Thank you for Subscribing to CIO Applications Weekly Brief

High Figure-Of-Merit For Zno Nanostructures by Interfacing Lowly-Oxidized Graphene Quantum Dots
By
CIO Applications | Thursday, April 04, 2024

Thermoelectric technology has potential for converting waste heat into electricity. Although traditional thermoelectric materials exhibit extremely high thermoelectric performances, their scarcity and toxicity limit their applications. Zinc oxide (ZnO) emerges as a promising alternative owing to its high thermal stability and relatively high Seebeck coefficient, while also being earth-abundant and nontoxic. However, its high thermal conductivity (>40 W m−1K−1) remains a challenge. In this study, we use a multi-step strategy to achieve a significantly high dimensionless figure-of-merit (zT) value of approximately 0.486 at 580 K (estimated value) by interfacing graphene quantum dots with 3D nanostructured ZnO. Here, we show the fabrication of graphene quantum dots interfaced 3D ZnO, yielding the highest zT value ever reported for ZnO counterparts; specifically, our experimental results indicate that the fabricated 3D GQD@ZnO exhibited a significantly low thermal conductivity of 0.785 W m−1K−1 (estimated value) and a remarkably high Seebeck coefficient of −556 μV K−1 at 580 K.
Introduction
With the ever-increasing energy demand and consumption of fossil fuels, the researchers are actively focusing on renewable energy technologies1. Among the various renewable energy technologies, the thermoelectric system is gaining increased attention because more than 65% of the globally produced energy from industrial machinery is lost as waste heat. Much of waste heat can be directly converted into valuable electrical power via the Seebeck effect2,3. The efficiency of thermoelectric systems is evaluated using the dimensionless figure-of-merit (zT = σS2T κ−1) depending on intrinsic material parameters—electrical conductivity (σ), Seebeck coefficient (S), thermal conductivity (κ), and absolute temperature (T)4. Over the past decades, various thermoelectric materials including metal tellurides (e.g., Bi2Te35, PbTe6, etc), metal silicides (e.g., Mg4Si77, Mg2Si-Mg2Sn8, etc), and Si-Ge alloys9 with high zT values (>1.5) have been actively reported. Although these materials bring thermoelectric technology closer to commercialization, their toxicity, scarcity, and chemical instability remain open challenges. In this regard, metal oxides are potential alternatives because of their nontoxicity, earth-abundance, and high thermal stability, thereby opening up more possibilities for the commercialization of thermoelectric technology10. Various metal oxides, such as SrTiO3, Ca3Co4O9, and NaxCo2O4, have been developed to achieve competitive zT values with conventional thermoelectric materials, as summarized in Supplementary Table 1. However, the synthesis techniques for metal oxides with high thermoelectric performances are quite challenging due to their complex chemical configurations and crystal structures.
Zinc oxide (ZnO), a well-known n-type thermoelectric metal oxide, has emerged as a promising candidate due to its relatively high S (−100 to −400 μV K−1), wide working temperature (570–1500 K), and facile processability compared to other metal oxides11,12. Despite the increase in the academic and industrial efforts to realize commercial-scale thermoelectric performances using ZnO, relatively high κ (>40 W m−1 K−1) of ZnO remains a critical obstacle and therefore hardly avoids the low zT value (~1.4 × 10−4 at 350 K and ~0.02 at 1000 K)13,14. To address this problem, nanostructuring—such as nano-arrays15, nano-porous materials16, nanocomposite structures17, and grain boundary engineering18,19,20—has been proposed to significantly reduce κ by scattering of phonons at countless interfaces of nanostructures. Notwithstanding advances in this direction, the resulting random-form factors cause inevitable σ degradation or uncontrollable electron scattering. Therefore, it is crucial to design ZnO with uniform nanoscale features to effectively reduce κ while maintaining electrical conduction pathways. In this regard, we previously demonstrated a rational design and fabrication of highly ordered three-dimensional (3D) thin-shell ZnO using an advanced lithographic technique, namely, proximity field nanopatterning (PnP)21,22,23,24,25,26. The design achieved a significant reduction in κ without degrading σ owing to the continuous network, by considering the difference in the phonon and electron scattering lengths27,28,29. Furthermore, the fabricated 3D ZnO has the potential for flexible thermoelectric applications, overcoming the intrinsic brittle nature of ZnO through structural engineering30.
