Photo by davide ragusa on Unsplash

IEEE International Conference on Microelectronic Test Structures

ICMTS 2024 Program

2024 Program Booklet

Day 1: Tuesday, April 16
08:00 Registration opens
09:00 Opening Remarks
S. Smith (General Chair), F. Driussi (Technical Program Chair)1
The University of Edinburgh, Edinburgh, UK
1University of Udine, Udine, Italy

Session 1: Layout Dependent Effects

09:10 1.1 Layout Dependent Effects Of Passive Devices And Their Impact On Analog Integrated Circuits
A. Jayakumar, M. van Dort, M. Vertregt, G.D.J. Smit, R. Lander, I. Liu, P. Volf
NXP Semiconductors B.V, Nijmegen, Netherlands
ABSTRACT: Analog blocks on products built using advanced CMOS technologies were seen to have deviating behavior on silicon than estimated by post layout simulations, especially in circuits used for biasing and reference generation. Circuit investigations pointed towards trimming resistor banks used for bias and reference current derivation. To detect possible unmodelled silicon effects experienced by these trimming resistor banks having different configurations in various analog blocks on products, a scribe-line test structure was implemented, aimed to capture major product design use-cases and study their effects. The outcome of this study is presented in this paper.
09:30 1.2 Test structures for studying the impact of the intrinsic contact metallization on the performance and stress sensitivity of SiGe HBTs
O. Dieball, H. Tuinhout, Jeroen Zaal1
NXP Semiconductors, Modeling, Eindhoven, the Netherlands
1NXP Semiconductors, Package Innovation, Eindhoven, the Netherlands
ABSTRACT: Based on various layout realizations of a typical SiGe HBT, the electrical impact of the intrinsic contact metallizations as well as their mechanical stress sensitivities are explored. This investigation provides valuable insights into performance shifts, important for design of modelling test structures as well as for high-precision circuit design.
09:50 1.3 A Step-by-step Layout Transformation Approach for Differentiating the Layout Dependent Effects on Device DC Performance
L. Lu, K. Xia, R. van Langevelde, C. C. McAndrew, W. Li
NXP Semiconductors, Front End Innovation, China
ABSTRACT: The ring oscillator is an important component of the electronic device. Its performance, however, is sensitive to parasitic effects when the device sizes scale down to sub-microns. We present a method, utilizing a series of test structures which step-by-step transforms the ring oscillator layout into a single-device modeling layout, to differentiate and quantify the impact of individual layout dependent effects (LDEs) on the device level DC performance. A specially tuned model that fits to the DC performance of the devices in a ring oscillator can give correct timing behavior.
10:10 Break
10:40 INVITED 1: What do we mean by measurement and can we trust it? - A commentary on the application of measurement process and its association with truth
P. Loftus
The University of Edinburgh, Edinburgh, UK
ABSTRACT: The topic of measurement underpins so much of our lives in science and everyday live that it can easily be taken for granted. As tests, and the equipment used in them, become more complex it is difficult to ensure the integrity of the conclusions that we draw from them. In this talk, Pete will draw on over 40 years working on aspects of measurement with a wide variety of stakeholders to take a step back from the applications and to consider measurement itself - what it is, why we do it, and whether we can trust it.

