By Dr Manas Kumar Haldar
“Impurities?” one might turn up one’s nose, “who wants them?” But as you might know from everyday life, impurities can be quite important. For example, when we introduce impurities in the form of spice to food, its taste may be greatly enhanced. But add too much spice and the taste can be horrible. So, one has to introduce the right amount of impurities. Control of “impurities” is important for food, but when it comes to electronics and communication engineering, the importance of this control cannot be overstated.
With all the electronic gadgets around us, it is not hard for one to appreciate that we live in the silicon age. Silicon has an interesting history. When scientists were exploring the electrical properties of materials it was very important to measure resistance of materials. To measure resistance, or more specifically resistivity, we apply a voltage to, say a block of a material of unit length and uniform unit cross-section, and measure the current. Then the resistivity is given by the voltage divided by the current.
When it came to measurement of resistivity of silicon, results varied from specimen to specimen even with the “highly pure” silicon produced in 1920s. As understanding of the process of electrical conduction improved, it was realized that extremely tiny amount of impurities as low as one impurity atom in a million (one million is one followed by six zeroes) silicon atoms can drastically change the resistivity of silicon. This led to the development of refining techniques to produce silicon with a few atoms of impurity in trillion (one million million, i.e. one followed by 12 zeroes) silicon atoms. Such highly pure silicon is employed as a starting point in the fabrication of integrated circuits which use diodes and transistors. Interestingly enough, to produce these diodes and transistors we need to have what is known as p and n type silicon. These are obtained by introducing two different types of impurities. So what was considered bad initially became very useful later. For proper operation of integrated circuits, the amount of impurity must be strictly controlled.
A similar story appeared more recently in the area of optical communications. Data and speech whizz around the world as light through optical fibres. To increase internet speed, optical fibres in some countries even connect to PCs at home. It is likely that in Malaysia, we will see them connecting our homes soon. Interestingly, the material of optical fibre is a relative of silicon – silicon dioxide or silica. We are quite familiar with this material called glass. Thinking of our glass windows, you will say that it is no wonder that glass is used for making optical fibres as it is “highly transparent”. But think again, are you right? Let us consider what we mean by transparent. It means that there is very little loss of light when it travels through a transparent material. But how little is little? For simplicity, consider the glass in your window to be 1mm thick, although it may be somewhat thicker. If 99% of the light passes through, your glass window may be considered to be transparent as only 1% is lost. Now consider 1m thick window glass to represent 1m length of an optical fibre. Then if we consider the power transmitted to reduce in proportion to the thickness, only 99/1000 or 0.099% of the light comes out because 1m equals 1000mm. In reality, the transmitted power decreases much faster with increase in thickness –mathematically it is called an exponential rate of decrease. If we calculate the transmitted power with such a rate of decrease, value will be about 0.004%. Even with our simple calculation of 0.099% of power transmitted, the one meter long fibre made of window glass can hardly be considered transparent. What is worse is that fibre lengths can be as large as 20km before light is amplified and retransmitted. Blame the impurities in glass for the loss. Hence, refining techniques for making highly pure glass were developed in the 1970s for making optical fibres. However, pure glass as such cannot be used to make optical fibres. The simplest optical fibre consists of a core and cladding. Light travels along the core suffering successive reflections at the boundary between the core and cladding. For this type of reflection (called total internal reflection) to happen, the refractive index of the cladding must be less than that the refractive index of the core.(For those who do not know, refractive index of materials determines how much a light ray bends when it travels from one material to another.) So pure glass cannot be used to make both core and cladding as the refractive indices will be the same. Now the refractive index of pure glass can be increased or decreased by adding small amounts of impurities. A small difference in the refractive indices of the core and cladding, say 1%, is good enough. As with silicon, the impurities were first considered undesirable but later put to use. Once again, the amount of impurities has to be controlled carefully and too much impurity is undesirable as it will increase loss.
Dr Manas Kumar Haldar is Associate Professor with the Faculty of Engineering, Computing and Science at Swinburne University of Technology Sarawak Campus. He is contactable at mhaldar@swinburne.edu.my