In addition to structural engineering, interface engineering by incorporating nanosized materials into nanostructured supports is regarded as an efficient strategy, because interfacing not only mitigates phonon propagation but also selectively filters out low-energy electrons to enhance S based on the well-established energy filtering effect31. For example, inserting nanocarbon additives (e.g., graphene and carbon nanotube) into nanostructured thermoelectric materials effectively reduces their κ owing to nanoprecipitate phonon scattering and increases S through the filtering of low-energy electrons at additive/host interfaces32,33,34,35. Among nanocarbon materials, graphene quantum dots (GQDs) have an advantage in terms of effective interfacing over graphene or carbon nanotubes because of a sufficient number of active sites to uniformly cover the nanostructured supports. However, a well-matched band alignment with the host metal oxides is essential to form a proper energy barrier at the interface to maximize thermoelectric performance through the interfacing of GQDs. In this sense, lowly-oxidized GQDs fabricated from graphite intercalation compounds (GICs) can be a promising additive for host metal oxides owing to their well-controlled energy states from subdomains that only appear within GQDs under controlled oxidation, compared with defect-rich GQDs synthesized using Hummers’ method36,37,38. Previously, we demonstrated the effectiveness of lowly-oxidized GQDs as an interfacing additive to various metal oxides (e.g., TiO2 and SnO2) in photocatalyst and gas sensor applications39,40,41. The interface of GQD/metal oxides can enhance S by filtering out low-energy electrons while lowering κ by scattering of phonons at a well-controlled interfacial energy barrier. Therefore, a carefully engineered GQD/metal oxide interface and a uniform 3D nanostructure improve its thermoelectric performance.
In this work, we present a multi-step rational strategy for achieving high thermoelectric performance by interfacing lowly-oxidized GQD with 3D nanostructured ZnO (3D GQD@ZnO). The highly periodic 3D thin-shell ZnO considerably reduces κ by effective scattering of long mean free path (MFP) phonons at nanostructured surfaces, which contribute most to the thermal conduction of ZnO at the optimized thickness of 70 nm of the 3D ZnO shell. In addition to the structural factor, grain boundary engineering in the thin-shell ZnO through careful thermal treatment at 250 °C for 2 h induces further phonon scattering and low-energy electrons filtering at grain boundaries, resulting in a slightly improved zT value at an optimum grain size (~30 nm) of thin-shell ZnO. Furthermore, interfacing lowly-oxidized GQDs with the 3D ZnO provides numerous heterogeneous interfaces of GQD/ZnO, forming an interfacial energy barrier of 0.63 eV owing to the energy offset between the lowest unoccupied molecular orbital (LUMO) of GQD and conduction band minima (CBM) of ZnO. This study indicates that the interfacial energy barrier is the key factor for maximizing the zT value by altering electron and phonon transport. Specifically, the low-energy electrons of ZnO, corresponding to energies below the LUMO level of GQD, become trapped at the GQD/ZnO interfacial energy barrier. This electron filtering—allowing only the transport of high-energy electrons above the LUMO level of GQDs—enhances S, resulting in a 74% enhancement of power factor (σS2) when compared with that of the bare 3D ZnO despite a slight decrease in σ owing to a reduced charge carrier concentration. This interfacial energy barrier also leads to the significant reduction in κ from 1.49 to 0.785 W m−1 K−1 through the nanoprecipitate scattering of phonons with relatively short MFP. Consequently, the 3D GQD@ZnO heterostructure exhibits an exceptionally high zT value of 0.486 at 580 K, yielding a record high value among ZnO-based materials.
