Polymer nanocomposites for high-energy-density capacitor dielectrics: Fundamentals and recent progress
Modeling partial discharge phenomena
Space charge behavior in polymeric materials under temperature gradient
As a background to this editorial, a simple search on Google was done on the term “capacitor”. It resulted in 146 million hits. Modifying the search to “power capacitor” resulted in 124 million hits, “high voltage power capacitor” gave 67 million hits, “low voltage capacitor” gave 55 million hits and “medium voltage power capacitor” gave 17 million hits.
There are several types of power capacitors and the focus will be on the types that basically use polymer dielectrics, and more specifically polypropylene (PP). These capacitor types can either be impregnated with some type of fluid (impregnated capacitors) or be of a dry type (dry capacitors). In the case of impregnated capacitors, it is only full film types that will be considered.
There are also special power capacitors, using mixed dielectrics, designed to provide some specific characteristics and performance. One example of a mixed dielectric system is PP/paper where the paper layer will reduce the capacitance change as a function of temperature. Some other mixed dielectrics are known for specific applications but will be omitted here.
General aspects of power capacitors
A power capacitor is a passive low-loss electrical apparatus with a multitude of applications in the transmission and distribution of electrical energy on all voltage levels up to and exceeding 1 MV (1 million Volt). In all their applications, capacitors lower CO2 emissions by reducing losses, securing high power quality and enabling integration of renewables. As a fact, a capacitor is an exceptional environmentally friendly component and its’ potential is far from fully exploited. With the growing awareness of CO2 emissions and global warming, one can expect an increased public demand for technologies and solutions with positive environmental impact. Capacitor-based solutions fall within that category.
A capacitor consists, in principle, of two conducting electrodes separated by a dielectric material, here considered single or multiple layers of PP film. The electrode is either a thin metal foil or a metallized layer evaporated onto the PP film. These two electrode technologies give different characteristics and performance of the capacitor – the metal foil capacitor is non-self-healing, the metallized layer capacitor (with at least one metallized layer) can be made self-healing.
The PP films, that are the dielectric material in between the electrodes, are specially made for capacitor applications. They are characterized by being able to withstand high electric stress, have very low losses, low degradation over time and with other properties meeting the requirements from capacitor manufacturing e.g. shrinkage and swelling. The films for impregnation have controlled surface roughness on both sides to facilitate a complete fluid impregnation. The films for metallization have a smooth surface on both sides and are corona treated to facilitate the metallization.
The impregnation of a power capacitor is done with a special synthetic fluid that interacts with the PP film as well as filling all voids in and around the dielectric. In the capacitor manufacturing, the impregnation process is of utmost importance to be done correctly in order to achieve the required performance. Also, for metallized capacitor types, there are special applications where impregnation is beneficial for the performance. The impregnant for such capacitors is different with another set of fluid characteristics and specifications.
Power capacitor design
The electric energy is stored in a capacitor statically by charge separation in an electric field between two electrode plates. The amount of charge stored per unit voltage is essentially a function of the area of the electrodes, the permittivity of the dielectric material, and the separation distance (i.e. dielectric thickness) between the electrodes. The voltage between the electrodes is limited by the voltage withstand of the dielectric material (including possible edge effects at the electrodes) and their separation distance. This means that a larger area combined with thin dielectrics will give the highest capacitance. The area of the electrodes is, to a large extent, limited by capacitor manufacturing constraints.
The development of the film manufacturing technology has resulted in today’s state-of the-art where standard high-quality capacitor film thicknesses are available from less than 2 µm and up to about 16 µm. The DC electric breakdown field of the film is in the range of several hundred V/µm. In impregnated capacitors, two or three layers of film are combined to reach the optimal design stress and voltage levels.
Build-up of power capacitor
For impregnated power capacitors, the electrodes (Aluminum foils 4 – 5 µm) and the dielectric PP films (total thickness 15 – 40 µm) are wound together to form a capacitor element. For metallized non-impregnated power capacitors, the build-up of elements is similar. Two layers of the metallized film (thickness 2 – 8 µm/layer), each with one electrode facing the non-metallized surface of the other film forms a capacitor element.
