The OEM desires identified in the 2002 – 2003
roadmap clearly identify,through their emulators,the
present and future needs of the products that the
emulators represent. There are a total of eight
emulators. Each emulator is broken down into the
technical drivers that help identify the OEM needs
for the different time frames established for this
Technology Roadmap. The emulator technical driver
information is organized into four major areas. These
are:
• Design Issues
• Printed Board Technology Issues
• Board Assembly Technology Issues
• Printed Board Purchasing Issues
The emulator attributes in design include such drivers
as on-chip rise time,minimum voltages,thermal
dissipation factors,reliability issues,and maximum
board temperature requirements. Printed board
technology attributes deal with materials,board size,
layer count,etc.; board assembly technology
attributes consider number of components,number of
solder joints,and type of assembly; purchasing issues
deal with the cost per interconnect and if assembly is
recyclable.
The emulators represent a quantitative summary of
the expected changes in board,component,and
assembly technology from 2002 to 2012. These
changes are addressed in two different product
categories: Revenue Center of Gravity (RCG) and
State of the Art (SoA). Revenue center of gravity
products represent the bulk of revenue and are
considered to be conventional technology; state of the
art technology is in production by only a few
manufacturers. SoA technology represents less than
5% of the world’s production. Table 1 is the Mid Size
System emulator for this new roadmap.

Tin whiskers found in electroplated deposits are known to be single crystals which spontaneously grow. Thus
whisker growth can be regarded as a grain growth phenomenon. In this paper we examine whisker grain growth in
the context of the well-developed principles of recrystallization process as applied to bulk metals that have
undergone deformation and annealing. As a grain grows in whisker form,recrystallization process must take place
as tin atoms rearrange in to the lattice structure of the elongating grain. Peculiarities of tin deposit structure that may
cause whisker growth are discussed. Frank-Read source of dislocations is proposed as a possible mechanism for
whisker formation. The effect of various factors on whiskering is analyzed.
Recrystallization theory postulates that shear strain introduced by plastic deformation is stored in the metal in the
form of dislocations (lattice defects). In bulk metals,produced metallurgically from the molten phase,these lattice
defects usually are not present in noticeable quantity unless the material is subjected to cold work (plastic
deformation at temperatures significantly below melting point). In electroplated tin,however,the metal is formed at
the temperatures much below melting point. During plating,energy is stored in the deposit in the form of crystal
defects such as vacancies and dislocations. This causes the crystal structure of metal deposits to resemble the
structure of cold worked metals,and thus forms the starting point for application of recrystallization principles.
The second important factor that justifies the application of recrystallization/grain growth principles to whisker
formation is related to the low recrystallization temperature of tin. Recrystallization temperature is defined as the
temperature at which a particular metal with particular amount of cold deformation will completely recrystallize
within one hour. Typically,it can be estimated as between 0.4 and 0.7 Tm (where Tm is the melting temperature). It
is a well-known fact that in most metals,recrystallization occurs at elevated temperatures. For tin,however,the
recrystallization temperature is approximately 30°C,which means that recrystallization will spontaneously occur
around room temperature (above and below 30°C),reforming a strain-free structure.1
These two factors – strain stored in the deposits in the form of dislocations and recrystallization at room
temperature,substantiate the application of recrystallization process principles to whisker formation. But before we
elaborate on this hypothesis,let us briefly summarize the principles of recrystallization process.

Over the past years there has been consistent growth in the use of electroless nickel/immersion gold (ENIG) as a
final finish. The finish is now frequently being used for PBGA,CSP,QFP and COB and more recently gathered
considerable interest as a low cost under-bump metallization for flip chip bumping application.
One of the largest users for this finish has been the telecommunication industry,were millions of square meters
of PCBs with ENIG have been successfully used.
The nickel layer offers advantages such as multiple soldering cycles and hand reworks without copper
dissolution being a factor. The nickel also acts as a reinforcement to improve through-hole and blind micro via
thermal integrity. In addition the nickel layer offers advantages such as co-planarity,Al-wire bondability and the
use as contact surface for keypads or contact switching. Especially those pads,which are not covered by solder
need a protective coating in corrosive environment – such as high humidity or pollutant gas.
This paper describes the influence of co-deposited Phosphorus within the Nickel layer,regarding the influence
to the ENIG process itself (especially the corrosive attack of the immersion gold reaction) and the survivability
of PCBs in corrosive atmosphere.
Within this paper,different test methods are described and discussed to check the protective performance of a
high Phosphorus ENIG layer.

