The effect of irradiation on the characteristics of MOS transistors

Technical Physics Letters, Vol. 30, No. 5, 2004, pp. 385–388. Translated from Pis’ma v Zhurnal Tekhnicheskoœ Fiziki, Vol. 30, No. 9, 2004, pp. 73–81. Original Russian Text Copyright © 2004 by Bormontov, Levin, Gitlin, Men’shikova, Tatarintsev. The Effect of Irradiation on the Characteristics of MOS Transistors E. N. Bormontov, M. N. Levin, V. R. Gitlin, T. G. Men’shikova, and A. A. Tatarintsev Voronezh State University, Voronezh, Russia Received August 12, 2003; in final form, December 22, 2003 Abstract—We have experimentally studied the effect of X-ray radiation on the parameters of MOS transistors. An analysis showed that correct evaluation of the density of surface states and the gate insulator charging by method of subthreshold current-voltage characteristics requires taking into account the planar inhomogeneity of a transistor. Some complication of the method is compensated by the increasing accuracy of determination of the surface parameters and the additional possibility of determining fluctuations of the surface potential. © 2004 MAIK “Nauka/Interperiodica”. Exposure to ionizing radiation can result in the formation of surface states at the semiconductor–insulator interface and a space charge in the gate insulator (oxide) of a metal–oxide–semiconductor (MOS) transistor [1, 2]. The irradiation can also give rise to significant fluctuations of the surface potential over the semiconductor–insulator interface, which reflects a random distribution of the radiation defects over the insulator surface [2]. The radiation-induced charging leads to a change in the threshold voltage. In addition, the surface states and charge fluctuations lead to a decrease in the slope of the transfer (drain–gate) current–voltage characteristics of MOS transistors in the region of subthreshold currents. In order to elucidate the mechanisms of radiationinduced instability and solve the problem of predicting the stability of irradiated large-scale MOS integrations, it is necessary to determine the radiation-induced charge, the density of surface states, and the degree of inhomogeneity of the surface potential by independent methods. The theory of MOS transistors with planar inhomogeneity and the methods of determination of the surface characteristics of such transistors were considered in [3]. This Letter presents the results of experimental investigation of the effect of radiation on the surface and fluctuation characteristics of MOS transistors and demonstrates the possibility of independent determination of these parameters. Experimental and calculation procedures. The samples were p-channel transistor structures with a channel length of 5 µm, a gate insulator (SiO2) thickness of d = 120 nm, and a dopant concentration in the substrate of Nsub = 1017 cm–3. The samples were irradiated with X-ray quanta at E ≈ 25 keV to a dose of D ≤ 106 rad (Si) in an IRIS-M3 setup. The effect of irradiation on the properties of MOS transistors was studied by measuring subthreshold current–voltage (I–V) characteristics on an automated setup. The effect of the radiation dose on the fluctuation and surface parameters of samples is evaluated within the framework of the model of an inhomogeneous planar MOS transistor. The subthreshold current–voltage characteristics of such transistors with the p channel are described by the relation [3] Z n kT 2 n I D = ---µ p ----  ------ ------i- C D ( Y S ) L m q  ND qV m q V g – V *g - 1 – exp  ---------D- ---- , × exp – ------ ----------------- kT n  kT n (1) where Z and L are the channel width and length, respectively; µp is the hole mobility in the channel; ni is the intrinsic carrier density; ND is the dopant concentration in the substrate; Vg and VD are the gate and drain voltages, respectively; V *g is the gate voltage corresponding to the middle of the region of weak inversion; Y S = 1.5lnλ; λ = ni/ND; q is the electron charge; k is the Boltzmann constant; T is the absolute temperature; Y