The design limiting factors in a capacitor element are defined by the electric field breakdown strength of the dielectrics (the combined film/fluid system) and the withstand at the electrode edges. In order to reduce the field enhancement at the edges of the Al foil, it is folded at the edge, giving a consistent round geometry. Another way to introduce a smoother foil edge is the use of laser cut foil whereby, in theory, the heat from the laser cut creates and leaves a “ball of melted material” at the foil edge. However, it is not only the geometry of the foil edge that defines the electric field limits that may yield partial discharges (PD), it is also the design of the film/fluid system around the foil edge. Availability of the free capacitor fluid at the foil edge reduces the probability of PD appearance. However, this effect is temperature dependent and must be considered in the design.
To form a capacitor unit, capacitor elements are stacked together, connected in series and/or in parallel, and inserted into a container with terminals brought out. By connecting capacitor elements in series, it is possible to accommodate voltage rating specified for the unit. By connecting capacitor elements in parallel, it is possible to accommodate the capacitance rating. In addition, the capacitor unit contains various electrical and thermal barriers for securing insulation requirements. One electrical insulation barrier is the insulation between the container and live parts inside it. The design may also have live parts at different potentials that will need electrical insulation.
Failure modes and protection
A capacitor element can store a lot of energy. If a breakdown of the dielectrics in a film-foil capacitor occurs, it creates locally an arc that releases this energy and melts the polymer layers as well as the Al foil electrodes. A breakdown in a film-foil capacitor always turns into a permanent shortage. All capacitor elements in parallel will also see this shortage and feed all their energy into it. The arc generates plasma and hot gases that may also initiate subsequent breakdowns in adjacent elements. The design of a capacitor unit must thus handle such situations as why there are restrictions on how much parallel energy can be tolerated; for guaranteeing that breakdowns in capacitor units and capacitor banks never cause any rupture of the container.
In a metallized capacitor, with self-healing properties, the failure mode in case of a breakdown of the dielectrics yields a fault current peak flowing only through the point of failure (< 1 mm2). This current almost instantly heats the thin metallized area around the failure point and evaporates the metal. If correctly designed, the area around the failure point becomes non-conducting (metalized layer removed) and the generated gas quenched, thus the fault current ceases.
The impact on capacitance in case of a dielectric breakdown in a film-foil capacitor corresponds to the number of affected elements versus the total number of elements (typically a few percent). In a metallized film capacitor, the impact is only loss of active area in the failure point (< 1 mm2) where the total area can be as large as hundreds of square meters. This means that a metallized film capacitor can experience a very large number of dielectric breakdowns before reaching its lifetime criterion, e.g. 5% loss of capacitance.
Several protection aspects have been developed in order to handle large amounts of energy, especially in failure modes of film-foil capacitors. The insulation barriers serve as both electrical and energy resistant barriers. Impregnated high density pressboard is used for this purpose. An additional advantage is that the impregnating fluid can penetrate not only electrically to strengthen the pressboard but also will reach the dielectrics in the elements.
To limit failures related to electrical breakdowns inside the capacitor units, the following designs have been adopted:
- internally fused capacitors – each individual capacitor element in a capacitor unit is internally connected via a current-limiting fuse that disconnects (isolates) the failing element,
- externally fused capacitors – each capacitor unit has its own outside fuse for disconnecting it from the bank after a certain number of elements fail,
- fuse-less capacitors (two types, conventional and internally stringed) – the capacitor units are designed to minimize the impact on capacitance change.
Film-foil capacitors, protected in the above described way, are characterized by different pros and cons and are intended for different applications and bank connections.
Metallized film capacitors have also different designs for accommodating various requirements. The metallization can be divided into smaller areas/segments connected via narrow bridges that act as fuses. The advantage is that such pattern metallization guaranties a safe end of life (capacitors can lose all their capacitance and remain electrically insulating). The drawback, however, is the cost associated with this safety feature – a certain fraction of the film area remains inactive and hence a larger film area is needed to obtain the required capacitance.
Power capacitors – specifics and applications
Profound knowledge and understanding in several disciplines and technologies are necessary to accomplish the characteristics and performance of high-quality power capacitors. A robust capacitor for a specific application is a result of the relevant design combined with the appropriate material selection and manufacturing processing.
