5703

DRAFT DOCUMENT FOR INDUSTRY CONSENSUS REVIEW ONLY
Page 1 of 38
IPC-5703 Cleanliness Guidelines for Printed Board Fabricators

PROPOSED STANDARD FOR BALLOT FEBRUARY 2013

1 SCOPE

1.1 Introduction In many IPC standards development meetings, those individuals responsible
for assessing the quality of incoming unpopulated printed boards have lamented the fact that bare
printed board cleanliness is often an unknown quality parameter, often with undesired results.
This is often due to the lack of understanding of materials and processes by the fabricator, or
more often because the industry has driven margins so low that experienced process
professionals cannot be retained by the fabricator.

Therefore, the IPC 5-32c Bare Board Cleanliness Task Group resolved to generate guidelines on
those factors in the printed board fabrication process which can directly or indirectly affect the
final cleanliness of packaged bare printed boards. This document is intended to be a general
tutorial on items in the printed board fabrication process which can affect, directly or indirectly,
cleanliness.

1.2 Why Printed Board Cleanliness is Important One of the aspects of printed board assembly
reliability is electrochemical reliability, which is related to the residues left on an assembly.
Some residues are benign while others are harmful. Electrochemical failure mechanisms are
comprised of three elements: (1) an ionic residue; (2) an electrical potential or voltage gradient;
and (3) moisture or humidity. All three must be present at some minimum level (which varies)
for electrochemical failures to occur. Electrochemical failures are evidenced in two primary
areas: (1) unacceptable leakage currents under humid conditions, and (2) electrochemical
migration (dendritic growth). An example of dendritic growth can be seen in Figure 1-1.

Figure 1-1 Examples of Dendritic Growth
The Venn diagram shown in Figure 1-2 is useful in understanding the relation between these
three factors. The amount of impact can be visualized by considering the diameter of the
individual circles to be proportional to the forcing function. As an example, if a device is
operating in a very humid environment, the diameter of the humidity circle grows larger as does

Page 2 of 38
the electrochemical failure region. In addition, operating temperature and time can compound or
accelerate these factors. Higher temperatures can increase dissolution of residues and can
accelerate electrochemical reaction rates. If service lives are long, the risk of electrochemical
failures increases.

Figure 1-2 Venn Diagram Illustrating Variables Affecting Electrochemical Failure
Every printed board, or printed board assembly made from such boards, has a different
cleanliness need. Unfortunately, not many assemblers or OEMs know how to determine how
clean their assemblies have to be. There have been many instances where the present printed
board cleanliness test methods, such as resistivity of solvent extract (ROSE) testing, have been
inadequate to determine the true cleanliness of printed boards or assemblies.
How clean a printed board or a printed board assembly needs to be for a particular application or
end use environment depends on a number of factors. Among them:
As circuit densities increase (e.g. smaller lines and spaces), the electrical bias across the
smaller space results in a higher voltage gradient across this space. Consequently, the bias
circle shown in Figure 1-2 expands, and the risk of failure increases.
For an assembler who utilizes some form of cleaning, the increased circuit density often
means a greater challenge to complete cleaning.
For an assembler who does not utilize any cleaning processes (i.e. no-clean), the
cleanliness of the bare printed board becomes critical as there is no opportunity to remove
harmful residues in the printed board assembly process.

From the perspective of the assembler or user of the printed board, the use of process control
methodologies in the manufacture of printed boards is very important, especially for high
reliability applications. Because printed boards for these critical applications are often very
sensitive to contamination, showing consistency in the fabrication processes involving cleaning
is often of great interest. To an experienced printed board assembly auditor, if there are no
metrics in the fabrication process which focus on board cleanliness, a red flag is generated.


Page 3 of 38
Such an auditor will also look for objective evidence that the resulting cleanliness of the printed
boards has been evaluated as part of the qualification process, whenever major changes in
materials or fabrication processes are made. New technologies, such as new plating solutions or
methods, may add undesirable residues to the printed board. It takes more than just a word from
a salesman to say a new material or process is clean. Factors which have never impacted
cleanliness before can become factors as technology advances. The vigilant fabricator keeps
evaluating the cleanliness of the process as new technologies come on line.

From the perspective of the assembler, tracing field failures to the cleanliness of incoming
printed boards is often exorbitant in terms of time and cost. Often, an assembler has to examine
the cleanliness of each step of the entire manufacturing process to determine where a cleanliness-
related fault originates. If the expertise or equipment to do this evaluation does not exist in the
assemblers organization, then the task must be subcontracted or there is a need to have
specialized consultants brought in. What may only be a difference in pennies in a printed board
fabrication process can cost an assembler hundreds of thousands of dollars to track down.

While this document goes over the various fabrication processes and how they impact printed
board cleanliness, the single greatest factor for printed board cleanliness is the quality of the
water used for washing and rinsing operations.

1.3 Terms and Definitions The definition of all terms used herein shall be in accordance with
IPC-T-50.

2 APPLICABLE DOCUMENTS

2.1 IPC
1

IPC-WP-008 Setting up Ion Chromatography Capability

IPC-TM-650 Test Methods Manual
2

2.3.25 Detection and Measurement of Ionizable Surface Contamination

2.3.25.1 Ionic Cleanliness Testing of Bare PWBs

2.3.28 Ionic Analysis of Circuit Boards, Ion Chromatography Method

2.3.28.2 Bare Printed Board Cleanliness by Ion Chromatography

2.6.14.1 Electrochemical Migration Resistance Test

IPC-HDBK-840 Solder Mask Handbook

IPC-1601 Printed Board Handling and Storage Guidelines

IPC-5704 Cleanliness Requirements for Unpopulated Printed Boards


Page 4 of 38
IPC-6012 Qualification and Performance Specification for Rigid Printed Boards

IPC-9191 General Guideline for implementation of Statistical Process Control (SPC)

2.2 American Society for Testing and Materials
3

ASTM F21 Standard Test Method for Hydrophobic Surface Films by the Atomizer Test

ASTM F22 Standard Test Method for Hydrophobic Surface Films by the Water Break Test

1. www.ipc.org
2. Current and revised IPC Test Methods are available on the IPC Web site (www.ipc.org/html/testmethods.htm)
3. www.astm.org

3 WATER QUALITY

Water quality varies tremendously around the world. Tap water, used by many fabricators for
rinsing and cleaning, may have minerals, particulate matter, pollutants, or additives like chloride
or sulfur that are used to reduce bacteria. Hard water (heavy in mineral deposits) is often limited
in the ability to clean. The minerals in the water can chemically interact with manufacturing
solutions in fabrication, resulting in a residue that may be difficult to remove. The cleaner the
water, the better the cleaning ability, which lowers the risk of unintended residues.

Effective aqueous cleaning in printed board fabrication is a combination of time, temperature,
and solvating ability of the water or cleaning solution:

Hot water cleans better than cold water.
Water quality may vary by season or by water source (e.g. river).
Clean water cleans better than polluted water.
Water with a detergent or cleaning agent additive works better than plain water, provided
you have adequate rinsing.
Printed boards exposed to a longer cleaning cycle will be cleaner than those with a shorter
cleaning cycle.
A thorough cleaning process, incorporating cleaning agents and rinsing, is better than a mild
or low cost cleaning process.

In consideration of overall water quality, the following grades of water progress from least
desirable as a cleaning agent to most desirable. As water quality rises, the cost per unit volume
also increases:

Unfiltered tap water: as indicated above, unfiltered tap water may have many materials
which are both detrimental to printed board reliability and which reduce the effectiveness as
a cleaning agent.
Tap water that has been pH altered: In some portions of the world, the natural pH of the
water may be alkaline and fabricators have sometimes added acids to reach a more neutral
pH. While this may balance the pH, the acid itself is detrimental to cleaning.