The wetting of alternative PCB surface finishes,including immersion silver (I-Ag) and immersion tin (I-Sn),by
Sn/Ag/Cu solder and Sn/Pb solder,was studied in this work,along with electroless nickel/immersion gold (Ni/Au)
and organic solderability preservative (OSP) finishes for comparison. Evaluation of wetting was carried out with
fresh boards and boards subjected to different pre-conditioning treatments which simulated the effects of aging,
storage and multiple reflow cycles. Selected conditions consisted of high temperature aging at 155°C for up to 6 and
12 hours,temperature-moisture exposure at 85°C/85%RH for 6,12 and 24 hours,and reflow treatments between 2
and 4 reflow cycles. Wetting was studied based on the IPC-TM-650 2.4.45 standard,by a wetting bar test,and by a
wave soldering test.
The results show that when the boards are fresh,the wetting of the I-Sn finish is excellent and comparable to that of
the Ni/Au finish,and the wetting of I-Ag is slightly better than that of the OSP finish. However,after the preconditioning
treatments,the wetting of the I-Sn finish degrades the fastest,whereas the wetting of the I-Ag and OSP
finishes decrease almost at the same rate after different pre-conditioning treatments,while the wetting of the Ni/Au
finish remains excellent through all the pre-conditionings treatments. In all cases,the wetting of surfaces is better by
the Sn/Pb solder than by the lead-free solder.

Lead-free soldering technology is still in its infancy with technical and cost issues posing major challenges for the
industry. It is expected that soldering in a nitrogen atmosphere might overcome some of the technical barriers and
provide soldered products comparable to those using conventional lead-containing materials processed in air. But
quantitative data regarding the soldering behaviour of lead-free solders under various atmospheres are sparse. As
part of an ongoing study on the effects of inerting on the solderability of lead-free alloys an examination has been
made of the solderability,as measured in a wetting balance,of three common board finishes using a 95.5/3.8/0.7
SnAgCu solder. The board finishes used were ENIG,HASL,immersion silver and a copper OSP. To simulate
typical lead-free soldering cycles the samples were subjected to multiple temperature cycles in a convection reflow
oven before solderability testing using a Multicore MUST II tester. A peak reflow temperature of 250°C and an R
flux was used for the ENIG and Ag finishes. The OSP finish performed poorly under these conditions and OSP
testing was done using a 235°C peak temperature and a 0.5% activated R flux. During testing the atmospheres were
controlled at levels of oxygen of 21% (air),10,000,1000 and 100ppm. Although inerting improved the solderability
of all three finishes,there were differences between the individual alloys. Aging by multiple reflow cycles adversely
affected the solderability of all finishes but the effects were less for ENIG than the other finishes.

The impact cooling rates exert on the reflow process is identified. The trends in shear strength and the
microstructural evolution of the solder joints are described. Lead free (Sn/3.5Ag/0.7Cu) and tin lead (63Sn/37Pb)
assemblies on copper organic solder preservative (Cu-OSP),electroless nickel-immersion gold (ENIG),and
immersion tin (Imm Sn) surface finishes were considered in this experiment. A Differential Scanning Calorimeter
(DSC) was used to simulate the heating and the cooling cycles undergo by solder in the reflow process. The strength
values and the analysis of microstructure were documented using an Instron Machine and Scanning Electron
Microscope (SEM):
• Slower cooling rates result in the increased formation of the Ag3Sn and Cu6Sn5 intermetallics. These will form
both in the bulk solder as well as at the pad – solder interface. Faster cooling rates inhibit the growth of these
intermetallics.
• Aging results in joint fatigue thus reducing the shear strength of the joint. This is the result of solder migration
and increased intermetallic thickness at the solder – pad interface.
• Time above Liquidus impacts the intermetallic thickness at the solder – pad interface.
• Faster cooling rates result in stronger joints as observed in those samples with a Cu-OSP board finish. The
opposite trend is observed for samples assembled on Imm Sn and ENIG board finishes.
• Failure modes for the shear test depended on board finish. Imm Sn samples failed by pad lifting while ENIG
and Cu-OSP failed by cracking in the bulk solder.
• Profile type influences joint quality but is greatly dependent upon all reflow process parameters,especially flux
– profile compatibility.

Flux management methods for reflow soldering have been debated for years. Current data suggests that multi-stage
filtration systems offer many benefits,particularly in dealing with the byproducts of new solder/flux paste
formulations. High system efficiencies result from targeting all areas of the oven,using multiple scrubbing units and
developing filtration units to capture volatiles released during the reflow process. These systems are also energy
efficient,require minimal facility resources and virtually operate maintenance-free by incorporating self-cleaning
cycles. This paper discusses these advantages in detail and shows the low cost of ownership that multi-stage flux
filtration systems can provide. The presented data is based on actual production performance within a mediumtohigh
volume reflow operation.

