Designing with LTCC
07/01/2004
MATERIALS AND TOOLS
BY GLENN E. OLIVER
As the U.S. Department of Defense and worldwide automotive manufacturers demand more functionality in electronic devices, electronic packaging solutions must be smaller, lighter, more complex, operate at higher frequencies and accommodate more components per unit area. This fact is not only a challenge to RF and microwave designers, but also to packaging engineers and electronic materials suppliers. Circuit designers, packaging engineers and materials suppliers must work together to meet these challenges.
This article discusses a low loss low-temperature cofired ceramic (LTCC) system and associated design tools. These tools are formatted as a design kit* that plugs into industry standard EDA design tools. The kit contains a library of verified circuit models that designers and packaging engineers can use to qualify low loss LTCC materials. Proposed development of design infrastructure for low loss LTCC materials are also discussed.
Challenges of Modern LTCC Design
LTCC technology has been used in multilayer circuits for decades. Automotive and military manufacturers already take advantage of LTCC's stable physical and electronic properties over a wide range of temperatures and environments. Electronic component and interconnect manufacturers benefit from LTCC's high process yield and relative ease of manufacture. Designers of system in package (SiP) modules take advantage of the high component density possible in LTCC designs. This is realized from applying conductive, dielectric or resistive thick film compositions between tape layers.
 Figure 1. Model cross section of die and substrate within a larger LTCC module. |
Figure 1 illustrates a model cross section of one die and substrate within a larger LTCC module. In this illustration, three inductors, three capacitors and two resistors are embedded within the LTCC tape layers. This diagram is illustrative of the passive components that can be integrated within the LTCC substrate. The top of the module is primarily dedicated to the die. This type of package design allows for maximum component density with minimum size and weight. The total system cost generally is lower, since passive components are screen printed as thick film composition instead of attached as discrete surface mount components. Although the cost of the substrate is slightly higher for LTCC than for FR4, significant overall cost savings are realized from the integration of passive components.
Military applications such as radar and secure communications use LTCC for its high reliability and ability to integrate both cavities and thick film compositions within the substrate. This allows for design flexibility and high reliability — both key factors in military applications. LTCC has also been used in the automotive industry for high component density circuits that demand high reliability in harsh environments. For example, LTCC is used in many control circuits for traction and stability, engines, transmissions, antilock and hydraulic brakes, electrical-assisted steering and sensor systems.
 Figure 2. Emerging automotive applications demand performance at frequencies higher than traditional LTCC substrate limits. |
Traditional LTCC substrates are optimized for performance below 15 GHz. At frequencies higher than this, dielectric loss increases and makes the material less suitable for many applications. Some emerging automotive applications demand performance at frequencies higher than the traditional limit of LTCC substrates. These automotive applications are short-range radar for side object detection and forward-looking radar for detection and adaptive cruise control. These are shown in Figure 2 and operate at frequencies in the range of 24 GHz and 77 GHz, respectively. For these applications, low loss LTCC materials must be used to meet frequency requirements. A similar evolution toward high microwave frequency levels is expected to occur in wireless telecommunications applications, but the recent burst of the tech bubble means that this evolution is more likely to be led by automotive and military applications.
 Figure 3. Typical design flow used to create multilayer circuits. |
For simple circuits at lower frequencies with little component integration, successful prototypes performing to specification on the first build are the norm. For complex, high frequency, densely integrated circuits, this goal is rarely realized. A typical design flow used to create multilayer circuits is shown in Figure 3. The main cause of frustration for high-frequency designers and packaging engineers lies in the "detail design" block. Several design iterations are generally required to create a successful design. There can be many causes for a circuit build to not meet specifications. Many times, the circuit build will exhibit resonance or parasitic effects that are unpredicted in the simulation, due to inadequate circuit models or a lack of experience with multilayer LTCC design. The material properties, processing capabilities and electronic design requirements become increasingly intertwined at higher frequencies. Unexpected electromagnetic coupling and crosstalk can plague designs. Many assumptions that rely on first-order approximations at lower frequencies become invalid as frequencies increase. As passives components are packaged more densely, there are interaction effects between features that are difficult to predict. Application of basic design methods and simulation tools are not enough to create successful designs on the first attempt. For designers to be confident of LTCC producing design "wins," significant improvements in the existing design infrastructure must be made.
Low Loss Materials System
LTCC is beginning to be used more commonly in high-frequency applications. This is because LTCC enables integration of passive components, which allows electronic packages to shrink significantly. Integration of passive components also leads to lower overall cost, because passive components can be screen printed instead of having to be soldered on as surface mounted elements. The next major advances in RF and microwave technologies will develop applications at higher frequencies than those currently in use. Traditional PCB technologies suffer from high dielectric loss as frequencies increase. As the functionality demands of high-frequency devices continue to increase, more signal loss cannot be tolerated.
The thermal expansion of LTCC closely matches that of an IC die. Thermal conductivity of LTCC is approximately 20 times better than standard PCB materials. LTCC also maintains stable properties over a wide range of frequencies, temperatures and environments. These factors explain why LTCC has long been popular in demanding military and automotive applications where stability and reliability are key design criteria. Many designers of emerging complex multilayer circuits are recognizing LTCC as a key to solving the challenges of high-frequency, high component density applications.
