GSCS Method Demonstration

Based on the models in section 3.1, we can directly track the “real-time” trapped charges during programming/erasing and reliability testing. In this section, we will describe in detail the preparation and the basic characteristics of GS and CS samples. The first simple results based on bulk trap model are also demonstrated. Finally, we will compare the results extracted by our method with other methods.

3.2.1 Sample Descriptions

Large area (500µm × 500µm) SONOS-type capacitors with various ONO conditions are fabricated. For each sample, both GS and CS capacitors are made, and we fabricate them at the same time except the poly gate/well doping processes to ensure they have identical ONO thickness. For all samples, the B.O. is grown by in-situ steam generation (ISSG) method which can minimize the impact of different substrate doping because ISSG is a radical-reaction process, but not a thermal-diffusion process. Nitride layers are deposited by the industrial standard LPCVD process, using SiH2Cl2 and NH3 at 680^oC. On the other hand, oxynitride (SiON) is also deposited by the same LPCVD process, but with additional N₂O gas flow. The dielectric constant of SiON is equal to 6.5, and optical index of refraction = 1.8 (nitride is 2.0). For the T.O., both wet oxidation of nitride (wet conversion) and HTO top oxide are used.

In this method, we adopted p-type well and p-type poly gate for both GS and CS capacitors. Typically, p-well doping densities for CS and GS are around 10¹⁷cm^-3 and 7×10¹⁸cm^-3, respectively while the p-poly gate doping densities for CS and GS are around 10²⁰cm^-3 and 5×10¹⁶cm^-3, respectively. For all the samples, n⁺-doped source/drain regions are formed so that channel inversion layer can be formed during +FN programming.

In this method, the well and poly gate doping do not need to be p-type. N-type well and n-type poly gate can be used as well. We have also used wells and gates with opposite

polarities. For example, a lightly doped p-poly gate with a heavily doped n-well was used for the GS capacitor. However the measured CV curve had a larger distortion, which affected the accuracy of VFB shift extraction. This is because both poly depletion and substrate depletion happened at the same time during –V_G bias for this sample, resulting in a complicated CV curve. Therefore, it is best to use the same polarity for both well and poly gate because only one of them has depletion during +/– V_G biases.

3.2.2 Basic Characteristics of GS and CS Capacitors

We first consider the device characteristics of SONOS (ONO = 54/70/90Å) under +FN programming. Device is programmed by +FN using various biases from the initial fresh state (VFB,GS and VFB,CS ~ 0V), and both the CS and GS capacitors are measured under identical electrical testing procedures. The corresponding CV curves are shown in Fig. 3.3.

In Fig. 3.3(a), the CV curves of CS capacitor are in accumulation at –VGate (or +Vwell), and become depletion and even inversion at +V_Gate (or –V_well). The inversion layer can be formed because n⁺ source/drain regions are introduced. On the other hand, the CV curves of GS capacitor are in accumulation at –V_well (or +V_Gate) and become depletion at +V_well (or –VGate). However, the inversion layer cannot be formed because the poly gate does not have source/drain regions hence deep depletion continues at larger +V_well. Moreover, the CV curves of GS capacitor have much more gradual slope during transition. This is possibly because there are more interfacial states at poly gate and T.O. interface. However, the CV curves still show parallel shift during +FN programming, which is sufficient for VFB shift extraction.

The extracted ∆VFB in Fig. 3.4 shows that the ∆VFB,CS and ∆VFB,GS are clearly different.

This is because SONOS (ONO = 54/70/90Å) has thicker T.O. and thinner B.O. According to Eqs. (3-1) and (3-2), ∆VFB,CS should be larger than ∆VFB,GS.

3.2.3 Extracted Trapped Charge Evolution and Vertical Location Evolution

In Fig. 3.4, the two VFB shifts (∆VFB,CS and ∆VFB,GS) at a given time can be transformed into Q and x by using Eqs. (3-3) and (3-4), and the results are shown in Fig. 3.5(a) and (b), respectively. The charge evolution (Q-t) and mean vertical location evolution (x-t) can be directly extracted by this GSCS method. In Fig. 3.5(a), Q-t has a similar shape as ∆VFB-t (Fig.

3.4), as expected. On the other hand, x-t in Fig. 3.5(b) shows that injected electrons migrate from the bottom interface toward the top interface.

In order to obtain a clearer idea about the charge evolution, x-Q curves are plotted, as illustrated in Fig. 3.5(c). x-Q is a simple transformation from Q-t and x-t curves. In Fig. 3.5(c), as electron density increases the mean charge vertical location migrates from the bottom interface toward the center of nitride. The final saturated x is around 40Å, which is very close to the center of nitride. Moreover, despite the large dependency of program speed to the program voltage, x-Q plots are independent of the program voltage. This clearly shows that the x-Q plot expresses an intrinsic property of the ONO structure (the charge trapping behavior) and is not affected by external factors such as e-field.

3.2.4 Comparison with Other Methods

In order to verify the charge vertical location in Fig. 3.5, we use two different methods to monitor the vertical location.

The first method is to apply our previous transient analysis method [3.2], as mentioned in section 2.3.2, where J-ETox curves of gate injection operated SONOS devices are compared with various charge location assumptions. However, for channel-injection operated SONOS devices the tunnel oxide is the bottom oxide instead of the top oxide in the gate-injection operated SONOS devices. Therefore, we should plot the J vs. bottom oxide e-field (E_Box) curves, and then compare the J-EBox curves at various program voltages by assuming different charge vertical locations. Since the E of CS capacitor is independent of charge vertical

location, the GS capacitor (∆VFB,GS) should be introduced to extract J-E_Box plots, and the detailed equations are shown in Appendix B. For the GS capacitor, since the channel and gate are effectively reversed, the equations to calculate the ETox in [2.36] can be utilized to calculate the E_Box in our case. The calculated results are shown in Fig. 3.6, in which Case 1 represents charges placed at T.O./nitride interface, Case 2 represents charges placed at the B.O./nitride interface, and Case 3 represents charges placed at the center of nitride. The results indicate that Case 3 has better consistency for every program voltage while it is much less consistent for Case 1 and Case 2. These results suggest that the centroid of electrons is more likely located near the center of nitride, which is consistent with Fig. 3.5.

The second method is to directly simulate the ∆VFB-t curves assuming various charge vertical locations. The theoretical calculation can be carried out by integrating the FN tunneling current with respect to time. Appendix C illustrates the detail of this simulation method for ∆VFB-t, and the results are shown in Fig. 3.7. For CS capacitor (Fig. 3.7(a)), it is very interesting that ∆VFB,CS-t are almost the same for various charge vertical locations.

Equation (C-4) shows that the EBox is independent of the charge vertical location (for Case 1, Case 2, or Case 3). Therefore, we cannot distinguish the real charge vertical location from

∆VFB,CS-t. On the other hand, for the GS capacitor (Fig. 3.7(b)), since the channel and gate are effectively reversed, the equations of E_Box are very different for different vertical locations, as shown in Eqs. (B-2), (B-4), and (B-6). Figure 3.7(b) shows the simulated ∆VFB,GS-t for various charge vertical locations, and only Case 3 (assuming the charges are placed at the center of nitride) is consistent with the measured data. Again, this result supports that the charge centroid is close to the center of nitride after +FN injection.

Although the above two methods can obtain results that are consistent with Fig. 3.5, these methods have limitations and can only provide a rough estimate of charge vertical location. On the other hand, our novel GSCS method is more powerful and provides

“real-time” monitoring of trapped charges.