Manufacturing Yield Research Articles

Multi-Domain Data Integration for Plasma Diagnostics in Semiconductor Manufacturing Using Tri-CycleGAN

The precise monitoring of chemical reactions in plasma-based processes is crucial for advanced semiconductor manufacturing. This study integrates three diagnostic techniques—Optical Emission Spectroscopy (OES), Quadrupole Mass Spectrometry (QMS), and Time-of-Flight Mass Spectrometry (ToF-MS)—into a reactive ion etcher (RIE) system to analyze CF4-based plasma. To synchronize and integrate data from these different domains, we developed a Tri-CycleGAN model that utilizes three interconnected CycleGANs for bi-directional data transformation between OES, QMS, and ToF-MS. This configuration enables accurate mapping of data across domains, effectively compensating for the blind spots of individual diagnostic techniques. The model incorporates self-attention mechanisms to address temporal misalignments and a direct loss function to preserve fine-grained features, further enhancing data accuracy. Experimental results show that the Tri-CycleGAN model achieves high consistency in reconstructing plasma measurement data under various conditions. The model’s ability to fuse multi-domain diagnostic data offers a robust solution for plasma monitoring, potentially improving precision, yield, and process control in semiconductor manufacturing. This work lays a foundation for future applications of machine learning-based diagnostic integration in complex plasma environments.

High‐speed and contactless inspection of defective micro‐LEDs through the photovoltaic effect

AbstractThe most significant challenge associated with micro‐light‐emitting‐diode (micro‐LED) displays, which are anticipated to be the next generation of display technology, is the high manufacturing cost. In order to reduce manufacturing costs, it is essential to improve yield. Improving the manufacturing yield of them necessitates the evaluation of micro‐LED chips prior to their installation onto substrates. However, the microsize and large quantity of these chips renders inspection difficult with conventional inspection methods. Herein, we propose a method for inspecting micro‐LED chips by measuring the voltage generated between the anode and cathode due to the photovoltaic effect using a developed proximity capacitance image sensor. As this inspection method does not require the use of probe pins to contact LED electrodes, it enables simultaneous inspection of multiple chips in a short time without causing any damage to the electrodes. In this paper, an experimental system equipped with this sensor was developed to demonstrate the basic measurement principle. Moreover, we demonstrated that more than 50,000 micro‐LED chips with a size of 60 μm × 34 μm can be simultaneously inspected in approximately 2 s.

Phase-separating resins for light-based three-dimensional printing of oxide glasses

Silica-based glasses can be shaped into complex geometries using a variety of additive manufacturing technologies. While the three-dimensional printing of glasses opens unprecedented design opportunities, the development of up-scaled, reliable manufacturing processes is crucial for the broader dissemination of this technology. Here, we design and study phase-separating resins that enable light-based 3D printing of oxide glasses with high-aspect-ratio features and enhanced manufacturing yields. The effect of the resin composition on the microstructure, mechanical properties and delamination resistance of parts printed by digital light processing is investigated with the help of printing experiments, compression tests and electron microscopy analysis. The chemical composition and microstructure of the cured resins were found to strongly affect the stiffness, delamination resistance, and calcination behavior of printed parts. These findings provide useful guidelines to enhance the reliability and yield of the DLP printing process of multicomponent silica-based glasses.

Challenges for Leading Edge node FINFET

The primary focus of this work is to find out the challenges associated with 7nm node finFET. To enable the current generation of gadgets/instruments, foundries are going toward smaller geometries (FinFET 7nm, 5nm, etc.) for manufacturing SOC/ASIC. To support the 7nm technology node without EUV, the Layout Design Rules have been scaled quite aggressively. As a result, obtaining satisfactory performance and yield in High Volume Manufacturing (HVM) has become a difficult undertaking. The gains in terms of power, performance, and other characteristics that become available with reduced geometries are the main drivers driving this movement. Analog/mixed-signal circuits, on the other hand, do not fully achieve these gains. They get increasingly difficult to design, with higher parasitic resistance and capacitance, more layout-dependent effects, and, in certain cases, layout growth. While in terms of fabrication the challenges are in node and cost, mask making, patterning, transistor formation, BEOL, MEOL, Technology parasitic elements and process control etc. This indicate what are the challenges for design of 7nm node FinFET.