Session 2: Reliability

11:20 2.1 A novel test structure with two active areas for eNVM reliability studies
K. Alkema, F. Melul, V. Della Marca1, M. Bocquet1, M. Akbal, A. Regnier, S. Niel2, F. La-Rosa
STMicroelectronics, Rousset, France
1Aix-Marseille University, CNRS, Marseille, France
2STMicroelectronics Crolles, France
ABSTRACT: This paper presents a test structure with a poly floating gate shared on two actives areas. Programming and erase can be split toward these two regions with a specific arsenic implantation. The aim is to study the tunnel oxide degradation and the injection efficiency of embedded charge storage memory cells.
11:40 2.2 Test Structures of Cross-Domain Interface Circuits with Deep N-Well Layout to Improve CDM ESD Robustness
H. -M. Huang, M. -D. Ker
National Yang Ming Chiao Tung University, Hsinchu, Taiwan
ABSTRACT: Charged-device model (CDM) electrostatic discharge (ESD) is a complex reliability issue for integrated circuits in advanced CMOS technology. Cross-domain interface circuits are particularly susceptible to CDM ESD during cross-domain ESD events. In this study, CDM ESD robustness of cross-domain interface circuits with deep N-well was investigated through test structures fabricated in a 0.18-μm CMOS technology.
12:00 2.3 A 4H-SiC Trench MOS Capacitor Structure for Sidewall Oxide Characteristics Measurement
H. -L. Huang, L. -T. Hsuesh, Y. -C. Tu, B. -Y. Tsui
National Yang Ming Chiao Tung University, Hsinchu, Taiwan
ABSTRACT: Test structure for evaluating gate oxide properties on the trench sidewall in 4H-SiC is proposed. Using the thick bottom oxide and poly-Si spacer structure, we are able to measure the C-V characteristics directly and extract the interface state density. It is observed that typical NO annealing process cannot passivate the trench etching induced defects effectively.
12:20 Lunch
13:50 Presentations by Exhibitors

Session 3: Cryogenic Characterization

14:20 3.1 Transistor Matrix Array for Measuring Variability and Random Telegraph Noise at Cryogenic Temperatures
T. Mizutani, K. Takeuchi, T. Saraya, H. Oka1, T. Mori1, M. Kobayashi, T. Hiramoto
The University of Tokyo, Tokyo, Japan
1National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan
ABSTRACT: Addressable transistor arrays using 65 nm bulk technology were fabricated and tested at cryogenic temperatures. It was confirmed that variability vs. 1/√𝐿𝑊 relationships at 1.5 K slightly degrades compared with 300 K. Random telegraph noise (RTN) was also measured and existence of extremely slow RTN at 1.5 K was confirmed using a quasi-parallel measurement technique.
14:40 3.2 Gaussian process-based device model toward a unified current model across room to cryogenic temperatures
M. Shintani, T. Iwasaki, T. Sato1
Kyoto Institute of Technology Matsugasaki, Kyoto, Japan
1Kyoto University Yoshida-hon-machi, Kyoto, Japan
ABSTRACT: We apply a Sparse Gaussian process to build a compact model for CMOS circuits operating at cryogenic temperatures. The Sparse Gaussian process prevents overfitting problem unlike neural networks. Evaluation using measurement results of nMOS transistor fabricated by 65 nm demonstrate that the I-V characteristics from 4K to 300K are accurately modeled.
15:00 3.3 A Testbed for Cryogenic On-wafer Noise Measurement Using Cold Source Method with Temperature-Dependent Loss Correction
G. -W. Huang, B. -Y. Chen, Y. -S. J. Shiao, C. -W. Chuang, L. -C. Shen, K. -M. Chen, C. -S. Chiu
Taiwan Semiconductor Research Institute, Hsinchu, Taiwan
ABSTRACT: On-wafer noise measurement at cryogenic temperatures is challenging due to complexity of temperature gradient distributions in cryostats. With proposed correction formulas, the cold source method may work properly for cryogenic on-wafer noise measurement and be a good fit for which involves of a large number of samples, e.g., Known-Good-Die Testing.
15:20 Break
15:50 INVITED 2: Characterization of ferroelectric HfOx-based devices – methods and test structures
S. Slesazeck
NaMLab, Dresden, Germany
ABSTRACT: The discovery of ferroelectricity in doped hafnium oxide that was firstly published in 2011 by Böschke et al. strongly increased the interest in ferroelectric memory devices. The polarization reversal in these thin films is used to store information in three flavors of ferroelectric memory devices – FeCAP, FeFET and FTJ. While the programming of all three devices is performed simply by applying an electrical field being larger than the ferroelectric coercive field, the read operation is very different. Hence, the electrical characterization methodology of these devices is strongly influenced by the whole device design and material stack, rather than being dictated by the properties of the ferroelectric layer itself. In my talk I will discuss the specific requirements for the ferroelectric device characterization and corresponding test structure designs.

