SOCKETS: IC Package Drives Contact Technology Innovation
BY ILA PAL, Antares Advanced Test Technologies
As ICs move toward high clock speeds with pin densities reaching 0.5-mm pitch and pin counts over 1000, packaging for such devices must feature finer interconnections and improved electrical and mechanical performance. Socketing these IC packages requires an innovative solution to the challenge of designing a high-quality, low-cost, high-speed contactor for development and production test stages. AC parametric and compliance testing of high-speed, high-pin-count ICs require great care in the design, layout, and manufacturing of the load board and test socket.
In a typical test system, the test socket is mounted to the load board, which is then interfaced to the automatic test equipment (ATE). A handler includes compartments for trays where devices under test (DUTs) are stored. A vacuum head/plunger inside the handler picks up the DUT and pushes it inside the test socket, where the ATE performs the necessary tests. The test socket is the critical link between ATE and DUT. A typical test socket consists of contacts for each pin on the DUT, housing to hold the contacts, and a precise alignment mechanism for the DUT.
A test socket should perform over one million touchdowns reliably to meet one of the main requirements. A variety of contacts are used in test sockets including spring pins, particle interconnects, and shaped contacts loaded in elastomer. In a broad sense, the socket contact requirements are driven by both electrical performance and mechanical reliability.
Photolithographic Plated Probe Technology
Low contact forces, long cycle life, low contact resistance (Cres.), and high bandwidth are some of the key drivers of interconnect technology. A novel interconnect technology was developed using a photolithographic pattern plating process (Figure 1). Unlike other contact technologies, photolithographic contact sets are fabricated as a complete unit (i.e., a full array of individual metal contacts isolated by polyimide insulator). This provides more accurate positioning of contacts than a socket with individually placed contacts, particularly at tight pitches (<0.5 mm) because of positional tolerances. The photolithographic contact set also allows for replacement without major disassembly of the socket. As the device is driven into the test position in the socket assembly, the contact tips are driven toward the load board, causing the contacts to rotate and slide simultaneously on the load board pads. This motion provides scrubbing/cleaning on the load board pad at the same time the tip is scrubbing along the DUT pad, providing a low resistance electrical path across the contact set (<20 mΩ) with lower force (<7 grams) than other contact technologies, resulting in minimal wear to the load board pads and long contact life.
Figure 1. A novel interconnect technology was made for 7×7 mm, 0.5-mm pitch, 48 QFN configuration to perform functional tests for qualification.
A key concept in this technology is the use of both the metal contact and the polyimide insulator as spring elements to generate the force necessary to insure low Cres. As the contact tip is pushed against the DUT pad, it bends and the insulator flexes simultaneously to provide the necessary contact force on the DUT pad. To evaluate the stresses/deflections, and optimize the design for ideal connection, numerous finite element simulations have been run. In addition to the stresses in the contact set, the deflection coupling of adjacent contacts also has been modeled.
A fundamental difference between this technology and conventional spring pin arrays is the physical connection between adjacent contacts. Photolithographic contacts are coupled by the polyimide insulator, which causes the deflection of a single contact to induce a deflection in adjacent contacts. For example, the center contact has been deflected 0.007 inch at the tip. There is a small amount of deflection in the adjacent contacts mainly due to the coupling effect of the polyimide insulator. From the modeling analysis, it was found that the amount of deflection coupling is different along the two axes of the contact set. In the X axis, the adjacent tip is deflected down 0.001 inch. In practice, this means that the adjacent pads along the X axis need to be planar within 0.002 inch to ensure adequate contact. Along the Y axis, the deflection coupling is small enough that the contacts are essentially independent. To minimize the coupling effect, a slit cut can be introduced between each contact leaf to isolate I; providing maximum compliance to the individual contact leaf.
A contact set was made for a 7×7 mm, 0.5-mm pitch, 48 QFN configuration to perform the following functional tests to qualify the technology.
The first test examines the relationship between deflection of the contact set and the Cres. Setup consists of a stand with precise Z-axis movement. The contact set was placed on a gold-plated base plate connected to the tester. The Z-axis moving plunger tip (gold plated) was also connected to the tester. A displacement gauge was used to measure the deflection at 0.001" increments (Figure 2). Cres. decreases with the increase in deflection, and stabilizes below 20 mΩ for the deflection of 3 mils to 7 mils. At different cyclic intervals, the deflection range is still 3 to 7 mils to obtain a stable 20 mΩ resistance. This helps to define the deflection variable to be designed in the part considering co-planarity of the DUT, board, and tip manufacturing tolerances.