Band flipping and bandgap closing in a photonic crystal ring and its applications

The size of the bandgap in a photonic crystal ring is typically intuitively considered to monotonically grow as the modulation amplitude of the grating increases, causing increasingly large frequency splittings between the "dielectric" and "air" bands. In contrast, here we report that as the modulation amplitude in a photonic crystal ring increases, the bandgap does not simply increase monotonically. Instead, after the initial increase, the bandgap closes and then reopens again with the two bands flipped in energy. The air and dielectric band edges are degenerate at the bandgap closing point. We demonstrate this behavior experimentally in silicon nitride photonic crystal microrings, where we show that the bandgap is closed to within the linewidth of the optical cavity mode, whose intrinsic quality factor remains unperturbed with a value ≈ 1×106. Moreover, through finite-element simulations, we show that such bandgap closing and band flipping phenomena exist in a variety of photonic crystal rings with varying unit cell geometries and cladding layers. At the bandgap closing point, the two standing wave modes with a degenerate frequency are particularly promising for single-frequency lasing applications. Along this line, we propose a compact self-injection locking scheme that integrates many core functionalities in one photonic crystal ring. Additionally, the single-frequency lasing might be applicable to distributed-feedback (DFB) lasers to increase their manufacturing yield.

Bioengineering of the Optimized Biosynthesis of Commercially Vital Carotenoids- Techno-Advanced Applications

Beta-carotene, a carotenoid found in plants, fungi, and algae, is a crucial antioxidant and anti-cancer agent. It is primarily derived from plants, algae, and microbes, but this method has drawbacks like high costs and low productivity. The growing demand for carotenoids has led to large-scale industrial manufacturing. However, extracting and synthesizing these chemicals can be costly and technical. Microbial synthesis offers a cost-effective alternative. Synthetic biology and metabolic engineering technologies have been used in various studies for the optimization of pathways for the overproduction of carotenoids. Four metabolic components are involved in carotenoid biosynthesis, central carbon (C), isoprene supplement, and cofactor metabolism. Metabolic engineering is a potential solution to enhance β-carotene production. This article explores the biochemical routes, methods used by natural microbial species, and metabolic engineering potential of microbial organisms for β-carotenoids production. Currently, Escherichia coli, certain euglena and yeast species are the primary microorganisms used in metabolic engineering, offering minimal environmental impact, cost-effective manufacturing, and high yield.

Mask-less Laser Direct Imaging & Adaptive Patterning for Fan-Out Heterogeneous Integration

Mask-less Laser Direct Imaging & Adaptive Patterning for Fan-Out Heterogeneous Integration

Phase 1 Study of CART-Ddbcma for the Treatment of Patients with Relapsed and/or Refractory Multiple Myeloma: Results from at Least 1-Year Follow-up in All Patients

