1 Introduction

It is well known that the basic damage due to ionizing radiation in any MOS device results from the generation of electron–hole pairs in the gate oxide region. When positive bias is applied to the gate, electrons drift rapidly under the influence of the applied electric field and most of them flow out into the external circuit. In this case, holes drift slowly to the Si/SiO2 interface, a fraction of them gets trapped and thus forming the radiation induced oxide trapped charge [1]. These positive charges induce a negative shift in the threshold voltage and increase the leakage current, which leads to increased power consumption. During the hole transport and trapping processes, hydrogen is released within the oxide and may be transported to the interface and react with Silicon dangling bonds, forming interface traps [2]. The density of these interface states is greatly enhanced by the positive bias voltage applied during the irradiation and they too can modify the overall charge dependent properties of MOS devices including the decrease in transconductance (g m ) and mobility (μ) [3, 4]. In many applications, particularly in mixed signal and digital technologies, g m affects the speed and output drive capabilities. The degradation of g m due to radiation at room temperature has been studied extensively and is attributed to the mobility degradation from increased interface trap densities [5, 6]. A few other investigators have observed degradation of g m from total dose at lower temperature and shown that both oxide and interface trapped charge densities (ΔN ot and ΔN it ) can modulate the resistivity and thus alter the value of g m [7, 8]. In the case of space applications, the MOS and bipolar devices along with circuits need to be radiation tolerant from 100’s of krad to a few Mrad of total doses. In high-energy physics experiments like Large Hadron Collider (LHC), the devices and circuits need to be radiation tolerant up to 100 Mrad of total doses in their 5 years lifetime. The testing of devices for such high doses, using 60Co gamma or proton facilities also need substantially large time of exposure (e.g. 166 h with dose rate of 167 rad/s to give 100 Mrad of 60Co gamma total dose). For this reason, one can reduce the irradiation time to more practical values by utilizing the swift heavy ions so that the time required to reach 100 Mrad of total dose is only about tens of minute [9]. Therefore, in this paper, we have made an attempt to compare the total dose effects of different high energy ions on the IV characteristics of N-channel depletion MOSFETs. The recovery in the IV characteristics of these devices is also studied by isothermal annealing at 200 °C up to 100 h and by varying temperature from 50 to 500 °C (isochronal annealing) [10].

2 Experimental details

The cross-sectional detail of the MOSFET was shown in Fig. 1. The gate metal (Al) thickness was ≈1.2 µm and gate length of used device was ≈5 µm. The special back-to-back diodes were diffused directly into the MOS pellet and are electrically connected between each insulated gate and the MOSFET source in order to avoid excess or transient voltage. The N-channel MOSFETs (3N 187) studied here were obtained from Bharath Electronics Limited (BEL), Bangalore, India. The MOSFETs were irradiated with the ion fluence varied from 2 × 108 to 1.53 × 1013 ions/cm2 at 300 K in an experimental chamber of diameter 1.5 m maintained at 10−7 mbar vacuum using the facility available at the 15 UD 16 MV Pelletron Accelerator at Inter University Accelerator Center (IUAC), New Delhi, India [11]. Different high energy ions viz., 175 MeV Ni13+, 140 MeV Si10+, 100 MeV F8+, 95 MeV O7+ and 48 MeV Li3+ were used to irradiate the experimental MOSFETs. The fluence on the sample kept in cylindrical secondary electron suppressed geometry was estimated by integrating the total charge accumulated on the sample using a current integrator and then counting by a scalar meter. The ion beam was scanned over the samples in an area of 10 mm × 10 mm by a magnetic scanner in order to get a uniform dose. During irradiation the gate terminals of the MOSFETs were biased at +2 V (V GS  = +2 V) and the typical beam current was set at one p-nA (one particle-nano ampere) for 48 MeV Li3+ ions, 0.29 p-nA for 95 MeV O7+ ions, 0.125 p-nA for 100 MeV F8+ ions, 0.1 p-nA for 140 MeV Si10+ ions and 0.03 p-nA for 175 MeV Ni13+ ions. The electrical characterization of the un-irradiated and irradiated MOSFETs were performed using computer interfaced Keithley 2636A Semiconductor Parameter Analyzer. The threshold voltage (V TH ) and transconductance (g m ) were determined from the I D V GS characteristics. The mobility (µ) of carriers in the channel was determined by g m measurements (V DS  = 0.1 V). The I DSat of irradiated MOSFET was extracted from I D V DS characteristics. While measuring the electrical parameters of the MOSFET, same voltage was applied to both the gates. The devices were characterized within 30 min after irradiation following MIL-STD-883H Method 1019.8 to avoid time dependent annealing, which changes the electrical effects of damage formation [11, 12].

