Microwave Journal
www.microwavejournal.com/articles/27069-lds-molded-interconnect-devices-fit-for-mmwave

LDS Molded Interconnect Devices Fit For mmWave

September 13, 2016

Figure 1

Figure 1 LDS process laser-substrate geometry (a) typical ENIG metallization (b) metallization compounds and thicknesses (c) and sample part at various steps (d).

Molded interconnect devices (MID) are the three-dimensional (3D) response to the increasing market demand for components to be smaller and allied to efficient and flexible circuit carrier manufacturing. For several years, different methods have been developed to allow integration of circuit patterns as well as radiating and shielding elements on nearly arbitrary shaped plastic parts, providing the full 3D design scope for the developer. 3D flexibility is becoming more and more important—indispensible— to address the challenges facing the design of modern radio frequency systems. Decreasing space for installation, increasing functionality, together with higher frequency coverage are contradicted by the uncompromising limitations of production costs.

The available manufacturing methods for 3D MIDs can facilitate optimal manufacturing processes depending on the specific application. For instance, when using these methods for RF device fabrication, the 3D plastic parts replace the RF laminates that are typically used. This means that the established requirements on the RF properties of the substrate material, as well as the applied metallization, have to be met by the 3D manufactured parts. This is especially significant for future applications in the millimeter wave (mmWave) frequency range. One manufacturing method for MIDs that is already common for large scale production of antennas in consumer devices (e.g., smartphones, tablets, laptops), is the laser direct structuring (LDS) method.1 Injection molded plastic parts combined with laser activated surfaces that allow selective metallization open up almost unlimited design scope. Because present LDS fabricated applications mainly cover frequencies up to 6 GHz, this article evaluates the suitability of the LDS method for RF applications up to the mmWave range.

LDS MANUFACTURING FOR 3D MIDs

The LDS process with its laser based selective metallization processes provides great flexibility for developing electronic devices. This includes the 3D design scope in the development process, as well as the adaptability of a design during ongoing manufacturing. The substrate to be metallized by the LDS method is typically fabricated in an injection molding process using thermoplastic or thermoset materials that are doped with a special filler of mixed metal oxides. Using conventional injection molding, high volume production of the desired 3D substrate is possible. Additionally, there are LDS capable materials available that can be processed with additive manufacturing processes such as fused deposition modeling (FDM) for development prototypes. Another method of making a substrate suitable for the LDS process is using an LDS varnish called LPKF ProtoPaint. With this varnish nearly arbitrary substrate materials can be metallized by the LDS method.2

Based on a CAD dataset including the mechanical and the electrical designs of the part, the circuit pattern is transferred to the molded part by laser structuring and forms the structures to be metallized. High speed 3D LDS laser systems can structure nearly arbitrary 3D shapes while achieving fine pitches down to 150 µm, as well as large circuit patterns for radiation, heat and current transportation. With the structuring parameters like laser beam width, overlap of the structured lines, pulse repetition rate, laser inclination angle (α) and laser power, the mechanical characteristics of the plastic surface, such as adhesion and roughness, as well as the electrical properties of the metal layer can be influenced (see Figure 1a). In that way, the specific requirements of an application can be considered in the structuring process, within certain limits. The laser beam causes a micro rough surface and activates the mixed metal oxide that serves as a catalyst for a subsequent metallization. In an electroless plating process, a first layer of copper is deposited only on these activated areas. The copper layer has a conductivity of about 30 MS/m. Due to the laser ablation, micro cavities are formed on the surface and ensure a strong bond between the copper and the substrates, with adhesion strengths of 0.8 N/mm or higher. In additional plating processes, the surface is typically finished using electroless nickel and immersion gold (ENIG). The resulting layer configuration is depicted in Figure 1b, showing its typical layer thicknesses. The different available metallization compounds (see Figure 1c) enable the metallization best suited for a specific application to be chosen.

Depending on the plastic resin used, an LDS MID can be used for state-of-the-art SMT processes by applying various low or high temperature soldering methods, such as reflow. Figure 1d shows an LDS part in different fabrication steps: injection molded plastic, laser structured, metallized with standard LDS ENIG and the final assembly (left to right). The recommended minimal line width is 150 µm with a minimal gap width down to 150 µm. Depending on the geometric shape of the substrate part and the laser focus size, smaller line and gap widths can also be achieved. Due to a 3D laser-based digital process, the laser beam can travel along nearly free formed shapes within an xyz scan volume of 200 mm × 200 mm × 80 mm. An optical z-axis allows the structuring of steep areas with high accuracy to an inclination angle of up to 70°.

Figure 2

Figure 2 Dielectric constant and dissipation factor of different LDS materials.

RF PROPERTIES

The typical LDS process is based on a thermoplastic material and a metallization of the laser structured surfaces. When developing RF devices, field simulation software is typically used to optimize the required characteristics. Since the exactness of this modeling increases the accuracy of the simulated results, the efficiency of the development process can be significantly influenced by knowledge of the RF properties. To provide this information for the user of the LDS method, a detailed evaluation of the RF properties is carried out: first, the dielectric properties of the available LDS substrate polymers are assessed. Second, the RF conductor characteristics for the different LDS metallization compounds are evaluated based on measurement and simulation.