Introduction: CART-ddBCMA, an autologous anti-B-cell maturation antigen (BCMA) chimeric antigen receptor (CAR) T-cell therapy with a unique, synthetic binding domain, is being studied in a first-in-human clinical trial in patients (pts.) with relapsed &/or refractory multiple myeloma (RRMM). One-year or more follow-up clinical data from all patients are presented in this report. Methods: Details of study design have been previously reported (Frigault et al Blood Adv 2023). Briefly, pts with RRMM who have received ≥3 prior lines of therapy were enrolled & received a single infusion of CART-ddBCMA following lymphodepletion chemotherapy (fludarabine: 30 mg/m 2/d & cyclophosphamide: 300 mg/m 2/d daily for 3 days). Two dose levels (DL1 & DL2 respectively) of 100 & 300 (±20%) ×10 6 CAR+ cells were evaluated. The primary endpoints included incidence of adverse events (AEs) & dose-limiting toxicities (DLTs). Additional endpoints included quality & duration of clinical response assessed according to the IMWG Uniform Response Criteria for MM, evaluation of minimal residual disease (MRD), progression-free (PFS) & overall survival (OS). Correlative analyses related to CART-ddBCMA expansion, characterization, patient & disease-related features were also evaluated. Results: As of June 2, 2023, 40 pts; median age 66 years (range: 44-76) were enrolled; 38 received CART-ddBCMA (32, DL1; 6, DL2) & 38 were evaluable for initial safety & clinical response. Two pts who were not dosed had cell product manufactured but were not eligible for cell infusion due to medical complications. Pts had a median of 4 (range: 3-16) prior lines of therapy. All infused pts (100%) were triple-refractory, & 26 (68%) were penta-refractory; 34 pts (89%) were refractory to last-line of treatment; 9 pts (24%) had high tumor burden with ≥60% bone marrow plasma cells; 13 (34%) pts had extramedullary disease; & 11 (29%) pts had high-risk cytogenetics (Del 17p, t(14;16), t(4;14)) at baseline. Median follow-up after CART-ddBCMA infusion was 22 months (range: 9-40 months). CAR+ cells comprised a median 70% (range: 48-87%) of total CD3+ T cells; median vector copy number was 2.2 copies/cell (range: 1.1-3.5); median cell viability was 98% (range: 90-100%), & median cell manufacturing yield was 1174 ×10 6 CAR + cells (range: 470-1626 ×10 6). CART-ddBCMA product characteristics were consistent with the specifications in all the lots, & there were no manufacturing failures. CRS occurred in 36/38 (95%) pts but only 1 pt in DL2 had grade (Gr) 3 CRS & all other cases were Gr≤2. ICANS occurred in 7 pts (5, Gr≤2; 2, Gr3), with 1 Gr3 case in each DL. All cases of CRS & ICANS resolved without further sequalae with management. No cases of off-tumor cell mediated toxicity, delayed neurotoxicity events (i.e., occurring after day 28), or Parkinsonian-like symptoms were observed. All 38 evaluable pts demonstrated investigator-assessed clinical response per 2016 IMWG criteria (ORR, 100%) with 22 sCR, 7 CR (≥CR rate, 76%), 6 VGPR (≥VGPR rate, 92%), & 3 PR. Responses deepened over time & conversion to CR/sCR was observed with longer follow-up (as late as month 12). Of those evaluable for MRD testing to date (n=29), 25 (86%) were MRD-neg at 10 -5. Median duration of response, PFS, & OS were not reached at the time of data-cut because 25 of 38 evaluable pts (66%) had ongoing response. The Kaplan-Meier estimated PFS rates for 6, 12 & 18 months were 92%, 74%, & 67% respectively. Durable responses were also observed in patients with high-risk features (EMD, BMPC ≥ 60%, or B2M ≥ 5.5 at baseline) & high-risk cytogenetics. PFS rates at 6-, 12-, & 18-months are shown in Table 1. Based on the results from the study, a dose of 115 ± 10 ×10 6 cells, consistent with DL1, was recommended for the phase 2 study. Conclusions: Adverse events with CART-ddBCMA, including CRS & ICANS, were manageable & no off-tumor tissue-targeted toxicity, delayed neurotoxicity, or Parkinsonian-like events were observed in the entire cohort at the time of data-cut. Ongoing efficacy results are encouraging, with 100% ORR, including 35 (92%) response of VGPR or better & 29 (76%) with CR/sCR. More importantly, clinical responses were durable with an overall estimated 18-mo PFS rate of 67% with comparable clinical responses seen in ‘high-risk’ patients known to have poor prognosis. Updated data with additional follow-up based on later data-cut will be presented.

Using image analysis techniques to enhance TFT-LCD manufacturing yield through timely detection of PI alignment layer defects

This study aims to develop a solution to the manufacturing process defects of thin-film transistor liquid crystal display (TFT-LCD) panels. Specifically, it focuses on online image analysis for detecting uneven coatings of the polyimide (PI) alignment layers, which are crucial in the TFT-LCD manufacturing process. The PI solution is typically applied to a substrate surface using an array of nozzles via droplet deposition. During this process, nozzle clogging can cause a series of defects in the form of spots on the alignment layer. These defects impact the alignment of liquid crystal molecules in the TFT-LCD, leading to decreased display quality, reduced visibility, uneven brightness, and compromised image details. If suspected signs of alignment layer spots are identified, the image analysis technology assists online inspectors with prompt responses to minimize TFT-LCD production errors. The results of this study demonstrate timely detection of alignment layer spots, prevention of their progression, and significant reduction of the time required for manual inspection to improve manufacturing yield. This newly developed approach is applicable to not only TFT-LCD production monitoring but also smart display panel manufacturing, thereby providing significant value in high-volume and large-area panel fabrication processes.

Enhanced mechanical property through high-yield fabrication process with double laser scanning method in two-photon lithography

Two-photon lithography (TPL) is a representative fabrication process for micro 3D structures with a sub-micron fabrication resolution. As the TPL process is developed and commercialized, the TPL process requires higher process yield to be a considered reliable manufacturing technique for mass production. However, delamination is one of the main factors that reduce the yield of the TPL process. In this study, a double-step laser scanning method is proposed as a reliable high-yield TPS process to reduce the delamination effect. The method entails a simple process that polymerizes the initial interfacial layer between the structures and substrate twice. The additional laser scanning step induces a strong crosslinking in the polymer, and thereby preventing swelling and delamination. Furthermore, the effectiveness of the method is evaluated with large-scale microlens arrays. Although the process time is slightly increased, the process yield is significantly improved through the use of the double-step laser scanning method. Hence, the double-step laser scanning process can be employed as a reliable additive manufacturing process for various 3D microscale devices, and more importantly, it can be used to increase the manufacturing yield through the reduction of defects caused by surface delamination.