Fig. 1
figure 1

Cross sectional view of the N-channel MOSFET

In order to understand the linear energy transfer (LET) dependence, in the present work five different ions viz., 175 MeV Ni13+ ions, 140 MeV Si10+ ions, 100 MeV F8+ ions, 95 MeV O7+ ions and 48 MeV Li3+ ions were chosen and the ratio of the LET of these ions in silicon was around 72:22:10:8:1. An identical cumulative total dose for all the ions was fixed. The total dose for each ion was varied from 100 krad to 100 Mrad. For the first time, in the present investigation, we compared the effect of LET of the different ions, by keeping the same total dose on the degradation of N-channel depletion MOSFET using current–voltage (IV) characteristics. The stopping and ranging of ions in matter (SRIM) simulations and calculations were also carried out to understand the effects of different high energy ions on electronic energy loss (ionization) and nuclear energy loss (displacement) in the MOSFET structure.

The recoveries in the IV characteristics of irradiated N-channel MOSFETs were studied by isothermal and isochronal annealing using high temperature oven and furnace. The MOSFETs irradiated up to 100 Mrad of total dose were subjected to isothermal annealing at 200 °C for different time periods up to 100 h and allowed for natural cooling before taking the IV characteristics. The MOSFETs irradiated up to 100 Mrad of total dose were also subjected to isochronal annealing from 50 to 500 °C at 1 h duration for each temperature and allowed for natural cooling in order to study the annealing of IV characteristics of irradiated MOSFETs.

3 Results and discussion

The values of electronic energy loss, nuclear energy loss and the range of different ions in the MOSFET estimated using SRIM-2011 Simulation Programme [13, 14] are given in Table 1. From this Table, all the different ions chosen in the present study can pass through the MOSFET and deposit energy through electronic excitations. During the transfer through MOS structure, the ions also create uniform damage. When a MOSFET is exposed to high energy ions, the incident ions pass through the metal contact, SiO2 region, N-channel and P-substrate, as shown in Fig. 1. In this work, the energies of Lithium, Oxygen, Fluorine, Silicon and Nickel ions are chosen such that all the ions pass through the active region (~5 μm from top surface) and create uniform damage in the MOSFET.

Table 1 The linear energy transfer (LET), non ionizing energy loss (NIEL) and range of different high energy ions in metal oxide semiconductor (MOS) structure

3.1 IV characterization

Figure 2 illustrates the transfer characteristics (I D vs. V GS ) of 175 MeV Ni13+ ion irradiated MOSFET at V DS  = 1 V. The threshold voltage (V TH ) is determined here by studying the subthreshold behavior of irradiated MOSFETs as a function of total dose. The V TH here is defined as the negative gate voltage for which the I D becomes 1 μA (V TH  ≡ V GS @ I D  = 1 μA). In this figure, it is clearly seen that the corresponding point shifts towards negative voltages with increase in dose. This means that the V TH decreases with increase in dose. Similar type of degradation has been observed for other types of ion irradiated MOSFETs. Similar types of results have been observed by Rathod et al. [15] and Zhang et al. [16]. Figure 3 shows the variation in V TH as a function of total dose for different ions. This figure also shows the recovery in V TH after annealing at 200 °C for different cumulative time. At higher doses the degradation in V TH is found to be more for lower LET compared to that of higher LET ions. Further, it also recovers with increase in annealing time and the observed recovery is found to be about 80 %. Figure 4 illustrates the variation in V TH with respect to total dose and annealing temperature from 50 to 300 °C after different ion irradiation. This figure shows that V TH recovers more at elevated temperatures and it becomes almost 100 % at 300 °C. The thermal annealing of the surface effects in MOS devices is slightly different from that of displacement related defects. Interface states anneal out at temperature in the range of 100–200 °C and trapped charges can be removed in the temperature range of 150–300 °C. Therefore complete recovery in V TH after annealing the ion irradiated MOSFETs up to 300 °C has been observed.