Page 5 of 38
Filtered tap water: Simple filtration with activated carbon can greatly reduce undesirable
materials in the water, as well as reduce particulates in the water. Undesirable
contamination levels can still remain.
Softened water: The water softening process will remove minerals from tap water and is
often effective at reducing particulates in water. The purpose of water softening is to
remove divalent metals like calcium and magnesium, exchanging them for sodium. Water
softening operations do not imply carbon filtration. The softening process involves
processing the water through salt brine. Consequently, softened water is often high in
sodium and/or chlorine. Both of these elements are known to be detrimental to printed
board cleanliness. If softened water is used in a cleaning process, it is highly recommended
that subsequent rinses be with reverse osmosis or deionized water.
Reverse osmosis (RO) water: RO water is generated using membrane technology. RO
water is free of particulates, minerals, and most ionic material, and is an effective cleaning
agent. RO water has a resistivity range of 100 kilohm-cm to 2 megohm-cm. One of the
drawbacks of RO water is that it usually takes 10 gallons of tap water to generate 1 gallon of
RO water.
Deionized (DI) water: DI water is the most desirable of water grades for purposes of
cleaning. Deionized water is generated by taking RO water and passing it through ion
exchange columns, removing any residual ionic material. Deionized water has a resistivity
range of 2 megohm-cm or higher. Laboratory grade DI water will be around 18 megohm-
cm. A resistivity of 2-8 megohm-cms is desirable for most aqueous cleaning. If water
resistivity is significantly higher than 10 megohm-cm, it can become too aggressive a
cleaning agent, especially for semiconductors and similar microcircuits.

An industrial aqueous cleaning operation will typically have tap water plumbed first into a
particulate trap and then into a carbon filtration column. The output of this column will feed into
a series of deionizing columns which make the water progressively cleaner. In-line resistivity
monitors can show either a numeric readout of the water resistivity or a simple red light / green
light arrangement if the resistivity falls below a set value, usually 2 megohm-cm. Most often,
these lights are located at the DI source. Monitoring lights or meters are most effective when
placed at the point of use (where the rinsing is performed). These monitors need some flow
through them to be effective.

As the ion exchange resins in the column become expended, the first column in line is removed
and taken for resin regeneration. Each column is moved up in the rotation and a fresh column
added at the end of the line.

In some cases, highly polished deionized water (18 megohm-cm at the source) has been used
successfully for specialty aqueous cleaning operations. It is difficult to maintain this high level
of resistivity. Carbon dioxide in the atmosphere will rapidly degrade water resistivity from 18
megohm-cm to 2-8 megohm-cm. Resistivity of the deionized water will also drop the farther it
is piped from the deionizing source (columns), placing the resistivity in the desirable 2-8
megohm-cm at the point of use. This resistivity drop should be considered when determining the
placement of deionizing columns and plating tanks or aqueous cleaning equipment. If the source
and destination are widely removed, and high resistivity water is required, a small polishing
column can be implemented at the destination.


Page 6 of 38
Piping for deionized water is recommended to be electropolished 316 stainless steel (not 304
alloy), or Schedule 80 gray polyvinyl chloride (PVC), chemically resistant PVC (CPVC) or
polyvinyldienefluoride (PVDF). It is recommended that the following NOT be used for piping
deionized water: copper, unpolished stainless steel, garden hoses, or Tygon tubing. Tygon
tubing can, over time, leach material from the plastic, and may contribute to organic growth (e.g.
algae).

A RO and DI water setup should incorporate ultraviolet (UV) irradiation of the water. This will
keep algae out of the water system and can greatly reduce the amount of bacteria in the water.
Algae and bacteria can have undesirable results in printed board fabrication and their presence
will not be detected by simple conductivity cells. An RO/DI system involves membrane
technology which requires periodic maintenance to the membrane to keep it clean.

It should be noted that the RO and DI process does not remove everything from the water. Most
public water treatment plants use sand as a filtration media. Consequently, most deionized water
contains some soluble silica. Standard filtration methods will not remove it and since silica is
not ionic, deionizing columns will not remove it. Soluble silica can accumulate with time,
making preventative maintenance problematic and can eventually transfer to circuit assemblies.

It has been debated in the industry on whether or not all rinsing in a fabrication process has to be
with DI water. Some fabricators use tap or RO water to knock off the bulk of the chemical
residues, followed by a hot DI water rinse. While this is possible in some operations, impurities
in the tap or RO water can cause some chemicals to become insoluble in a later rinse.

While many of the chemical analysis methods focus on the presence of ionic contaminants on
printed boards, the presence of non-ionic contaminants can be just as harmful to board quality.
These non-ionic materials can cause problems in plating, with solder mask adhesion, mass
lamination layer separation, solderability problems with plated metals, and conformal coat
adhesion.

Highly deionized water does not negatively impact most electronic assemblies. DI water in
manufacturing rapidly drops from the 18 megohm-cm level to the 2-8 megohm-cm level.
Repetitive cleaning with DI water is not detrimental. Long term exposure to highly polished (18
megohm-cm) DI water (e.g. soaks for extended periods) may microetch electronic surfaces, but
this is a rare occurrence.

One exception to this may be the cleaning of microelectronics, such as wafers and semiconductor
die, where the wire bonding pads are pure aluminum. Highly deionized water can quickly attack
such surfaces, oxidizing them and interfering with subsequent wire bonding. In such cases, it is
common to bubble carbon dioxide through 18 megohm-cm water to bring it down to 2-3
megohm-cm resistivity. This assures that the water is very clean but not so aggressive as to etch
the aluminum pads.

In short, with some limited exceptions, all water used in printed board fabrication should be DI
water.


Page 7 of 38
3.1 Cleaning Additives There are a variety of chemical additives designed to improve the
effectiveness of water in an aqueous cleaning process. These additives are generally detergents
or saponifiers and should be chosen for the soils involved. While commercial grade detergents,
such as dishwashing detergent, are attractive from a cost standpoint, they are not designed for
cleaning electronic materials and often leave undesirable residues such as surfactants.

Some cleaning agents can be corrosive and may pose a risk to some board finishes if not
completely removed during the rinsing process. Care must be taken in selecting the cleaning
agents for the particular application.

4 PREVENTATIVE MAINTENANCE
As with any production facility, periodic preventative maintenance is critical to consistent
production and consistent quality. Plating tanks can accumulate chemicals and other residues.
Pumps in cascade rinse systems must be maintained to ensure proper and adequate flow rates.
In-line water resistivity meters should be periodically checked and calibrated. Older systems
which are not used in continuous production may be more at risk than new systems used on a
daily basis. The newer systems may have numerous self-checks built in.

5 HANDLING AND ASSOCIATED SOILS

Ideally, printed boards should not be touched or handled at all during the manufacturing process.
However, in practice this is often impractical. The best that can be done is to minimize the
accumulation of human-related contaminants, by reducing handling to a minimum and ensuring
that various cleaning steps are capable of removing the contamination before it can do any harm.
Handling occurs during shipping and receiving, transferring of printed boards through process
steps, inspection and manual assembly. Contaminants introduced by personnel handling are
perhaps the most complex to define. Human contact with the surface can introduce a variety of
contaminants that include:

Finger oils
Salts (from perspiration)
Dust and airborne debris
Food particles
Flaking skin
Hand creams and cosmetics
Lint or debris from gloves
Bodily fluids (may require disinfection)

Such contamination may be water-soluble, solvent-soluble, or insoluble in either water or
solvent. For this reason, handling contaminants often present challenging cleaning problems.