Capillary Underfill Flow Simulation as a Design Tool for Flow-Optimized Encapsulation in Heterogenous Integration.

As the power electronics landscape evolves, pushing for greater vertical integration, capillary underfilling is considered a versatile encapsulation technique suited for iterative development cycles of innovative integration concepts. Since a defect-free application is critical, this study proposes a capillary two-phase flow simulation, predicting both the flow pattern and velocity with remarkable precision and efficiency. In a preliminary performance evaluation, Volume of Fluid (VOF) outperforms the Level-Set method in terms of accuracy and computation time. Strategies like HRIC blending, artificial viscosity, and implicit Multi-Stepping prove effective in optimizing the numerical VOF scheme. Digital mapping using physical experiments and virtual simulations validates transient flow predictions, achieving excellent agreement with deviations as low as 1.48-3.34%. The accuracy of flow predictions is thereby greatly influenced by non-Newtonian viscosity characteristics in the low shear range and time-dependent contact angle variations. The study further explores flow manipulation concepts, focusing on local flow speed adjustment, gap segmentation, and the use of arcuate shapes to influence interface confluence near the chip. Experimental validation corroborates the effectiveness of each design intervention. In conclusion, this research highlights the potential of predictive engineering to develop flow-optimized package designs that enhance reliability while supporting high manufacturing yields.

A novel bispecific antibody dual-targeting approach for enhanced neutralization against fast-evolving SARS-CoV-2 variants.

The emergence of new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants has caused unprecedented health and socioeconomic crises, necessitating the immediate development of highly effective neutralizing antibodies. Despite recent advancements in anti-SARS-CoV-2 receptor-binding domain (RBD)-specific monoclonal antibodies (mAbs) derived from convalescent patient samples, their efficacy against emerging variants has been limited. In this study, we present a novel dual-targeting strategy using bispecific antibodies (bsAbs) that specifically recognize both the SARS-CoV-2 RBD and fusion peptide (FP), crucial domains for viral attachment to the host cell membrane and fusion in SARS-CoV-2 infection. Using phage display technology, we rapidly isolated FP-specific mAbs from an established human recombinant antibody library, identifying K107.1 with a nanomolar affinity for SARS-CoV-2 FP. Furthermore, we generated K203.A, a new bsAb built in immunoglobulin G4-(single-chain variable fragment)2 forms and demonstrating a high manufacturing yield and nanomolar affinity to both the RBD and FP, by fusing K102.1, our previously reported RBD-specific mAb, with K107.1. Our comprehensive in vitro functional analyses revealed that the K203.A bsAb significantly outperformed the parental RBD-specific mAb in terms of neutralization efficacy against SARS-CoV-2 variants. Furthermore, intravenous monotherapy with K203.A demonstrated potent in vivo neutralizing activity without significant in vivo toxicity in a mouse model infected with a SARS-CoV-2 variant. These findings present a novel bsAb dual-targeting strategy, directed at SARS-CoV-2 RBD and FP, as an effective approach for rapid development and management against continuously evolving SARS-CoV-2 variants.

Formulation And Evaluation of Quercetin Loaded Nanosponges of Abiraterone Acetate

A novel idea in medicine delivery is the use of nanosponges. In the current research, an effort was made to create in-situ gel based on nanosponges that contained abiraterone acetate loaded with quercetin. The major purpose of nanosponges is to offer a prolonged release, lessen the negative effects of conventional dosage forms, and enhance the bioavailability of abiraterone acetate by using quercetin as a natural bioenhancer. Combining ethyl cellulose and eudragit in varying quantities, the quasi emulsion solvent diffusion technique was used to create nanosponges and quercetin as a natural bioenhancer. The formulation properties of the produced nanosponges were assessed. All batches of nanosponges had their manufacturing yield and trapping effectiveness assessed. Formulations F 03-P1 and F 07-P2 were found to be the optimum formulations based on results. Both F 03- P1 and F 07-P2 were used to create the gels G1 and G2. Viscosity, spreadability, drug content, and in-vitro tests showing 28.88% drug release after 24 hours all favored G2 as the optimal formulation. A gel base was created using the optimized formulation after it had been tested for scanning electron microscopy (SEM) analysis. Nanosponges containing gel had evaluated for different tests. According to the study, adding drugs to nanosponges can increase the rate of drug release, improve medication targeting at a particular spot, and hence lessen systemic toxicity.

