Low Temperature Co-fired Ceramics offer quantum leap forward in Bluetooth device miniaturization.

By Shoichi Sekiguchi, Taiyo Yuden Co., Ltd. and Hiro Kawada, TRDA Inc.

Bluetooth-emerging wireless communication standard or technical oddity of questionable commercial viability? Perhaps the answer lies just around the next curve on the technology highway, in the form of Low-Temperature Co-fired Ceramics (LTCC) — an emerging technology that greatly increases the platform's commercial potential by opening the door to a new generation of smaller, less expensive, lower-power-consumption module designs. This is actually happening right now, through a combination of high-frequency circuit design, advanced materials science and state-of-the-art process technology. In contrast to standard fabrication techniques in which discrete capacitors, inductors and other passive components are mounted onto the surface of the PCB, LTCC technology embeds equivalent electronic functionality directly into the substrate material itself, generating significant advantages in module size, performance, cost and manufacturability.
Fabrication Advances
In the past decade, extensive research has resulted in the development of monolithic surface-mount capacitors, inductors, LC filters and other passive components that are now used everywhere. Employing multilayer construction, this new breed of passives contributed to the development of several spin-off technologies, including ultra-thin green sheet formation, fine conductor patterning, precision green sheet stacking and via-hole formation. These technologies have helped to not only decrease package size, but also to reduce development lead times. It is expected that component reduction requirements will only grow more stringent as mobile phones, WLAN systems and other communications devices operating on multiple frequency bands gain wider usage.

     For the digital processing sections of these devices, IC and module size enlargement due to increased functionality can be avoided to a certain degree by refining the design rules and using higher processing speeds. However, it is a totally different case for the RF section where added functionality usually causes the circuit to become more complex, sophisticated and larger. In some cases this even results in doubling the size of the RF circuit, which totally compromises device portability. LTCC technology provides a promising new alternative method of resolving the conflicting agenda of increased functionality without a corresponding size increase.

Figure 1. The Taiyo Yuden Bluetooth RF module (right) shows the dramatic size reduction achieved by LTCC technology versus an equivalent conventional-type module (left).

Bluetooth LTCC Module
Figure 1 shows the Bluetooth RF module (right) newly developed at Taiyo Yuden. For comparison purposes, the size of the PCB required to mount all passives as discrete components is also shown (left). Although the two modules display identical performance, the LTCC module is significantly smaller compared to the discrete module. Major size reduction is achieved by embedding the bandpass filter into the LTCC substrate. Size is further reduced by using a flip-chip IC instead of a BGA packaged device (Figure 2).

     This particular LTCC substrate was layered using 50 μm green sheets. Passives installed on top of the substrate are the capacitors, inductors and resistors in either 0402 or 0201 case sizes creating the matching network and loop filter circuits. Data shown in Figure 3 describes the frequency characteristics of the bandpass filter that is embedded into the LTCC substrate. The frequency characteristics indicate that the performance is equivalent to that gained by conventional discrete filters and/or multilayered filters. The module consumes only minimal power yet achieves high sensitivity and interference. The interference and spurious radiation characteristics in Bluetooth and cellular phone systems are highly preferred. All signals generated from this module, except those sent to the antenna, are digital and require only bypass capacitors for proper functional operation.
Evolution of the Multilayer Manufacturing Process from MLCC to MLCI to Integrated Module
Figure 4 describes the manufacturing process of LTCC substrates. Besides the SMD component assembly process, the first half of the procedure is identical to the methods used in manufacturing multilayer ceramic capacitors (MLCCs) and multilayer ceramic inductors (MLCIs). MLCCs have recently undergone a significant capacitance increase and are now entering a market traditionally dominated by tantalum capacitors. There are two methods to increase capacitance on MLCCs: increase the dielectric constant of the ceramic material, or create thinner ceramic films and increase the number of stacked layers. More recently, we have seen MLCCs using sheets as thin as 1.8 μm stacked 900 times, as shown in Figure 5. The manufacturing methods of MLCCs have not only led to significant performance and other improvements, but have also contributed to the development and refinement of key technologies such as ultra-thin green sheet processing, multiple layer stacking and ultra-high conductor printing.

     MLCIs were developed after MLCCs because of the need for additional technological breakthroughs. Because the internal conductor of MLCIs must be built as an interconnected three-dimensional spiral, the development of conductor technology required more time than MLCCs. There were additional difficulties when silver was chosen as the material for the conductors. Although the chosen metal was necessary for improving the Q value, its 960C melting point required the development of a new co-firing method. In contrast to the high melting point of metals such as nickel and palladium frequently used in MLCCs, silver conductor-based MLCIs needed a new ceramic material that could be co-fired at lower temperatures. These technological challenges related to the construction and material properties of the MLCIs caused these components to be introduced into the market at a much later date than MLCCs.

Figure 2. Structural cross-section of an LTCC module showing the approximate location of the passive component functions embedded within the substrate.