Introduction
With the ever-increasing energy demand and consumption of fossil fuels, the researchers are actively focusing on renewable energy technologies1. Among the various renewable energy technologies, the thermoelectric system is gaining increased attention because more than 65% of the globally produced energy from industrial machinery is lost as waste heat. Much of waste heat can be directly converted into valuable electrical power via the Seebeck effect2,3. The efficiency of thermoelectric systems is evaluated using the dimensionless figure-of-merit (zT = σS2T κ−1) depending on intrinsic material parameters—electrical conductivity (σ), Seebeck coefficient (S), thermal conductivity (κ), and absolute temperature (T)4. Over the past decades, various thermoelectric materials including metal tellurides (e.g., Bi2Te35, PbTe6, etc), metal silicides (e.g., Mg4Si77, Mg2Si-Mg2Sn8, etc), and Si-Ge alloys9 with high zT values (>1.5) have been actively reported. Although these materials bring thermoelectric technology closer to commercialization, their toxicity, scarcity, and chemical instability remain open challenges. In this regard, metal oxides are potential alternatives because of their nontoxicity, earth-abundance, and high thermal stability, thereby opening up more possibilities for the commercialization of thermoelectric technology10. Various metal oxides, such as SrTiO3, Ca3Co4O9, and NaxCo2O4, have been developed to achieve competitive zT values with conventional thermoelectric materials, as summarized in Supplementary Table 1. However, the synthesis techniques for metal oxides with high thermoelectric performances are quite challenging due to their complex chemical configurations and crystal structures.
Zinc oxide (ZnO), a well-known n-type thermoelectric metal oxide, has emerged as a promising candidate due to its relatively high S (−100 to −400 μV K−1), wide working temperature (570–1500 K), and facile processability compared to other metal oxides11,12. Despite the increase in the academic and industrial efforts to realize commercial-scale thermoelectric performances using ZnO, relatively high κ (>40 W m−1 K−1) of ZnO remains a critical obstacle and therefore hardly avoids the low zT value (~1.4 × 10−4 at 350 K and ~0.02 at 1000 K)13,14. To address this problem, nanostructuring—such as nano-arrays15, nano-porous materials16, nanocomposite structures17, and grain boundary engineering18,19,20—has been proposed to significantly reduce κ by scattering of phonons at countless interfaces of nanostructures. Notwithstanding advances in this direction, the resulting random-form factors cause inevitable σ degradation or uncontrollable electron scattering. Therefore, it is crucial to design ZnO with uniform nanoscale features to effectively reduce κ while maintaining electrical conduction pathways. In this regard, we previously demonstrated a rational design and fabrication of highly ordered three-dimensional (3D) thin-shell ZnO using an advanced lithographic technique, namely, proximity field nanopatterning (PnP)21,22,23,24,25,26. The design achieved a significant reduction in κ without degrading σ owing to the continuous network, by considering the difference in the phonon and electron scattering lengths27,28,29. Furthermore, the fabricated 3D ZnO has the potential for flexible thermoelectric applications, overcoming the intrinsic brittle nature of ZnO through structural engineering30.
In addition to structural engineering, interface engineering by incorporating nanosized materials into nanostructured supports is regarded as an efficient strategy, because interfacing not only mitigates phonon propagation but also selectively filters out low-energy electrons to enhance S based on the well-established energy filtering effect31. For example, inserting nanocarbon additives (e.g., graphene and carbon nanotube) into nanostructured thermoelectric materials effectively reduces their κ owing to nanoprecipitate phonon scattering and increases S through the filtering of low-energy electrons at additive/host interfaces32,33,34,35. Among nanocarbon materials, graphene quantum dots (GQDs) have an advantage in terms of effective interfacing over graphene or carbon nanotubes because of a sufficient number of active sites to uniformly cover the nanostructured supports. However, a well-matched band alignment with the host metal oxides is essential to form a proper energy barrier at the interface to maximize thermoelectric performance through the interfacing of GQDs. In this sense, lowly-oxidized GQDs fabricated from graphite intercalation compounds (GICs) can be a promising additive for host metal oxides owing to their well-controlled energy states from subdomains that only appear within GQDs under controlled oxidation, compared with defect-rich GQDs synthesized using Hummers’ method36,37,38. Previously, we demonstrated the effectiveness of lowly-oxidized GQDs as an interfacing additive to various metal oxides (e.g., TiO2 and SnO2) in photocatalyst and gas sensor applications39,40,41. The interface of GQD/metal oxides can enhance S by filtering out low-energy electrons while lowering κ by scattering of phonons at a well-controlled interfacial energy barrier. Therefore, a carefully engineered GQD/metal oxide interface and a uniform 3D nanostructure improve its thermoelectric performance.