The dielectric insulation in a capacitor is exposed to stress levels far above those typically used in other HV-components. The operating stresses in AC capacitors are in the range of 60 – 100 kVrms/mm and even higher in DC capacitors. Several hundreds of square meters of the dielectric is exposed to such high field levels in capacitor units. The appearing technology challenges include the obvious need for securing long lifetime expectancy (20-30 years) and limiting ageing due to the exposure to high field and relatively high operating temperature. These must be addressed at the material manufacturing process, i.e. when selecting properties of PP granulate and controlling morphology, thickness variation, surface roughness, mechanical properties, number of pinholes and other defects of the film. It is the strict control of these parameters that differentiates the large area performance of PP films supplied by various manufacturers.
There are certainly also key operations in the capacitor manufacturing process itself that have major impact on the performance of the final product. Setting proper winding parameters and their control is one among those operations. This is valid for both impregnated film capacitors as well as for metallized film capacitors. For impregnated capacitors, the drying and impregnation process is vital. In-depth understanding of the physics involved as well as fluid dynamics is necessary to secure long-term high performance. Very low internal losses are one of the characteristics formed in this process.
Impregnated film capacitors are used in AC applications for reactive power compensation. They are also used in power quality and harmonic filter applications. Impregnated capacitors are also designed for and used in DC applications. Such capacitors can be designed for ambient temperatures ranging from -50 0C to +55 0C, for high altitudes >1000 m and for subsea applications down to 3000 m in depth.
Metallized film capacitors are mainly used in DC applications and for AC applications for voltages <1000 V. It is the self-healing feature of metallized films that makes a significantly higher DC operating stress possible. With higher stresses, the energy density (Joule/litre) increases and the footprint (the required volume) reduces, which is beneficial for DC applications and has made some new DC solutions possible. As regards the temperature range for metallized film capacitors, some manufacturers define the minimum operating temperature at -40 0C, while others indicate its higher limits.
Market trends and future opportunities
Similar scenarios are indicated by different sources regarding future trends in the development of electricity generation, transmission and distribution, where the common features include increased demand, more penetration of renewables, local generation, changes in transmission and distribution infrastructure, electrification of transportation. In addition to the economic growth, the climate change, environmental concerns and sustainability are the additional drivers. In almost all technical solutions that are discussed or proposed for the future electricity systems as well as for the upgrade of the existing systems, capacitors is one of the needed building blocks, as the future needs will require a great deal of flexibility to handle fast changes in the electricity system conditions. One example is the power quality requirements in industrial low voltage networks where the power factor correction or reduction of harmonics can be done using semiconductor-based active apparatuses containing capacitors, but the specifications will be different. Such active solutions are nowadays widely used in low voltage applications (<1000 V). However, the development of semiconductor components’ performance continuously proceeds, making the demand for higher voltage solutions grow as well.
One can see, based on analyzing historical developments, that leaps in technology for power capacitors has been driven by new demands combined with opportunities arising from successful material development. The dry type metallized film DC capacitors is a good recent example of the steps in the energy density increase and the future power electronic circuits are expected to require sources of energy characterized by high power density as well. Today there is a difference between characteristics of a capacitor and a battery. The battery has high energy density (J/kg) and low power density (W/kg). In the capacitor, these parameters are reversed. The time constants are also different; the capacitor can be charged/discharged at a time scale 10,000 times faster than the battery. Supercapacitors are in this context filling the gap between the battery and the capacitor, but some of their parameters remain not yet mature. The car industry can, in this context, serve as an example of where there are demands for better performing batteries. Similar demands to have fast responding systems that also can store large amounts of energy also come from energy industry. Even if one can expect that a future combined technology for batteries, capacitors and supercapacitors will be developed and a new family of ”SuperBatCaps” components will evolve, it is the author’s conviction that the need for power capacitors and high-performance dielectrics will continue to increase.
From The Editors
From the Editors’ Desk March-April 2020
In this issue of the Magazine the focus is on materials for energy storage. We start with an extensive review on this subject and an editorial written by Birger Drugge, who explains power capacitor design and elucidates challenges that need to be considered during manufacturing. Thereafter, limitations that are overlooked in modelling of partial discharge activity under AC excitation are discussed for taking them into consideration in future research. We finally present the effect of temperature gradient on space charge accumulation in dielectrics.
