Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 15

3-1

Chapter 3
Basics

A Few important definitions and terms Wavelength Unlike electrical engineers who use frequency, opticians use wavelength. When nothing is specified, the wavelength refers to free space wavelength,

0 = c/f f = c/ 0
where c is the velocity of light in free space ( 3 10 8 m / s ) and f is the frequency.

(3.1)

Example 3.1: Consider a light emitting diode emitting a wavelength of 800 nm. The corresponding frequency is 3 10 8 /(800 10 9 ) Hz = 375 THz (1 THz = 1012 Hz) a very high frequency indeed. Coherence A wave in time and space has a variation in space and time as e j( t z ) . If it is a perfect sinusoidal function of time of frequency, rad/s all the time, it is called perfectly coherent. This is depicted in Fig.3.1 (a). There is of course no perfectly coherent source. t , as In practice, the source may emit a burst of a pure sine wave over a period of time, shown in Fig,3.1(b). Then this time period is called the coherence time. After this time, the source may emit the same frequency but with an abrupt change in phase over another t . If we split light from the source into two parts with equal power, let length of time them travel along different paths and combine them, the intensity of the combined light

3-2 will be between dark and very bright depending on whether the path difference is an odd or even multiple of the wavelength this the well known interference of light. However, interference giving dark and bright spots is possible for coherent light. If the source is not t , or corresponding to a path c t in perfectly coherent, i.e., it is coherent over time free space, the interference phenomenon can be obtained (in free space) only if the path t is called the coherence length difference does not exceed this length. The length c length and is illustrated in Fig.3.1.b. The time variation of an incoherent light, such as white light is random as shown in Fig.3.1c. The coherent characteristic of light can also be seen in the frequency domain (blue RHS plots in Fig.1.1. The amplitude variation in the frequency domain is called spectrum. Thus, a perfectly coherent light will have a single line spectrum as shown in Fig.1.1a, a partially coherent light will have a band limited spectrum as shown in Fig.1.1b, and completely incoherent light will have a flat spectrum as shown in Fig.1.1c. You are familiar with the last one in electronics as the spectrum of white or Gaussian noise. In fact, the term white noise comes from optics where it is the spectrum of white light.
Amplitude

(a)

P Time Amplitude t l = ct Q

(b)
Time Field Space

= 1/t

(c)
P

Q Time

Amplitude

(a) A sine wave is perfectly coherent and contains a well-defined frequency o. (b) A finite wave train lasts for a duration t and has a length l. Its frequency spectrum extends over = 1/ t. It has a coherence time t and a coherence length l. (c) White light exhibits practically no coherence.
1999 S.O. Kasap,Optoelectronics (Prentice Hall)

Fig.3.1

3-3

Spectral width of a source We just saw that no source is perfectly coherent. Hence they are characterized by their spectral width. If the spectrum is plotted as power against wavelength the difference in wavelengths of the points at which the power is half the peak power is called the spectral width (also called full width at half maximum or FWHM in optics). Again this is familiar to electrical engineers who call it the 3dB bandwidth.

Fig.3.2 : represents optical power. 0 is the nominal or mean wavelength and 1 / 2 is the spectral width (FWHM) In optics, the spectral width is expressed in wavelengths, while in electrical engineering, we are more familiar with the quantity expressed in Hz.. The relationship between spectral width, f1 / 2 in Hz and spectral width, 1 / 2 in wavelengths can be easily obtained by differentiating (3.1). 1 / 2 = (c / f 2 ) f1 / 2 (3.2) 2 f1 / 2 = (c / 0 ) 1 / 2 Example 3.2 The light source in example 3.1 has a spectral width of 30 nm. Its spectral width in Hz is
3 10 8 30 10 9 Hz = 14.06 THz. (800 10 9 ) 2

