The manufacturing of modern semiconductor devices and chips requires specialized equipment, detailed analysis, and ultra-high precision. The processes are complex, with stringent customer response requirements and demand uncertainty.
Image Credit: Elizaveta Galitckaia/Shutterstock.com
Precision manufacturing is essential due to the nanometer-scale dimensions and intricate structures of modern devices.1 Atomic force microscopy (AFM) is crucial for studying and streamlining nanoscale fabrication, as well as characterizing semiconductor properties.
A Brief Overview of AFM
AFM techniques are optimal for measuring defects and in-line local responses in semiconductor materials, making them ideal for applications in material science, optics, and nanotechnology.
AFM provides detailed information about semiconductor nanostructure by employing a cantilever with a sharp tip at one end. The sample exerts a force on the tip, maintaining the cantilever in equilibrium. As the tip scans the sample surface (raster scanning in x and y directions), the cantilever moves to follow the surface contours and deflects a laser beam proportionally to the tip's displacement.
This process records the sample surface's profile or topography as a function of the lateral raster displacement in the x and y directions.2
Experts apply different forms of AFM during various stages of semiconductor manufacturing. Conductive AFM (C-AFM) is popular for investigating nanoscale transport properties during semiconductor processing, especially in semiconducting transition metal dichalcogenides (TMDs).
C-AFM allows for accurate nanoscale current-voltage (I-V) mapping, enabling precise analysis of 2D semiconductor materials. The use of a conductive tip records local voltage and current values on TMDs used in semiconductor fabrication.3
Another distinct type of AFM is Kelvin probe force microscopy (KPFM), which is extensively used by semiconductor manufacturers worldwide. KPFM involves passing an AC voltage through the tip, which is in contact with the specimen, and recording the cantilever's deflection as a feedback signal.
The KPFM process involves positioning a metal tip at a pre-specified distance from the semiconductor specimen. Instead of ordinary DC voltage, both AC and DC voltages are passed through the metal tip. The optimized DC voltage output is recorded as the KPFM signal, denoted as "V k."
For semiconductor samples, the tip-sample system forms a metal-insulator-semiconductor (MIS) capacitor, making the tip-sample interaction in KPFM more complex compared to metal samples.4
Applications of AFM in Semiconductor Manufacturing
Surface Characterization
With the advancement of nanotechnology, understanding nanoscale and sub-nanoscale surface properties has become increasingly critical. Methods such as coherence scanning interferometry, laser microscopy, and AFM are available for evaluating surface topography.
Modern semiconductors and electronic devices need highly accurate nanoscale surface properties to ensure efficiency, making AFM the top choice for providing 3D nanometer-resolution data.
The surface roughness of silicon wafers is critical as it can degrade certain electrical characteristics. Researchers conducted a study using two different wafer conditions, each with varying average roughness measured by AFM, to determine the performance differences between the semiconductors.
AFM images, consisting of 512 × 256 pixels, were used to calculate the root mean square plane roughness (Sq) for each sample. The study found that local narrow patch roughness of thin films influences semiconductor device performance parameters, such as film reliability, more intensely than long-region roughness.5
Thin Film Analysis
AFM is also used to analyze thin films utilized in semiconductor devices. A prime candidate is nanoscale TiO2 thin films, which are used in semiconductor devices and modern photocatalytic applications. The effectiveness of AFM for studying thin films was demonstrated when experts developed different TiO2 thin films with varying thicknesses and investigated each film's properties.
The surface morphology and roughness of the films were evaluated using AFM measurements. An increase in grain size and roughness of TiO2 thin films was observed with increased thickness, leading to roughening of the film surface due to changes in surface morphology.
The increased thickness also introduced additional resistance to the material, which is undesirable for electro-optic device applications.6 These films offer flexibility and precise control, crucial for improving the overall performance and attributes of semiconductor devices.
Failure Analysis
Another important application is semiconductor failure analysis. Failure analysis is crucial at every stage of semiconductor manufacturing, from product development to fabrication, assembly, and packaging.
C-AFM and Scanning Capacitance Microscopy (SCM) are two widely used AFM techniques for this purpose. C-AFM is particularly effective for analyzing bulk silicon devices, while SCM operates by inducing capacitance variations in the sample near the tip using a small AC bias.7
Nanomechanical Property Measurement
Over the past several decades, AFM has evolved from primarily imaging surface topography to characterizing various chemical, mechanical, electrical, and magnetic material properties with sub-nanometer resolution.