Session 4: Dielectrics and Ferroelectrics

16:30 4.1 Analysis and compensation of the series resistance effects on the characteristics of ferroelectric capacitors
M. Massarotto, F. Driussi, M. Bucovaz, A. Affanni, S. Lancaster1, S. Slesazeck1, T. Mikolajick1, D. Esseni
DPIA, University of Udine Udine, Italy
1NaMLab gGmbH, Dresden, Germany
ABSTRACT: Ferroelectric device optimization requires a dependable characterization of the ferroelectric (FE) material. Here, we highlighted how series resistance (RS) impacts the I-V characteristics of Metal-Ferroelectric-Metal (MFM) stacks with peculiar distortions, possibly leading to an inaccurate extraction of the FE parameters and a misleading interpretation of its switching dynamics. For the first time to our knowledge, we here propose a procedure for an improved extraction of the FE parameters even in presence of a significant series resistance.
16:50 4.2 Impedance Measurement Platform for Statistical Capacitance and Current Characteristic Measurements of Arrayed Cells with Atto-order Precision
K. Saito, T. Suzuki, H. Mitsuda, T. Nozaki, T. Mawaki, R. Kuroda
Tohoku University, Sendai, Japan
ABSTRACT: A novel impedance measurement platform that enables high-precision statistical measurements of C-V and I-V characteristics is presented. The platform consists of 366H × 228V cell array and a common readout circuit, and the pre-formed device under test (DUT) cells were measured to verify its measurement performance. The results demonstrated that for about 5,000 DUTs, the C-V measurement can measure fF-order capacitance with 0.06 % precision and the I-V measurement can measure fA-order current with 0.6 % precision.
17:10 4.3 Compact expression to model the effects of dielectric absorption on analog-to-digital converters
S. Saro, P. Palestri1, E. Caruso2, P. Toniutti2, R. Calabro2, S. Terokhin2, F. Driussi
DPIA, University of Udine Udine, Italy
1University of Modena and Reggio Emilia, Modena, Italy
2Infineon Technologies, Villach, Austria
ABSTRACT: An analytical model for the dielectric absorption on capacitors and its impact on the errors induced in ADC conversion is here proposed. The reported simulations are consistent with the results of the R-C model widely used in the literature and demonstrate that a large set of experiments can be fitted just calibrating a single model parameter.
17:30 End of Day 1
17:30 Exhibitor and Sponsor Reception
Day 2: Wednesday, April 17
09:00 BONUS TALK: How to write excellent papers
C. Cagli
STMicroelectronics, Crolles, France
09:30 INVITED 3: Empowering MEMS Innovation: Bridging Access through EU Access Programmes
V. Sandeep Nagaraja
Tyndall National Institute, Cork, Ireland
ABSTRACT: The talk will give a brief introduction to the PiezoMEMS research work being carried out in Tyndall National Institute, Ireland. It will dwell into the technology being developed by the group in various areas like space technology, communication systems and packaging. To empower research, it is essential to bridge the gap that exists between researchers and infrastructure / technology platforms. The talk will give an overview of different EU infrastructure and technology access programme like EUROPRACTICE, ASCENT+ and INFRACHIP.
10:10 Break

Session 5: MEMs and Sensors

10:40 5.1 Test structure for chemiluminescence measurement in aqueous solutions: initial design
A. A. Moreno-Guerrero, C. Dunare, A. Walton, P. Lomax, J. Terry, I. Underwood
The University of Edinburgh, Edinburgh, UK
ABSTRACT: On-chip integration of sensitive sensors is in high demand due to their extensive applications in chemical analysis. This research introduces a novel test structure for easy opto-electrical integration in microfluidic systems. The architecture has proven versatile for validating the fabrication process and as a final sensor for Luminol-based analysis.
11:00 5.2 Droplet Impact Sensing with Low Noise Coplanar Reverse-Electrowetting Test Structures
A. Moyo, M. W. Shahzad, S. Smith1, J. G. Terry1, Y. Mita2, J. Lewis, Y. Li
Northumbria University, Newcastle, UK
1The University of Edinburgh, Edinburgh, UK
2The University of Tokyo, Tokyo, Japan
ABSTRACT: Reverse electrowetting on dielectric (REWOD) has emerged to be a promising sensing technology for low frequency vibrations and deformations. This study aims to use test structures to study the parasitic elements in characterising the coplanar electrode (CREW) configuration of REWOD. This allows for better understanding in device design and optimization.
11:20 5.3 Analysis methodology of Deep Trench Isolation Field-Effect Passivation techniques for Image Sensors through dedicated test structures
S. Tlemsani, S. Ricq, O. Marcelot1, P. Magnan1
STMicroelectronics Crolles, France
1ISAE, Toulouse, France
ABSTRACT: In this work a new methodology is proposed to characterize the Deep Trench Isolation field-effect passivation thanks to dedicated test structures. Two deep trench isolations are studied and quantitatively compared: an electrically active trench and an innovative passive trench using a charged Al2O3/SiO2 stack. The results show that the best passivation is achieved here with the active trench.
11:40 5.4 Parametric Optimization of RF MEMS Variable Capacitor with high linearity for C-Band Application
S. Shaheen, P. Lomax, T. Arslan
The University of Edinburgh, Edinburgh, UK
ABSTRACT: A design methodology is proposed to obtain a linear C-V response and large capacitance tuning ratio by avoiding the pull-in effect. The study includes structural optimization of RF Microelectromechanical system (MEMS) varactor, using simulation tools ANSYS and HFSS. The optimized design shows better isolation at 4.2 GHz which makes it appropriate for C-Band applications.
12:00 Lunch
13:30 ICMTS 2025 Presentation