Note that the frequency was 375 THz. So the spectral width is 3.75% of the nominal frequency (its the same in terms of wavelength, i.e., 30/800). We will see later that this width is not good for a high speed optical communication system. Refractive index If v p is the phase velocity of light in a medium, the refractive index, n of the medium is defined as:

3-4
n= c vp

(3.3)

where, c is the velocity of light in free space. In an unbounded medium, light as an electromagnetic wave travels as a TEM wave. Recall from electromagnetics that its velocity is given by v p =
1 0 r 0 r

r and r are called relative permittivity and relative permittivity respectively. r is also

called the dielectric constant. In free space, r = 1 = r and the velocity is c, while in a dielectric, r = 1 and the velocity is v p . So c = and v p = 0 0 0 0 r Hence
1 1

n = r r = n 2
Example 3.3

(3.4)

We must be careful to note that dielectric constant varies with frequency. For example, at low frequencies, the dielectric constant of water is about 80. At optical frequencies (around 800 nm), the refractive index of water is about 1.3 and so its dielectric constant is 1.69 ! Reflection and transmission coefficients In electromagnetics, we are familiar with voltage reflection coefficients Voltage reflection coefficient, = Reflected wave voltage/Incident wave voltage Voltage transmission coefficient, = Transmitted wave voltage/Incident wave voltage. From the continuity of voltage, Incident wave voltage + Reflected wave voltage = Transmitted wave voltage 1+ = Or, (3.5) In optics, the reflection and transmission coefficients refer to power unless stated otherwise. This can be a source of confusion while reading literature in optics. coefficient Power reflection coefficient, R= Reflected wave power/Incident wave power Power transmission coefficient, T= Reflected wave power/Incident wave power From the conservation of power, Incident wave power - Reflected wave power = Transmitted wave power

3-5 Or, (3.6) Relationships between the coefficients Suppose that the wave is traveling from medium 1 to medium 2. Let the intrinsic (or characteristic) impedance of the two media be 1 and 2 respectively.

1 R = T

( reflected wave voltage) 2 / 1 R= Reflected wave power/Incident wave power= (incident wave voltage) 2 / 1
(reflected wave voltage) 2 = (incident wave voltage) 2

Or, (3.7)

( transmitted wave voltage) 2 / 1 T= Transmitted wave power/Incident wave power= (incident wave voltage) 2 / 2
( reflected wave voltage) 2 2 = (incident wave voltage) 2 1

Or, T = Decibels

2 1

(3.8)

Communication engineers measure loss and gain in decibels. Gain in dB= 10 log10
Power output Power input

This will be a positive number for an amplifier because the power output is greater than the power input. But when the device has loss, this number will be negative, because power output is less than the power input. Now nobody likes negative numbers (like negative bank balance!). So we express attenuation in dBs as Attenuation, in dB = 10 log10
Power input Power output

(3.9)

which will be a positive number. Thus for a lossy device, power output =

10 in in dB / 10 :power input

(3.10)

3-6

Nepers Although decibel is widely used, it does not come naturally in the formulae of physics. This is just like degree which is widely used as measure of angle but the angle measure that comes in formulae is radians. In a lossy medium, fields (voltages and currents in transmission lines) decay exponentially as e per unit length, where is called the attenuation constant in Nepers per unit length and z is the direction of propagation. As power is proportional to square of the field, it decays as e 2 per unit length. Then from the definition of attenuation in decibels, in dB = 10 log10 e 2 in Nepers = (20 log10 e) in Nepers i.e., in dB = 8.6859 in Nepers (3.11) Thus there are more decibels to a Neper(~8.7 decibels for a Neper) just as there are more degrees for a radian(~57.3 degrees for a radian) or just as there are more sens for a Ringgit (how many for a Ringgit?). dBm, dBW These are units of power often used in fiber optic communications. 1 dBm= 10 log10(Power/1 mW) Thus dBm is the number of dBs by which the power exceeds 1 mW. I find it more convenient to remember 1 dBm= 10 log10(Power in mW) Hence if the power is given as x dBm, Power in mW= 10x/10 mW Similarly, dBW is defined relative to powers of 1W . 1 dBW = 10 log10(Power/1 W) = 10 log10(Power in W) if the power is given as x dB W, Power in W = 10x/10 W (3.12) (3.13) (3.14) (3.15) (3.16)