The first article in this issue is an extensive review on “Polymer Nanocomposites for High-energy-density Capacitor Dielectrics: Fundamentals and Recent Progress”. It is written by Qi Li and Sang Cheng of Tsinghua University in Beijing, China. Dr. Qi Li is the awardee of the 2018 IEEE Caixin Sun and Stan Grzybowski Award for young-professional achievements and the article has been solicitated to summarize his excellent presentations at recent DEIS conferences. The article concentrates on polymer dielectrics. These are characterised by high breakdown strength, low loss and unique self-clearing behavior but suffer from low dielectric constant, while ceramics exhibit high loss, low breakdown strength, along with high dielectric constant. To combine the advantages of polymers and ceramics as well as to suppress their drawbacks, nanocomposite dielectrics consisting of a polymer matrix and embedded nano-sized ceramic fillers have been developed and this review introduces the recent progress in dielectric polymer nanocomposites toward high energy density capacitors, with the focus placed on interfacial engineering of the nanofillers, controlling their spatial arrangement, employing high-aspect-ratio nanofillers and multiple nanofillers, as well as implementing topological modulation of the composite film. Nanocomposites for high-temperature applications are also covered in this review. The main purpose for using nanocomposites in high-temperature applications is to promote the thermal stability and suppress the leakage current, which can be realized by either introducing insulating nanofillers or by coating a dense layer of wide-bandgap materials. Finally, the existing challenges for commercialization of the nanocomposite dielectrics are briefly discussed.
The second article of this issue, entitled “Modelling Partial Discharge Phenomena” is authored by George Callender and Paul Lewin of Southampton University, UK. It refers to the seminal work of Niemeyer and others in the early 1990’s, which is the basis for the majority of PD activity modelling to date. However, many of the equations used in these works stem from experimental investigations conducted in the 1960’s and 1970’s. The purpose of this article is to introduce this earlier work to the reader, discuss possible limitations and provide a link to the physical concepts and assumptions used in PD models. The concepts discussed include the inception electric field required for a PD to take place; the residual electric field immediately after the discharge is extinguished; and the mechanisms of electron generation to seed discharges. It was found in all cases that the experimental investigations are not necessarily directly applicable to typical PD systems. This is especially true in the case of the residual electric field, where there seems to be limited evidence that the electric field within a void will be reduced to a fixed residual value after any PD event. Moving forwards models of plasma dynamic simulations appear to be a useful tool to provide additional insight into the physical mechanisms behind discharges and could be used to inform PD activity models. There is still significant scope for additional investigations before a detailed physical interpretation of PD activity in operational high voltage plant can be achieved.
The third feature article in this issue is entitled “Space Charge Behavior in Polymeric Materials under Temperature Gradient” and authored by Kai Wu, Rui Su and Xia Wang of Xi’an Jiaotong University, China. This article reports on significant effects imposed by the existence of temperature gradient on space charge distributions in various practically important dielectrics. It is shown that the presence of temperature gradient produces in various material systems heterocharge accumulation adjacent to the electrode with a lower temperature. However, addition of nanoparticles can strongly suppress this accumulation. The authors also present results of simulations showing that the space charge behavior under temperature gradient becomes highly dependent on the balance between injection, drift and extraction processes and thus notably related to the electric field and temperature.
To awaken interest in a new and promising field of dielectric studies, we also decided to publish in this issue a report on an interesting workshop, entitled “Next-generation dielectric materials for microelectronics/electrical applications”, which was recently organized by researchers from University of Texas at Austin. This research area is, in our opinion, very important for being extensively developed at the interface between power electronics and high voltage technology disciplines. We thus look forward to receiving articles elucidating it.
News From Japan
World’s First DC 400-kV XLPE Cable System
The completion of installation and the commencement of operation of a new high-voltage (HV) DC power interconnection system connecting between two major islands in Japan, Honshu and Hokkaido, using cross-linked polyethylene (XLPE) insulated cables, was reported in “News from Japan” in the November/December issue of the Electrical Insulation Magazine . As a continuing article, the outline of a 400-kV HVDC XLPE cable system connecting between the United Kingdom (UK) and Belgium is introduced in this article .
As mentioned in , building a large circular power grid is very important for achieving efficient, stable, and robust operation of transmission of electric power. This is especially the case for the UK and countries in the European Continent, where a tremendously large number of facilities for electric power, such as wind power and solar energy, have been installed. With this background, a joint venture, called the NEMO Link interconnector, was founded between National Grid Interconnector Holdings Limited, a subsidiary company of the UK’s National Grid Plc., and Elia Group in Belgium, aiming at installing the first electricity interconnection between the UK and Belgium.