Fingerprints are a complex and potentially harmful source of contamination. Fingerprints contain
not only oils, salts and acidic body products, but also residues from soaps, hand lotions, and
previously handled objects. Since they are so complex, it is impossible to ensure complete
removal by any specific cleaning process. Personnel should be trained to handle printed boards
as little as possible and then only by the edges. Potential damage depends on the location of the
fingerprints. On bare copper and on laminate surfaces between copper conductors, they reduce

Page 8 of 38
insulation resistance. Fingerprints can also reduce solderability. Fingerprints on boards may do
no damage, but any accumulation of fingerprints will give higher ionic contamination readings in
the ROSE test. See IPC-TM -650, Method 2.3.25.

The use of gloves will protect against fingerprints, but gloves also can act as a source of
contaminants. Plasticizers, talcum powder, dust, and lubricants can be transferred from rubber
gloves. Woven fabric gloves may contribute fibers. All gloves could have been previously
contaminated, if used for other purposes. Gloves which come into contact with other process
materials become contaminated and so must be kept clean or changed frequently. Operators must
be trained that if wearing gloves, the gloves themselves must be kept clean as well. This is
frequently overlooked or misunderstood, leading to a false sense of security. If using disposable
gloves, frequent glove changes can become a significant expense, leading to pressures to change
gloves less often, leading to contamination.

Gloves should be chosen to be resistant to the process chemicals they contact. They also should
be chosen to be comfortable to the operator. If not, they will either be used less or not at all. An
alternative is to use cotton gloves with rubberized palms and fingers, but with the back of the
gloves untreated cotton. This allows for a glove that breathes and so is more comfortable for the
operator, while maintaining a tactile surface that is cleanable with suitable solvents. When
properly cared for, such gloves can last for months.

Food, drink, and tobacco products should be prohibited in all process, storage, and inspection
areas.

6 FABRICATION PROCESSES

As in all manufacturing processes, control of the process and process parameters is critical for
continuing quality. Statistical process control (SPC) methods will reduce risks and allow faster
response to out of control situations. More information on SPC can be found in IPC-9191.

Figure 6-1 gives a generic process flow for fabrication of a multilayer rigid printed board. Each
step in the fabrication process has the potential for adding undesirable residues to a printed
board, or for changing the properties of the materials such that they become more susceptible to
residue retention.


Page 9 of 38

Figure 6-1 Generic Fabrication Process for Multilayer Rigid Boards

6.1 Raw Material Selection Laminate Laminates vary widely in terms of resin materials,
additives to the resin, and porosity of the resin to manufacturing fluids. Lower grade laminate
resin systems can absorb processing fluids, such as hot air solder level (HASL) fluxes, especially
after repeated exposures to temperatures above the glass transition temperature (Tg) of the
laminate.

Some resin systems contain brominated fire retardants, such as tetrabromobisphenol A, to gain a
particular flammability rating for the laminate material. Some laminates may carry other ionic
elements with them, such as sodium (from alumina fillers). This may be a concern if these
chemicals are pulled from the laminate matrix during assembly operations (e.g. reflow or
cleaning). At this point, the chemicals may be chemically aggressive towards surface
metallizations, or may be free to fuel electrochemical failure mechanisms.

Additives to laminates can be either chemically bound into the resin system, or may not be
bound. When laminates are exposed to reflow operations, which takes the resin systems above
the laminate glass transition temperature (Tg), the laminate can become more porous. As such,
they may absorb process residues. A more porous laminate is likely to have internal additives
leach from the bulk of the resin in a rigorous chemical extraction test, such as ion
chromatography.

The predominant resin system used for rigid printed boards is epoxy-glass, more commonly
known as FR-4. Other resin systems, such as polyimides or polytetrafluoroethylene (PTFE) may
have different porosities and so have different risks for absorption of process chemicals. Those
resin systems that are more porous and at a higher risk of absorption are also at a higher risk of

Page 10 of 38
process chemicals coming out of the resin during assembly operations. Ion chromatography is
most often used to test for this condition.

6.2 Raw Material Storage Most fabrication shops start with packages of copper clad laminate
and pre-preg material. The storage environment can affect the quality of the material. Residue
considerations in this process include:

Dust Fabrication shops are dusty places. Particulate matter from drilling and routing is
common, even when vacuum capture of dust is used. Dust and particulates can have
detrimental effects in many phases of fabrication and lamination.
Environmental gases and pollutants Many packaging materials are permeable to chemical
vapors. Vapors may include methane, salt air, fuel combustion byproducts (e.g. natural gas
powered forklifts), etc. Many cardboard materials are heavily loaded with sulfur compounds,
which can detrimentally affect surface cleanliness of materials.
Moisture if the storage of base materials is not controlled, laminate materials can absorb
water from the air, effecting later lamination operations.
Infrastructure Residues (e.g. wallboard /ceiling particulates).
Air Handling Systems - very seldom do facility air handling (HVAC) systems get cleaned.
Dust and other particulates can accumulate in HVAC ducts and unless filtered out, become a
source of particulate contamination in production.
Bird droppings many fabrication sites have loading and shipping docks with open doors.
As such, birds and insects can enter the facility and affect product quality.
Food and drink, smoke it does not take much of these kinds of residues to detrimentally
affect board quality, either chemically or from a particulate standpoint.
Fingerprints and handling dirt All human hands carry oils and salts, regardless of how often
hands are washed. In addition, you may have residues from hand sanitizers or hand lotions,
which add chemical residues.
Fiber from cardboard and textiles Fibers from packing material, dust from packing
materials , and threads from clothing are all particulates which can get on the raw material
before it starts the fabrication process.
Material handling racks Individual panels may have low levels of contaminants, but if the
racks used to move them between processes are not periodically cleaned, they can
accumulate residues and cross contaminate everything they come in contact with.
Conveyors - Conveyorized systems should be periodically cleaned as part of a preventative
maintenance program. Failure to do so can result in accumulation of particulates, copper
slivers, oils and maintenance fluids. Prolonged contact with plating chemicals can result in
the breakdown of the plastic material for conveyor rollers, resulting in undesirable residues
such as silicones.
Separation Sheets - Many laminate materials come with Mylar separation sheets. Some
commercially available sheets have coatings on them to minimize contamination. While it is
a common practice to re-use these separation sheets, it can represent a source of cross
contamination. If sheets are reused, they should be periodically cleaned.

6.3 Copper Clad Laminate Preparation Copper will oxidize with time and so most processes
start with removal of the surface oxides in the copper. Residue considerations in this process
include:

Page 11 of 38

Pumice residues particulate and chemical. Pumice is crushed volcanic rock that is used as
a mechanical abrasion removal of the copper oxide. Failure to properly rinse the pumice can
leave behind particulate material as well as the minerals that pumice contains.
Aluminum oxide slurries are also used to remove copper oxides, similar to pumice.
Some copper surface prep rinsing will contain low levels of citric acid in the DIwater,
because citric acid can improve image resist adhesion. Citric acid can also delay the re-
oxidation of the copper for a more robust process. Pumice abrasives sometimes incorporate
citric acid for this reason.
If the process is not well controlled and well cleaned, it can lead to copper mousebites and
extraneous copper going into the resist lamination area, which is generally a cleanroom
setting.
Finger salts and oils at this point can also jeopardize photoresist adhesion in subsequent
operations.