Dielectric Properties

The complex permittivity is the main influencing factor of typical dielectric materials in an RF application. Besides the frequency related geometric dimensions, the losses induced in the substrate material can be derived from the complex permittivity. Because the dielectric properties are frequency related, the evaluation has to reflect the frequency range covered by the specific application. Figure 2 shows the permittivity values (Dk) for five different LDS materials with their associated dielectric losses (DF). The measurements used an open resonator method.3 The permittivity values are in a range of about 3 to 5.5 for frequencies from 1 to 67 GHz. These values are in the same range as for the basis polymer and do not differ significantly from that of typical RF substrate materials. The measured dissipation factor ranges from 0.003 to 0.014. The dissipation factor of some LDS materials is slightly higher than for typical RF materials.

With materials like Vectra E840i LDS, a liquid crystal polymer (LCP); polyetheretherketon (PEEK), Cyclic Olefine Copolymer (COP); and also polycarbonate (PC) based materials such as Xantar LDS 3720/3732/3730, the induced material losses are comparable to those of typical RF laminates. LDS materials are generally suitable for RF applications and similar to a typical RF laminate, the choice of a LDS substrate material must be made by considering the specific application, taking into account both the RF and mechanical requirements.

Conductor Properties

Another important aspect for a circuit carrier in RF applications is the applied metallization. In addition to the electrical properties, the mechanical properties and the geometrical shape have to be considered. These characteristics influence each other. The metallization of an RF device is typically used to guide the electromagnetic waves. The condition of the metal structure influences the losses induced in the conductor. The main influencing aspects are the electric conductivity of the metal material, the geometric dimensions and the geometric surface condition. Since LDS metallization typically consists of more than one layer (e.g., Cu and ENIG), the effective conductivity is influenced by the specific surface current distribution. Especially for higher frequencies, the main surface currents will be in the outer cross section of the metal structure due to the skin effect. However, this is not automatically the outer surface of the metalized structure. The main current distribution can also be displaced to the cross section connected to the substrate material.

Table 1

Evaluating the condition of a surface, roughness parameters like the arithmetic mean roughness, Ra or Sa, and the maximum height, Rz or Sz, are usually used. Typical values of these surface roughness values for LDS metallized surfaces are shown in Table 1. For comparison, the same parameters for a Rogers 4003C substrate with a 17.5 µm rolled copper layer are included. The measurements were made using the Keyence One-Shot-3D-Measurement Macroscope VR-3000. As shown, the Sa and Sz values of the LDS fabricated metal layer are considerably higher than for the RF substrate. When evaluating the surface properties of the LDS metallization, these higher values defined by Rz/Sz and Ra/Sa are not complete because the exact shape of the surface will influence the losses. This fact has already been evaluated for other rough metal surfaces,4, 5 where parameters other than the surface roughness values are defined to model the conductor losses. In the case of LDS manufacturing, the laser causes grooves where the laser beam activates the surfaces to be metallized. This leads to a waviness that depends on the laser beam width, the overlap, the pulse repetition rate and the laser power used for the structuring. Figure 3 depicts one example of a laser structured plastic part where the grooved surface from the laser is analyzed. A 3D profile view of the surface shows this slight surface modulation. This modulation has a high impact on the resulting surface roughness values, while the influences on the RF losses may be lower due to the specific shape of the surface.

Figure 3

Figure 3 Laser structured test sample with surface profile.

Figure 4

Figure 4 Simulated insertion loss for the laser induced surface modulated CPW and MSL.

Therefore, the RF losses are evaluated based on a field simulation carried out with the Ansys Electromagnetics Suite 16.2.0. The grooved surface structure is modeled with a sinusoidal surface curvature for a coplanar waveguide (CPW) and a microstrip line (MSL). The modulation is evaluated for a laser structuring parallel (x-direction) and across (y-direction) the transmission line. The transmission lines are evaluated without this surface modulation for comparison. All metal surfaces are defined as sheets with a finite conductivity boundary condition and the conductivity of LDS copper (30 MS/m). The substrate material used is an LDS polycarbonate MEP Xantar LDS 3730 with a dielectric constant of about Dk = 3 and a dissipation factor of about DF = 0.005 at 25 GHz. The resulting simulated insertion loss in dB/cm and the corresponding simulation models are shown in Figure 4. For a surface modulation with a periodicity of 150 µm and a peak-to-peak amplitude of 40 µm, the conductor losses are increased to about 0.11 dB for the MSL and 0.18 dB for the CPW for structuring in the y-direction at 50 GHz. In case of structuring parallel to the transmission line (x-direction), no apparent difference can be observed compared to the simulated flat conductor. This shows that this part of the LDS surface roughness does not have the high impact expected. To reduce the losses, structuring in parallel to the direction of propagation (x-direction) is recommended.