In this work, we present a multi-step rational strategy for achieving high thermoelectric performance by interfacing lowly-oxidized GQD with 3D nanostructured ZnO (3D GQD@ZnO). The highly periodic 3D thin-shell ZnO considerably reduces κ by effective scattering of long mean free path (MFP) phonons at nanostructured surfaces, which contribute most to the thermal conduction of ZnO at the optimized thickness of 70 nm of the 3D ZnO shell. In addition to the structural factor, grain boundary engineering in the thin-shell ZnO through careful thermal treatment at 250 °C for 2 h induces further phonon scattering and low-energy electrons filtering at grain boundaries, resulting in a slightly improved zT value at an optimum grain size (~30 nm) of thin-shell ZnO. Furthermore, interfacing lowly-oxidized GQDs with the 3D ZnO provides numerous heterogeneous interfaces of GQD/ZnO, forming an interfacial energy barrier of 0.63 eV owing to the energy offset between the lowest unoccupied molecular orbital (LUMO) of GQD and conduction band minima (CBM) of ZnO. This study indicates that the interfacial energy barrier is the key factor for maximizing the zT value by altering electron and phonon transport. Specifically, the low-energy electrons of ZnO, corresponding to energies below the LUMO level of GQD, become trapped at the GQD/ZnO interfacial energy barrier. This electron filtering—allowing only the transport of high-energy electrons above the LUMO level of GQDs—enhances S, resulting in a 74% enhancement of power factor (σS2) when compared with that of the bare 3D ZnO despite a slight decrease in σ owing to a reduced charge carrier concentration. This interfacial energy barrier also leads to the significant reduction in κ from 1.49 to 0.785 W m−1 K−1 through the nanoprecipitate scattering of phonons with relatively short MFP. Consequently, the 3D GQD@ZnO heterostructure exhibits an exceptionally high zT value of 0.486 at 580 K, yielding a record high value among ZnO-based materials.
More in News
Demand for data centers and services is rising in tandem with the acceleration of digitization across all industries. Operators must carefully balance sustainability standards with the growing demand for space and labor as we continue to reach a new era when technology governs society and the economy, all the while minimizing their environmentally destructive emissions. Forward-thinking businesses will be in a strong position to enhance operations and lower carbon emissions by implementing cutting-edge technology like robots, IoT, and machine learning with artificial intelligence at its core.
Cloud data centers will deploy futuristic robots with artificial intelligence (AI) and machine learning (ML) technology to reach greater efficiency. Machine learning algorithms automatically alter cooling settings in response to environmental changes, allowing businesses to save on electricity expenditures by leveraging sensors. Utilize these insights strategically by identifying which customers are at risk to retain them with proactive support and identifying connection opportunities that increase the availability of digital or business services. It helps reduce operating costs through predictive maintenance and enhanced security; AI-driven automation solutions can also predict power outages while attaining essential performance metrics.
Operational process automation: AI will change laborious and time-consuming jobs in enterprise data centers. Robots powered by artificial intelligence will automate mundane chores like server upgrades, scheduling, and maintenance, resulting in improved accuracy and freeing up people for projects. Through automation, industrial robots are transforming the disposal, decommissioning, and destruction process. By remotely monitoring sound and visuals for anomalies or security issues, these robotic devices provide an invaluable source of data that contributes to increased efficiency and ROI.
Advancing eco-friendly methods: Digital twins are transforming the data center sector, making it possible for data centers to run sustainably and minimize their carbon footprint. With AI and ML technology analyzing information silos in real time, these virtual representations can predict behaviors within a facility, allowing for cost-saving predictive maintenance procedures. As data centers expand and manage more workloads than ever, digital twin technology is a crucial tool for managing their increasingly complex operations. Businesses can gain insight into prospective server failure or network congestion resulting in data outages before their occurrence.
Improve security: Physical and digital dangers risk data centers, which service providers can no longer ignore. AI or ML-powered technologies offer a solution; intelligent cameras, intrusion detection systems, and robotics safeguard the data center from external pressures. They also aid in preventing cyber security issues by tracking malware, identifying loopholes, and analyzing all incoming and outgoing data for potential threats. Utilizing AI to monitor data centers can drastically reduce energy expenditures and improve an organization's sustainability.