6.4 Clean and Apply Inner Layer Photoresist. Photoresist is a coating of photo sensitive
material to be used to place the image of the inner layer circuit pattern on the clad material. This
step is most often done in a clean room environment, because small amounts of particulates, such
as dust or human hair, can result in defects in fine circuitry. As such, image resist should be
applied in a clean room environment (class 10,000 or better recommended) with HEPA air
filtering and positive air pressure. Failure to use these kinds of facilities dramatically increases
the probability of problems from particulate residues. Residue considerations in this process
include:

Cleaning of the inner layers can be done mechanically or chemically. Mechanical scrubbing
is often used to remove chromate materials.
If the inner layer core material arrives direct from the manufacturer, the materials could have
a conversion coating on them which must be removed, usually with an acid-based process,
followed by a DI water rinse. If the material is supplied in double-treat condition, it can go
into the lamination process without cleaning.
Chemical cleaning most often comes in the form of a microetch for improved adhesion.
Handling should be done with gloved hands as finger salts and oils are detrimental. Failure
to remove particulates and oils will result in areas where the oxide treatments will not stick or
where the resist will not stick.
There are a number of cleaning processes for surfaces, or chemical preparation steps prior to
resist application. In general, cleaning should be with DI water.
To address surface particulates, printed board surfaces pass under a tacky roller to pick up
particulates. Most of these rollers are clean and do not leave chemical residues, but this
should always be verified. These rollers have to be changed periodically or they can cross-
contaminate the copper surface. The frequency of such changes depends on how clean the
material is coming in and the volume of laminate passing through the rollers.

6.5 Expose and Develop Photoresist After the light-sensitive photoresist is laminated to the
surface of copper clad laminate, it must be imaged. Older processes used high precision
phototools, while newer processes are based around laser direct imaging (LDI). The photoresist
exposure is the process of transferring the circuit image, which is on working tools, to the copper

Page 12 of 38
surface. Some have reported cross contamination issues, where undesired residues on the
photoresist transfer to the phototool and then are subsequently transferred to other surfaces. In
the developing process, the unexposed portion of the resist is chemically removed.

For the developing process, there are a number of different chemistries used, depending on what
resist is being used.

A common developing solution uses potassium carbonate (around 1.25%). For this
chemistry, tap water as a quench is used more often than DI water because the minerals in
tap water can help to harden the image resist walls. This is one of the exceptions to the
general recommendation of using deionized water.
In developer operations, it is strongly recommended that there be dedicated equipment for
mask developer chemistries. These chemical processes are very sensitive to chemical
impurities, and so changing chemistries in the same developing tanks often results in cross-
contamination affecting the quality of the process.
Final rinsing in this step should be done with DI water and should be thorough enough to
remove residual etching and developing chemicals.

6.6 Etch Inner Layers After photoresist imaging, extraneous copper is etched from the copper
clad laminate. The copper protected by the developed photoresist is not removed, while all other
copper is stripped away. In general, etching inner layers is different than etching what will be
the outermost layer of a printed board. For the outmost layer, the photoresist is generally 2-2.5
times thicker than the photoresist used on inner layers. Inner layers of a printed board often go
through a process to induce an alternative oxide structure in the copper to improve adhesion in
lamination. This oxide process (not a true oxide) is the result of a microetch. Laminate material
should be age controlled. Older resins can affect resin flow. Glass fibers can absorb humidity.
Both of these can affect stackup.

When considering inner or outer layer imaging, there are four considerations between developing
and stripping of layers as shown below:

Developing Stripping
Inner Layer Incomplete stripping can leave
islands of copper
Residual resists affect oxide
uniformity (mottling)
Outer Layer Residual resist interferes with
subsequent plating or stripping
processes
Incomplete etching can result
in hard shorts or incorrect line
widths

Residue considerations in this process include:

This etch process is often done with cupric chloride etchant and hydrochloric acid, usually
with a pH of around 1, and often with a sodium chlorate oxidizer. It is important that these
etching chemicals be fully neutralized or they continue to attack the copper.
If particulates were under the photoresist during lamination of the resist, etching solution can
get into the space between the resist and the copper, attacking/removing the copper.
If punch machines are used, they can leave oils and greases on surfaces.

Page 13 of 38
Finger oils or residues from hand lotions make surfaces resistant to mild microetch processes.
This can effect etching processes and interfere with bonding processes.

6.7 Outer Layer Imaging The outermost layer of the printed board undergoes more process
treatments than do the inner layers. Additional plating processes are often used. Residue
considerations in this process include:

Pure tin or tin-lead alloys are often used as an etch resist material for outer layers. Tin lead is
usually not used for inner layer etch resists.
After etching, the tin or tin-lead is stripped from the surface. Residues can remain from the
tin-lead stripping process. Residues can include intermetallics if left too long in the stripping
process. Some all-in-one strippers go after both residues and intermetallics while two-
stage strippers strip the tin or tin-lead, then go after the intermetallics.
Nitric acid is often used to strip tin or tin-lead. Failure to rinse can lead to nitrate residues.
In many fabrication processes, the tin-lead is applied and stripped several times, resulting in a
cumulative attack on polymeric materials on the board.
In the treatment transfer process, the resin is very rough at the microscopic level where the
copper tooth has been removed. If the etching process is not complete, residual copper can
remain in the bottom of the resin crevices, and is difficult to remove. This residual copper
can reduce dielectric strength, increase the risk of arc-over and support electrochemical
migration failures.
In printed boards that go through sequential lamination, the outer layers and subsequent
residues on those layers become inner layers on the laminated printed board.
If the layers are not clean and kept clean, then resin spots can form on the top of the copper
material. These resin spots do not come off with microetching and can effect subsequent
process steps.
Be aware that even in acidic plating baths, algae can grow, resulting in biological
contamination of the boards.

6.8 Strip Photoresist After the developed photoresist has performed its function, it must be
removed from the board for subsequent operations. This is often an in-line process, using highly
alkaline materials (pH 13) such as monoethanolamine (MEA). Fortunately, MEA rinses very
well. Residue considerations for this process include:

Residual stripping agents.
Incomplete removal of photoresist leaving particles of resist or a thin film of resist. This can
affect adhesion.

6.9 Layup for Lamination This is the procedure by which a multilayer package is
assembled. The circuit layers (laminates) are stacked over pins in plates, in numerical circuit
order. Between each laminate is placed a pre-defined thickness of bonding material, usually a
multiple of pre-preg sheets. Residue considerations in this process include:

Particulates - layup should be done in clean room environments with laminar air flow and
filtering. Particulates will interfere with layer to layer adhesion.

Page 14 of 38
Handling residues finger salts and oils can degrade layer to layer adhesion. Layup should
be done with powder-free gloved hands.
Mixed resin chemistry in the layers some grades of epoxy laminate and polyimide laminate
do not work well with each other. This most often occurs when the epoxy selected cannot
take the heat of lamination needed for polyimides (or similar high temperature resin
systems). This most often manifests itself in dark spots or charring of the epoxy resin. Cross
contamination can also result in delamination, as epoxy dust trapped in polyimide has often
been the cause of delaminated boards.
Post-etch punch particulate residues or optically-aligned punches This occurs when the
tooling holes for the layup are punched, producing particulates. The duller the punch, the
more ragged the hole and the more particulates that get generated. Some materials are more
susceptible to generating particulates or debris than others. Punch presses can also transfer
oils and greases to surfaces if not properly maintained.
Copper fibers from post-etch punch flakes of copper or copper slivers can result in inner
layer shorts and delamination.