Stability and repeatability are imperative in any
reflow oven today. If oven operation is not validated
periodically,all subsequent profiles and adjustments
are not reliable in finding or maintaining an optimum
reflow process window for products. This is
especially true for lead free where the actual process
window has shrunk to a fraction of what everyone is
used to when using eutectic solders.
Ovens have been validated using mock-up FR4
assemblies,where the assembly is wired with
thermocouples and used with a profiling unit.
Unfortunately,the thermocouples lift,the board
suffers warping and material degradation after only a
short number of passes through the oven. Stainless
steel and aluminum plates have been configured to
resemble a board,and are able to withstand repeated
thermal cycles. However the material is so alien
compared to FR4,the radically different thermal
conductivity of the material masks potential
weaknesses in oven design. If you want a true
characterization of all oven parameters,it’s important
to match these physical properties within a single test
tool. Therefore,to do proper thermal validation,it is
best to use a tool specifically designed for that job
rather than to try to create either type of vehicle
described above. The composites used to create the
pallet for thermal characterization and validation
equipment known generically as Machine Quality
Management tools (MQM) have highly insulating
properties similar to typical FR4 material as well as a
high tolerance for long endurance use in high
temperature environments.
For this experiment,we took a typical MQM tool and
investigated the performance of four different ovens.
The tool measured the uniformity of gas temperature
and the heat transfer characteristics and saved the
data from each oven which was downloaded into the
accompanying MQM software program for
comparison and analysis. Our goal was to find out
how easy or difficult it is to characterize specific
physical properties of a given oven design as they
relate to the soldering process and to see if the results
indicate a dramatic differences in oven performance,
specifically,the uniformity across the width and
length of the tunnel as well as the heat transfer
capacity.
Using an MQM tool,it was possible to characterize
specific physical properties of each oven to see if oven design significantly impacted the steady state
operation of the individual zones as well as the entire
heated length. We assumed the uniformity or lack
thereof in the measured air temperature along the
length of the tunnel and across the width of the tunnel
would have an impact on the ability to consistently
and uniformly heat the product to be run. The
difference in temperature as measured between the
air and the high mass sensors of the MQM tool
demonstrated the efficiency of each oven to transfer
heat,or heat transfer value. A higher heat transfer
value means tighter control over the reflow process.
This proportionality is related to the heat transfer
coefficient,however,is not the exact heat transfer
coefficient for any particular oven. Several tools exist
to characterize the heat transfer coefficient as well as
defining the portion of heating associated with
infrared energy.

One of the electronic assembly markets that has
been emerging in recent years is the automation
equipment associated with odd form component
placement and final product assembly. As with
insertion mount and surface mount equipment,the
procurement process of odd form equipment
typically includes having the machine provider
guarantee certain performance levels and provide
evidence from a machine qualification process that
the machine being shipped is actually capable of
performing to specification. Due to both the slower
tact times and higher component prices in many
typical scenarios (relative to most surface mount
and insertion mount applications),the sample sizes
and test hours required to verify odd form
equipment to the high quality levels that the
electronic assembly industry has come to expect -
with defects measurable in parts per million (PPM)
units - can be quite burdensome. Having to insert
tens or hundreds of thousands of parts during
machine qualification can be quite costly and timeconsuming.
For example,a test involving the
insertions of 20,000 parts with only one defect
would yield a performance estimate of 50 PPM.
But a test with this many real odd form parts would
be very costly.
One alternative method to demonstrate odd form
machine accuracy has relied upon placing surface
mount slugs instead of inserting odd form parts.
Slugs (pieces of glass with the image of an SMT
part etched on it,such as those defined in IPC-
9850),are easily placed on a board with sticky tape
and measured with a coordinate measuring
machine (CMM) in order to obtain very precise
variables data. The benefit of such a variables
approach (as opposed to a go/ no go “attribute”
approach,where each insertion or placement is
only characterized as good or bad,rather than being
precisely measured) is that a very small sample size
of parts can be placed in a slug run to gain a PPM
estimate with a high degree of confidence.
The downside to the slug approach is that these
parts are dissimilar from odd form parts in that they
cannot be inserted into holes,and even if they
could,without adhesive they would move around
after insertion,making the variables data for each
part incorrect. So this difference in the nature of
board assembly (between surface mount and odd
form assembly),combined with the fact that odd
form assembly equipment users use a wide range of
odd-shaped parts that are handled quite differently than slugs,has led us to consider alternative
approaches to estimating PPM that could either
replace or supplement the slug approach.
In this paper,a method used internally by Universal
Instruments Corporation to qualify its Polaris
assembly equipment to 50 PPM or better will be
presented. The method is statistics-based,and
involves insertion rather than placement. Yet it still
gives results,PPM estimates,comparable to slug
runs. This method will be shown to be quite
efficient and,since it requires no operator once it is
started,it is even less labor-intensive than a series
of slug runs. And it does not require the use of
CMM,so it could easily be replicated at a
customer’s facility.

The advent of fine pitch area array components and the constant drive to reduce the cost of electronic products
mandate the enhancement of manufacturing systems and assembly yields. In fine pitch area array packaging,
significant cost benefits can be obtained with yield prediction at the design stage and allowing control of the
variations in the process for good assembly yields. This includes optimizing the substrate designs to compensate for
the process variations to reduce the formation of electrical opens and bridgings in the assembly. The present paper
discusses the development of a yield prediction model for area array assemblies using Monte Carlo simulation. The
simulation model was developed with the objective function of predicting the defect levels. The yield prediction
involved several critical parameters including the substrate pad dimensions,solder bump height,solder volume,
warpage and pad layout on the component. A public domain program,“Surface Evolver”,was also used in the
course of this yield modeling effort to calculate the solder joint shape. The Monte Carlo simulation based model can
be used for conducting sensitivity analysis of various design parameters and their effects on the assembly yield.