Cetyltrimethylammonium bromide reformed ceria nanocomposites of chemical mechanical planarization for silica wafers

BackgroundChemical-mechanical planarization (CMP) is the most important process in the semiconductor industry, as it affects the product manufacturing yield. Of the several factors involved in planarization, the choice of abrasive slurries affects the extent to which impurities and roughness are removed. Cerium oxide (CeO2) is considered an outstanding polishing material because of its excellent hardness, well-controlled structure, and chemical force, and CMP requires well-dispersed slurry nanoparticles (NPs) with uniform structure, size, and shape. In this study, urea/alkaline-induced nanocubic MethodsCeO2 was synthesized in the presence of cationic surfactant cetyltrimethylammonium bromide (CTAB)-generated cubic CeO2 NPs, which were then used in a silica wafer polishing test. Dynamic laser scattering and transmission electron microscopy indicated that the cubic plate NP size distribution and zeta potential ranged from 18 to 30 nm when measured between –35 and –45 mV. Elemental composition via X-ray photoelectron spectroscopy was studied to demonstrate the mechanism of Ce3+/Ce4+ with crystal information, which matched the X-ray diffraction grain size calculation and finite element–scanning electron microscopy measurements. We also performed a simulation of cubic CeO2 NP formation and pillar CeO2 inhibition under an additional amount of structure-modifying agent (e.g., CTAB). Fourier transform infrared spectra indicated a successful CTAB reaction with CeO2. Small angle X-ray scattering spectra were used to obtain simulated NP shapes and parameters via the SasView model. A confocal laser scanning microscope was then used to measure the roughness and 2-D plane view of an 8-inch tetraethyl orthosilicate coated silica wafer. Significant findingsOur results showed that the bared silica wafer planarize with CTABx-CeO2 NPs reduced the ISO grade number from Level 9 to Level 7, which corresponds to a 70.7–74.5% roughness planarization of the raw material. This study presents a fast, uniform, low-roughness CMP slurry material synthesis and characterization method that can enhance the technique of slurry nanoparticle synthesis from stable materials.

A General Layout Pattern Clustering Using Geometric Matching Based Clip Relocation and Lower-Bound Aided Optimization

With the continuous shrinking of feature size, detection of lithography hotspots has been raised as one of the major concerns in Design-for-Manufacturability (DFM) of semiconductor processing. Hotspot detection, along with other DFM measures, trades off turn-around time for the yield of IC manufacturing, thus a simplified but wide-range-covered pattern definition is a key essential to the problem. Layout pattern clustering methods, which group geometrically similar layout clips into clusters, have been vastly proposed to identify layout patterns efficiently. To minimize the clustering number for subsequent DFM processing, in this paper, we propose a geometric-matching-based clip relocation technique to increase the opportunity of pattern clustering. Particularly, we formulate the lower-bound of the clustering number as a maximum-clique problem, and we have also proved that the clustering problem can be solved by the result of the maximum-clique very efficiently. Compared with the experimental results of the state-of-the-art approaches on ICCAD 2016 Contest benchmarks, the proposed method can achieve the optimal solutions for all benchmarks with very competitive run-time. To evaluate the scalability, the ICCAD 2016 Contest benchmarks are extended and evaluated. And experimental results on the extended benchmarks demonstrate that our method can reduce the cluster number by 16.59% on average, while the run-time is 74.11% faster on large-scale benchmarks compared with previous works.

12‐4: Invited Paper: Enhancing MicroLED Manufacturing with New Horizons in Metrology and High‐Throughput EL Inspection

microLEDs present a number of unique challenges in testing and inspection because of their unique structure and tiny size. In the earlier days of microLEDs, the greatest hurdles to overcome were miniaturization, chip design, and mass transfer. And while solutions for these challenges continue to evolve, today's most critical obstacles are of a different sort. The most fundamental problem in microLED manufacturing today is the very low yield in manufacturing. As a result, the industry is in need of effective testing and high‐throughput inspection equipment in order to enable high volume production.