6.10 Lamination This process is a temperature and pressure method, typically performed under
vacuum. During lamination the resin in the bond material liquefies and then cures back to a solid
state creating a homogenous package. Although different bonding and laminate materials may
be interchanged, the best results are usually found when using common dielectrics that are
compatible. Residue considerations include:

Overcooking up to and including laminate degradation. Some of the older epoxy resin
systems, with low Tg values, when subjected to too much heat or to heat for too long will
severely darken or char. Such degradation cannot be reversed and may result in undesirable
chemical products breakdown on the surface of the laminate.
Undercooking this means that the resin system has not attained the desired degree of cure.
This can mean that the resin structure is more porous than it should be, resulting in
absorbance of harmful process solutions later in the fabrication process. Undercured resins
may also result in more smear in subsequent drilling operations, requiring a more aggressive
desmear operation. Drilling an undercured resin system can also result in birdnesting of
epoxy shavings during the drilling operation. In most cases, the severe smearing resulting
from drilling an undercured resin is too great to be removed by desmear processes, resulting
in scrapping of the printed board.
Undercured resins can also smear during routing operations and leave edges of the printed
board more porous, which can carry plating chemistry from one process to the next.
Under curing resins may allow for migration of chemical constituents in the resin to migrate
out as a contaminant.
If inner layers are not clean prior to lamination, they may carry harmful acidic residues.
After lamination, those acids may still corrode the copper, resulting in opens. After a printed
board assembly has gone through lamination, there is no possible way to clean these inner
layers.
Oxide treatments of inner cores traditional black oxides vs. the new oxides, propensities for
conductive anodic filament (CAF) and other such failures increases with the new materials.
If Press Pads (silicone rubber pads) are used, after so many uses the silicone will break down
and form a residue that needs to be cleaned.

Page 15 of 38
Gloves used in fabrication are most often untreated latex (powder free) or nitrile based. Be
aware that if the external surfaces of the glove become contaminated, they should be
replaced.
Gloves advertised as antistatic or electrostatic discharge (ESD) compliant often gain their
antistatic properties using chemicals, such as amines, which can transfer from the gloves to
the board surface. ESD or treated gloves are typically not used in fabrication. If they are
used, gloves should be investigated for this prior to use.

6.11 Drill Holes After lamination, or for simple double-sided boards, precise holes must be
drilled to connect the surface layers and sub-layers. The drilling of holes uses the design data for
size and location at the start of the interconnect process. Drilling provides the mechanical
opening that contacts or clears internal features. Residue considerations include:

This is a very dusty process generating large amounts of particulate matter, even when
vacuum systems are employed.
Drill vacuum is critical to ensure complete removal of drill debris from holes, which could
otherwise obstruct or inhibit plating.
Drill sharpness is important. A sharp drill shears away the resin. A dull drill tends to smear
the resin and leave a more ragged drilled hole, which can entrap plating residues later on.
If there are foreign objects between the printed board being drilled and the backing sheet,
then the board will not sit flat on the backing sheet. This leads to the formation of copper
burrs on the bottom side of the board.
High helix drill bits are often used due to greater material ejection and less heat generated.
These drill bits tend to smear the resin less.
Backing material can be pulled up into the hole as the drill exits.
Laser drilling of vias. This is not commonly done for large holes, but is common for creation
of microvias. If this is not tightly controlled, the resin can pyrolize and re-deposit on nearby
surfaces. The resin system will have been degraded by heat and the redeposited material may
interfere with subsequent plating. This is often a problem with laser drilling of holes.
Some holes are produced by punching rather than drilling. Hole punching can result in
ragged hole walls and entrapped plating solution, and fractured resin which may allow
absorption of plating chemistries into the bulk of the resin or along the glass fiber bundles (a
conductive anodic filament, or CAF, risk).

6.12 Deburring, Etchback, and Desmear Following the drilling of the hole, the hole must be
cleaned for subsequent plating operations. Deburring operations remove copper burrs from the
top and bottom of the hole. Desmearing removes the smeared resin, opening up the copper faces
for electroplating. Etchback will chemically remove some of the resin material such that inner
layer copper traces will be exposed during plating. Residue considerations include:

Vacuum systems are commonly used to remove particulate matter from the holes, such as
backing material. The particulates that remain are often but not always removed during the
subsequent deburring process.
Plasma desmear operations have to be tightly controlled in terms of gas composition and
flow, power levels, temperature, and duration to get a consistent amount of etch. If

Page 16 of 38
uncontrolled, the plasma can over-etch metal surfaces and re-deposit the metal on nearby
surfaces, resulting in possible shorts or low resistance paths later.
Desmear is sometimes done using solutions of potassium permanganate. If this process is
not well controlled, the resin can be swelled and the potassium permanganate might not be
totally removed in subsequent rinsing operations. In addition, if the resin-to-glass bond is not
good in the hole side walls, then wicking of the permanganate solution into the glass bundles
can occur. This can lead to CAF growth in subsequent electronic assemblies.
Older chemical processes utilized hydrofluoric acid (HF) to etch glass after the plasma
desmear operations. More modern plating processes use a safer sodium bifluoride material to
perform this function when needed. Both materials need to be carefully handled and
thoroughly rinsed.
High aspect ratio holes, which have high length or depth ratios in comparison to the holes
preplated diameter, are quite often hard to clean, and can entrap plating residues. This is
especially true if the via is capped on one side. Blind microvias, which are not drilled all the
way through the printed board, may be hard to clean. Drilling operations sometimes have
difficulty producing a clean bottom to the hole.

6.13 Making Holes Conductive After the holes have been drilled, cleaned and made ready for
plating, the holes are then made conductive. There are a wide range of chemical processes used
for making holes conductive with hundreds of variables to control. Some of the more common
methods include electroless copper, black hole, and direct metallization

Electroless Copper chemically plates an initial thin copper layer over the whole panel surface.
"Direct Metalization" chemically deposits a thin layer of one of several alternative conductive
materials (carbon, palladium, graphite) in a binder over all panel surfaces, then removes the
excess from the top and bottom surfaces, leaving a surface that can be electrically plated.
Strike Plate is the electroplate process that deposits additional amounts of copper in the holes
and on the surface. (This process is optional.)

All of these chemical processes involve copious amounts of water. Here it may be
permissible to use lower grades of water to knock off most of the harmful chemistry, with DI
water to do the final rinsing and not the bulk washing.
The primary cleanliness consideration for all of these processes is entrapped plating residues.
Most of these processes incorporate an alkaline cleaning step, which rinses fairly well.
Persulfate or peroxide microetch solutions are also sometimes used and can be entrapped if
rinsing is not thorough.
In the black hole process, where carbon fills the hole, if too much carbon is left in the hole,
it dries and skins over. This subsequently results in voids in the plating, which can entrap
other residues.

6.14 Application of Outer Layer Photoresist This is similar to the inner process of surface
preparation except this time the goal is to define the outer features. The clean and photoresist
application steps are very similar. The coating process can be used with varying thicknesses of
resists to allow for specific photo and plating process operations. The resist thickness is usually
greater when multiple, subsequent plating steps are involved. In this process, oxides are not
used, and tin-lead is often used as an etch resist. Residue considerations include:


Page 17 of 38
If surfaces are not prepared well, or cleaned, the photoresist will generally have adhesion
issues, either too much or too little. Too much adhesion and the presence of resist will
interfere with the final plating. Too little adhesion and the resist will delaminate and allow
plating residues to get under the resist.
This process may include acidic microetch processes to remove oxides. Microetch processes
must be well rinsed and thoroughly dried.