     Since the internal conductor of an MLCI must form an integrated spiral, via-holes had to be created in the sheets to connect the separated conductors. In early production methods, the original MLCIs offered by Taiyo Yuden incorporated metal punching equipment to form these via-holes on the sheets. However, with the development of 0201 case size products, a new method using lasers was employed. The conventional method could only create holes 150 μm in diameter, which is significantly larger than the 60 μm hole required for 0201 products. Forcing the metal punch to create holes smaller than 150 μm would damage the punch and render the equipment unsuitable for mass production. The new laser equipment provided the means to not only punch ultra-small, high-precision via-holes at the desired location of the sheet, but also reduce lead-time as well. The metal punch method required a fixture that necessitated about one to two months lead-time for delivery. However, the flexibility of the laser punch equipment virtually reduced that lead-time to zero.

Figure 3. The frequency characteristics of the bandpass filter integrated into a LTCC substrate show equivalent performance, compared to a conventional discrete component.      Manufacturing LTCCs not only requires the technology for green sheet processing and high precision layer stacking commonly seen in MLCCs, but demands the precision via-hole capabilities seen in MLCIs as well.

     LTCCs typically use materials with an approximate dielectric constant of 8 and a tangent delta value of 0.002. Ceramic materials with dielectric constants that are too high would generate unnecessary capacitive coupling between conductors and may pose significant difficulties in obtaining the desired performance. On the opposite end, materials without enough dielectric constant would not create enough capacitance in desired locations and would make circuit designs impossible. Obtaining the optimum performance in LTCC circuits relies heavily on the control of the dielectric constant. Co-firing temperature is also another issue because the materials must be fired at temperatures lower than 96°C. Internal conductors are formed using silver, which has a melting point of 96°C. Co-firing temperature must be kept below this melting point in order to create precise patterns. Other considerations include the development of materials that do not diffuse the silver.

Figure 4. The multi-step LTCC manufacturing process exploits techniques first developed for MLCC and MLCI fabrication. Designing the Internal Conductor
In the world of high-frequency signals, interference between two conductor materials may cause performance to degrade significantly. To avoid this, simulations of three-dimensional electromagnetic distributions must be conducted at the earliest design stages. LTCC modules with integrated passive functions are developed in the following steps:

1. Determination of the equivalent circuit and required component values of the embedded passive elements using the microwave design system (MDS).

2. Design of the conductor layout required to generate the desired component values using pre-determined factors, such as green sheet properties (dielectric constant, dielectric dissipation factor), post co-fired sheet thickness and conductor thickness. Cost and minimum size of the substrate are generally determined at this stage.

3. Pre-determination of reliability performances of the conductor and terminations such as resistance to board warping and bonding strength using finite element method (FEM) structural simulations.

4. Optimization of the component values of the passives by feeding back to the MDS the S-parameters obtained with the high-frequency structural simulator (HFSS) electromagnetic field distribution analysis. This process is conducted on the pre-finalization version of the design and is repeated as necessary.

5. Pre-determination of the manufacturing yield rate using the Monte Carlo experimental method on substrate designs with embedded passives compensated for LTCC tolerance. Figure 2 shows a cross-sectional view of the LTCC module, indicating the location of passive component functionality embedded within the substrate.

Figure 5. Cross-section of an MLCC, composed of up to 900 layers of 1.8 µm-thick film.      Furthermore, the component values are again refined through yield rate expectations obtained through simulations based on the Monte Carlo statistical experimental method.
Recently, PC manufacturers have recognized the value of having an 802.11a/b and Bluetooth combo module to support a full range of personal area network applications. The combination of these wireless standards greatly increases the complexity of the RF design and would enlarge the circuitry if existing integration methods were used. The great potential of LTCC technology in this area suggests that such tri-system RF sections will incorporate integrated LTCC modules sometime in the future. However, historically, such integration typically occurred one system at a time rather than all at once. For example, as was seen in antenna switches and VCO modules, peripheral passives were integrated gradually. Likewise with tri-system RF modules, integration would only progress with the release of new design and construction technology.

     New technologies like Low Temperature Co-fired Ceramics show great promise as platforms upon which the next generation of Bluetooth solutions will be built. By integrating passive components like filters, balun, and bypass capacitors into the substrate, LTCC technology offers a quantum leap forward in miniaturizing module design. The technology needed to manufacture LTCC substrates is based on those used to produce multilayer capacitors and inductors. As such, it has been under rigorous study and development for over 15 years and is now the most widely used method for mass producing discrete passive devices. RF modules using LTCC substrates can also be cost-effectively produced by integrating high-frequency design technology into this process. The many benefits of LTCC technology are not limited to Bluetooth alone, but will be widely applied in the design and development of scores of other devices, including WLAN equipment and mobile handsets.
Shoichi Sekiguchi is General manager, Development Division, Wireless Info. Network Group, for Taiyo Yuden Co., Ltd., located in Gunma Japan. Mr. Sekiguchi can be reached at Hiro Kawada, Ph.D., is the Chief Executive Officer of TRDA Inc. (Taiyo Yuden R&D Center of America), located in San Diego, CA. Dr. Kawada can be reached at