Impact of high DC bias on RF LDMOS reliability for radar application M.Gares (1) , H.Maanane (1) , M.Masmoudi (1) , P.Bertram (2) , J.Marcon (1) , K.Mourgues (1) and Ph.Eudeline (2) (1) LEMI, University of Rouen, IUT Rouen, 76821, Mont Saint Aignan, France. (2) THALES Air Defence, ZI du Mont Jarret, 76520 Ymare, France. Abstract—The reliability of power RF LDMOS submitted to RF and DC accelerated tests under high temperature has been investigated by DC and dynamic electrical characterization. After DC and RF ageing, some electrical parameters changed, with a significant drift of the feedback capacitance (C RS ). In order to explain qualitatively electrical parameter shifts, a 2D RF LDMOS structure was implemented and simulated with ATLAS- SILVACO. Simulation results tend to show that, for a given bias conditions, all conditions are set for hot carrier degradation. I. INTRODUCTION DMOSFET are widely used in RF power applications. In the radar field, users are concerned by the reliability of RF LDMOSFET submitted to RF pulses with high drain-source DC bias for maximum output power with elevated base plate temperature conditions. In order to apply all those strong requirements for RF LDMOS reliability assessment, we designed and implemented a reliability bench [1] as shown on Fig.1. This bench allows to keep track of up to 8 devices’ performances under variable bias voltages and temperatures under RF pulse conditions. Therefore, the transistor is constantly biased under high DC voltage while delivering RF power only during the RF pulse. Consequently, it’s crucial to understand the effects involved by both high DC bias and high temperature during the time the transistor is not operating. That’s the reason why we carried out a DC accelerated test. This paper also attempts to understand the physical degradation mechanisms induced by RF and DC life- tests. II. LIFE-TEST CONDITIONS The discrete RF device 1 has been submitted during 1000 hours to RF life-test on the reliability bench (see Fig. 1), with the following conditions: Frequency =2.9 GHz Pulse width/ Duty cycle= 500µs/50% Drain voltage= 44V, Device Base plate Temperature=110°C, PIN = 30.5 dBm, POUT ≈ 43 dBm. Device 2 has been aged during 1000 hours with the same settings (bias, temperature) as device 1, but with no RF input power applied The permanent current value is I DS =7mA. Such a weak value corresponds to the quiescent current also set to device 1 without RF power (out of pulse current). III. ELECTRICAL CHARACTERIZATIONS OF THE DEVICE AFTER AGEING TESTS Our experiments are based on the comparison between devices, before and after ageing, for the two kinds of life-test considered. For this reason, both transistors are characterized in static and dynamic mode in order to extract a critical electrical parameter set. This work should allow us to correlate RF and DC ageing to any electrical parameter drift. I-V and C-V measurements were performed at ambient temperature, respectively by an E5270 DC analyzer and a HP 4194 impedance analyzer, derived by a commercial software package IC- CAP [2]. A. DC parameter evolution The most significant result observed after the RF life- test experiment is an increase of the drain-source on-state resistance (obtained by an output characteristics extrapolation with V DS =[0,2V] and V GS =[4,10V]) from 1,31 Ohms to 1,63 Ohms (+24%), see Fig. 2. L Computer Pulse Generator Pref Pin Tuner 40 dB Attenuator DUT Average Power meter Peak Power meter Pout Att X 8 Switchs RF 8 > 1 20 dB Attenuator 30 dB Coupler Switchs RF 8 > 1 100W HPA Computer Pulse Generator Pref Pin Tuner 40 dB Attenuator DUT Average Power meter Peak Power meter Pout Att X 8 Switchs RF 8 > 1 20 dB Attenuator 30 dB Coupler Switchs RF 8 > 1 100W HPA Computer Pulse Generator Computer Pulse Generator Pref Pin Tuner 40 dB Attenuator DUT Average Power meter Peak Power meter Pout Att Att X 8 Switchs RF 8 > 1 20 dB Attenuator 30 dB Coupler Switchs RF 8 > 1 100W HPA Fig. 1. Synoptic of a RF power reliability bench in pulsed mode Fig. 2. R DSON evolution after 1000h ageing, for device 1 4 5 6 7 8 9 10 0 2 4 6 8 10 R DSON before ageing R DSON after ageing R DSON [Ohms] V GS [V]