6.15 Outer Layer Expose and Develop These steps again are similar to the inner layer
processes. The expose process, although similar, has the basic concept reversed. The photo
tooling or laser image has been made to allow the polymerized areas to remain where copper will
be later removed. The develop process will chemically remove the un-polymerized resist on the
surface features. This now opens a pocket for additional plating and reveals the copper. Residue
considerations include:

One of the biggest concerns in this area is peeling plating, where plating does not adhere to
the substrate. If the substrate is not clean, peeling plating may result.
Residual acids can carry through into this process, but not often. Subsequent processes are
very good at removing such acid residues.
Water quality here can affect plating adhesion. The use of surfactant chemicals, if not fully
rinsed from the printed board can lead to peeling plating.
If you have hydroscopic residues on the printed board, usually from poor rinsing after
development processes, you can have both peeling plating and mousebites (missing copper).
If there are exposed glass fibers, which is a rejectable condition, the glass fibers can absorb
many process chemicals by capillary action and are impossible to remove. These residues
cannot be detected by most production tests and contribute to later failures like electrical
leakage and CAF growth.
Residual resist material can affect etching. If the resist, regardless of chemistry, remains
where it is not desired, and there will be incomplete etching of the copper.
In this process, incomplete rinsing in any step leads to subsequent process problems:
o Leaving etchant residues too long results in over-etching, with the copper conductors
being undercut, and a copper lip on top. Etching goes down and sideways. When
this happens, residues become harder to rinse out of the area.
o Solder strip over-exposure
o Pumice lines
o If rinsing is poor or poor quality water is used, chemicals which may be fully soluble
using hot DI water may form a different complex if rinsed with cold/poor water. The
resulting salts are then insoluble and contaminate subsequent processing steps.
Poor rinsing is the most common source of ionic contaminants and entrapped materials.
HASL and fused tin-lead use very aggressive fluxes. Some fusing fluids (e.g. hot oil) have
acids incorporated or have fluxes added. In both cases, failure to clean these flux/acid
residues results in high levels of contamination and corrosion.

6.16 Pattern Plate and Outer Resist

Pattern Plate means the plating of defined features. It is the process by which additional copper
is electrically deposited in the developed cavities to bring hole sizes and surface feature

Page 18 of 38
thicknesses up to customer specifications. This copper plated surface is then over plated,
traditionally with a type of tin/lead, but other over-plates such as nickel, gold, etc. may be used
as specified by design. Many of the over-plates have a dual role. They act as the protector of the
plated copper in the fabrication process. In the printed board assembly process they become the
areas to which components can be mounted/soldered.

Strip Resist is a chemical process of removing the photo resist which is over the unwanted
copper. Note that the inner sequence of etch and strip is now reversed. Over-plate has taken the
place of the photoresist as copper circuit protector.

Residue considerations in this step include:

The same elements listed previously for plating.
If there is a significant time period between process steps, oxides can build up between
process stations. If the oxides dry out, especially the nickel in an ENIG process, then
subsequent plating and etching processes will be adversely impacted. This may affect resist
stripping, leaving hard bridges and etch resist metals where they are not desired.

6.17 Etch Copper and Etch Resist Metal For the outer circuitry, the etch chemical process
now removes all the unwanted copper as supplied by the raw material vendor and plated on at
electroless and strike (panel) operations. Similar to the inner etch, a surface cross section view
of the circuitries will show hour glassing. At this point in the build process the design is
electrically completed. The tin-lead strip process involves either completely removing the etch
resist metal or using another photoresist process to selectively strip the tin-lead. This prepares
the copper lines for solder mask to become the protective coating over bare copper lines. The
only steps remaining are finalizing the surface requirements, mechanical processes (milling, etc.)
and final inspecting.

Residue considerations in this step include:

All of the elements covered in Outer Layer Expose and Develop also apply to the surface
finish and copper etch steps.
Copper etching is often done in an alkaline system and uses flood rinsing. A replenisher
chemical removes the copper salts. The amount of tin-lead overhang can affect the ability of
the rinsing to get rid of undesirable copper salts.
For the resist metal, many metals can be used as an etch resist, with tin being the most
common. Some metal salts rinse easier than others.
As with other rinse steps, it is important to remove all of the metal salts and the etch
chemicals.
Surfaces have to be clean prior to solder mask application. After solder mask application, the
residues cannot be removed without completely stripping the solder mask. Usually it is more
cost effective to scrap the printed board.

6.18 Solder Mask Application and Development Solder mask is a protective layer of acrylic-
epoxy resin applied over bare copper lines, known as Solder Mask over Bare Copper (SMOBC).
The process to apply solder mask is to clean, coat, expose, develop and cure. In many modern
designs, the solder mask also serves as the processing mask for the application of immersion

Page 19 of 38
finishes. Most solder masks are one of two varieties dry film solder mask or liquid
photoimageable solder mask. As a photosensitive material, it is applied, imaged, and developed
much the same as other photoresists discussed previously. See IPC-HDBK-840 for more
information on solder mask composition and processing.

During the imaging process, ultraviolet (UV) light is used to cure the solder mask in the
desired areas. This process is highly dependent on the wavelength and intensity of the UV
light, and the consistency of the light with time. UV light is provided by mercury-vapor
lights, and the intensity of the light degrades with time. If the light intensity is not monitored
frequently, then the solder mask does not completely cure, even in subsequent thermal
bump secondary cure processes.
It is important that the UV bulb intensity be monitored frequently (at least weekly). Process
belt speeds can be slowed to give the mask the same amount of UV irradiation, but printed
boards get very hot in a UV oven/tunnel. If the belt speeds are too low, the boards can get
too hot. UV bulbs need to be replaced periodically as part of a preventative maintenance
schedule.
Most facilities do UV first, then thermal. Some of the immersion finishes (tin, gold) are very
hard on solder masks. UV then thermal is more effective in most cases. The thermal cure is
important because the UV energy may not get to the bottom of the solder mask at the critical
mask/laminate interface. Thermal curing completes the bonding process.
Excessive thermal cure can turn the solder mask brown and brittle.

Additional residue considerations include:

One of the most important operations in the solder mask process is the pre-clean step. It is
essential that printed board fabrication shops use DI water in this step and not city water.
City water quality is highly variable and often contains very high levels of chlorine and other
contaminants which can affect solder mask adhesion to the printed board, as well as
contributing to corrosion problems on the printed board in the field. See Figure 6-2 and
Figure 6-3 for examples of solder mask adhesion problems caused by inadequate rinsing
prior to the solder mask application process.

Figure 6-2 Solder Mask Adhesion Problems Caused by Inadequate Rinse Prior to
Application

Page 20 of 38

Figure 6-3 Solder Mask Lifting from Conductor Due to Inadequate Rinse Prior to
Application

In many cases, the solder mask bubbling illustrated in Figures 6-2 and 6-3 will result in
flaking solder mask, as shown in Figure 6-4.

Figure 6-4 Solder Mask Flaking Due to Inadequate Rinse Prior to Application

Figure 6-5 shows a similar bubbling of the solder mask, but is due to entrapped volatiles
from the solder mask resulting from improper solder mask curing, and not from improper
rinsing prior to solder mask application.