Building a power link between the UK and Belgium has many advantages. First, the electricity transmission network in Belgium is densely connected all over Europe. Secondly, holding a large power grid helps the UK to make its power system more robust. Thirdly, the distance between the two countries is relatively short.
The NEMO link is composed mainly of the following systems. The first system is two sets of symmetrical monopole 400 kV HVDC converter stations that convert AC to DC, and vise versa, with a capacity of 1000 MW, one in the UK and the other in Belgium. The second system is a DC 400 kV XLPE-insulated cable that connects the two converter stations. Figure 1 shows the cable route of the NEMO HVDC link. Since power transmission systems in the UK and continental Europe are operated in AC, two AC/DC converter stations, equipped with a voltage sourced converter (VSC), are set in Richborough, Kent, UK and Zeebrugge in Belgium. These two sites were chosen for the following reasons. Many offshore wind farms and other renewable power generating facilities have been built in the coastal area of Belgium, while the electricity demand is very high in the south east region in the UK. The main part of the link, an approximate length of 130 km, is under seawater. This subsea section is connected to the two converter stations via two land sections, 2 km in the UK and 9 km in Belgium.
Figure 2 shows a profile of the seabed along the cable route. The seawater depth is less than 40 m in most parts, although it is around 55 m at the deepest portion near the UK. The highest sand wave accumulation was estimated to be 12 m. The structure and material of the cable system was sensitively decided by taking account of these conditions.
In June 2015, a contract for engineering, procurement, and construction (EPC) was made between the J-Power Systems Corporation, Tokyo, Japan and NEMO Link Ltd. The J-Power Systems Corporation is a subsidiary of Sumitomo Electric Industries, Ltd., which transferred the transmission cable and conductor business to Sumitomo in 2016.
Figure 3 shows the cross section (a) and a photo of a real model (b) of the submarine cable. The conductor has a cross section of 1100 mm2 and is watertight for the installation at sea depth of 100 m. The cable is equipped with 16 single-mode optical fibers for data communication and 16 multimode optical fibers for temperature distribution measurement. The optical fiber system allows online monitoring of the occurrence of damages to the cable during its installation and operation.
Figure 4 shows the cross section (a) and a photo of a real model (b) of the land cable. The land cable has a large cross section of 1600 mm2 compared to the submarine cable, while it has 16 single-mode optical fibers for data communication and 16 multimode optical fibers for temperature distribution measurement similar to the submarine cable.
The manufacture of the cable was started in April 2016 in Sumitomo’s Ibaraki Works and completed in February 2018. The submarine cable was manufactured by dividing it into four lots. The first two lots, 69 km each, and the second two lots, 71 km each, were loaded out in May 2017 and in February 2018, respectively. Figure 5 is a photo that shows a scene of embarking the cable into a transport ship from an onshore cable yard in Ibaraki Works, while Figure 6 shows a bundle of cable being loaded into the ship. After the finish of very difficult work for cable installation into the sea and on the land, the NEMO link started its commercial operation on the 31st of January 2019. The designed maximum cable operating temperature is 90°C, much higher than other DC XLPE cables.
This article was completed in cooperation with Mr. Satoshi Nishikawa of Sumitomo Electric Industries, Ltd.
-  Y. Ohki, “News from Japan – A New 250-kV HVDC XLPE Cable System in Japan”, IEEE Electr. Insul. Mag., Vol.35, No.6, pp.43- 45, 2019.
-  T. Igi, S. Asai, S. Mashio, S. Nishikawa, S. Tomioka, T. Miyazaki, and T. Kazama, “Qualification, installation and commissioning of world’s first DC 400kV XLPE cable system”, JICABLE’19, A6.1, Versailles, France, 23-17 June 2019.
Fellowship in IEEE is intended to recognize major contributions to engineering, science, and technology, of significant value to society. No more than 0.1% of total membership may become fellows in any given year. Senior members who have been members of IEEE for at least five years may be nominated for fellowship. The nominator prepares the appropriate forms, determines the IEEE Society or Technical Council which will evaluate the nominee, and must find between three and five fellows who will provide references. Up to three endorsements may also be provided by individuals, who do not have to be fellows.