Fig. 2
figure 2

I D V GS characteristics of 175 MeV Ni13+ ion irradiated MOSFET with different doses shown

Fig. 3
figure 3

Threshold voltage shown as a function of total dose and annealing time after different ion irradiation

Fig. 4
figure 4

V TH shown as a function of total dose and elevated annealing temperature after different ion irradiation

During irradiation, the different electrical characteristics of MOSFETs such as threshold voltage (V TH ), transconductance (g m ) and mobility (μ) decrease with increase in radiation dose and densities of interface and oxide trapped charges (ΔN it and ΔN ot ) increase with increase in radiation dose. Further, the irradiated devices are annealed by isothermal and isochronal methods. In isothermal annealing method, the 100 Mrad irradiated MOSFETs have been annealed at 200 °C up to 100 h whereas in isochronal annealing the irradiated devices have been annealed at different temperatures starting from 50 to 500 °C for 1 h duration at each temperature. In isochronal annealing, 100 Mrad irradiated MOSFET has been annealed at 50 °C for 1 h then allowed the device to cool down to room temperature to be electrically characterized. The same device is again annealed at 75 °C for 1 h and then allowed for cooling to the room temperature. The same procedure has been followed at different temperatures say 100, 150, 200, 250, 300, 350, 400 °C etc. At temperature 300 °C during isochronal annealing, almost all the defects created during the irradiation process have been completely annealed.

The leakage current (I L ) is the current flowing in the source to drain junction when MOS transistor is in OFF condition. Besides the change in the V TH due to the build up of N it and N ot , the other major effect of ionizing radiation on MOS transistor is the increase in the I L and is related to the N it and N ot build up. Increase in I L leads to an increased current flow in the circuit and therefore increased static power dissipation. This can have a serious impact on electronics, not only on the affected components but also on the other components connected to the same power supply. The increase in I L can also be seen in the subthreshold characteristics of the irradiated MOSFET, as shown in Fig. 2.

The net threshold voltage shift (ΔV TH ) and contribution to that shift due to the interface traps (ΔV Nit ) and the trapped oxide charge (ΔV Not ) have been calculated from the subthreshold measurements using the technique proposed by McWhorter and Winokur [17]. The ΔN ot is calculated using the standard expressions ΔN ot  = ΔV Not C ox /q and ΔN it  = ΔV Nit C ox /q, where q = 1.6 × 10−19 C. ΔN ot is the effective oxide-trapped charge density in cm−2 as projected to the interface. Here C ox is the oxide capacitance per unit area. The variation of the ΔN it and ΔN ot as a function of total dose and annealing with cumulative time are shown respectively in Figs. 5 and 6 for irradiated MOSFETs. These two Figs. 5 and 6 also show about 60 % recovery in ΔN it and 80 % recovery in ΔN ot after annealing for 100 h at 200 °C [18]. Further the effect of annealing temperature on ΔN it and ΔN ot for different ion irradiated MOSFETs are shown graphically in Figs. 7 and 8. It is observed that the recovery in ΔN it and ΔN ot with annealing temperature approaches 100 % at around 300 °C. It is important to note from the Figs. 58 that ΔN ot is higher than ΔN it and because of this V TH shifts towards negative voltage.

Fig. 5
figure 5

Interface traps density shown as a function of total dose and annealing time after different ion irradiation

Fig. 6
figure 6

Oxide traps density as a function of total dose and annealing time after different ion irradiation

Fig. 7
figure 7

Interface traps density as a function of total dose and temperature after different ion irradiation

Fig. 8
figure 8

Oxide traps density as a function of total dose and temperature after different ion irradiation

