This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination. IEEE ELECTRON DEVICE LETTERS 1 Frequency-Modulated Charge Pumping: Defect Measurements With High Gate Leakage Jason Thomas Ryan, Member, IEEE, Richard G. Southwick III, Member, IEEE, Jason Paul Campbell, Member, IEEE, Kin P. Cheung, Senior Member, IEEE , Anthony S. Oates, Fellow, IEEE , and John S. Suehle, Fellow, IEEE Abstract— Charge pumping is one of the most relied techniques used to quantify interface defects in metal-oxide-semiconductor devices. However, conventional charge pumping is easily hindered by excessive gate leakage currents, which render the technique unsuitable for advanced technology nodes. We demonstrate a new frequency-modulated charge pumping methodology in which we transform the quasi-dc charge pumping measurement into an ac measurement. The ac detection scheme is highly resistant to gate leakage currents and extends the usefulness of charge pumping as a defect monitoring tool for future technologies. Index Terms— Charge pumping, defects, leakage current. I. I NTRODUCTION A DVANCED research and development relies heavily on the ability to “gauge” defects, which limit the perfor- mance and reliability of ultrascaled devices. Ironically, suc- cessful device scaling actually limits the ability of many device characterization techniques to “gauge” these defects. Charge pumping (CP), e.g., is a seemingly ubiquitous characterization technique often used to study the density and energy distrib- utions of interface defects in metal-oxide-silicon field-effect- transistors (MOSFETs) [1]–[8]. However, an unfortunate side effect of successful scaling is increasingly large gate leakage currents. In modern devices, leakage currents are so large that they can easily mask the CP signal and render conventional CP techniques essentially unworkable. Even with advanced high-k gate stacks, the leakage current can be sufficiently large to completely obscure the CP signal. Without a feed- back mechanism (such as CP) to determine the relationship between processing changes and defect densities, advanced development teams are left to blindly drive further research efforts. In this letter, we demonstrate a new CP methodology called frequency-modulated charge pumping (FMCP), which robustly treats the leakage current issue and extends the usefulness of Manuscript received November 21, 2012; accepted February 28, 2013. The review of this letter was arranged by Editor K.-S. Chang-Liao. J. T. Ryan, K. P. Cheung, J. P. Campbell, and J. S. Suehle are with the Semiconductor and Dimensional Metrology Division, National Insti- tute of Standards and Technology, Gaithersburg, MD 20899 USA (e-mail: kin.cheung@nist.gov). R. G. Southwick was with the National Institute of Standards and Technol- ogy, Gaithersburg, MD 20899 USA. He is now with IBM Research, Albany, NY 12205 USA. A. S. Oates is with Taiwan Semiconductor Manufacturing Corporation, Hsinchu 30844, Taiwan. Color versions of one or more of the figures in this letter are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/LED.2013.2251315 the CP technique to future highly scaled technology nodes. The key concept of FMCP is transforming traditional “quasi- dc” detected CP into an ac detected measurement, thereby exploiting the 1/f noise curve and effectively suppressing the dc leakage background. This is accomplished via lock-in amplifier detection. In conventional CP, a square wave voltage pulse (50% duty cycle) is applied to the gate electrode such that the MOSFET is pulsed between strong accumulation and strong inversion at some frequency (typically in the kHz to MHz range) [1]–[3]. This gate pulse scheme cyclically populates interface defects with both electrons and holes. The source, drain, and substrate electrodes are grounded while the (quasi-dc) CP recombination current ( I CP ) is measured at the substrate terminal. In the absence of bulk dielectric defects, I CP scales linearly with the CP gate pulse frequency [1], [2]. Often, I CP is measured as a function of CP frequency with the slope of this line proportional to the number of defects in the device [1], [2]. This simple view of CP is complicated by the presence of a gate leakage current component in I CP . In devices with thick gate dielectrics, the leakage component is negligible. How- ever, in modern devices with highly scaled gate dielectrics, the leakage component is the overwhelming majority of the measured substrate current. This issue is often treated by presuming that the gate leakage current component is independent of CP frequency. Typical gate leakage corrections usually involve either: 1) a multifre- quency CP measurement in which the extrapolated value of the substrate current at 0 Hz CP frequency (assumed to be entirely due to leakage) is subtracted from the data [5] or 2) a subtraction of a low-frequency swept base voltage/constant amplitude CP measurement (Elliot curve [4]) from a higher frequency Elliot curve [6], [7]. Other methodologies have also been proposed [8]. However, when the gate leakage component becomes too large compared to the CP signal, these conventional correction approaches begin to fail due to measurement precision issues. Simply speaking, it is difficult to precisely measure a very small signal ( I CP ) riding on a very large background (leakage) utilizing the above approaches. II. EXPERIMENTAL PROCEDURES The key innovation of FMCP is the utilization of lock-in amplifier detection to eliminate unwanted signals (dc leakage current, drift/offset, and other noise) from the measurement. In FMCP, the gate electrode voltage pulse is modulated 0741-3106/$31.00 © 2013 IEEE