Photocurrent measurements for oxide charge characterization of high-κ dielectric metal oxide semiconductor capacitors

Abstract

Photocurrent-voltage (photo-I-V) measurements for oxide charge characterization of high-κ metal-oxide-semiconductor (MOS) capacitors is a viable technique for quantifying both the oxide charge density and the centroid of charge. The latter information is not obtainable from conventional capacitance-voltage (C-V) measurements. In this paper, the theoretical background for photo-I-V measurements is reviewed and extended to address high-κ dielectric stacks. Experimental results comparing C-V and photo-I-V measurements for oxide charge characterization of Al gated, high-κ HfO2 MOS capacitors is given. For electron injection from Al, it was found that similar oxide charge densities were calculated using photo-I-V (between 5.5×1012 and 6.7×1012cm−2) and C-V (between 4.5×1012 and 5×1012cm−2) measurements. The centroid of trapped charge, x¯∕tox, as defined as a distance from the gate electrode normalized to the dielectric stack thickness was measured to be ∼0.2, which is near the Al electrode and suggests that significant charge trapping is occurring at defects within the HfO2 and not at the HfO2 interface with the thin oxide. Similar measurements using electron injection from the substrate shows that the centroid of trapped charge shifts from near the gate electrode (x¯∕tox=0.22) to the substrate electrode (x¯∕tox=1) as the applied stress bias increases from 0.75to2V. This result is consistent with increased field-assisted detrapping of oxide charge. From photocurrent measurements, it was also determined that the effective barrier height for electron photoinjection from the Al gate to the HfO2 conduction band was 2.6±0.1eV, which is consistent with current knowledge of the HfO2 band structure.

Interfacial Reaction and Electrical Properties of HfO2 Film Gate Dielectric Prepared by Pulsed Laser Deposition in Nitrogen: Role of Rapid Thermal Annealing and Gate Electrode

“…The substantial increase in the leakage current from the dark to the photo flux is attributed to the internal photoemission of the carriers into the dielectric conduction band. The shift at negative gate voltage is much larger than at positive gate voltage, which implies that the centroid of the oxide charge is located closer to the gate than to the substrate, − which is consistent with the formation of parasitic interlayer between the Ti electrode and HfO 2 .…”

Charge trapping effect at the contact between a high-work-function metal and Ta2O5 high-k dielectric

“…The characteristic time for detrapping in the absence of the field can be estimated to be of the order of 10 4 s, which is not much higher than the relaxation times for trapping obtained here. Therefore, the detrapping rate does not substantially increase when applying the field, in contrast to the case of field assisted detrapping [11] typical of bulk traps. The above finding can be attributed to detrapping of electrons from traps located near the metal electrode into the metal conduction band, which is expected to be weakly dependent on the applied field.…”

Physical and optoelectronic characterization of reactively sputtered molybdenum-silicon-nitride alloy metal gate electrodes

“…However, the photocurrent measured at photon energies below 3.6 eV is dominated by leakage current and is there not useful for extracting the interfacial barrier height. The experimental procedure used to measure the values of effective barrier height ͑⌽ b eff ͒ has been outlined in detail by Nicollian and Brews, 13 Afanas'ev et al, 14,15 and Felnhofer et al 16 As an example, for a capacitor ͑R N =10%͒, the measured effective barrier height ͑⌽ b eff ͒ is plotted as a function of the applied oxide electric field in Fig. 5.…”