A Best Balance Ratio Ordered Feature Selection Methodology for Robust and Fast Statistical Analysis of Memory Designs

Recently, machine learning yield models for Integrated Circuit (IC) have gained widespread prominence in the EDA community, and are very promising in terms of emulating memory design functionality and thereby speeding up circuit simulation based variance reduction methods. A main challenge that arises in this area is class imbalance that occurs naturally due to the high targeted manufacturing yield. Thus, the imbalanced nature of the sampled memory datasets can compromise the model performance. In this work, we attain deep insights into the memory classification problem for modeling rare fail events in the context of importance sampling based yield analysis. We propose a comprehensive and computationally efficient method that addresses the joint considerations of the best combination of relevant features and class balance ratios, which are key for classifier generalization capability. The methodology relies on synthetic minority oversampling techniques to enforce the minority class while probing for the best data balance ratio in conjunction with an iterative L1-SVM based approach that qualifies as an approximation to the L0-norm regularization for the best feature subset selection. We compare the proposed methodology against standalone L1-SVM solutions, unbalanced L0-norm approximation as well as an algorithmic data balancing method in the context of yield estimation methodology. The methodology is shown to result in high fidelity classifiers as demonstrated when analyzing the yield of a 14nm FinFET SRAM cross-section with speedup of 179x for the importance sampling simulations compared to pure circuit simulation based approaches and an average error of 0.19 sigma.

CmpCNN: CMP Modeling with Transfer Learning CNN Architecture

Performing chemical mechanical polishing (CMP) modeling for physical verification on an integrated circuit (IC) chip is vital to minimize its manufacturing yield loss. Traditional CMP models calculate post-CMP topography height of the IC’s layout based on physical principles and empirical experiments, which is computationally costly and time-consuming. In this work, we propose a CmpCNN framework based on convolutional neural networks (CNNs) with a transfer learning method to accelerate the CMP modeling process. It utilizes a multi-input strategy by feeding the binary image of layout and its density into our CNN-based model to extract features more efficiently. The transfer learning method is adopted to different CMP process parameters and different categories of circuits to further improve its prediction accuracy and convergence speed. Experimental results show that our CmpCNN framework achieves a competitive root mean square error ( RMSE ) of 2.7733Å with 1.89× reduction compared to the prior work, and a 57× speedup compared to the commercial CMP simulation tool.

Environment and economic analysis of reverse supply chain scenarios for remanufacturing using discrete-event simulation approach.

The study covers the concepts involved in reverse supply chain modeling using the case of a manufacturing company. The purpose of this study is to build a sustainable reverse supply chain model for resource conservation through remanufacturing of stator shafts by using a discrete-event simulation approach. The simulation studies in the reverse supply chain have taken up cases of either plastic or electronic waste remanufacturing, while very limited studies deal with simulation of sustainable reverse supply chains using a manufacturing industry case study from international customers. In this study, reverse supply chain using simulation study in manufacturing sector is carried out using Arena Rockwell simulation software. The simulation model is built using discrete-event simulation for returns from customers of two developed countries, i.e., Germany and the USA to Chennai, India. The study emphasizes full container load and less than container load modes of shipment scenarios and multiple return cases. The comparative analysis suggests that the value-added and non-value-added time of the reverse supply chain is slightly greater in the less container load scenario. The wait time per entity in remanufacturing processes similar for both shipment scenarios varies significantly based on return cases. The cost and carbon emission associated with transportation, in the reverse supply chain inclusive of social carbon cost, have also been estimated. Therefore, the study proposes a possible sustainable reverse supply chain framework that could be adopted by different manufacturing industries and yield opportunities for performance improvement.

The Impact of Changes in Crude Oil price on the Yield and Volatility of Manufacturing: Evidence from the US and China

This study is based on the data of WTI international futures crude oil price, China manufacturing Index and the United States Dow Jones Manufacturing Index from November 23, 2021 to June 16, 2022. This paper uses VAR and ARMA-GARCH models to analyze the changes in yield and volatility of manufacturing in the US and China. The results show that the change of crude oil price under the situation of Russia-Ukraine conflict has a negative effect on the return rate of manufacturing industry in China and the United States, but the impact on the return rate of manufacturing industry in China is greater than that of Dow Jones, and the impact on the return rate of manufacturing industry in the United States will gradually disappear as time goes by. In order to reduce the adverse impact of oil price fluctuations on manufacturing industry, this study puts forward some suggestions, such as adopting diversified ways of oil import and export trade and actively developing and utilizing other energy sources.