Figure 6-5 Solder Mask Bubbling Due to Entrapped Volatiles

Page 21 of 38

Solder mask is often required to be within a specific thickness range. Tests used to verify
solder mask thickness can often result in significant levels of handling residues.
Some customers require testing of surface energy of the cured solder mask. This is most
often accomplished either by water drop contact angle testing, or the use of calibrated alcohol
solutions applied by dyne pens. Both of these methods have the opportunity to introduce a
contaminant to the solder mask surface which must be cleaned prior to subsequent
operations.
Solder mask is often used to seal over holes/vias on the printed board. This operation is often
called tenting. When solder mask is used to cap only one side of a hole/via, then it creates
an entrapment site from the opposite side of the printed board. If that hole has a high aspect
ratio, it is next to impossible to thoroughly clean out, resulting in entrapped acidic residues
from the final plating operation. Such residues can etch away the through-hole walls.
If the solder mask is used to seal the top and bottom of the hole, but not the bulk of the hole,
it leaves a thin film unsupported at either end of the hole. It is not uncommon for that thin
film to crack during thermal operations, allowing flux or other process solutions to wick into
the hole. Such residues cannot be removed.
If solder mask is used to fill the hole/vias, it may result in a very thin film of solder mask on
the surface pad, interfering in later plating operations. In this case, the undesired residual
solder mask is itself a contaminant.
A partially cured solder mask is an undesirable condition as the mask becomes a sponge for
subsequent process fluids, such as plating solutions, fluxes, or fusing fluids. Present
cleanliness test methods often cannot detect this problem.
As with other photoresists, it is important to completely clean and rinse the developing
solution.
Solder masks vary greatly in terms of porosity and absorbance of process chemicals. This
should be a consideration when choosing a solder mask. Absorbed chemicals are often
difficult to detect in standard production processes.
Dry film solder masks often have a problem with immersion in alcohol-based solutions, such
as the 75% isopropanol / 25% deionized water used in ROSE testing or IC testing. Extract
solutions may acquire a greenish cast to them for dry film masks. The degree of cure of the
mask often is directly related to the amount of green in the extract solution. The less cured a
mask is, the more green the solution. Several test labs have determined impromptu criteria
for the amount of green seen in the extract vs. the cure of the mask.
Solder masks often have a high degree of filler material, such as calcium. Filler material can
be ionic and can be leached out, especially if the mask is not fully cured.
The most common contaminant on the outer surface of HASL/HAL printed boards is that of
flux residues from the fluxing or fusing process. Here, it is recommended that hot DI water,
with a mild surfactant, be used to clean the HASL flux residues. Cold tap water followed by
a DI water rinse is seldom effective. If the mask is not fully cured prior to the HASL/HAL
step, the flux residues will be absorbed into the solder mask and the exposure to the molten
solder essentially seals off the surface, making the condition undetectable until it gets to the
OEM.
Solder mask cures from the top down. Dust on top of the solder mask during the UV
irradiation can scatter the UV light, affecting image quality. FOD can be imaged into the

Page 22 of 38
mask and results in undesirable mask quality. The energy absorbs differently in cured vs.
uncured mask.
If the development process is not well controlled, a thin layer of solder mask may remain on
copper surfaces. This residual mask is very thin and very difficult to detect. If the residual
mask is present prior to final finish, the surface will not plate correctly. If the residual mask
is present on the board that goes to the OEM, the surface will not solder properly, can effect
electrical test, and can interfere with component bonding.
If the air going into the UV oven is not filtered, then dust or other foreign objects can deposit
on the wet solder mask surface. There should be a solvent drying step prior to going into UV
curing to minimize this.
If legend or marking inks are being applied to the mask, or if the mask is being used as a
legend or marking material, then residues can transfer board to board if the screens are not
cleaned well. Residual ink or mask can be very thin and can effect final finish plating.
Screens do wear out and become for difficult to clean with age. The emulsion coating can
break down and come off into the ink, compromising the material properties. The screen
registration may change and the ink edge definition may be feathered resulting in more
FOD.

6.19 Processing Aids This can refer to a wide array of materials and tooling:

Temporary masks (e.g. peelable latex masks)
Residues from the adhesives from masking tapes. Many are acrylic based but some are also
silicone based, resulting in later adhesion problems.
Gloves can often contribute a false sense of security. Gloves can contribute fibers and
many white cotton gloves are bleached to whiteness, and so have chlorine residues. Can also
contribute conductive media such as talcum, which is heavily ionic.
Local cleaning media such as wipes, cotton swabs, etc.
Slip sheets used as inner layer separators during transport between process stages.
Transport equipment and tooling
Are silicone greases used as sealants for vacuum-based test fixtures? If so, this is a common
source of cross-contamination.

6.20 Final Finish One of the last steps in the fabrication process is the application of the final
finish, usually a solderable metal such as gold, tin, tin-lead, silver, or copper with an OSP finish.

The earliest final finish was Reflowed Tin/Lead where tin/lead etch resist was melted using
hot oil to form a solderable finish. This process often contributed residual fusing fluids,
such as peanut oil. If the solder mask was not fully cured, these corrosive fusing fluids
carried through to the OEM.
Next came solder coating using a process called Hot Air Solder Leveling (HASL) where
boards are dipped in molten solder and withdrawn through forced air which leveled the
solder coating. The most common residue from the process was residual water soluble
fluxes, which were highly aggressive. Simple water cleaning is not enough to remove these
residues.

Page 23 of 38
More modern immersion finishes have gained favor due to the planar deposit they
provide. Planar deposits help placement and soldering of tiny components. The immersion
finishes available today include:
Immersion Tin
Immersion Silver
Electroless Nickel / Immersion Gold (ENIG)
Electroless Nickel / Electroless Palladium / Immersion Gold (ENEPIG)
Copper Organic Solder Preservative (OSP)
These plating chemistries are generally cleaner than HASL or fused tin-lead, but are very
chemically aggressive. As such, they can chemically attack solder masks and carry corrosive
residues to the OEM.
These plating processes are a long series of: treatment, then rinse; next treatment, then rinse.
If there is inadequate rinsing in any of the process sequences, then poor plating may result.
For example, poor rinsing in an ENIG process can lead to passivated nickel, which leads to
peeling plating.
If inadequate rinsing occurs at the end of the plating line, conductive metal ionic material can
result. Process lines are generally set up for thorough rinsing. If a rinse line shuts down
during the process, you get all kinds of problems, cross contamination of baths, etc. Final
board washes with DI water are usually inadequate to clean these insoluble salts. All process
chemistries rinse well when wet, but generally do not rinse well when dry.

6.21 Outer Layer Marking After solder mask curing, the outer surfaces of the solder mask are
often printed with legend and marking inks to indicate part numbers, serial numbers, component
identification marks, etc., with the most common form of this being silk screen printing. The
inks used are typically black, white, yellow or red. Residue considerations include:

Be aware that some of these legend and marking inks are electrically conductive and can
result in conductive paths if the ink is smeared or is applied where unintended. Cross
contamination of screens may be an issue if the screens are not properly screened or the same
screen is used for different kinds of inks.
Permanent ink markers (e.g. Sharpies) can be conductive and in some cases corrosive. The
non-black and yellow versions should be viewed with caution. These inks are often easily
soluble in isopropyl alcohol and so may be removed unintentionally.
Sometimes the wrong ink is used, and the ink will not cure properly or have the necessary
chemical resistance in OEM cleaning processes. Part markings that get washed away at the
OEM level is a problem.
Temporary maskants (e.g. tapes) can leave residues on the printed board which can interfere
with later coating adhesion at the OEM.
Very seldom do people consider preventative maintenance on ovens used to cure legend and
marking inks. All of these inks have volatile constituents and outgas certain levels of
contaminants. They accumulate on oven surfaces and can eventually contaminate anything
that goes into that oven.
Labels used to mark defects on a printed board can often have undesirable residues from the
adhesive used for the labels. Adhesives can be acrylic, but many are also silicone based,
which can interfere with later coating adhesion.
Residues from grease pencils can seldom be completely removed. These residues can
interfere with later adhesion. Writing on boards is not recommended.

Page 24 of 38
Black legend inks are available but are often conductive. When applied across conductors,
they are often sources of arcing failures in high voltage testing. Conductive inks are more
often a problem on subsequent runs if the inks are not well cleaned from the screen.
Ovens used for curing legend and marking inks should be cleaned periodically. Legend and
marking inks often outgas small amounts of material. Over time, these residues accumulate
on oven surfaces and can cross contaminate, or cure-inhibit other items like adhesives.
Inspection stickers (arrows) can leave residues on a surface. Most adhesive residues are
soluble with isopropanol.