The SRIM simulation reveals that 48 MeV Li, 95 MeV O, 100 MeV F, 140 MeV Si and 175 MeV Ni ions create uniform ionization and uniform displacement or vacancies in the active region of the MOSFETs. The high energy ions are implanted or stopped at the end of the ion range in p-substrate. The vacancies or displacements per ion for 175 MeV Ni ions are found to be around 250 in the active region of the MOSFET. Similarly, for 140 MeV Si, 100 MeV F, 95 MeV O and 48 MeV Li ions are 38, 17, 14 and 3 respectively. In case of annealing, as shown in Figs. 5 and 6, it can be seen that the recovery in N it and N ot for lower LET ion like 48 MeV Li ions is more at 200 °C for 100 h (isothermal annealing), when compared to higher LET ions like 175 MeV Ni ions. As explained above, the 175 MeV Ni ions create more displacements or vacancies and these cannot be annealed at 200 °C. Therefore the complete recovery in the N it and N ot for higher LET ion like 175 MeV Ni ion irradiated MOSFET can be found at 300 °C, as shown in Figs. 7 and 8.

The mobility (µ) of electrons in the channel has been estimated from transconductance (g m ) measurements [19]. The mobility of carriers in the MOSFET can be calculated using the formula shown in Eq. (1):

$$ \mu_{FE} = \frac{{Lg_{m} }}{{ZC_{OX} V_{DS} }} $$
(1)

where g m is the peak transconductance of the MOSFET. The representative plot for the behavior of g m for 175 MeV Ni13+ ion irradiated MOSFET is shown in the inset of Fig. 9. It can be seen from this figure that there is a significant decrease in g m after ion irradiation [20]. Similar type of results are observed in other types of ion irradiated MOSFETs. The µ of carriers in the channel has been estimated from the peak g m and is called field effect mobility (μ FE ). The variation of peak g m as a function of total dose and annealing time for different ion irradiated MOSFETs is shown in Fig. 9 and it can be observed around 50–60 % recoveries in peak g m after annealing the ion irradiated MOSFETs at 200 °C for 100 h. The variation of peak g m as a function of total dose and annealing up to 400 °C for different types of ion irradiated MOSFETs is shown in Fig. 10. It can be seen around 100 % recovery in peak g m after annealing the MOSFETs till 300 °C.

Fig. 9
figure 9

The peak transconductance as a function of total dose and annealing time after different ion irradiation (Inset figure represents transconductance characteristics of 175 MeV Ni13+ ion irradiated MOSFET)

Fig. 10
figure 10

The peak transconductance as a function of total dose and elevated temperature after different ion irradiation

The variation in the µ with total dose is shown in the inset of Fig. 11 for different ion irradiated MOSFET. The degradation in the μ of carriers in the channel is mainly due to radiation-induced trapped charges and these act as Coulomb scattering centers [8]. Earlier, the μ degradation has been attributed mainly due to the interface-trapped charge (N it ), while the effect of oxide-trapped charges (N ot ) which lies further away from the inversion layer has been considered to be negligible. Following the μ model of Sun and Plummer [8], μ is related to the build up of interface-trapped charge (ΔN it ) through the Eq. (2):

Fig. 11
figure 11

Normalized mobility degradation during 175 MeV Ni13+ ion irradiation as a function of interface trapped charge. The top axis shows the values of the dimensionless product of α it ΔN it (Inset figure represents mobility versus total dose for different ion irradiated MOSFETs)

$$ \mu = \frac{{\mu_{o} }}{{1 + \alpha \varDelta N_{it} }} $$
(2)

where, μ o is the pre-irradiation value of μ and α is a constant. This model is observed to be fairly successful in explaining the observed data of decrease in μ. Manic et al. [21] and Dimitrijev et al. [22] have suggested the use of a modified version of Eq. (2) as:

$$ \mu = \frac{{\mu_{o} }}{{1 + \alpha_{it} \varDelta N_{it} + \alpha_{ot} \varDelta N_{ot} }} $$
(3)

