Chapter 14.4 - Types of Photodiodes

14-4. TYPES OF PHOTODIODES

PIN Photodiode

The analysis of the transit time response in the previous section assumed that the photoex-cited electron-hole pairs are created in the depletion region, where there is a strong E field. This is not always the case, however, as illustrated in Fig. 14-1 1a. Some light that is incident on the highly doped p⁺ side passes completely through the depletion region, and is absorbed in the n-type region on the other side. In this latter region, the E field is very small, because the high free-carrier concentration makes the electrical conductivity high (and E is small inside a good conductor). According to Eq. (13-9), there is little contribution to the photocurrent when E is small. Therefore, the photoexcited charge carriers created outside the depletion region do not contribute significantly to the photocurrent, as long as they remain outside the high-field depletion region.

The charge carriers created outside the depletion region do not remain motionless, however. Like any particles subject to random thermal motion, they spread out from their initial position in a process known as diffusion. Diffusion proceeds much more slowly than drift, which is the term given to the directed motion induced by an electric field. After spreading out by diffusion for a certain time, some holes initially generated in the n-type region will enter the depletion region, where they are quickly swept across by the high E field there. This results in an additional component to the photocurrent, delayed by the diffusion time. Holes generated at different distances from the edge of the depletion region have different diffusion times, which results in a diffusion "tail" in the photocurrent response to a square-wave light pulse. This type of signal distortion, depicted in Fig. 14-11b, is generally undesirable.

The solution to the problem of diffusing charge carriers is to simply eliminate the diffusion region. This can be accomplished by decreasing the donor concentration in the n

Figure 14-11 (a) In a simple p-n junction photodiode, charge carriers may be created in a high-field drift region or a low-field diffusion region. (b) Charge carriers created in the diffusion region give rise to a distortion in the photocurrent waveform.

region until the depletion region occupies nearly the entire space between the electrodes. As shown in Fig. 14-12, the E field then extends nearly all the way to the far electrode, so charge carriers generated anywhere in the material will be subject to drift rather than diffusion. At the far end is a thin, highly doped n region, needed to make good ohmic contact with the electrode. Since the middle region is very lightly doped (nearly intrinsic), it is labelled I, and the device is termed a PIN photodiode.

The PIN photodiode is the most commonly used photon detector today. It not only eliminates carrier diffusion, but also has the advantage that the depletion width d is fixed by the geometry of the device. The ability to adjust d by design, rather than applied voltage, allows the photodiode's performance to be optimized for specific applications. For example, making d larger increases the path length for absorption of light, which increases the efficiency _abs with which light is absorbed. This is especially important for wavelengths near the semiconductor's bandgap, where . In this regime, Eq. (13-15) gives

On the other hand, a larger d degrades the time response by increasing the transit time. In Eqs. (14-31) and (14-32), d should now be considered a constant, independent of applied voltage. The transit time is therefore minimized by increasing E to the saturating value E_s.

Figure 14-12 (a) In a PIN photodiode, charge carriers are mostly created in the high-field drift region, which extends almost to the far electrode. The lightly n-doped "intrinsic" region has a nearly constant E field, which sweeps charge carriers through the device without diffusion.

EXAMPLE 14-3

A silicon PIN photodiode has an intrinsic region of thickness 0.1 mm. Determine the minimum rise time for the detector, its corresponding bandwidth, and the required reverse-bias voltage. Repeat for an intrinsic region of thickness of 10 m.

Solution: The minimum transit time for holes in silicon is

Taking this as the rise time, the corresponding bandwidth is

The required E field is E = E_s2 × 10⁶ V/m, so

which is an inconveniently high voltage. Repeating the calculation for d = 10 m gives t_tr = 0.1 ns, B = 3.5 GHz, and V= 20 V. This is a much improved time response, and occurs with a more convenient bias voltage.

The above example shows that in terms of transit time, a thinner intrinsic region is preferable. However, if d is made too small, capacitive effects become important. There is, therefore, an optimum value of d that minimizes the overall response time (see Problem 14.8). Another problem with small d is illustrated by the following example.

EXAMPLE 14-4

For the silicon PIN photodiodes of Example 14-3, determine the absorption efficiency for 860 nm light. At this wavelength, the absorption coefficient in silicon is 335 cm ¹and the reflectivity (from air) is 32%.

Solution: For d = 0.1 mm,

ad = (335 cm^-1)(10^-2 cm) = 3.35

Eq. (13-15) then gives

Note that this neglects light reflected back from the far end of the Si material, so it somewhat underestimates _abs. Repeating the calculation for d = 10 m gives d = 0.335, so

The thinner intrinsic region is seen to be less efficient at absorbing the incident light. There is, therefore, a trade-off between detector speed and sensitivity. A PIN photodiode can be optimized for either of these, depending on the application.

< Previous Excerpt Next Excerpt >

Purchase This Book

Preface

TABLE OF CONTENTS

Chapter 14.4 - Types of Photodiodes

Preface

TABLE OF CONTENTS

Chapter 14.4 - Types of Photodiodes

Contact Preferences

This is embarrasing...

Customize Your GlobalSpec Experience

Select Your Free Newsletters

Industry Newsletters

Select Your Free Product Alerts

This is embarrasing...