Session 6: Wafer Measurements

13:40 6.1 Efficient Characterization Methodology for Low-Frequency Noise Monitoring
L. Pirro, T. Chohan, P. Liebscher, M. Juettner, F. Holzmueller, R. Jain, Y. Raffel1, K. Seidel1, R. Olivo1, A. Zaka, J. Hoentschel
GlobalFoundries, Dresden, Germany
1Fraunhofer-IPMS, Dresden, Germany
ABSTRACT: In this work, a novel methodology to characterize the Low-Frequency Noise LFN on large device statistics is proposed. The maximum drain current fluctuations over time are measured. The slope of the LFN distribution is modeled with physical equations related to the basic device properties. The approach is validated by studying the impact of transistor geometry (different gate width, number of fingers and length) as well as gate oxide thickness and characterization temperature. In conclusion, the proposed methodology is tested evaluating different process integration elements. The outcome is compared to classical 1/f read-out for final confirmation.
14:00 6.2 Rapid MOSFET threshold voltage testing for high throughput semiconductor process monitoring
M. H. Herman, T. T. Nguyen, K. Wong, J. Johnson, B. Morris
Advantest Corporation, San Jose,USA
ABSTRACT: We describe a rapid method for testing MOSFET threshold voltage (Vt). Multiple spot Ids measurements are compared to stored reference data. Each spot measurement gives an independent Vt, and multiple Vt estimates are used to generate quality metrics. Two spot measurements permit an accurate Vt within 7 msec.
14:20 6.3 Statistical investigation of SnOx RRAM memories for switching characteristics
A. G. Panca, A. Serb, S. Stathopoulos, T. Prodromakis
The University of Edinburgh, Edinburgh, UK
ABSTRACT: Resistive Random Access Memory (RRAM) has seen significant developments in the past years. An important improvement is to solve the device-to-device variability that limits RRAM implementation. A proposed solution is to measure a high number of devices to statistically identify trends and evidences behind the cause. Here, we present a method to characterise the RRAM switching behaviour and analyse a large dataset of SnOx RRAM, in a statistical fashion for further optimization and modelling.
14:40 6.4 Making Accurate and Consistent Wafer Measurements with Next Generation Guarded True-Kelvin MEMS DC Probes
C. B. Sia, Y. Funatoko, I. Kunioka, M. Watanabe, P. Andrews, K. Dawson, M. Sameshima, T. Saeki, J. Yang, J. Li, X. Li, S. Guo, L. Fan, W. M. Lim, E. Wilcox, A. Lord, S. Lastella, N. Kawamata, J. Klattenhoff, J. Kister, M. Losey, M. Slessor
FormFactor Inc., Singapore
ABSTRACT: As silicon-based transistor scales down, it becomes increasingly challenging for test engineers to perform precise and repeatable waferb measurements. Reducing aluminium-capped copper test pads to 30μm×30μm and smaller means that new test strategies with novel true Kelvin MEMS probes are needed to overcome these emerging test challenges.
15:00 Break
15:20 INVITED 4: 3D interconnects characterization: basic test structures and electrical measurement
M. Stucchi
imec, Leuven, Belgium
ABSTRACT: 3D stacking technology is based on vertical interconnects, as Through-Silicon Vias (TSV), Hybrid Bonding (HB) pads, microBumps (uB). Their function is providing electrical connections among dies stacked vertically for power supply delivery, signal propagation, clock distribution. Low-resistive electrical continuity is the primary requirement of these vertical interconnects; however, parasitic components as resistance at metal interfaces, capacitance between interconnects, leakage current, need to be carefully characterized as well and benchmarked with the requirements of system-level applications of the 3D stacked dies. Furthermore, results of the 3D interconnect characterization are essential information for the optimization of 3D architectures (interconnect dimensions and materials) and processing. This talk presents examples of 3D interconnect test structures and their relative measurement methodologies; their important role for 3D technology optimization is highlighted by examples of applications and case studies.