The Fellows Committee of each appropriate Society or Technical Council evaluates its nominees, and provides a numerical rating and an explanation of its evaluation, for each of its nominees. These are submitted, along with the nomination form, references, and endorsements, to the IEEE Fellow Committee, which considers service to professional engineering societies, in addition to the criteria which have previously been evaluated. This committee then provides numerical ratings for all of the nominees to the IEEE Board of Directors, which names the new fellows.
DEIS congratulates two members who have been elevated to IEEE Fellow – Yuji Suzuki and George Chen.
Yuji Suzuki (M’01-SM’11-F’20) was born in Tokyo, Japan in 1965. He received the B.S., M.S. and Dr. Eng. degrees in mechanical engineering from The University of Tokyo in 1987, 1989, and 1993, respectively. He joined Nagoya Institute of Technology, Japan as a lecturer in 1994, and moved to The University of Tokyo in 1995. He is currently with the Department of Mechanical Engineering, The University of Tokyo, as a full Professor. His research interests include energy harvesting using electrets, development of new polymer electrets based on quantum chemical analysis, microscale combustion, and optimal design of micro thermo-fluids systems. He is a Fellow of JSME (The Japan Society of Mechanical Engineers), a Senior Associate Editor of IEEE Sensor Letters, Steering Committee Members of International Conference on Micro and Nanotechnology for Power Generation and Energy Conversion Applications (Power MEMS), and Organizing Committee Member of IEEE International Symposium on Electret (ISE). He is also working for standardization of energy harvesting devices through IEC (International Electrotechnical Committee) as co-convenor of IEC TC47/WG7. He also served as General Co-chair of IEEE MEMS 2010, General Chair of PowerMEMS 2017, and Chair of Micro-Nano Science & Technology Division (2015), and Chair of Thermal Engineering Division (2020), JSME.
In 2004, he firstly proposed MEMS-friendly amorphous fluorinated polymer CYTOP (Cyclic Transparent Optical Polymer, AGC) as an electret material, which can give high surface charge density and long-term stability of charges. The new nano-cluster-enhanced CYTOP polymer is commercialized through collaboration with industry. Recently, he employed quantum chemical analysis to investigate the effect of end group of CYTOP polymer on the charging performance. He has discovered a range of compositions with greatly improved charge density and stability compared to earlier formulations, and realized a record high surface charge density of 4 mC/m2 for 15 um thick films. He has also developed a novel charging technology using soft X-ray, and proposed MEMS electret power generators with different designs for vibration energy harvesting.
George Chen was born in China in 1961. He received his BEng (1983) and MSc (1986) degrees in electrical engineeringfrom Xi’an Jiaotong University, China. He took a teaching assistant position at Xi’an Jiaotong University before receiving a scholarship for his PhD study in the UK in 1987. After he obtained the PhD degree (1990) from the University of Strathclyde on the work of permanent changes in electrical properties of irradiated low-density polyethylene, he joined the University of Southampton as a postdoctoral research fellow and became a senior research fellow subsequently working on an industrially sponsored project of HVDC submarine cables. In 1997 he was appointed as a lecturer and promoted to a reader in 2002. He is now a professor of high voltage engineering at the University of Southampton and a visiting professor of Xi’an Jiaotong University and Chongqing University. Professor Chen is currently the head of Electrical Power Engineering Research Group at the University of Southampton. Professor Chen has supervised 30+ PhD students to the completion and collaborated with researchers/engineers from both academic institutions and electrical power industries. Over 30 years of his professional life, Professor Chen has attracted research funding from government and industry. He has developed a wide range of interests in high voltage engineering and electrical properties of materials. Professor Chen has made significant contributions to space charge measurement and interpretation for improving dielectric performance. He has been an active member in IEEE, CIGRE and IEC and published more than 180 peer reviewed journal papers.
