1. www.globalsmt.net
The Global Assembly Journal for SMT and Volume 9 Number 4 April 2009
Advanced Packaging Professionals
ISSN 1474 - 0893
Krassy Petkov
Interview Inside
Hot air solder leveling
in tHe lead-free era
NEW PRODUCTS
vapor pHase vs. convection reflow in
roHs-compliant assembly INDUSTRY NEWS
conquering smt stencil printing INTERNATIONAL DIARY
cHallenges witH today’s miniature
components

2. Vapor phase vs. convection reflow in RoHS-compliant assembly
Vapor phase vs. convection
reflow in RoHS-compliant
assembly
by Dan Coada, EPIC Technologies, Norwalk, OH USA
introduction to consider in volume manufacturing.
The contract manufacturing in- Vapor phase (VP) reflow technology has Today’s VP reflow process makes use of
dustry is changing rapidly from been in existence since the early 1970s as a the heat produced by a boiling fluorinated
lead-based soldering to lead-free reflow process for surface mount technol- polymer or fluid. This boiling fluid pro-
soldering. There is no stopping ogy (SMT) assemblies. While used for some duces a uniform temperature zone (vapor
defense applications and in smaller volume blanket) in which the PCBA is exposed for
the transition or the reality that production settings, the disadvantages
lead-free components are going solder purposes. Heat is transferred to the
associated with the initial processing tech- PCBA as it is immersed into the vapor area
to be introduced in lead-based nology limited its widespread acceptance. until the PCBA reaches temperature equi-
processes. This challenge to These disadvantages included environmen- librium with the boiling point of the fluid.
engineering and quality is a tal concerns about the fluids being used, The primary soldering benefits of VP in
huge concern and one that needs throughput limitations, applicability only comparison to infrared (IR) or convection
scrutiny and a watchful eye. to single-sided printed circuit board as- include an oxygen free (inert) environment
EMS providers rely on compo- semblies (PCBAs) and an inherent problem without the need for nitrogen, fixed upper
nent suppliers to ensure that the with tombstoning. temperature exposure and superior heat
lead-free transition on the com- Advances in VP technology have transfer on thermally challenged PCBAs.
ponent terminations is seamless addressed many of these shortcomings VP also offers distinct advantages
through continued development of im- in the realm of lead-free soldering. Key
to their soldering processes, but proved machines, chemical selections and
that rarely happens. Termina- benefits include a lower peak reflow
process controls. Doublesided PCBAs are temperature, an inert environment without
tion changes require additional easily processed in current equipment. As a nitrogen, improved solder wetting and flow
modification to solder profiles result, VP is becoming a viable alternative and a reduction in profiling time
and flux chemistries in order to
ensure proper wetting of the sol-
der to the lead-free termination.
The need for nitrogen to be used
in convection reflow is becoming
a requirement more than an op-
tion, and nitrogen is costly.
This paper will look at the
advantages and disadvantages
of vapor phase (VP) and convec-
tion reflow in RoHS-compliant
processing and discuss associ-
ated design for manufacturing
(DFM) issues.
Keywords: Vapor Phase,
Convection Reflow, IR Reflow,
Lead-Free
This paper was originally published in the
proceedings of the SMTA International
Figure 1. Vapor phase soldered, lead free, ENIG surface fnish (U26 partially removed in shear/tensile test).
Conference, Orlando, Florida, August 2008.
20 – Global SMT & Packaging – March 2009 www.globalsmt.net

3. Vapor phase vs. convection reflow in RoHS-compliant assembly
Figure 3. Vapor phase lead-free profile.
Figure 4. Vapor phase tin-lead profile.
and complexity to get the profile close so
that process development time is kept to a
minimum.
Conversely, VP reflow profiling can
be classified by process type to the point
Figure 2. Convection lead-free profile. where there are fewer profiles to develop.
For instance, standard multi-layer 0.062”
thick PCBAs can follow the same lead-
lower peak reflow processing improved solder wetting and flow
based VP profile regardless of component
VP reflow requires a lower processing Visual inspection of micro sections
complexity. With advances in VP process
temperature. In convection or IR process, indicates that VP creates good solder joint
systems, monitoring of heat load during
temperatures can reach 245˚C to 265˚C performance. When larger thermal load
the soldering process allows the systems to
at the component level. VP temperatures components or clusters of components are
profile almost automatically. Ramp rates
stay at the boiling point of the fluid, present, time above liquidus (TAL) should
and soak times at peak temperature can be
typically established between 230˚C and be increased beyond the 60-90 seconds rec-
defined by the engineer and controlled by
240˚C. The lower temperatures make it ommended by solder paste manufacturers
the systems regardless of the product mix
possible to reduce the cost of the PCB by to accommodate thorough heat transfer.