Session 7: S-Parameters and De-embedding

16:00 7.1 Understanding the substrate effect on de-embedding structures fabricated on SOI wafers using electromagnetic simulation
B. Neckel Wesling, M. Deng1, C. Mukherjee1, T. Mikolajick, J. Trommer, C. Maneux1
NaMLab, Dresden, Germany
1IMS, Talence Cedex, France
ABSTRACT: In this paper, we present the fabrication, characterization and electro-magnetic (EM) simulation of open pad test-structures, emphasizing the impact of substrate properties on RF performance. We demonstrate that a high-resistivity substrate is essential to minimize losses and capacitances in RF measurements for technologies using Silicon on Isolator (SOI) wafers.
16:20 7.2 Modified Bisection Thru-Only Deembedding Algorithm for Long Test Fixtures
A. Quint, L. Valenziano, J. Hebeler, T. Zwick, A. Bhutani
Karlsruhe Institute of Technology, Karlsruhe, Germany
ABSTRACT: Thru-only deembedding is a simple deembedding method using a single deembedding structure. Common thru-only deembedding methods often have overshoots in the deembedded S-parameters when the deembedding structure is λ2 long. This is caused in the bisection step. A modified approach is presented, which extracts the phase of the S-parameters of the deembedding structure before the bisection step, performs the bisection on the magnitude of the S-parameters and applies the phase afterwards again, therefore eliminating the overshoots. The accuracy of the proposed algorithm is verified on measured test structures.
16:40 7.3 Evaluation of Lab-based Lithium Niobate Surface Acoustic Wave Test Structure Using Efficient Maskless Lithography and SMA Connection Approach for Microfluidic Applications
M. S. Parvez, S. Hussain, M. Goeckner, C. D. Young, J. -B. Lee
The University of Texas at Dallas, Richardson, USA
1Baylor University, Waco, USA
ABSTRACT: Surface Acoustic Wave (SAW) devices, particularly those made of Lithium Niobate (LiNbO3), are extensively used in telecommunications and microfluidic applications. This work describes the fabrication of a LiNbO3-based SAW test structures using maskless photolithography for rapid device dimension changes as well as introduces a cost-effective technique to solder SMA connectors to delicate substrates such as LiNbO3-based SAWs, without the need to fully create printed circuit boards. The SMA connected device facilitated improved characterization results compared to simple copper tape connections. The characterization accuracy is then validated through simulation.
17:00 End of Day 2
19:00 Banquet - Dovecot Studios
Day 3: Thursday, April 18

