|
Leadless Plastic Packages, Such As the DFN and QFN, Have Inspired a Renaissance in a Mature Technology
|
By Edward G. Combs, QPL Limited, Fremont, Calif.
|
Historically, packaging engineers have considered etching to be either a short-term solution for high-volume packages or a "necessary evil" for high-end or niche products. With the growing popularity of leadless packages based on etched leadframes, however, etching has entered a renaissance, breathing new life into an old technology.
|
 |
| Figure 1. Shown in strip form, leadframes may be taped prior to assembly. |
Etched leadframes have been in the mainstream of plastic packages for many years. These frames, however, were traditionally employed for prototyping and small volume production, while larger volume runs were assigned to stamped leadframes as soon as possible.
Stamping Tools
It was easy to justify stamping tools when the cost of the tools was "reasonable," and the rule of thumb for justifying a tool was measured in volumes that exceeded 100,000 frames/month.
This rule of thumb changed abruptly when package leadcounts began to exceed 200 leads, and the cost of the stamping tools escalated even as their useful life decreased.
As internal lead pitch narrowed and semiconductor die size decreased, stamping tools could not meet the increased demands.
New Packages
New leadless packages, designated by JEDEC as the QFN (Quad Flat No-lead) and DFN (Dual Flat No-lead) packages have arrived. These are widely identified by the various manufacturers' nomenclature such as Amkor's MLF, Carsem's MLP and many others, shown in Figures 1-4.
One result of the popularity of these leadless packages has been the decline of stamp tools. No longer can the determination of a stamp tool be made simply based on leadcount, monthly volumes or lead pitch.
|
No longer can the determination of a stamp tool be made simply based on leadcount, monthly volumes or lead pitch.
|
In most cases, stamped leadframes simply will not work in these applications because of internal locking mechanisms that are designed into the individual leads and die-attach pads.
The only cost-effective and viable solution today for these packages is etched leadframes, arguably the fastest growing package format in the history of plastic-packaged ICs.
Reduction in Board Area
With higher speeds, higher power and smaller footprints, these packages are fast becoming the format of choice for wireless and handheld applications. In many instances, the board area required for mounting similar I/O counts is reduced dramatically.
For example, a 48-lead device in a standard TQFP with a body size of 7mm x 7mm occupies a footprint of 91mils2, while the 48-lead 7mm x 7mm body size QFN requires only 49 mils2, (a 40 percent reduction). Moreover, the QFN package will accommodate a larger die size. Additionally, the reduction is even greater in higher I/O count devices.
 |
| Figure 2. Amkor's QFN part, now in volume production, is the MLF (Microleadframe). |
Standards
JEDEC standards for both QFNs and DFNs exist (JEDEC MO-220 and MO-229), and upcoming standards for multiple row devices are being developed. Still, the variations in packages and leadcounts are probably greater than any other non-array package type that has ever been introduced.
There are variations in package configurations, lead pitch, lead length, pad size and package thickness-to name a few-and each manufacturer of the package has developed its own leadframe strip dimensions to accommodate specific tool sets.
These variations alone might appear sufficient to preclude any standard leadframe configuration development; however, there are other technical issues for consideration, as well.
Manufacturability
Each package manufacturer retains its own "intellectual property" in the design of the leadframes that determines manufacturability, performance and the reliability of the finished device. Additionally, each supplier has developed its own lead-frame strip configuration for efficient manufacturing.
Because the leadless packages possess both exposed lands (or leads) and exposed die pads (and are asymmetrically molded), internal locking mechanisms must be designed into the leadframe to ensure robustness.
 |
| Figure 3. This cutaway illustrates Amkor's MLF package. |
These locking mechanisms are typically half-etched areas on both the lands and the die paddle. Without the locks on the sides of the exposed features, the leads and paddle would demonstrate little mechanical integrity and might simply "fall out" during temperature cycling, thermal shock or mechanical shock tests.
Minimizing Moisture
These locking mechanisms also serve to minimize the ability of moisture to penetrate the package.
Traditional stamped leadframes typically lack the features needed to "lock" the leadframe into the mold compound. Stamped frames also lack the features required to accommodate hidden features, such as non-exposed tie bars.
The new leadless packages are all manufactured in a matrix format; therefore, the external dimensions of the leadframe strip are normally fixed regardless of the leadcount or package outline configuration.
|
Selective and/or ring plating for some very small packages presents a significant challenge to leadframe manufacturers...
|
Each vendor determines its own "standard overall dimensions" for the external strip depending on their molds and backend equipment; therefore, the total number of packages per leadframe strip will vary depending on a specific strip design.
The density of the design determines the individual package cost, so it is no longer appropriate to consider the cost of the frame as simply the "cost per package" or "cost per I/O."
The cost of the leadframe strip depends more on the size of the overall strip and the complexity of design than it does on the number of leads per package. Since the number of units/strip varies, depending on design density, one manufacturer may enjoy an advantage over the competition simply by optimizing the strip design.
Material Costs
Some of the other direct material costs (such as molding compound), are comparable whether a strip of 10mm x10mm packages or a strip of 3mm x 3mm packages is being produced. As a result, the design of the backend processing equipment, as well as the strip design becomes most critical for these packages.
The more densely populated the strip, the lower the cost per package. This cost concept is slightly different from the way leadframes were formerly priced.
In addition to cost advantages, features that are easily incorporated into etched frames include half-etched leads that extend inward to the pad, thus reducing wire length. Moreover, half-etched lands (or leads), are desirable with saw singulation.
Leadframe designs for leadless packages tend to be more critical than for leaded packages. Etching tolerances are tighter. Requirements for top-to-bottom etching ratios may be different, and plating dimensions and requirements are more critical.
 |
| Figure 4. Carsem's MLPQ, shown with a 16-lead device, is its QFN package. |
MSL Requirements
Ring plating, selective plating and NiPd or NiPdAu plating are becoming popular ways to meet the Moisture Sensitivity Level (MSL) requirements-including lead-free temperature profiles-of the new JEDEC standards.
Selective and/or ring plating for some very small packages presents a significant challenge to leadframe manufacturers, who are developing and installing new plating techniques and equipment. To meet cost objectives of pre-plated frames, new plating equipment for the NiPd and/or NiPdAu processes is a necessity.
Leadless packages often require that leadframe makers provide the strips with backside taping for the molding process. This taping is needed where pocket (or individual) molds are not used, and the package is either singulated with saws or water jets.
Conclusion
Historically, packaging engineers often viewed etching as either a short-term solution for high-volume packages, or as a "necessary evil" for some high-end or niche products.
Now that picture is changing. For leadless packages, which employ etched leadframes, these engineers will have to appreciate etching as the only viable solution to produce this package format in the foreseeable future. It now appears that this "old" technology will continue to lead a long and useful life!
|
Mr. Combs joined QPL as vice president of technical sales last year. A semiconductor packaging industry veteran, he was most recently vice president of engineering and earlier he was executive vice president of customer engineering at ASAT Inc., Fremont, Calif.
From 1985 until he joined ASAT in 1987, he was vice president of sales and marketing at Swire Technologies (since absorbed by Hana Semiconductor). Earlier, he was operations manager at Amkor Technology Inc. [ed_combs@qpl.com]
|
|