during the process. In a true one-piece flow
using lower Tg/Td material selections for Comparatively, in convection processing it
on a prototype, it is much easier to get it
SMT assemblies. Savings of 10-15% and can be more difficult to ensure good joints
right the first time using VP processing.
even more could be seen on laminate costs on components with high thermal mass be-
The days of inadequate reflow tempera-
alone. VP also offers processing advantages cause achieving TAL in larger components
ture or over temperature on the first piece
with large mass components, such as con- may result in smaller components overheat-
are virtually eliminated by use of the VP
nectors, because the thermal equilibrium ing. There is no chance of overheating
systems.
is better. In convection reflow, particularly smaller or isolated components with VP
in higher temperature lead-free processes, because VP cannot heat a component
comparative data on lead-free and
correctly soldering large mass connectors higher than vapor temperature.
tin-lead solder joint creation
may overheat the rest of the PCBA.
Standard test boards (Figure 1) populated
reduction in profiling time
with common components, including
inert environment Another advantage of VP’s heat transfer
BGAs and QFPs, were built with vapor
The inert environment and consistency of characteristics and the uniformity at which
phase and convection soldering technology
heat transfer allow VP to be more forgiving it accomplishes heat transfer is that it
and tested at an EPIC facility. Reliability
with lead-free component terminations. makes it easier to understand the profiling
testing demonstrated that lead-free and
Less active no clean flux chemistries have relationship between PCBAs. In traditional
tin-lead joints produced by vapor phase to
proven to be adequate in soldering lead- profile development for a new PCBA,
be equally robust as those from convection
free terminations that demand high activ- sample assemblies with thermocouples are
reflow. The controllable, lower peak solder
ity fluxes in the convection reflow process. run through the reflow process numerous
temperature makes vapor phase ideal for
There is also cost savings due to lower times in order to get the right profile, with
soldering complex assemblies having sensi-
energy consumption. In addition to the manual inspection of the solder joints
tive lead-free SMT components.
elimination of nitrogen, electricity usage and flux residue used to determine if the
with VP is much lower. target profile is correct. Engineers develop
matrix charts on board size, layer count
www.globalsmt.net Global SMT & Packaging – March 2009 – 21

4. Vapor phase vs. convection reflow in RoHS-compliant assembly
2000 1000 hrs. Visual inspection for solder balls,
Cycles Accel. Age tombstones, bridging, voids and dewetting
SnPb No Clean Convection
indicated no apparent difference between
19.5 * 9.34 *
the two methods of solder joint creation.
Vapor Phase 6.67 16.8 No tombstones were experienced on the
Lead-free No Clean Convection JOCY test vehicle boards in either case.
2.11 9.16
Visual inspection indicates that while
Vapor Phase 2.59 5.5 vapor phase created solder joint perfor-
SnPb Water Soluble Convection 14.5 No data mance and micro-section appearance on
the board is very good, it might be a good
Vapor Phase 10.2 7.96 idea to explore increasing the lead-free
Lead-free Water Soluble Convection 1.31 23.7 TAL above the 60 to 90 seconds recom-
Vapor Phase
mended by solder paste manufacturers
4.16 5.16
to accommodate thorough heat transfer
to larger components or clusters of large
* Tin-Lead No-Clean resistance measurements tended to decrease slightly. Others in-
creased or were mixed hence, absolute values to were used to assess changes. components. Larger thermal load compo-
nents, especially in clusters, tend to retard
Table 1. Average percent resistance change (absolute value). the complete melting of lead-free paste.
It is more difficult to ensure good joints
on components with high thermal mass
Initial 2000 Cycles 1000 hrs. in convection processing because while
Thermal Shock Accel. Age trying to achieve a sufficient TAL on larger
SnPb No Clean Convection 19.1 23.1 23.6 components, smaller components in less
populous areas may tend to overheat.
Vapor Phase 22.8 21.6 19.0 Much discussion in trade magazines
Lead-free No Clean Convection 27.6 25.6 26.6 and forums such as the IPC TechNet has
focused on the question of soldering tin-
Vapor Phase 27.3 27.6 26.2
copper and SAC-alloy-terminated BGAs
SnPb Water Soluble Convection 23.7 20.3 and other components with standard tin-
Vapor Phase lead solders. Using a 230˚C vapor phase
23.5 20.5
system, even liquification of these termina-
Lead-free Water Soluble Convection 26.3 27.6 25.0 tions ceases to be a problem while posing
Vapor Phase 26.9 28.0 25.7 little chance of overheating heat sensitive
components. Similarly, risks associated
Table 2. Shear/tensile force required to remove SOIC16 (pounds of force). with lower Tg substrates and temperature
sensitive components is reduced relative to
lead-free convection processing.