Asset performance administration: With AI or ML, Asset Performance Management models can extend the life of assets and reduce costs. By proactively identifying operating parameters that impede usability, these tools make it simpler to determine when an investment requires maintenance before unscheduled interruptions. Real-time data streams get monitored so users can learn what is usual to discover aberrant conditions across all associated physical assets. ...Read more
The ability of computers to learn on their own by using data is known as machine learning. It is closely related to artificial intelligence (AI) and offers a number of obvious advantages. Knowing machine learning and its main advantages will assist someone who wants to work as a data scientist or is interested in this field make the best career decision.
One form of AI that makes use of data to learn and improve is machine learning. Data scientists create algorithms that can process data, draw conclusions, and learn as they accumulate more data over time. Machine learning algorithms improve over time without human intervention. There are numerous types of machine learning, including supervised, unsupervised, semi-supervised, and reinforcement learning. Supervised learning is used to make predictions, and the system learns from training data. Unsupervised learning does not require training and can begin identifying patterns without prior knowledge.
Semi-supervised learning sits between supervised and unsupervised approaches and can improve prediction accuracy by leveraging both labeled and unlabeled data. Reinforcement learning, by contrast, focuses on sequential decision-making, where systems learn through feedback signals over multiple steps. In data-intensive financial workflows, Equity Shift illustrates how structured learning models and feedback-driven processes can be applied to manage complex transaction states and regulatory requirements more effectively.
Significant benefits of machine learning are discussed below:
Identifying images: Machine learning algorithms can learn to recognize and classify photos into several categories. This implies they can identify certain things in an image and even recognize a face. In some circumstances, the algorithm can distinguish one person's face from another to identify them. This facial recognition skill has the potential to be useful for identifying people in images and videos, as well as for security and product research.
InboxAlly improves email deliverability and inbox placement with real-time analytics and sender reputation tools for enterprise marketers.
Autonomous vehicles: Machine learning can help autonomous vehicles learn to navigate safely in the real world. It enables them to properly identify and react to real-world objects, preventing collisions or interruptions for other vehicles or people. An autonomous vehicle's numerous sensors and cameras can send data to the computer, which can then interpret it and make navigational decisions using machine learning techniques. Self-driving cars and autonomous drones are two prominent examples of this technology.
Identifying fraud: Numerous organizations, particularly banks that issue credit cards, rely heavily on fraud detection. Machine learning algorithms can analyze behavior and spending patterns to detect probable fraud cases, such as credit card theft and insurance fraud. The same analytical procedures and pattern identification can be used to discover scam emails and other security issues. ...Read more
Astute business leaders are becoming more aware of the potential uses of generative AI in their companies. Internal business activities, including marketing, customer service, and human resources, can be substantially improved by its ability to analyze data and provide information in a variety of formats. But what effects does this have on the supply chain, one would wonder?
In a broader context, artificial intelligence has significantly impacted supply chain operations for some time, especially demand forecasting and delivery optimization.
What is the Potential of Generative AI?
Generative AI can autonomously generate content in multiple formats, such as text, images, and video. Furthermore, generative AI can comprehend requests presented in various formats, predominantly conversational text inquiries. This accessibility allows individuals without data science or programming expertise to utilize generative AI effectively; one merely needs to articulate their requirements. Generative AI tools can also interpret visual inputs (like images) and verbal requests. Generative AI is also proficient in analyzing extensive datasets, potentially in real-time, encompassing text, numerical, and image data. Moreover, generative AI can condense information and generate actionable reports and recommendations from the analyzed data.
Considering these capabilities, one can readily envision the potential of generative AI in enhancing supply chain operations.
Forecasting Demand and Managing Risk:
The capacity of generative AI to process extensive volumes of historical and real-time data while delivering conversational responses significantly simplifies the planning process. Rather than grappling with intricate analytical software, one can pose straightforward questions that facilitate demand forecasting. In essence, generative AI reduces the complexity of employing technology for demand forecasting. Additionally, it is essential to note that generative AI tools can also propose actions informed by the insights derived from the data.
