Carbon Thin Films
• Chapter 5 Silicon Nitride and Other Insulator Films
• Chapter 6 Photovoltaic Structures by Plasma Deposition
• Chapter 7 Xerographic Applications of Plasma-Deposited Amorphous Tetrahedral Materials
• Chapter 8 Hydrogenated Amorphous Silicon Electronic Devices and Their Arrays
• Index