By analyzing tip-sample contact mechanics, AFM can convert this contrast into quantitative measurements of various nano-mechanical properties such as elastic modulus, shear modulus, wear rate, adhesion, and viscoelasticity.
The intermittent contact resonance AFM (ICR-AFM) mode is crucial for this purpose. It captures the changes in contact mechanics with indentation depth during each tip-sample interaction, providing a three-dimensional map of the characteristic contact.
ICR-AFM enables direct measurements of adhesive tip-sample interactions, targeting a quantitative 3D mechanical characterization of the sub-surface region. The depth-dependence of contact stiffness can uncover sub-surface inhomogeneities, facilitating a 3D structure-property characterization of the sample.8
In this way, AFM can be utilized for a variety of tasks during semiconductor fabrication, ensuring their high quality and effective nanoscale characterizations.
Future Prospects
AFM is widely used for material subsurface mapping and characterization, though its precision is limited by the size of the microscope's probe. Recently, researchers have utilized Artificial Intelligence (AI) to address this issue, developing a deep learning (DL) approach using an image-to-image translation methodology.
The algorithm employs datasets of tip-convoluted and deconvoluted image pairs to train an encoder-decoder-based deep convolutional neural network.9 The model was trained using AI-generated 3D structures, and then the model simulated and analyzed AFM readings.
The DL algorithm has been designed to incorporate the effects of probe size and then use AFM principles to characterize the semiconductor material at the nanoscale. This novel AI-integrated method is the first of its kind, far superior to traditional techniques, and successfully generated 3D-surface characterizations at much smaller nano-scales with greater accuracy.10
Researchers are expected to integrate more modern technologies to enhance the capabilities of modern AFM modes. In addition, new AFM techniques are being developed with much faster data acquisition rates and higher efficiencies.
These modern AFM techniques will be optimized for semiconductor materials like 2D perovskites and quantum dots. The application of novel AFM techniques will likely lead to future improvements in semiconductor design, nanoscale characterizations, and failure analysis.
More from AZoOptics: X-Ray Diffraction vs. X-Ray Fluorescence
References and Further Reading
[1] GF Hub Precision Technology Co. Ltd. (2023). How Precision Manufacturing is Driving Advances in the Semiconductor Industry. [Online] LinkedIn. Available at: https://www.linkedin.com/pulse/how-precision-manufacturing-driving-d4ioe#:~:text=Precision%20manufacturing%20plays%20a%20critical,and%20high%2Dperformance%20semiconductor%20components (Accessed on June 12, 2024).
[2] Reijzen, M., et al. (2022). Recent Advancements in Atomic Force Microscopy. Metrology, Inspection, and Process Control for Semiconductor Manufacturing XXXV. doi.org/10.1117/12.2595426
[3] Giannazzo, F., et al. (2020). Conductive Atomic Force Microscopy of Semiconducting Transition Metal Dichalcogenides and Heterostructures. Nanomaterials. doi.org/10.3390/nano10040803
[4] Xu, J., et al. (2021). Interpreting Kelvin probe force microscopy on semiconductors by Fourier analysis. J. App. Phys. doi.org/10.1063/5.0024073
[5] Mori, K., et al. (2020). Influence of silicon wafer surface roughness on semiconductor device characteristics. Japanese Journal of Applied Physics. doi.org/10.35848/1347-4065/ab918c
[6] Güzelçimen, F., et al. (2020). The effect of thickness on surface structure of rf sputtered TiO2 thin films by XPS, SEM/EDS, AFM and SAM. Vacuum. doi.org/10.1016/j.vacuum.2020.109766
[7] Chin, J., et al. (2011). Fault isolation in semiconductor product, process, physical and package failure analysis: Importance and overview. Microelectronics Reliability. doi.org/10.1016/j.microrel.2011.06.061
[8] Stan, G., et al. (2020). Atomic force microscopy for nanoscale mechanical property characterization. Journal of Vacuum Science & Technology B. doi.org/10.1116/6.0000544
[9] Bonagiri, L., et al. (2024). Precise Surface Profiling at the Nanoscale Enabled by Deep Learning. Nano Letters. doi.org/10.1021/acs.nanolett.3c04712
[10] Bonagiri, LKS, et al. (2024). Precise Surface Profiling at the Nanoscale Enabled by Deep Learning. doi.org/10.1021/acs.nanolett.3c04712
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.