New Materials for Emerging Electrical Environments—Workshop Report
V. Bahadur, A. Ouroua, P. Acharya,
M. Lokanathan, S. Strank, and R. Hebner
University of Texas at Austin
Wide band gap (WBG) semiconductors are delivering on the promise of higher power and energy density circuits. In doing so, they are also opening an opportunity for a new generation of dielectric materials. Researchers from the University of Texas at Austin organized a 1.5 day workshop on “Next-generation dielectric materials for microelectronics/electrical applications” at the Massachusetts Institute Technology on December 4-5, 2019 focusing on this emerging opportunity. The invitation-only workshop brought together the materials researchers with those attempting to improve power electronics (PE) packaging to compare information that would accelerate the field (Figure 1).
Electrification, particularly transportation electrification, is placing a premium on higher power and energy density. Increasing the power and energy density reduces the parasitic mass, making electrified vehicles more attractive. An enabling technology to meet this need is the development of wide band gap (WBG) semiconductors. WBGs enhance power and energy density by operating at higher temperatures, voltages, and frequencies than legacy semiconductors. While this capability enables new applications, it also poses considerable challenges, since this new operating environment can significantly reduce system life. It is noted that accelerated life testing routinely involves increasing voltage, frequency and/or temperature. Overall, replacing legacy dielectric materials is likely a key step if these devices are to meet their full potential.
Responding to this opportunity, many novel materials are being developed. Specific efforts include the development and characterization of novel filler-insulating material combinations, the evaluation of approaches to incorporate fillers in a neat polymer, and improvements in the performance of traditional material systems.
Holistic multifunctional assessments of these materials have not received sufficient attention, however. In addition to shortfalls in information on material properties, other challenges that are usually not reported in the basic materials literature include scalability, cost and resistance to harsh environments. But all of these are important for designers.
Material Development Considerations
Experience has made it clear that the life of a dielectric material depends on the electrical, mechanical and thermal environment in which it is used. Most researchers in the field have apparently concluded that improvement in thermal behavior is the foremost priority. However, it is also vital to preserve electrical and mechanical properties as the thermal properties are improved.
An important opportunity arises because the majority of recent studies and commercial products consider polymers that are not well-suited for high temperature applications (> 200 °C). The development of composites based on polymers better suited for high temperature applications will benefit packaging as well as energy storage applications. In the domain of energy storage, high temperature materials is a bigger challenge than high dielectric constant materials. In addition, enhancing the properties of the polymer (without adding nanoparticles) may be an alternative to composite development. In particular, the upper limits on the potential enhancement of the thermal conductivity of polymers is not well understood.
Another possible opportunity to extend thermal ranges in packaging applications may be among those materials in which operation above the glass transition temperature (Tg) is feasible. It is important to develop an understanding of specific performance degradation mechanisms (thermal, mechanical, electrical) that occur above Tg and consider whether they can be remediated/alleviated for specific materials (e.g. self-healing materials) or applications.
Enhanced thermal conductivity of the dielectric materials is important to permit the dissipation of the heat generated in the WBG device to the thermal management system, while maintaining adequate electrical insulation. Commonly used polymers and epoxies have thermal conductivities ranging from 0.1-0.4 W/mK. There do exist commercially available encapsulation materials with thermal conductivities approaching ~ 10 W/mK. Recently reported novel materials report thermal conductivities > 50 W/mK (with some studies reporting values > 100 W/mK).
While the focus of much of the research remains the development of composite materials with a thermal conductivity > 10 W/mK, it should not be lost that a thermal conductivity of 10 W/mK itself represents an order of magnitude of enhancement over commonly used materials. As such, many advancements can be realized with such materials if other properties can be controlled. One example is that the benefits of advanced encapsulants can be magnified by using them in conjunction with advanced cooling techniques (e.g. double sided cooling or phase change cooling).
Many composite materials have out-of-plane thermal conductivity being one order of magnitude less than the in-plane thermal conductivity; this is a direct consequence of the nano-materials being added. While this anisotropy may serve well for heat spreading applications, isotropic properties are appropriate for other applications like polymeric heat sinks.
Another area that may yield improved performance is the use of carbon-based nanomaterial as additives. Most recent research for composite development for packaging applications involves the use of electrically insulating fillers (a majority involve the use of boron nitride) and not carbon-based materials. There are emerging approaches, however, to develop carbon-based materials, which have high thermal conductivity but remain electrically insulating. Additionally, there may be benefits to better managing the bonding of the polymer and the nanomaterial.
