Variation Diode I-V Characteristics with Band Gap of the Semiconductor Material

Aim of the Experiment To determine whether the forward I-V characteristics of a diode depends on the band gap of the semiconductor material with which the diode is made of. Methodology To do this experiment, we need: 1. A simple method to identify that the band gap of diode A is dierent from diode B 2. To easily determine the band gap of a given diode We will now show that Light Emitting Diodes (LEDs) satisfy both criteria as mentioned above. A Brief Introduction to p-n Junction Diodes A p-n junction is obtained when a p (hole) doped semiconductor is brought into metallurgical contact with an n (electron) doped semiconductor. Figure 1 shows an energy diagram for a p-n junction at zero bias. At the interface there is a depletion region sandwiched between two neutral regions- p and n respectively. The electron carrier concentration in the neutral n region is given by the shallow dopant concentration ND . Similarly, the hole carrier concentration in the neutral p region is given by the acceptor concentration NA . depletion region p-type n-type

EC EF EV Figure 1: Energy band diagram of unbiased p-n junction Focusing on the n side, application of a forward bias results in hole injection into the neutral n region. The conductivity of the n region is increased by minority carrier injection (holes are minority carriers in N and vice versa in the P region). (As an aside, this is what makes semiconductors special- by application of voltage, the conductivity of the material can be changed and can be changed by modulating the majority or minority carrier density depending on the device architecture.) A similar argument is valid for holes. Again focusing on the n side, on application of forward bias, the hole concentration on the n side increases exponentially with applied voltage in an ideal diode. This enables the electrons to recombine 1

with holes. If the recombination is radiative, then it is accompanied by light emission. We see from this argument that the energy of the light emitted should be equal to the band gap. So materials with dierent band gaps will emit light of dierent frequency and hence dierent colour. At one stroke, we have solved the needs 1 and 2 (methodology above) for the experiment. 1. By choosing dierent coloured LEDs, choice of diodes with materials of dierent band gap is easily made. 2. The peak emission wavelength of the LED is a measure of the band gap i.e. Eg = 1240 (1)

where Eg is the band gap in electron Volts (eV) and is the emission wavelength in nanometers (nm). We should emphasize that this simple picture is exact for ideal diodes- which are only grown in text books! Figure 2 shows the spectra of dierent coloured LEDs driven at dierent current levels. As expected, the intensity of light emission increases with current as minority carrier injection increases. Notice that the white LED shows two wavelengths. Why? I-V Characteristics of an Ideal Diode The I-V characteristics of an ideal diode is given by the equation ID = I00 e kT (e
Eg qVD kT

1)

(2)

Where VD and ID indicate voltage across the diode and current through the diode respectively. The Eg saturation current IS is given as IS = I00 e kT . Assuming qVD kT , equation (2) can be rewritten in log form as Eg ID qV ln + = (3) I00 kT kT Thus, for constant ID /I00 , VD is proportional to Eg . If I00 does not vary very much from material to material, then for a constant ID , VD will increase as the band gap increases. Hence one way to test equation (3) is to determine VD for a constant value of ID for LEDs of dierent colours and plot VD v/s Eg of the LED obtained from its emission spectrum and look at the correlation. This precisely, is the experiment. Does the Value of ID Chosen Matter? Equation 2 is valid for ideal diodes, which are grown in text books. Real diodes use technology for both materials growth and processing. We list below some reasons for the non-ideality of real diodes. (a) In an ideal diode, the applied voltage is assumed to drop entirely across the depletion region. In a real diode, the neutral region has a nite resistance. There is hence a non-zero voltage drop across the neutral region which can only be inferred and not measured. Clearly this is important for large currents. (b) At low currents there is a departure from ideal behavior. This arises due to defects in the material which gives rise to generation currents rather than diusion current- which requires a modication of (2).

(a) Red LED

(b) Blue LED

(c) Green LED

(d) White LED

Figure 2: Emission Intensity v/s Wavelength of various LEDs for dierent currents (c) What is important is the current density J0 . Dierent diodes used here may have dierent areas (A) and hence the actual current i.e. J0 A may vary. Thus, it is important not to choose a very low current (reason (b)) or a very high current (reason (a)). A possible choice could be 0.5 mA ID 1 mA. The Experiment 1. You will be given four LEDs- red, green, blue and white. Use the circuit shown in Figure 3. 2. Vary the 1 k pot to accurately measure ID with change in VD . In the 0.5 mA ID 1 mA range, measure VD for ID = 0.5, 0.6, 0.7, 0.8, 0.9 and 1 mA. 3. Repeat steps 1-2 for all LEDs.

100

10 V +

1 k

VD

ID 100

Figure 3: Circuit to determine LED I-V characteristics Post Experiment Work- Obtaining Results and Interpreting Them 1. Plot a graph of ID v/s VD for all LEDs. Let us call this as Plot 1. Since there are multiple plots, you are free to use any plotting software such as GNUplot/Scilab/MATLAB/MS Excel etc. 2. Now plot a graph of ln ID v/s VD for all LEDs. Call this Plot 2. Recall from Experiment 1 that the slope of the graph is given by 1 ln ID2 ln ID1 = (4) VD 2 VD 1 VT Calculate the ideality factor of each LED from the slope. Also calculate the saturation current IS from the y-intercept. 3. Calculate the bandgap Eg for each LED using the emission wavelengths from Figure 2 and putting them in equation (1). 4. From Plot 1, choose a constant value of ID , say 1 mA. For each LED, nd out the value of VD corresponding to ID = 1 mA. 5. Refer back to the data you obtained in Experiment 1, for silicon diode 1N914. Find out the value of VD corresponding to ID = 1 mA. Assume that for silicon, Eg = 1.1 eV. 6. Now plot a graph of VD v/s Eg for the LEDs as well as 1N914. For the chosen value of ID , you should get one point (VD , Eg ) on the graph for each diode and hence you can plot all ve points (for the dierent diodes) on a single graph. 7. From the graph, try to nd a relation between VD and Eg . What is the expected correlation? Do you observe any variation practically? If yes, why? 8. In addition, answer the following questions. (a) Google about a white LED to study what it is made of. This will help you to answer (b). (b) What value of Eg will you choose for the white LED? (Hint: Look at the spectrum closely. Which is the stronger emission wavelength?) 9. Additionally you may want to have some fun and do the following. (a) From Plot 2, see whether equations (2) or (3) are satised for the entire range of VD . (b) Look at the correlation between VD and Eg by choosing a current level like 50 A and/or 5 mA and see how non-ideality of the I-V aects the experiment.