S + 3σ 1 C D ( Y S ) = -------------2πσ ∫ q εS N D   -------------------------------- 2kT ( – Y S – 1 ) 2 1/2 Y S – 3σ (2) 2 (YS – YS) - dY S × exp – ----------------------2 2σ is the capacitance of the depleted layer for the average surface potential Y S ; C OX + C *SC + qD SS -, n = ------------------------------------------C OX 1063-7850/04/3005-0385$26.00 © 2004 MAIK “Nauka/Interperiodica” C OX + C *SC -, m = -----------------------C OX 386 BORMONTOV et al. transistor, whereby all capacitances in (1) are averaged with respect to the surface potential. The method of determination of the surface parameters of MOS transistors is essentially as follows. By measuring the slope of the rectified input characteristic (equal to m/n in the ln(1 – ID/ID max) versus qVD/kT coordinates) and the transfer characteristic (1/n in the lnID versus qVg /kT coordinates), we determine the m value. Then, using the definition of m, we determine the experimental capacitance of a space charge region in the semiconductor corresponding to the middle of the region of weak inversion, C *SC = (m – 1)COX . For a given dopant concentration in the substrate ND , C *SC depends only on the fluctuation parameter σ. Using the theoretical dependence of C *SC on σ described by rela* value, we can detertion (3) and the experimental C SC mine the fluctuation parameter. Using the m and n values, we can also calculate the spectral density of states by the formula ID , A (a) 10–4 10–6 4 3 Vg, V –20 log(1 – ID /ID max) –4 (b) –15 2 1 –10 10–8 –5 –6 1 –8 4 2 3 –0.4 C OX + C *SC  n C OX + ( m – 1 )C OX - ---- – 1 = -------------------------------------------D SS = -----------------------  q m q –0.3 VD , V tan α D C OX  1 1 - ---------- --------------- – 1 . ×  --------------- – 1 = --------------- tan α D  tan α g q  tan α D  Fig. 1. The transfer (a) and output (b) current–voltage characteristics of a p-channel MOS transistor (1) before irradiation and (2–4) after X-ray irradiation to a dose of D = 104, 105, and 106 rad, respectively. Finally, upon additionally determining the threshold voltage VT of the MOS transistor (e.g., by extrapolating COX is the capacitance of the gate insulator; DSS is the spectral density of the surface states; εS 1 - -------------C *SC = ------------2L D 2πσ 1.5 ln λ + 3σ ∫ 1/2 the I D (Vg) curve to the region of strong inversion until intersection with the Vg axis) and using the known DSS and σ values, we can calculate the charge Q0t in the oxide as * – C OX  V T – φ ms – 2 kT ------ ln λ Q 0t = – Q SC   q 1.5 ln λ – 3σ exp ( Y S ) – 1 – λ ( exp ( – Y S ) – 1 ) × ----------------------------------------------------------------------------------------------------2 exp (YS ) – YS – 1 + λ ( exp ( –YS ) + YS – 1) (4) 2 (3) ( Y S – 1.5 ln λ ) × exp – ---------------------------------- dY S 2 2σ (5) kT + 2qD SS ------ ln λ, q 2 is the total capacitance of the space charge region in the semiconductor for YS = 1.5lnλ; and σ is the fluctuation parameter. The form of relation (1) is similar to that of an analogous expression for a homogeneous planar transistor [4], but the parameters entering into these formulas are substantially different. In particular, the parameters m and n in (1) depend on the total capacitance CSC of the space charge region (including the contribution from an inversion layer) rather than on the capacitance CD of the depleted layer. In addition, the adopted model [3] takes into account the planar inhomogeneity of an MOS where 2qN D L D Q *SC = -----------------------2πσ 1.5 ln λ + 3σ ∫ [ ( exp Y – Y – 1 ) 1.5 ln λ – 3σ 1/2 ( Y – 1.5 ln λ ) 2 - dY + λ ( exp ( – Y ) + Y – 1 ) ] exp – -------------------------------2 2σ 2 is the total space charge of the semiconductor in the middle of the region of weak inversion. Results and discussion. Figure 1 presents the typical experimental current–voltage characteristics measured in the region of subthreshold currents. Figure 2a shows an experimental plot of the threshold voltage TECHNICAL PHYSICS LETTERS Vol. 30 No. 5 2004