Reporting Properties of Novel Materials
An important impediment to faster progress is that a majority of studies report a small set of properties/parameters of the novel materials developed. Of late, with the focus on high thermal conductivity materials, frequently only the thermal conductivity gets reported. Single parameter data, however, is inadequate to draw any meaningful conclusions about material suitability.
A significant challenge to better reporting of the novel material properties is that well documented or standardized test procedures for measurements of properties are rarely employed. Establishing standard tests (where they do not exist) and emphasizing their use in academic/fundamental research (e.g. at the time of peer review of papers) would help address this issue.
But implementing such an approach is a challenge due to the lack of multi-instrument characterization facilities. Measuring thermal, mechanical, electrical properties along with characterization of structure and composition of materials involves several pieces of equipment, which usually do not co-exist at a single location. Having characterization facilities for dielectric materials development can address this issue.
Establishing a centralized, dedicated and widely accessible center on “Development and characterization of novel dielectric materials” would ensure the creation of a long term caretaker of this field. Such an entity could be established at a government laboratory to enhance the likelihood of funding continuity or at a university to provide the additional benefit of a continuous development of new talent in the field.
In addition to reporting a wider set of parameters for the finished materials, the sensitivity to important processing parameters of the measured properties are often underreported. Similarly, the dispersion of filler particles is underreported.
Materials Database and Machine Learning–Based Approaches for Materials Development
A well maintained and open source database of newly developed materials and their properties would be important, particularly for transferring the results of laboratory research to practice. While there do exist databases, the data sets need to be broadened and made more useful and more widely available.
In addition to the industrial benefit of databases, they can also be used to explore machine learning and QSPR (Quantitative structure property relationship)- based approaches to predicting properties of composites. This aspect of research can benefit from similar research aimed at the development of materials for other unrelated applications.
Road Mapping Materials Development
It was clear that relevant materials would likely be available more quickly with a roadmap for the development of new di- electric materials. Such roadmaps would likely be area-specific (e.g. power electronics packaging, energy storage, oil-gas etc.). Government laboratories, professional societies (e.g. IEEE) and industry groups all have track records in the development of appropriate roadmaps that guide investments and development activities. A major challenge in the development of such a roadmap is to attract and fund the appropriate team that can develop and maintain the roadmap.
Helping Bring New Materials to Market
The emerging composite dielectrics provide new challenges in the transition from bench-top to commercial availability. One new manufacturing step imposed by the new materials is the cost-effective appropriate distribution of the nanomaterials in the polymer matrix. One possible approach is to adapt the roll-to-roll processing, a widely used manufacturing technique for high volume production. Controlling defects (e.g. voids) in large volume production is important. Through continuous improvement, manufactures are doing this well on legacy materials. But new materials may well be less amenable to conven- tional quality control approaches.
Additive manufacturing is promising for novel dielectric materials, particularly because additive manufacturing has shown itself to be a way to make functionally graded materials. This manufacturing approach could provide materials that are unobtainable in any other way, opening a potentially useful new design option.
What may be the most valuable emerging capability is model-based material development. A major impediment to the adoption of a new material is ignorance about the lifetime of the material in service. The ability to very accurately perform multi-physics modeling of power electronics packaging was highlighted at the workshop. If this can be combined with the continuously improving modeling capability for insulation life and computational material science, it may be possible to reduce risk significantly using high performance computing resources. This is particularly appealing for small batch boutique solutions or for specific applications of WBG semiconductors.
From Concept to Commercialization
Significant leadership in the novel composite materials is in academia. This research is frequently more guided by exploring what is possible than by a focus on what is needed. So, better synergy between academic and industry may stimulate successful development and commercialization efforts by leveraging the strengths of both parties. In the US, intellectual property issues frequently impede valuable collaborations. Broader sharing of successful approaches to mitigate this impediment would likely improve productivity.
It must also be acknowledged that in the transition from experimental development and commercial use, sustainability should be a key criteria in the development of new materials. Recycling and upcycling need to be considered in lifecycle analysis of new materials. For example, when fillers are added to polymers, the resulting composites can’t be recycled and development of bio-source materials is preferred in that regards.
The workshop illuminated the importance of new dielectric materials (particularly for WBG semiconductors), which, if microprocessors are the brains, are expected to be the muscles of our emerging economy. For those interested in the final report of the workshop, please contact Prof. Vaibhav Bahadur vb@ austin.utexas.edu