Test Description profiles (Figure 2) provided a peak tem- Since cleanliness had been studied
Tests were conducted using a VP reflow perature of 245˚C and a TAL in the 60-90 using ion chromatography for a previously
process. Vehicle boards were used with tin- second range recommended by paste sup- published report2, a cleanliness compari-
lead HASL or lead free immersion silver, pliers. Vapor phase soldered boards were son was made for this report using ROSE
immersion tin or ENIG surface finishes, as soldered in an EPM-IBL SLC500 vapor techniques. An Omegameter operating
appropriate. Boards were populated with phase soldering chamber using Galden above 100˚F was employed. No differences
tin-lead or lead-free components, printed, LS/230 Perfluorinated heat transfer were detected in ionic cleanliness between
assembled and soldered using standard fluid. The vapor phase profiles developed boards soldered using convection reflow
reflow or VP production equipment. The provided a TAL of about 90 seconds and and those soldered in vapor phase. Lead-
solder pastes selected for testing included a maximum temperature of 230˚C, a free no-clean samples tended to have 50%
tin-lead and lead-free no-clean and water temperature that is governed by the vapor higher contamination levels than standard
soluble formulations. Assembled test temperature. After a vapor phase profile tin-lead boards due to the type and level
boards were thermal shocked between is established, TAL can be modified to of flux used in lead-free pastes. All results
-45˚C and +125˚C with 20 minutes dura- achieve any time required without exceed- were well below IPC limits.
tion at each limit for 500, 1000 and 2000 ing the 230˚C maximum temperature. Resistance across soldered BGA daisy
cycles in EPIC’s Failure Analysis Labora- The vapor phase equipment first chain arrays of 40 and 80 joints were the
tory. Other test boards were subjected to preheats the board using infrared. Next, same for convection and vapor phase
accelerated aging at 85˚C and 85% relative the work is lowered into the vapors at a reflowed test boards within the limits of
humidity for 1,000 hrs. The JOCY test programmed rate to regulate ΔT and TAL. experimental measurement. (Daisy chained
vehicle is populated on only one side, al- After the work reaches the maximum va- dummy 169 and 352 termination BGAs
though it is equipped with plated through- por temperature, the duration of its expo- containing four daisy chains each were
holes (PTH) for mixed technology tests. sure is preprogrammed. Several soldering used.) Solder joint conductivity did not ap-
Test boards were populated with dum- programs can be developed by the engineer pear to deteriorate measurably after either
my 402, 603, 805, 68 pin PLCC, TSOP32, and stored in memory to suit the needs of 2000 thermal shock cycles or 1000 hours
SOIC TQFP QFP208 and daisy-chained different lead-free or tin-lead board types. of accelerated aging at 85˚C/85RH. Dur-
BGA169 and BGA352 components. ΔT and TAL are controlled by the program ing the 2000 thermal shock cycles and
Standard lead-free convection reflow developed by the engineer.
Continued on page 34
22 – Global SMT & Packaging – March 2009 www.globalsmt.net

5. Conquering SMT stencil printing challenges with today’s miniature components
Minimum Aperature Vapor phase vs. convection reflow in
surface registration RoHS-compliant assembly, continued
Technology area ratio Cost Material accuracy from page 22
accelerated aging, the average absolute
Chemical etch 0.66 Low SS, Alloy Moderate
change in resistance on measured daisy
Traditional laser-cut 0.66 Low SS, Alloy Very High chains is summarized in Table 1. The differ-
Traditional laser-cut 0.55 Low Slic™ Very High
ence between the performance of VP sol-
dered and convection soldered test boards
Electroformed 0.5 High Electroformed High is insignificant considering the limited data
Nickel set. The resistance change for each sample
Advanced laser-cut 0.45 Medium Fine Grain Very High is reported as the average of absolute values
of the changes in resistance for a set of sam-
Table 1: Stencil technology summary. ples. No special preparation or seasoning of
samples was performed. Resistance values
were recorded ‘blind.’ A small amount of
mounted into a stencil frame, tension The answer to these important ques- ohmmeter drift was experienced at the low
is applied to electroformed foil by the sten- tions is in our view an unequivocal ‘yes.’ resistances measured.
