Supervisors info:
Αγγελική Αραπογιάννη, Καθηγήτρια, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Summary:
The objective of this thesis is the study of modified Oxide-Nitride-Oxide (ONO) devices as a charge trapping memory. The modified ONO structures were formed by high dose (~ 1x1016 Si+/cm2) and low energy (1 keV) implantation of silicon ions into oxide-nitride structures followed by wet oxidation process (850 °C, 15 min). The aforementioned process leads to the formation of a triple layer stack of oxide-nitride-oxide.
Specifically, three implantation doses were used, i.e. 5x1015, 1x1016, 1.5x1016 Si + / cm2 labeled as low, medium and high doses respectively at an energy of 1 keV. The implantation of silicon ions causes swelling of silicon nitride in order to accommodate the large number of silicon ions. Moreover, the implantation results in a high concentration of silicon atoms close to the free upper surface of the nitride, as it has been observed by transmission electron microscopy analysis (TEM). Furthermore, TEM observations showed a slight swelling of the bottom oxide, located above the silicon substrate, a fact that is attributed to implantation secondary phenomena. The above swelling of nitride as well as bottom oxide is enhanced with the dose.
After the oxidation process, X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (ToF-SIMS) revealed that the top layer of the stack is silicon dioxide with almost no presence of nitrogen. Also, the remaining nitride is enriched with oxygen and the lower oxide is enriched with nitrogen.
The sequence of the above processes results in the formation of triple-layer structures, which are tested as charge trapping memory elements. High frequency capacitance-voltage measurements (C-V) revealed particularly small standard deviations of the equivalent oxide thickness of just about all structures. Moreover, the dependence of this thickness upon the dose of the implantation was studied. Furthermore, high-low frequency capacitance technique was implemented in order to determine the density of interfacial traps Dit, which proved that Dit has a weak dependence with the implantation conditions.
The charging of the modified three layer ONO structures were studied with incremental step pulse charging (ISPC). The write/erase parameters of the modified memory are determined by the slope of the characteristic curve of the voltage shift versus the applied pulse to the gate electrode (ΔVFB=f(Vp) ). The slope of the curve indicates the trapping efficiency of the nitride and determines the voltage shift that occurs while increasing the applied pulse amplitude by 1 V. The low-dose-ion-implantation modified structure can trap only electrons and its trapping efficiency is determined at 0.83 V/V. It was not possible to verify the hole trapping. The medium-dose structure shows electrons and holes trapping by applying the proper voltage pulse with efficiencies 0.67V/V and -0.35V/V to -0.44V/V respectively. Finally, in the high-dose sample, electron and hole trapping is observed under the proper applied pulse. The electron trapping efficiency was calculated 0.57V/V and hole trapping efficiency from -0.47V/V to -0.56V/V.
The detrapping of charges and in specific the decay rate of trapped electrons in the modified structures was extracted by the slope (V/decade) of the retention characteristic curve at 50 ºC, 90 ºC, 120 ºC, 150 ºC και 180 ºC. In the low-dose modified ONO structure the decay rate of electrons is 0.08 V/decade at 50 ºC and 90 ºC, whilst it increases to 0.167 V/decade at 120 ºC, 150 ºC and 180 ºC. After 10 years, the initial voltage shift (3 V) will range from 1.9V to 2.5 V. Similarly, the modified medium-dose structures, showed an increasing decay rate of the trapped electrons of 0.11 V/decade at 50 ºC to 0.32 V/decade at 180 ºC. After 10 years, the initial 3 V voltage shift will range from 2.4 V to 0.9 V. Finally, in the high-dose modified structures, the decay rate is increased with temperature. The decay rate increases from 0.18 V/decade to 0.27 V/decade by increasing the temperature from 50 ºC to 180 ºC. The initial 3V voltage shift will be after 10 years reduced within the range of 0.7V to 1.6V.
The determination of the density of traps as a function of their energy was performed for an unimplanted MONOS and for the low-dose samples, using the Yang-Wang model for the retention characteristics. A comparison between the two samples showed that the low-dose implanted sample has a higher density of deep traps within the nitride layer.
The above results indicates that the modified MONOS (metal-oxide-nitride-oxide-semiconductor) structures of high and medium dose can act as charge trapping memories particularly at high temperatures due to the fact of the increased density of deep traps in nitride layer.
Keywords:
charging of modified structures ONO, trapping efficiency, decay rate, density of traps