Power in dBm or dBW is very convenient when gain or loss has to be taken into account. Example 3.4 (a) What is the power output of an amplifier with a gain of 10 dB if the power input is 2dBm? (b) If the power output and input of a device are -8 dBm and -2 dBm respectively, what is the attenuation of the device?

3-7 (a) Power output in dBm=Power input in dBm+ Gain in dB=2 +10=12dBm ( =10 12/10 mW= 15.8 mW) (b) Attenuation in dB= Power input in dBm Power output in dBm = -2-(-8) = 6 dB Note powers in dBm can not be added to get total power. Example 3.5 Two devices have power outputs of 20 dBm each. What is the total power output? (The devices could be loudspeakers each with audio power output of 20 dBm) The power supplied by each device is 1020/10 mW or 100 mW. Hence the total power supplied by two devices = 200 mW = 10 log10 (200 mW) = 23 dBm If we add powers in dBm, we get the wrong answer 40 dBm.

3-8

Optical fiber basics How does a fiber work?

Fig.3.2 An optical fiber used in fiber optic communication consists of two different types of highly pure, solid glass (mainly silicon dioxide), composed to form the core and cladding. A protective acrylate coating then surrounds the cladding. This coating protects the glass from dust and scratches that can affect fiber strength. The operation of an optical fiber is based on the principle of total internal reflection. Light reflects (bounces back) or refracts (alters its direction while penetrating a different medium), depending on the angle at which it strikes a surface. That reflection is usually caused by creating a higher refractive index in the core of the glass than in the surrounding cladding glass, creating a waveguide. Usually, The refractive index of the core is increased by slightly modifying the composition of the core glass generally by adding small amounts of a dopant, while the cladding is pure glass. Alternatively, the waveguide can be created by reducing the refractive index of the cladding using different dopants. How do dopants change the refractive index of glass? Fig.3.3 shows the refractive index against the percentage of dopants added to silicon dioxide (glass). From the diagram, it is seen that the refractive index can be increased or decreased by very small amounts by the addition of the dopant. Noting that the core must have higher refractive index, some of the possibilities are for example: GeO2- SiO2 core and SiO2 cladding P2O5- SiO2 core and SiO2 cladding SiO2 core and B2O3- SiO2 cladding.

3-9 GeO2 -B2O3- SiO2 core and B2O3- SiO2 cladding

Fig.3.3(Fig.4.6 Senior) Effect of various dopants on the refractive index of glass Total Internal Reflection (TIR): Basic physics

3-10 Fig.3.4(Fig.2.2 Senior): Light rays incident on a high to low refractive index interface (a) refraction (b) limiting case of refraction showing the critical angle c (c) total internal reflection where > c Consider a ray of light traveling from a medium of higher refractive index n 1 to a medium of lower refractive index, n 2 . Snells law: n 1 sin 1 = n 2 sin 2 (3.17)

As the angle of incidence, 1 increases, the refracted ray moves away from the normal till as in Fig.3.4(b), it is along the interface, i.e., 2 = 90 0 . The corresponding angle of incidence is called the critical angle, c . From (3.17), n sin c = 2 (3.18) n1 If the angle of incidence exceeds this value, the ray is turned back. No light is transmitted in the medium of lower refractive index and light is described to have suffered a total internal reflection.. Example 3.6 For the glass-air interface, n1 = 1.5, n2 = 1.0, and the critical angle is given by
c = sin 1 1 = 41.8 0 1.5

On the other hand, for the glass-water interface, n1 = 1.5, n2 = 1.33, and
c = sin 1 1.33 = 62.5 0 1.5

If light can be guided by a material surrounded by a material of lower refractive index, why not just use glass and air? What is the purpose of the cladding in the fiber? For transmission of light from one place to another, the fiber must be supported. Supporting structures, however, may considerably distort the fiber, thereby affecting the guidance of the light wave. A sufficiently thick cladding. protects the total-reflection surface from contamination. Further, in a fiber bundle, in the absence of the cladding, light can leak through from one fiber to another. The idea of adding a second layer of glass (namely, the cladding) came (in 1951) from OBrien (USA) and Van Heel (Holland).

