Reading time ( words)
By Craig Hillman and Cheryl Tulkoff, DfR Solutions, Inc. This article is Part I of a two-part series on QFNs.
While the advantages of QFNs are well documented, they can be considered as a next-generation technology for non-consumer-sector OEMs. This article is designed to review specific concerns in relation to design and manufacturability by CMs and OEMs and provide possible mitigations or solutions to allow the reliable introduction of QFN components into products.
One of the fastest growing package types in the electronics industry today is the quad flat pack no lead (QFN). Non-consumer electronics OEMs can omit QFNs in current product generations due to concerns with design and manufacturability, compatibility with other OEM processes, and reliability. Acceptance of this package type, especially in long-life, severe-environment, high-reliability applications, is currently limited as a result. QFN-specific reliability concerns will be addressed in Part II, available on smtonline.com and in the SMT WEEK e-newsletter on May 6, 2009.
What is a QFN?The quad flat pack no lead or quad flat non-leaded (QFN) package has been referred to as the poor man's ball grid array (BGA) and is also known as a leadframe chipscale package (LF-CSP), micro-leadframe (MLF), and other names such as MLP, LPCC, QLP, and HVQFN.
It comprises an overmolded leadframe with bond pads exposed on the bottom and arranged along the periphery of the package in one or two rows. Commonly available in two- or four-sided configurations with either sawed or punched leads, it was developed by multiple component manufacturers in the 1990s and standardized late in the decade by JEDEC/EIAJ.
Figure 2. Example showing pad extension and visibility of solder fillet.
QFN Design and Manufacturability IssuesMultiple areas of concern regarding the use of QFNs have been identified including bond pad design, stencil design, reflow profile control, rework and inspection, and board flexure. For each of these key areas, a brief explanation of the issues and some recommendations are provided.
Bond pads. Non-soldermask-defined pads (NSMD or copper-defined pads) are preferred since the PCB copper etch process provides for greater dimensional repeatability and control than the solder mask process does. NSMD pads also allow the solder to bond to both the top and the sides of the pad resulting in a stronger, more repeatable joint. If a design does require soldermask-defined (SMD) pads, be aware that the pads can grow or vary in size significantly based on the PCB supplier's capabilities. This may leave the assembly vulnerable to solder bridging. Additionally, consider extending the bond pads 0.20.3 mm beyond the package footprint. Solder may or may not wet to the extended edge, but it allows for easier inspection since the outer joints are more clearly visible to an inspector. The extra pad length can also reduce bridging due to excess paste deposition by giving the solder a pathway out from under the component.
Exercise caution if vias in pads are used. Open vias act as solder drains funneling solder away from the joint down into the via and to the opposite side of the PCB. In the worst-case scenario, there is insufficient solder on the pad and a "solder bump" on the opposing side. This can result in marginal joints on the topside and printing difficulties and solder shorts on the opposing side. Most PCB manufacturers offer several options for closing vias plugging, capping, tenting which help keep sufficient paste volume on the pad. Tenting or plugging, the cheaper methods for via closure, are both prone to placement and chemical entrapment issues due to voiding and lack of planarity. Capping is a more robust, more expensive process that eliminates these two concerns.
Stencil design. Appropriate stencil thickness and aperture design are crucial for reliable soldering of QFNs. Follow the QFN manufacturer's design guidelines when they are provided; the general goal is to provide approximately 2 to 3 mils of solder thickness. Excessive amounts of paste can induce large-scale voiding of the thermal pad and component float, where the QFN is literally lifted off the board. Use multiple, smaller windowpane apertures to avoid large bricks of solder paste that can cause larger and greater numbers of voids. Windowpanes also reduce the propensity for solder balling. For the thermal pad, the general rule of thumb for the ratio of aperture to pad is approximately 0.5:1.
Moisture absorption and reflow profile. Current industry standards J-STD-020 and J-STD-033 have been useful for documenting moisture sensitivity levels (MSL) in older component packages, but there are increasing indications that moisture absorption in the thinner QFN package can drive excessive warpage during reflow. In one case study, a military supplier experienced solder separation under QFNs. The QFN supplier admitted that the package was more susceptible to moisture absorption than initially expected. This resulted in transient swelling during reflow soldering, which induced vertical lift and caused solder separation. This was not a popcorning phenomenon since no evidence of cracking or delamination in the component package was seen. To minimize this potential, larger and thinner QFNs should be treated as an MSL of 3 or higher and reflow profiles should be carefully controlled with a slow, steady ramp rate.
Rework and inspection. QFN rework can be difficult, since the thermal pad and any inner row joints are neither accessible nor visible to the repair technician. Mini-stencils, rebumping, or solder performs may be used to replace or add solder volume. Small, portable preheaters can help provide sufficient heat to reliably resolder the larger thermal pad and inner row pads in hand soldering operations.
X-ray inspection equipment is crucial for QFNs as this technique allows for inspection for good joints, adequate solder coverage, and void percentage under the package. There is currently a lack of good, universally accepted criteria for allowable voiding of the thermal pad. Robust, concave solder fillets are possible with either sawed or punched QFNs. Punched QFNs, however, are more prone to concave fillets since more copper is exposed on the component pads. Unfortunately, convex fillets or the absence of fillets are more likely since etching of the leadframe can prevent the bond pads from reaching the edge of the package and the edge of the bond pads are not plated for solderability.
Large, convex fillets also can be an indication of soldering problems such as poor wetting under the QFN; tilting due to excessive solder paste on the thermal pad; or elevated solder surface tension, from insufficient solder paste under the thermal pad, pulling the package down.
Board flexure. Area array devices are known to have board flexure limitations. For SAC alloy attachment, the maximum microstrain can be as low as 500. QFNs have an even lower level of compliance and may be more susceptible to flex-induced joint and laminate cracks. Since there is currently limited quantifiable knowledge in this area, be very conservative during board builds. Special focus should be placed on any in-circuit test (ICT) and depanelization processes and on hand assembly operations, since these areas typically induce the most strain. The IPC is currently working on a QFN strain specification similar to the one in use for BGAs.
QFN Risk MitigationFor a successful, reliable introduction of QFN components into design and manufacturing, a serious assessment is required. At a minimum, the QFN design and manufacturing capability review must include a review of bond pad designs and any design constraints; design of experiment (DOE) on stencil design to optimize thickness and apertures; the degree of reflow profiling control available; assessment of rework and inspection capabilities; and the control of board flexure.
Review, implementation, and control of best practices in these key areas can ensure a smooth, successful QFN introduction. Once this is successful, the user is ready to consider the reliability challenges of QFN.
Craig Hillman and Cheryl Tulkoff, DfR Solutions Inc., may be contacted at firstname.lastname@example.org and email@example.com, respectively. Look for Part II: QFN-specific reliability concerns, available on smtonline.com and in the SMT WEEK e-newsletter on May 6, 2009.