Conductor Losses of a Coplanar Waveguide

Measurements on a CPW prove the results evaluated by simulation. Different laser angles (α) are used to structure the test samples metallized with different metallization compounds afterwards. Each metallization compound is based on a layer of electroless LDS copper. With the different laser angles, the three dimensional fabrication process can be evaluated, although a planar transmission line has to be used due to the limitations in the measurement setup. Each test sample consists of four transmission lines with different lengths, a short and an open line to allow for multiline TRL calibration.6 The test structures are metallized on a plate made of MEP Xantar LDS 3730 with a height of about 1 mm. Five samples are fabricated for each configuration. The measurements are carried out on a wafer prober using GSG measurement probes with a pitch of 400 µm. The frequency range evaluated is from 1 to 67 GHz. Figure 5 depicts the measurement setup with a test sample as described.

Figure 5

Figure 5 CPW structure (a) and measurement setup (b).

Considering the skin depth (δ1 GHz = 2.9 µm for LDS copper) and the electromagnetic fields of a coplanar waveguide, it becomes obvious that the main surface currents will be concentrated in the outer cross section of the structure. The current distribution is dependent on the specific RF structure. Figure 6 shows the insertion loss in dB/cm for all evaluated test samples, with the metallization compound and laser parameter depicted in the legend. The laser structuring is done in parallel to the direction of propagation (x-direction) with a laser angle of 45º and 0º. For comparison, a CPW fabricated in a photolithographic process out of a single rolled copper layer on Rogers 4003C is shown. The LDS substrate material used has a loss tangent (DF) of approximately 0.005 at 20 GHz, compared to approximately 0.0027 at 10 GHz for the Rogers 4003C. Consequently, the higher losses are also due to the dielectric losses as well as conductor losses. At 60 GHz, the CPW test samples with copper show a loss of about 0.9 dB/cm, while the nickel surface finish causes a loss of about 1.7 dB/cm. The sample on Rogers 4003 with the rolled copper foil induces a loss of about 0.75 dB/cm. The results show that there is only a slight difference between the losses of the typical RF substrate and the LDS sample with the copper metallization. The LDS samples with the nickel and nickel/gold layer induce higher losses due to the poor conductivity of nickel.

The investigations of the different metallization compounds of LDS fabricated RF test samples show that the surface roughness of LDS copper does not have the high impact on the losses as may have been expected. This is due to the specific shape of the laser structured surface, with its laser induced grooves, leads to a kind of surface modulation that only slightly influences conductor loss. Because the coplanar samples are manufactured with different laser inclination angles (α= 0º and α= 45º), this applies to 3D fabricated structures.

CONCLUSION

3D manufacturing methods are becoming more important for the fabrication of RF devices. The flexibility of an arbitrarily shaped and selective metallized injection molded part can be used to meet often contradictory requirements. When replacing the typical RF substrate with a 3D manufactured plastic part, it becomes clear that the RF properties of the 3D fabricated components have to be evaluated to guarantee proper functionality, an efficient development process and a long lifetime. As these properties are frequency dependant, evaluation depends on the specific frequency of application. In this article, the 3D LDS process was evaluated for RF applications up to 67 GHz. The results show that the complex permittivity of the LDS polymer materials is only slightly changed due to the LDS additive. There are materials available that have comparable properties to those of typical RF substrate materials.

Figure 6

Figure 6 Measured CPW insertion loss on Xantar LDS 3730 compared to Rogers 4003 with a rolled copper layer.

The conductor losses in different metallization compounds were evaluated. The surface roughness values of the LDS metallization may lead to the assumption that the roughness will cause high conductor losses. A detailed mechanical surface evaluation together with different field simulations showed that the laser structuring causes a grooved surface that has only a slight impact on the resulting RF losses. Measurement of the insertion loss of LDS manufactured CPW structures compared to typical rolled copper metallization on an RF laminate verified these results. The metallized surfaces were evaluated for structuring with different laser angles. In that way, the results are adaptable to 3D fabrication, even though the test samples were planar, due to limitations in the measurement setup.

LDS fabrication is suitable for RF applications in the mmWave range. The evaluated data of the frequency related losses and the available design scope show that 3D LDS technology is a good solution for fabricating RF devices, helping to meet future challenges.

References

  1. N. Heininger, “3D LDS Components for New Production Opportunities,” Microwave Journal, Vol. 55, No. 2, February 2012.
  2. LPKF Laser & Electronics AG, “A Shortcut to 3D-MID Prototypes,” Microwave Journal, Vol. 56, No. 11, November 2013.
  3. www.damaskosinc.com/cavitiy.htm, 2016.
  4. P. G. Huray, “The Foundations of Signal Integrity,” New York, NY, USA: Wiley, 2009, Ch. 6, pp. 216–277.
  5. S. Hall, S. G. Pytel, P. G. Huray, D. Hua, A. Moonshiram, G. A. Brist and E. Sijercic, “Multigigahertz Causal Transmission Line Modeling Methodology Using a 3-D Hemispherical Surface Roughness Approach,” IEEE Transactions on Microwave Theory and Techniques, Vol. 55, No. 12, pp. 2614–2624, December 2007.
  6. R. B. Marks, “A Multiline Method of Network Analyzer Calibration,” IEEE Transactions on Microwave Theory and Techniques,Vol. 39, pp. 1205–1215, July 1991.