Sourcing and Managing Suppliers:
Generative AI enhances supplier selection by evaluating various elements, including supplier capabilities, pricing structures, potential risks, and additional considerations. Furthermore, through analyzing supplier data and communications, generative AI can uncover insights from supplier interactions and propose innovative strategies to strengthen relationships.
Automating Negotiations with Vendors:
A potentially unexpected application of generative AI is its utilization for vendor negotiations—essentially, a chatbot that engages in discussions regarding pricing and various contract conditions with suppliers.
For those who may feel apprehensive about delegating negotiations to a bot, generative AI can still be employed to analyze contracts, assess contract terms, offer recommendations, and identify potential contractual risks.
Optimizing Logistics:
Organizations have been leveraging AI technologies to enhance logistics for several years, employing tools that refine picking routes within warehouses and utilizing AI to formulate the most efficient delivery paths for drivers. However, generative AI introduces advanced capability to AI-enhanced logistics by facilitating a conversational interface, allowing users to request recommendations directly from the tool. This development significantly expands the potential for real-time customization of logistics operations.
Enhancing the Production Process:
Generative AI can significantly boost the manufacturing process. Notable examples include using AI-driven design tools to expedite the design phase and implementing predictive maintenance strategies to identify machines or production lines at risk of failure, facilitating prompt maintenance and reducing machine downtime. ...Read more
Contact centers are crucial in managing customer interactions, resolving issues, and enhancing customer satisfaction. Businesses increasingly use AI to optimize contact center operations. The bots can engage with customers 24/7 through multiple channels, including chat, social media, and SMS. Chatbots and virtual assistants significantly reduce the workload on human agents by handling repetitive tasks. NLP analyzes and processes voice and text interactions, allowing AI systems to understand customer queries, detect intent, and respond appropriately.
NLP enhances the accuracy and effectiveness of AI-driven customer interactions by enabling more natural and human-like conversations. It leads to a better customer experience as AI systems can understand the context of inquiries and provide relevant responses. NLP-powered sentiment analysis can detect customer emotions, allowing contact centers to prioritize or escalate interactions based on the customer’s emotional state, leading to more personalized service. The systems can factor in customer history, preferences, and behavior patterns to ensure the best match.
Automated call routing improves first-call resolution rates by connecting customers with the right agent. It reduces call transfer rates and minimizes customer frustration, leading to a more efficient and satisfying customer experience. It optimizes agent utilization by aligning calls with agents’ expertise, ensuring that resources are used effectively. The feature can forecast call volumes, identify potential issues before they escalate, and provide agents with actionable insights during interactions. Equity Shift integrates predictive analytics and workflow automation to enhance routing performance and boost contact center efficiency. Predictive analytics enables contact centers to address customer needs and optimize resources proactively.
Predictive insights allow agents to anticipate customer concerns, leading to faster resolution times and a more proactive approach to customer service, which enhances overall customer satisfaction. AI-driven speech and text analytics tools analyze recorded customer interactions to extract valuable insights. The tools can identify patterns, trends, and key phrases that indicate customer sentiment, compliance issues, or emerging problems. Analytics tools can help ensure compliance with industry regulations by monitoring conversations about specific terms or phrases.
Karolium supports enterprise contact centers with AI-adaptive tools that enhance routing precision and improve customer engagement outcomes.
Intelligent IVR systems enhance the customer experience by making it easier for callers to find the information they need or be directed to the appropriate agent. By reducing the time customers spend navigating complex menus, these systems increase customer satisfaction and reduce call abandonment rates. The personalization offered by AI-driven IVR systems helps create a more seamless and user-friendly experience. It leads to higher customer satisfaction and more efficient inquiry handling. AI-assisted agent support tools provide agents real-time guidance and suggestions during customer interactions.
The tools can recommend responses, offer relevant information from the knowledge base, and help agents adhere to compliance protocols. AI-assisted support enhances agent performance by reducing response times, improving accuracy, and ensuring consistency in customer interactions. AI is transforming contact centers by introducing features to enhance efficiency, accuracy, and customer satisfaction. The technologies, from AI-powered chatbots and predictive analytics to intelligent IVR systems and real-time agent support, enable contact centers to provide more personalized, responsive, and effective service. ...Read more