cil frame’s polyester mesh. This tension Stencil laser and material technologies The shear force required to cause SOIC
pulls on the foil causing slight shifts in the have advanced to the point where laser-cut joint failure was measured and found to
locations of the stencil apertures. In most stencil performance is beyond that of be the same for convection and vapor
cases, the electroformed stencil aperture current electroformed technology. Using phase reflowed test boards. Shear force was
locations will be long, or further away the new LPKF high-power short-pulse fiber measured on an SOIC16 (U25 and U26)
from their expected locations. If the PCB laser technology and the new Fine Grain exerting a combination shear and tensile
has SMT pad locations that are short of material, stencil performance is significant- force that pushed the component parallel
expected locations and the electroformed ly improved over electroformed, especially to the plane while lifting the component
stencil has aperture locations further away when printing miniature components. by means of a 30˚ wedge. These measure-
than expected, there can be a significant Improvements in stencil laser and mate- ments did not deteriorate after shock and
shift, or misalignment, between the stencil rial technologies have lead to significant accelerated age. While thermal shock re-
apertures and PCB pads. improvements in solder paste release down sults were measured at 500 and 1000 shock
A shift between the stencil aperture to a surface area ratio of 0.45 as well as cycles, only those from the 2000 cycle test
are reported here. Results are summarized
and PCB pad reduces the amount of solder improved aperture registration accuracy.
in Table 2.
paste in contact with the surface of the These improvements are critical to meet-
PCB pad. This lowers the adhesive force ing future requirements when printing
conclusion
between the solder paste and PCB pad, miniature components like 01005s. The VP key benefits include:
effectively reducing the ability of the board technology summary is as follows: • Lower peak reflow temperature
to pull the paste from the stencil. Minia- At a cost savings of 30-50 percent • Inert environment without
ture components already have very low compared to electroformed, the ability to nitrogen
surface area ratios. The lower the surface produce multi-thickness (step) stencils, • Improved solder wetting and flow
area ratio, the more critical the alignment and the option of same day turn times, • Reduction in profiling time.
between the stencil aperture and PCB Fine Grain stencils, cut with the new
pad. The Fine Grain stencil in this DOE fiber lasers, are a marked improvement As the data above indicates, thermal
was cut in the frame on the new LPKF compared to the high-performance stencil profiles using vapor phase soldering equip-
high-power short-pulse fiber laser. The solutions available today. OEMs and CMs ment are controllable with the maximum
intent was to minimize stencil aperture can get the performance they need while temperature dictated by the specific ther-
registration errors, thereby increasing the reducing costs and meeting critical delivery mal transfer fluid employed. The tin-lead
alignment accuracy between the stencil and schedules. The new stencil laser and mate- and lead-free solder joints created using
PCB. The results (27 position errors for rial technologies available today give stencil vapor phase technology have equivalent
the Fine Grain stencil and 2,307 position manufacturers the tools and materials performance to those created using convec-
errors for the electroformed stencil) below needed to supply an ever-changing industry tion equipment, while offering a uniform
show a marked improvement in aperture for many years to come. fixed maximum temperature of controlled
registration when compared to an electro- duration. Vapor phase solder joint creation
Acknowledgement offers a viable alternative to convection
formed stencil.
reflow. Convection reflow has less uniform
The authors would like to thank Stephan
maximum temperatures over complex
conclusion Schmidt and Sebastian Gerberding of
circuit board surfaces.
As advancements continue in component LPKF Laser Electronics (www.lpkfusa.com)
and PCB technologies, will the stencil tech- for their contribution to this article. references
nology of today provide current and future 1. Munroe, C., “Beating the RoHS Heat,”
solutions to the challenging assembly issues Circuits Assembly Magazine,” March
faced by OEMs and CMs? Is electroformed Robert F. Dervaes is V.P. technology and engi- 2005, pp. 38-47.
technology the right solution or have new neering for Fine Line Stencil, Inc. Jeff Poulos, 2. Fraser, S. and Munroe, C, “Lead-Free
developments in stencil laser and material is V.P. of manufacturing and sales, Alternative Using Vapor Phase Reflow in Lead-
technologies caught up with and surpassed Solutions, Inc. Scott Williams is product/ac- free Processing,” SMT Magazine, April
the electroformed technology of today? count manager with Ed Fagan, Inc. 2005, pp. 48-49.
36 – Global SMT & Packaging – March 2009 www.globalsmt.net