Manufacturing in cleanroom environments

Clean environments are required for manufacturing modern electronics devices, in particular semiconductor devices, but also hard disks, flat panel displays (FPDs), and solar panels. Wafer processing in the semiconductor industry includes some of the most demanding processes in terms of complexity and cleanliness, due to the submicron dimensions of modern semiconductor devices. This book focuses on industrial cleanroom robotics in semiconductor and FPD manufacturing. Both industries experienced phenomenal technical advancement and growth in the 1980s and 1990s and have established manufacturing facilities in several geographic regions: North America, Europe, and Asia/Pacific Rim. India may emerge as another manufacturing region. The market for semiconductor manufacturing equipment was valued at approximately US$45.5 billion in 2007. The market for FPD manufacturing equipment surpassed the US$1 billion mark in 1997 for the first time. In 2008 it was estimated at US$10 billion.

Cleanroom requirements

Cleanrooms are isolated environments in which humidity, temperature, and particulate contamination are monitored and controlled within specified parameters (SEMI standard E70). Particulates are fine particles, solid or liquid, that are suspended in a gas. Particulate sizes range from less than 10 nm to more than 100 µm. Particulates of less than 100 nm are called ultra-fine particles. Here the term ‘particle’ is used throughout, representing particles of all applicable sizes, either suspended in a gas or attached to a surface. Cleanroom environments are required if particle contamination is a concern, as is the case, for example, in semiconductor manufacturing.

This chapter presents procedures and relevant standards for testing and characterizing robots designed for electronics manufacturing in cleanroom environments. Several robot performance parameters are of interest: airborne particle contamination, surface particle contamination, positioning accuracy and repeatability, path accuracy and repeatability, vibration, and axis decoupling. The presented examples apply directly to substrate handling in semiconductor manufacturing and related industries. In the 1970s a typical semiconductor product yield, the number of semiconductor devices suitable for sale compared to the total number manufactured, was 10–15% in the USA. In Japan 60–90% was achieved through rigorous contamination control (Donovan, 2001). Today (2009) the product yield in modern semiconductor factories exceeds 90%, the result of contamination control equipment and procedures. Sections 7.1 and 7.2 present particle contamination tests.

Airborne particle contamination

Airborne particle contamination (APC) generated by substrate-handling robots can be a critical performance index, depending on the cleanliness requirements, and must be tested and characterized under normal robot operating conditions. The number of airborne particles is measured at selected locations in the immediate vicinity of the robot. For example, Sematech standard 92051107A-STD recommends measuring ‘particles per wafer pass,’ a performance criterion specified in SEMI standard E14. Substrate-handling robots should be tested and certified as subassemblies, and their contamination contribution specified as part of the total tool contamination allowance. An APC test procedure for individual robots is described below. Note that APC is different from airborne molecular contamination (AMC), a non-particle gaseous substance that can also be detrimental to a product or process.

Introduction

Throughout history, there has been a never-ending effort to develop materials with higher yield strengths. However, a higher yield strengths is generally accompanied by a lower ductility and a lower toughness. Toughness is the energy absorbed in fracturing. A high-strength material has low toughness because it can be subjected to higher stresses. The stress necessary to cause fracture may be reached before there has been much plastic deformation to absorb energy. Ductility and toughness are lowered by factors that inhibit plastic flow. As schematically indicated in Figure 13.1, these factors include decreased temperatures, increased strain rates, and the presence of notches. Developments that increase yield strength usually result in lower toughness.

In many ways, the fracture behavior of steel is like that of taffy candy. It is difficult to break a warm bar of taffy candy to share with a friend. Even children know that warm taffy tends to bend rather than break. However, there are three ways to promote its fracture. A knife may be used to notch the candy bar, producing a stress concentration. The candy may be refrigerated to raise its resistance to deformation. Finally, rapping it against a hard surface raises the loading rate, thus increasing the likelihood of fracture. Notches, low temperatures, and high rates of loading also embrittle steel.

There are two important reasons for engineers to be interested in ductility and fracture. The first is that a reasonable amount of ductility is required to form metals into useful parts by forging, rolling, extrusion, or other plastic working processes.

Introduction

With elastic deformation, the strains are proportional to the stress, so every level of stress causes some elastic deformation. However, a definite level of stress must be applied before any plastic deformation occurs. As the stress is further increased, the amount of deformation increases, but not linearly. After plastic deformation starts, the total strain is the sum of the elastic strain (which still obeys Hooke's law) and the plastic strain. Because the elastic part of the strain is usually much less than the plastic part, it will be neglected in this chapter and the symbol e will signify the true plastic strain.

The terms strain hardening and work hardening are used interchangeably to describe the increase of the stress level necessary to continue plastic deformation. The term flow stress is used to describe the stress necessary to continue deformation at any stage of plastic strain. Mathematical descriptions of true stress—strain curves are needed in engineering analyses that involve plastic deformation, such as predicting energy absorption in automobile crashes, designing of dies for stamping parts, and analyzing the stresses around cracks. Various approximations are possible. Which approximation is best depends on the material, the nature of the problem, and the need for accuracy. This chapter will consider several approximations and their applications.

Introduction

This book is concerned with the mechanical behavior of materials. The term mechanical behavior refers to the response of materials to forces. Under load, a material may either deform or break. The factors that govern a material's resistance to deforming are quite different than those governing its resistance to fracture. The word strength may refer either to the stress required to deform a material or to the stress required to cause fracture; therefore, care must be used with the term strength.

When a material deforms under small stresses, the deformation may be elastic. In this case, when the stress is removed, the material will revert to its original shape. Most of the elastic deformation will recover immediately. There may be, however, some time-dependent shape recovery. This time-dependent elastic behavior is called anelasticity or viscoelasticity.

Larger stresses may cause plastic deformation. After a material undergoes plastic deformation, it will not revert to its original shape when the stress is removed. Usually, high resistance to deformation is desirable so that a part will maintain its shape in service when stressed. However, it is desirable to have materials deform easily when forming them by rolling, extrusion, and so on. Plastic deformation usually occurs as soon as the stress is applied. At high temperatures, however, time-dependent plastic deformation called creep may occur.

Fracture is the breaking of a material into two or more pieces. If fracture occurs before much plastic deformation occurs, we say the material is brittle. In contrast, if there has been extensive plastic deformation preceding fracture, the material is considered ductile.

Introduction

Tensile properties are used in selecting materials for different applications. Material specifications often include minimum tensile properties to ensure quality so tests must be made to guarantee that materials meet these specifications. Tensile properties are also used in research and development to compare new materials or processes. With plasticity theory (Chapter 5), tensile data can be used to predict a material's behavior under forms of loading other than uniaxial tension.

Often the primary concern is strength. The level of stress that causes appreciable plastic deformation is called its yield stress. The maximum tensile stress that a material carries is called its tensile strength (or ultimate strength or ultimate tensile strength). Both measures are used, with appropriate caution, in engineering design. A material's ductility may also be of interest. Ductility describes how much the material can deform before it fractures. Rarely, if ever, is the ductility incorporated directly into design. Rather, it is included in specifications to ensure quality and toughness. Elastic properties may be of interest, but these are measured ultrasonically.

Tensile Specimens

Figure 3.1 shows a typical tensile specimen. It has enlarged ends or shoulders for gripping. The important part of the specimen is the gauge section. The cross-sectional area of the gauge section is less than that of the shoulders and grip region, so the deformation will occur here. The gauge section should be long compared to the diameter (typically, four times).