Zero bias detector diodes for the RF/ID market - HP's HSMS-285x - includes related article on backscatter RF/ID systems - Product Announcement

Hewlett-Packard Journal, Dec, 1995 by Rolando R. Buted

[MATHEMATICAL EXPRESSION OMITTED] where co = 27[eta]f, [C.sub.j] is junction capacitance, [I.sub.s] is diode saturation current, [R.sub.s] is the series resistance, and [R.sub.v] is the junction resistance.

Load Resistance. The diode resistance [R.sub.v] at zero bias is usually not small compared to the load resistance [R.sub.L]. If the diode is considered as a voltage source with impedance [R.sub.v] feeding the load resistance [R.sub.L], the voltage sensitivity will be reduced by the factor [R.sub.L]/([R.sub.L] [R.sub.v]), or:

[gamma]2 = [gamma]1 ([R.sub.L]/([R.sub.v] [R.sub.L])) For example, a typical load resistance is 100 k [omega]. If [R.sub.v] is 5 k [omega] then

[gamma]2 =[gamma]1 (0.952).

Reflection Loss. Further reduction in voltage sensitivity is caused by reflection losses in the matching circuit in which the diode is used. In Fig. 3, the package capacitance [C.sub.pkg] and package inductance [L.sub.pkg] can be used to determine the packaged diode reflection coefficient. If this diode terminates a 50 [omega] system, the reflection coefficient [rho] is:

[rho] = ([Z.sub.D] - 50)/([Z.sub.D] 50), where [Z.sub.D] is a function of frequency and the package parasitics. If there is no matching network, the voltage sensitivity can be calculated as:

[gamma]3 = [gamma]2(1 - [[rho].sup.2]).

The chip, package, and circuit parameters all combine to define an optimum voltage sensitivity for a given application. Our design goal was to develop a diode to operate in the frequency range used in RF/ID tags. Using the voltage sensitivity analysis described above, we hoped to produce an optimum, low-cost, manufacturable part in the shortest time possible.

Implementation and Fabrication

Hewlett-Packard's preeminent zero bias detector diode (HSCH-3486) already provides excellent detection sensitivity in an axially leaded glass package, particularly at high frequencies. To meet our design goals, the project team decided to leverage the HSCH-3486 technology.

We chose the plastic SOT-23 package because of its low manufacturing costs for high-volume products. Using the SOT-23 package, several modifications were possible that we hoped we could take advantage of Before building prototype devices, we made a detailed device model for the HSCH-3486. The model helped us fabricate an optimum device with minimum design iterations.

Two-dimensional process and device simulators were used to model and predict the performance of the HSCH-3486. Diode parameters such as silicon doping, area, epitaxial layer thickness, metal pad size, and passivation thickness were included to study the effects that these process parameters had on diode electrical performance and ultimately on detector performance. A typical sensitivity analysis (Fig. 4) showed the effect of contact area and epitaxial thickness on voltage sensitivity y3, assuming an ideal matching circuit. The model was also used to check for sensitivity to parameters that are not directly measurable during processing, such as surface states and recombination velocities. The model was good for trend analysis but could not be used to predict absolute values until devices were fabricated and tested.