Introduction

In 1958 it was discovered that mercury cadmium telluride alloys Hg1 -xCdxTe are semiconductors, with a bandgap which can be varied from approximately -0.3 to 1.6 eV as the variable composition x goes correspondingly from 0 to 1 [1], This has opened up the possibility of designing infrared detectors in mercury cadmium telluride tuned for the 3-5 μm or the 8-14 μm wavelength atmospheric window. The material has been thoroughly studied since its appearance and overviews of its properties have been published [2-4], In figure 1, the energy bandgap/cut-off wavelength is shown versus temperature for various values of x.

Methods of manufacturing mercury cadmium telluride material have evolved from bulk melt growth to liquid phase epitaxy (LPE) technology, vapor phase epitaxy (VPE) and metal-organic chemical vapor deposition (MOCVD) [5-7]. These new methods have made it possible to manufacture large two-dimensional focal plane arrays [8-11]

Designing infrared detectors and infrared focal plane arrays is however a demanding task. The task is not made easier by the fact that thermal background radiation at room temperature is many orders of magnitude greater in the infrared range than in the visible range.

Monolithic arrays, in which read-out means are integrated in the same mercury cadmium telluride chip as the infrared detectors are presented in part one of this book. The chapters in this part, chapter 1.1 to 1.5, correspond to the operation of the arrays. The arrays discussed are charge coupled device imagers [12-14], ambipolar drift field imagers, static induction transistor imagers, charge injection device imagers [12, 15, 16] and charge imaging matrices [17, 18], An example of a charge imaging matrice is shown in figure 2 [19].

A problem with the monolithic arrays is that the techniques for building metal-oxide-semiconductor (MOS) devices in silicon cannot be transferred intact to narrow bandgap materials such as mercury cadmium telluride, mainly due to tunneling and avalanche breakdown occuring at very low voltages. A monolithic array, in which read-out electronics is integrated in the same mercury cadmium telluride chip as the infrared detectors, is therefore difficult to achieve.

An alternative approach is to form an hybrid array [20]. The detector array is then formed in a mercury cadmium telluride substrate and the readout circuits are formed in a silicon substrate. The hybrid arrays are presented in part two. Detector arrays having detector elements which are provided with individual read-out leads formed in or on a non-active supporting substrate are presented in chapter 2.1. The detector elements may however be directly connected to a read-out chip which is bonded to the detector chip. A flip-chip bonding technique using indium bumps may be used as shown in figure 3 [21].

These detector arrays, as well as detector arrays for which no connectors at all are disclosed, are presented in a chapter 2.2. Sections for cross-talk preventing measures and passivation and leakage current preventing measures are comprised in this chapter.

When a silicon read-out chip and a mercury cadmium telluride detector chip are bonded together by the flip-chip bonding process, the chips are exposed to a mechanical stress which may lead to damage, especially of the fragile detector chip. Damage may also occur due to mechanical stress generated when the temperature of the detector array is changed. Due to the small energy bandgap of mercury cadmium telluride, the detector array is normally cooled to cryogenic temperatures during operation to reduce thermally generated noise. The temperature cycles which are generated when the array is cooled from room-temperture to an operational temperature of for example 77 K create mechanical stress in the array due to the difference in thermal expansion coefficients between silicon and mercury cadmium telluride. The coefficients of thermal expansion for silicon and mercury cadmium telluride are 1.2 × 10- 6 m/mK and 3.8–4.5 × 10- 6 m/mK, respectively. Flip-chip arrangements for mercury cadmium telluride detector arrays are presented in chapter 2.3. The chapter comprises a section for thermal stress preventing measures.

A detector array having a more compact structure is formed by using the Z-technology [22-24]. These detector arrays are presented in chapter 2.4. An example of such an array is shown in figure 4 [25].

In another approach to form a detector array a mercury cadmium telluride detector chip is directly contacting or directly glued to a read-out chip. These detector arrays are presented in chapter 2.5. An advantage is that the assembly of mercury cadmium and read-out chip may be processed in a silicon-like fashion. A separate section is provided for detector arrays having through holes formed in the mercury cadmium telluride through which the detector elements are connected to the read-out chip [26]. Such a detector array is shown in figure 5 [27].