 Figure 4. Low loss LTCC compared with other materials. |
A low loss LTCC materials system has advantages over alternative materials for high-frequency, high component density applications. Obviously, low loss LTCC will have a lower loss tangent than alternate materials. Figure 4a compares low loss LTCC with other materials that have low loss tangent values at high frequencies. Low loss LTCC compares favorably to ceramic-filled PTFE material, which is another low loss substrate material used at high frequencies. Note also that at high frequencies, circuit elements are packaged closer together. Preventing thermal and mechanical failure of these high component density modules becomes much more of a challenge, so properties like thermal conductivity become more significant. Figure 4b compares typical values of thermal conductivity of various materials. These values can be further improved by addition of thermal vias to the design, enabling higher package density. Another often overlooked advantage is that the thick film compositions (conductor, resistor, etc.) are matched to the LTCC dielectric. This is especially important to packaging engineers, since thermal mismatch between different thick film materials can result in significant dimensional variation of key features on finished parts. Attempts to optimize thick film materials of various types from various manufacturers often lead to limited process capabilities that are next to impossible to control. LTCC materials systems with strong research and development foundations demonstrate suitable process repeatability without forcing the packaging engineer to perform constant qualification and optimization.
Low loss LTCC materials are now being qualified for use with next-generation microwave applications. As operating frequencies of applications increase, a stable, reliable process becomes critically important. Dimensional tolerances of embedded components, especially filters, become tighter as features approach the quarter-wavelength of the signal.
Passive Component Library for LTCC
Although LTCC is used significantly for high-performance, high-density packaging, low loss LTCC has not yet realized its potential as a material for high-volume manufacturing requirements of emerging microwave designs. One factor holding low loss LTCC back from reaching this potential is the lack of design infrastructure. Silicon and gallium arsenide technologies have well-developed libraries of component models in various EDA packages for designers to use. The reason that low loss LTCC does not yet have this level of design infrastructure is that for simple, low frequency, low component density applications, these models are unnecessary. Coupling between signal lines, effects of quarter-wave resonance effects, and dielectric loss were generally minor considerations that could largely be ignored. For modern microwave designs, however, these factors cannot be ignored.
To make the design process as efficient as possible, a library of passive circuit components was developed and made available to designers, LTCC fabricators and packaging engineers. This library is formatted as a design kit for use with one company's** circuit design platform in use today for RF and microwave SiP designs. The library was developed for a low loss LTCC materials system, because it lends itself to emerging high-frequency applications due to its lower loss tangent at high frequencies (Tan δ= 0.002 at 40 GHz).
 Table 1. Summary of design kit components. |
Components in the design kit include inductors, capacitors, couplers, filters, resonators, transmission lines, discontinuities and a balun. Details about these components are summarized in Table 1. Each component design is based on a simulated model developed using commercially available electromagnetic simulation tools. Mechanical layouts were generated based on these models and parts were manufactured. S-parameter files included in the design kit reflect verified test data taken from these manufactured parts. Some components were not verified experimentally, due to limitations in the test fixture and equipment. These models are still useful, because the designs were based on electromagnetic simulations.
The design kit consists of 47 different fixed artwork designs, most of which have been verified by measurement. These library elements can be used now by designers and packaging engineers to verify and qualify low loss LTCC materials. In their current form, the library elements cannot be easily dropped into a microwave design, because these models reflect specific, measured components in a specific layer stackup. They also include ground-signal-ground probe pads as part of the model. Most components are not de-embedded; they include probe pad ports as part of the component. Finally, library elements reflect fixed artwork without parameters that can be used to scale the individual component to a specific application.
More Proposed Solutions
Designers generally use models with selectable parameters that can be modified based on the individual application. For complex, high component density circuits, nearly infinite number of parameter combinations can be possible.
To move LTCC materials from the realm of specialty applications to the realm of high-volume wireless markets, a similar level of commitment is required. This must be a shared commitment between LTCC materials suppliers, packaging engineers and electronic designers.
Materials suppliers provide leadership and expertise regarding LTCC's capabilities and processing guidelines, but must work with microwave designers and packaging engineers regarding circuit applications and design. One of the main purposes of introducing this design kit is to facilitate stronger interaction between materials suppliers, packaging engineers and microwave designers. The focus of this interaction is to determine what basic structures need to be modeled in LTCC and the details of the parameters that can be varied in the geometries of these structures. The payoff to designers in this type of collaboration is to obtain models of LTCC structures that can be quickly integrated into circuits and are critical to meeting specification on the first design build. The benefit to packaging engineers is more efficient processing of LTCC designs, because circuit models are verified according to established design rules. LTCC materials suppliers want to make designing with LTCC as efficient as possible. This will help unlock the untapped potential of LTCC as an attractive solution to high-volume, high-frequency applications.
References
- Raymond L. Brown, "Miniaturization of a C-Band Frequency Synthesizer Using Multilayer-ceramic Integrated Passive Components". IMS Workshop WME. June 2000. IEEE International Microwave Symposium (MTT-S).
- Mary Cramer, "Ceramic Technologies for Microwave: Multilayer Ceramics for Bluetooth?", IMAPS Advanced Technology Workshop on Ceramic Technologies for Microwave, Handsets, Bluetooth, Broadband and LMDS. March 2001, Denver, CO.
- Dan I. Amey et al., "Low Loss Tape Materials System for 10 to 40 GHz Application". IMAPS International Symposium on Microelectronics. Boston, MA. September 18-22, 2000.
- Howard A. Morgenstern et al., "RF Response of DuPont 943 and Ferro A6 from 2.7 to 14.2 GHz". IMAPS Conference on Ceramic Interconnect Technology. Denver, CO. April 26-28, 2004.
*DuPont's Low Loss Green Tape™ Materials System.
**Agilent's EEsof EDA.
GLENN E. OLIVER, research engineer, may be contacted at DuPont Microcircuit Materials, PO Box 13999, Research Triangle Park, NC 27709; e-mail: [email protected].