where, ΔN ot is the radiation induced oxide trapped charge. The experimental results of Zupac et al. [23] on irradiated and annealed MOSFETs convincingly have demonstrated that taking ΔN it only into account does not adequately fit the data and ΔN ot also has to be considered at low temperature measurements. There has been some discrepancy in the reported values of α it and α ot . The different methods of fabrication of devices may be the cause of this discrepancy. From the detailed analysis of experimental data of MacLean and Boesch [24], a good fit has been obtained with α it  = 3.1 × 10−12 cm2 and α ot  = 7.0 × 10−15 cm2. On the other hand, studies by Zupac et al. [23] provide α it  = 3.9 × 10−12 cm2 and α ot  = 7.0 × 10−13 cm2 as good fit values for Eq. (3). In both the cases the α ot value is quite small compared to α it value, although in the case of MacLean and Boesch [24] studies, the value is significantly smaller. The smaller value of α ot reflects the fact that ΔN ot , further away from the interface, is less effective in scattering inversion layer carriers than ΔN it . The best fit values of α it and α ot for the experimental data on mobility for 175 MeV Ni13+ ions, 140 MeV Si10+ ions, 100 MeV F8+ ions, 95 MeV O7+ ions and 48 MeV Li3+ ions irradiated MOSFETs are presented in Table 2. It can be seen from the Table 2 that the value of α it is ≈10−12 cm2 and the value of α ot is not consistent and it is even negative besides being very small compared to α it values. Similar results have been observed by Dimitrijev et al. [22] and Gnana Prakash et al. [4, 5]. Hence, as inferred by the earlier investigators, the effect of ΔN ot on mobility is negligible at room temperature. The values of α it obtained in the present work are consistent and agree with the results reported by earlier investigators [4, 5, 22]. Therefore, normalized mobility of carriers in the channel has been correlated with ΔN it and it is found that the decrease in mobility is indeed significant after irradiation. To facilitate better illustration, the results obtained for the normalized mobility degradation as a function of ΔN it (circles) and the values of dimensionless product α it ΔN it (solid line) are shown graphically in Fig. 11 for 175 MeV Ni13+ ion irradiated MOS devices. From this figure, it can be seen that μ FE of the irradiated devices decreases with increase in dose and about 80 % degradation has been found after a total dose of 100 Mrad.

Table 2 The effect of different ion irradiation on the mobility degradation coefficients of N-channel MOSFETs

I DSat of irradiated MOSFET has been extracted from I D V DS characteristics. The drain current (I D ) of the MOSFET becomes saturated after a certain drain voltage (V DS ) and can be referred as I DSat . The representative plot of I D V DS for 175 MeV Ni13+ ion irradiated MOSFET is shown in the inset of Fig. 12 and the effects of ionizing radiation on I DSat of the MOSFETs have been studied for various doses. The high-energy radiation may produce defects and their complexes in inbuilt channel and produce carrier removal effect and this in turn increases the resistance of the channel. In addition this Coulomb scattering between the free carriers in the channel and radiation induced trapped charge decreases I DSat . The variation in the I DSat with respect to total dose and annealing time is shown graphically in Fig. 12 for different ion irradiated MOSFETs. It can be seen from this figure that I DSat decreases drastically with increase in ion doses. About 20–30 % recovery in I DSat has been observed for the devices annealed at 200 °C for 100 h (isothermal annealing) and about 60 % recovery in the I DSat , annealed up to 500 °C for 1 h duration at each temperature (isochronal annealing).

Fig. 12
figure 12

The drain saturation current as a function of total dose and annealing time after different ion irradiation (Inset figure represents I D V DS characteristics: before and after 175 MeV Ni13+ ion irradiation)

4 Conclusions

The electrical characteristics of N-channel MOSFETs have been studied before and after exposure to different swift heavy ions in the dose range of 100 krad–100 Mrad. The annealing effects on different electrical parameters of ion irradiated MOSFETs have been also studied systematically. The important parameters such as V TH , ΔN it , ΔN ot , g m and μ of the irradiated MOSFETs have been found to decrease significantly after ion irradiation. The degradation in electrical parameters is almost same up to 10 Mrad of total dose for all the ions, but beyond 10 Mrad, the degradation is more for lower LET ions. Around 80 % degradation is observed in the I–V characteristics of the irradiated MOSFETs after receiving 100 Mrad of total dose. Almost 100 % recovery in the V TH , ΔN it , ΔN ot , g m and μ of the irradiated MOSFETs is observed after annealing. This shows that both interface and oxide trapped charges, created during irradiation, have been completely annealed.