Figure 2. Cres. decreases with increased deflection, and stabilizes below 20 mΩ for the deflection of 3 to 7 mils.
Using a similar test set-up, the second test was performed to determine the current carrying capacity of the probe. A current was supplied constantly through the gold-plated plunger that passed through the contact leaf and returned back via the gold-plated base plate. Resistance was measured at 1-minute increments for 10 minutes under load (Figure 3). The graph illustrates that the Cres. remains less than 20 mΩ over time for a 3 Amp current supply. The data was repeated on multiple probes. This experiment helps to qualify the technology for specific current applications.
Figure 3. Cres. remains less than 20 mΩ over time for the supply of a 3 Amp current.
The third test was an AC parametric test. The test set up consisted of a 8722D network analyzer, probe station, and ground-signal-ground (GSG) probes with 450-µm pitch. The signal integrity measurement results demonstrated the high frequency performance of this contactor. The -1dB insertion loss bandwidth is 36 GHz. The contactor is also well-behaved beyond 40 GHz. With such a high bandwidth, this contactor suits all RF/microwave applications requiring low insertion force, and high density, pin count, and bandwidth. For high clock speed applications, considering 3rd harmonic, it can support clock rates beyond 12 GHz, which is higher than most digital IC clock speeds.
The fourth test focuses on force analysis of the individual contact. The test set-up consists of a vertical plunger connected to a load cell and displacement gauge. The contact set is placed on the flatness calibrated base station. The vertical plunger is moved down in 0.001" increments, and force required was measured using a force gauge and documented.
The primary intention is to reduce the force per contact. The objective is to maximize pressure where pressure equals force divided by area. This way, load board pads won’t be damaged due to the high force resulting from the contact. For a 7 mil deflection, force-per-contact is less than 8g compared with other technologies (typically 30g force per contact). 7 mil deflection was chosen based on device co-planarity, load board pad height variations, contact tip height variations, etc. The last test measures the endurance characteristics of the contact. The socket with contact set was mounted on a load board which in turn was mounted to the handler. A fixture holding a device simulator was used for cycling on and off. A compression pressure of 5 PSI was applied with 1 second ON and 1 second OFF compression cycle. This compresses the contact to 7 mil deflection. The deflection will not exceed 7 mils because of a hard stop, designed in to prevent over compression. Over compression will generate stresses in some areas and leads to the breakage of contact. It can be seen from the graph that the contact is robust over a million cycles (Figure 4). The contact resistance is below 20 mΩ when tested with device simulators. The endurance test was performed at -20°C and +130°C for 50K cycles. The contact resistance is below 20 mΩ.
Figure 4. Endurance test results demonstrate that the contact is robust over a million cycles.
ICs must be functionally tested using ATE system before being sent for field installation. The test socket is the brain of an ATE system that provides the signal flow from the DUT to the tester. The ATE system needs to be up and running without interruption for shorter throughput time, which means little or no cleaning required for contactors and also easy replaceable contactors allow users to change independently without changing the entire socket. Photolithographic plated contacts meet these requirements and also possess sound electrical and mechanical characteristics for reliable and repeatable IC testing. A properly designed socket using appropriate contacts will connect the dots and complete the chain of any ATE system. AP
The interconnect shown and discussed in this paper was developed in collaboration with: David Bogardus, senior mechanical designer; Praba Prabakaran, senior thermal engineer; James Zhou, senior staff engineer; Resty Querubin, advanced engineer and Steve Fahrner, staff engineer.
ILA PAL, program manager, may be contacted at Antares Advanced Test Technologies 1150 N. Fiesta Blvd., Suite 101, Gilbert, AZ 85233; (480) 682-6230; email@example.com.
The Short Story
Socketing high-speed, high-density IC packages requires an innovative solution to the challenge of designing a suitable contactor for test stages. Photolithographic contact sets, fabricated as a complete unit using a photolithographic pattern plating process, improves positioning accuracy.