3-11

The Numerical Aperture (NA) or Acceptance Angle of a fiber

(a)

(b)

Figure 3.5 (a) A glass fiber consists of a cylindrical central core clad by a material of slightly lower refractive index. (b) Light rays impinging on the core-cladding interface at an angle greater than the critical angle are trapped inside the core of the fiber.

For a ray entering the fiber core at its end, if the angle of incidence at the internal corecladding interface is greater than the critical angle c [= sin-1 (n2/n1)], the ray will undergo TIR at that interface. Further, because of the cylindrical symmetry in the fiber structure, this ray will suffer TIR at the lower interface also and therefore be guided through the core by repeated total internal reflections. Even for a bent fiber, light guidance can occur through multiple total internal reflections Consider a ray that is incident on the entrance face of the fiber core, making an angle i with the fiber axis. Let the refracted ray make an angle with the same axis. Assuming the outside medium to have a refractive index n0 (which for most practical cases is unity), we get from Snells law (3.17) sin = (n 0 / n 1 ) sin i (3.19) The critical angle is given by (3.10) as sin c = n 2 / n 1 and hence
cos c = 1 ( n 2 / n 1 ) 2

In order for the light to travel by TIR, > c . But since = 90 0 , 90 0 > c or < 90 0 c , or sin < sin(90 0 c ) = cos c = 1 ( n 2 / n 1 ) 2

3-12 Hence from (3.11)


( n 0 / n 1 ) sin i < 1 ( n 2 / n 1 ) 2 ,
2 Or, n 0 sin i < n 1 n2 2 The maximum value of i for which light can be guided along the fiber is

(3.20) Only the cone of light formed by this angle can be accepted by the fiber and guided along it. Any light outside this cone is radiated out of the fiber. Hence the angle i defined by 3.20 is called that acceptance angle. The numerical aperture (NA) is defined as: 2 NA= n 0 sin i = n 1 n2 (3.21) 2 It is convenient to define a quantity 2 n1 n2 2 = (3.22) 2 2n 1 (n 1 n 2 )(n 1 + n 2 ) (n 1 n 2 ) 2n 1 As n 1 n 2 , = 2 2 2n 1 2n 1 n1 n 2 Or, (3.23) n1 Thus can be viewed as the fractional refractive index difference between the core and the cladding (expressed as a fraction of the core refractive index). From (3.21) and (3.22), NA= n 1 ( 2)1 / 2 (3.24)
2 n 0 sin i = n 1 n2 2

Example 3.5 For a typical fiber with n1 = 1.45 and = 1%, we get
NA = n 1 2 = 1.45 2 (0.01) = 0.205

If the medium through which light is accepted is air so that n 0 = 1 , the acceptance angle Typically for commercial fibers glass NA = 0.29 plastic NA~ 0.5 Skew rays and acceptance angle In the analysis of acceptance angle, we considered a ray passing through the fiber axis. Such rays are called meridional rays. However there are rays which may not pass through the axis. Such rays are called skew rays. An example is shown in Fig. 3.6. Such rays are accepted at larger axial angles than meridional rays. Moreover the acceptance angle obtained for meridional rays is the minimum input angles for skew rays. The meridional rays tend to propagate only in an annular region near the outer surface of the core. So these rays actually increase the light gathering by the fiber and could be important for large NAs. The analysis is more involved (see Senior) and is hardly justified as for most purposes, the expressions for meridional rays are adequate for design.
i = sin 1 0.205 = 11.830