Session 8: GaN Technology

09:30 8.1 Test Structures to Investigate ESD Robustness of GaN Devices for Applications of Circuit Integration
W. -C. Wang, M. -D. Ker
National Yang Ming Chiao Tung University, Hsinchu, Taiwan
ABSTRACT: ESD robustness of E-HEMT GaN devices was investigated through test structures fabricated in a GaN-on-Silicon process. The experimental results showed that the ESD robustness is proportional to the device dimension when it was operating in the forward mode. In addition, with the gate-coupled design, the ESD level of E-HEMT GaN device can be further improved.
09:50 8.2 Method for GaN HEMT IV Characterization Without Trap-Related Memory Effects
J. Bremer, N. Rorsman, M. Thorsell
Chalmers University of Technology, Gothenburg, Sweden
ABSTRACT: A new method to perform IV output characterizations of FETs is presented, which minimizes the influence of trap-related memory effects. This is achieved by minimizing the charging effects caused by the electric fields of preceding measurement points. The proposed method provides rudimentary IV characteristics, useful for technology evaluation and modeling purposes.
10:10 8.3 A Neural Network-based Manufacturing Variability Modeling of GaN HEMTs
F. Chavez, D. Bavi, N. C. Miller1, S. Khandelwal
Macquarie University, Macquarie Park, Australia
1Michigan State University, USA
ABSTRACT: A new technique to accurately model the manu- facturing variability of GaN HEMT using neural network(NN) is presented. PCA is used to automatically generate parameters from variations in I-V data, which are then used to train the NN. The trained NN model captures the I-V behavior of 115 GaN HEMT with excellent accuracy.
10:30 8.4 Use of DC Probes for Multi-MHz Measurements of Crosstalk and Substrate Coupling in Gallium Nitride Power Integrated Circuits
M. Cui, S. Lam
Xi’an Jiaotong-Liverpool University, Suzhou, China
ABSTRACT: With simple compact pads, DC probes were used for measurements of crosstalk and substrate coupling in gallium-nitride power integrated circuits. By proper calibration, a crosstalk voltage down to 4.4 mV and substrate coupling up to -38 dB were measured up to 25 MHz, using an oscilloscope and spectrum analyzer respectively.
10:50 Break

Session 9: Process/Material Monitoring

11:20 9.1 An add-in Test Structure Chip to Unitedly Assess PVD Material Properties in University Open Nanotechnology Platform
S. Yasunaga, K. Misumi, A. Mizushima, A. Toyokura, E. Ota, Y. Inoue, M. Fujitawa, N. Kawai, M. Yoda, S. Tsuboi, T. Sawamura, A. Higo, R. Nakane, Y. Ochiai, Y. Mita
The University of Tokyo, Tokyo, Japan
ABSTRACT: Open nanotechnology platform at universities and research institutes is becoming an essential tool for agile and rapid microelectronic devices research and development. To further extend its capability with more users, process data acquisition and digital transformation (DX) is of high priority. In order to uniformly acquire process data in an open platform, where users’ process charts are all different, we are developing an add-on test structure chip. In the conference, we would like to share a chip design and measurement results to assess material parameters deposited by physical vapor deposition (sputtering and evaporation) materials.
11:40 9.2 Electrical behaviour of ALD-molybdenum films in the thin-film limit
K. van der Zouw, S. D. Dulfer, A. A. I. Aarnink, A. Y. Kovalgin
University of Twente, Enschede, The Netherlands
ABSTRACT: The thin-film electrical properties of molybdenum layers grown by atomic layer deposition were determined. Among others, four-point collinear probe, Van der Pauw, and (circular) transfer length method [(C)TLM] structures were designed and fabricated, which allowed to determine the sheet resistance (Rsh), temperature coefficient of resistance (TCR), contact resistance (Rc), transfer length (LT), and the field effect in a few nm thin Mo films.
12:00 9.3 Test Structure to Assess Bump Shape Influence on Hybrid Bonding
A. Mizushima, K. Misumi, S. Yasunaga, A. Higo, R. Nakane, K. Tsumura1, K. Higashi1, Y. Ochiai, Y. Mita2
The University of Tokyo, Tokyo, Japan
1TOSHIBA CORPORATION, Kawasaki, Japan
2The University of Tokyo, Tokyo, Japan; 
ABSTRACT: Bump bonding is widely used in CMOS-MEMS integration. As compared to the other technology (planar bonding), bump bonding is more interesting for its simplicity and applicability to wide range of device types. Towards higher density and reliability improvement, the authors have designed and fabricated a test structure. Our new contribution is to intentionally structuralize shapes of bumps to not only a flat head but concave as well as convex ones, expecting better occlusion and passive alignment. To experiment our new idea, the first test structure has been fabricated to assess alignment accuracy and resistance.
12:20 Best Paper Award, Closing
12:35 Lunch for Excursion Attendees
13:30 Excursion

 ICMTS Sponsors:
 Top