3-13

Fig.3.6(Fig.2.7 Senior): Ray path within the fiber core for a skew ray A more rigorous explanation of light guidance using plane waves So far, we considered only a simple ray pictures. From this picture, it appears that for total internal reflection, light is reflected from the interface of the core and cladding and there is no light in the cladding. This picture is incorrect because some light does penetrate the cladding. This can be understood from the wave picture. For the ease of reading, Fig.3.4(a)is redrawn as Fig.3.7. The plane of incidence is taken as the y-z plane. Each ray represents the direction of propagation of a uniform plane wave.

Fig.3.7 The phase velocities in the two media are c / n1 and c / n 2 . The propagation constant is /phase velocity, where = 2f is the angular frequency in radians/s and f is the frequency. Then the propagation constants in the two media are 2f /(c / n1 ) and 2f /(c / n 2 ) or n1[2/(c / f )] and n 2 [2 /(c / f )] , i.e. n1k 0 and n 2 k 0 , where

3-14
k 0 = 2 / 0 is the propagation constant of free space (note c/f = 0 is the free space . r where r is in the wavelength). Each ray is associated with fields that vary as e j direction of the ray. y y cos 1 + a z z sin1 ) For the reflected ray, = n1k 0 (a
y y cos 2 + a z z sin2 ) . For the refracted ray, = n 2 k 0 (a y and a yy +a zz z are unit vectors in y and z directions and r = a Where a From Snells law, sin 2 = (n1 / n 2 ) sin 1

and hence, cos 2 = 1 (n1 / n 2 ) 2 sin 2 1 So the fields of the reflected ray vary as e jn1k 0 cos 1 ( y ) e -jn1k 0 sin 1 z and the fields of the refracted ray vary as
e jn 2 k 0
1 ( n1 / n 2 ) 2 sin 2 1 y

(3.25)

e jn 2 k 0 ( n1 / n 2 ) sin 1 z

(3.26)

The critical angle (for incident ray is given by sin c = n 2 / n1 What happens for the condition of wave guiding, 1 > c ? For 1 > c , sin 1 > sin c = n 2 / n1 So (n1 / n 2 ) 2 sin 2 1 > 1 and 1 (n1 / n 2 ) 2 sin 2 1 < 0 , i.e., negative. Hence 1 (n1 / n 2 ) 2 sin 2 1 is an imaginary quantity, say jd , where d is real. From (3.26), the fields of the refracted ray vary as e n 2 k 0 d y e jn 2 k 0 (n1 / n 2 ) sin 1 z This means that the fields in medium 2 (with refractive index n 2 ) are decaying in the y direction but are propagating in the z direction. On the other hand, referring to (3.25), the reflected ray is propagating in both y and z directions. Of course, the reflected ray has to be reflected back from the opposite side for wave guiding (see Fig.3.5b). Thus we see, that when the angle of incidence exceeds the critical angle, (i)the fields in medium 1 (with refractive index n1 ) travel in the z direction and oscillate (sinusoidal) in the y-direction. (ii) the fields in medium 2 (with refractive index n 2 ) are decaying in the y direction, but propagate in the z-direction. It is not true that there is no field in medium 2 when the angle of incidence exceeds the critical angle. Whatever the structure of an optical fiber, we except the fields in the core (medium 1) to be oscillatory along the cross-section and propagating in the axial (z) direction, while the fields in the cladding (medium 2) decay along the cross-section but travel in the axial direction. This concept will be used when we come across a more realistic theory of the step index fiber in Chapter 4.

3-15 The consideration given here is simplistic. A more rigorous (and hence more mathematical) analysis of uniform plane waves incident at the boundary between two dielectrics can be found in books. It was also covered in your electromagnetics course. For the present course, the simple explanation is enough.

You might also like