6.22 Depanelization and Routing Following the completion of the printed board build process,
the multiple boards in a standard panel must be removed from the panel and separated into
individual boards. This may be done through the use of high speed routing, V-scoring, or punch
presses. Residue considerations include:

The primary consideration here is the dust and debris that is generated. Even with vacuum
systems, some material will escape into the shop environment.
Handling residues are a more important consideration here, because there is seldom a
cleaning or rinsing step following depanelization.
Dull equipment (drill bits, router bits, and punches) greatly increase the amount of dust and
debris in the air. Dust and debris can also come from backing materials.
Pre-routing operations:
o Printed boards as individual pallets within the manufacturing panel will have pre-
routes defined to facilitate removal of the final printed board assembly. These are
usually where connectors over-hang or tall components near the edge that would
hinder the router on final route out. There could also be internal routing defined. If
this is done as an in-process step and is followed by final test and a final cleanliness
requirement, then there is less of a concern.
o Same as above, except this is done as a post-process step. Handling, the routing
process itself and/or assembly could contaminate the printed board.
o Same as above but done at a sub-tier printed board assembly house. This may include
the punching, or scoring or a true engineered pre-route. Concerns are same as above
only more so due to potential unknown (or lack thereof) of critical processing steps.
The packaging and re-packaging are added concerns.
o If the final delivered entity is the single printed board image that had pre-routes and a
post-route and there are requirements for a final test, final clean, SIR data, IC data or
equivalent, then there is less concern.
Post routing operations. This usually refers to deburring and detab operations. The primary
residues are handling residues and dust.
During depanelization and routing, it is not recommended to wear gloves as the gloves may
get caught in the rotating equipment, pulling hands into the cutting mechanism. In such
cases, care should be taken to handle the boards by the edges only. A mild water rinse can
usually remove finger oils and mild dust levels.
Pallet residues or printed board stacking residues. In many fabrication operations, the
individual printed boards are separated by an interleave material, such as silver saver paper,
copper saver paper, or a re-use of interleave material from incoming laminate packages.
Such materials are designed to be rugged and low residue. Standard copier paper should not

Page 25 of 38
be used as bleach is used to get the vibrant white color. Bleach is heavy with chloride
residue. Sulfur free paper should be used for any media that contacts the printed board
surface.

6.23 Electrical Test After depanelization, printed boards are tested electrically for opens and
shorts. This can be done manually, or more commonly by bed-of-nails (BoN) testers, which test
many circuit networks at once, or by flying probe testers that are slower than BoN testers, but
more flexible for low rate production. Residue considerations for this operation include:

Handling Residues. There is a large amount of handling that occurs during this operation.
Finger salts and oils can accumulate on printed board assembly surfaces. A mild water rinse
is usually sufficient to address these.
Residue transfer from pins. This may include oils, greases, inks, residual flux, solder mask
residues, or flakes of metal from soft solderable surfaces such as tin-lead. While this kind of
transfer is possible, it is not common. The in-circuit test pins would not work if such was the
case and pins are replaced routinely.
Oils and other lubricants from the testing machine.
Particulates.
It is not uncommon to have collections of assemblies which have failed electrical test, which
may have been reworked, and are awaiting to run a re-test. These have a high tendency
towards the accumulation finger salts, rework chemicals, etc.
An electrical test area should be in a separate room, but not necessarily a standard clean
room. This room should have a higher than normal humidity level to avoid arcing in high
voltage electrical testing that can occur if the air is dry.

6.24 Packing / Board Storage / Shipment The last materials that come into contact with the
printed board are the materials used for packaging and shipping to the assembler. Some
materials are benign with respect to printed board cleanliness, and others can be very
detrimental. The cleanest fabrication process in the world will be rendered ineffective if the
wrapping or shipping materials contaminate the product.

The reader is referred to IPC-1601 for additional of information on this topic.

7 ASSESSING CLEANLINESS

All OEMs desire clean printed boards; however, the issues of how cleanliness is measured and
how adequate cleanliness is determined for any particular board have been around for decades.
References [1] and [2] are articles from the IPC Review that give a summary of IPC and military
cleanliness specifications as of 2000. Those articles show that the most prevalent form of
measuring printed board cleanliness is the use of ROSE testing, discussed earlier. In addition, a
common form of cleanliness testing for the telecommunications industry has been the use of the
Bell Communications Research (Bellcore) SIR testing.

Because the issue of cleanliness by ROSE testing is so central in the industry, the reader needs to
understand some of the history surrounding this test and its intended uses, and the misuses and
abuses commonly found.


Page 26 of 38
7.1 History of ROSE Testing In the 1960s, the U.S. Department of Defense (DoD) became
increasingly concerned about the failure of printed board assembly cards due to ionic
contamination. At that time, there was no recognized test for cleanliness in existence, either for
printed boards or for manufactured assemblies.

Bill Hobson and Bob Denoon, material scientists at the Naval Avionics Center in Indianapolis,
Indiana, developed a test method that involved isopropyl alcohol, a beaker, and a conductivity
cell. High solids rosin fluxes (i.e. R/RMA/RA), commonly used in military manufacturing at
that time, were used as the base of the testing. This is an important thing to remember when
comparing modern materials to materials of that era. The predominant ionic materials of interest
were amine hydrochloride or amine hydrobromide activator materials in fluxes. When a mixture
of 75% isopropyl alcohol and 25% DI water, at a high resistivity level, was poured over a printed
board, the conductivity of the captured wash solution would increase as ionic residues from the
activators would dissolve in the extract solution.

Hobson and Denoon wanted to relate the conductivity of the solution against some known
standard. They chose sodium chloride (NaCl) as a simple chemical salt that ionized easily in
isopropanol and water. As a comparison, they developed the relationship between the number of
micrograms of NaCl and the corresponding decrease in solution resistivity. They then equated
the drop in resistivity of the captured solution to the amount of NaCl required to give the same
drop in solution resistivity. Therefore, the result of the test was a measure expressed as
micrograms of NaCl equivalence per square inch of extracted printed board assembly surface. It
is important to understand that this measure has no relation to the amount of sodium or chloride
(or any other ion) on the actual printed board assembly. This is a frequent misunderstanding in
the industry.

Consequently, the test was termed the Resistivity of Solvent Extract, or ROSE test. It is
important to understand that this methodology was developed as a process control tool only, not
for product acceptance. The research by Hobson and Denoon was published as NAC Materials
Research Report (MRR 3-72) in 1972.

In subsequent research, Hobson and Denoon worked to correlate the results of the ROSE testing
with some measure of electrical performance degradation. What did a drop of 6 megohm-cm in
solution resistivity mean to electrical insulation? They worked to correlate the results of the
ROSE testing with surface insulation resistance (SIR) testing. Their work was published in NAC
Materials Research Report (MRR 3-78) in 1978. They found that, in general, when ROSE
results indicated a minor drop in solution resistivity, measured SIR under humid conditions was
high. When major drops in ROSE testing occurred, SIR levels dropped. They established a
general breakpoint between acceptable SIR levels and unacceptable levels of SIR, which was
determined to be 10.06 micrograms of sodium chloride equivalence per square inch. Again, it is
important to remember that this testing was done to develop a good process control tool, and not
to determine acceptable or unacceptable cleanliness of electronic assemblies as that depends on a
high number of factors.

Because the DoD, then a major factor in the electronics manufacturing industry, had no measures
for acceptable or unacceptable ionic cleanliness, and such a measure was desperately needed, the
10.06 micrograms of sodium chloride equivalence per square inch metric was placed into the

Page 27 of 38
Navys Weapon Specification WS-6536. This was done against the strident objections of
Hobson and Denoon. The DoD soon incorporated this value into other specifications, such as
MIL-STD-2000, MIL-P-28809 and MIL-P-55110.

Hobson and Denoon also worked with industry partners to develop an automated ROSE test
instrument. The resulting instrument was called the Ionograph. Similar instruments followed,
but each manufacturer went about the process differently, which meant that for the same
contamination, each unit would read a different result. An equivalence factor was determined
for each machine, as shown in Table 7-1.

Table 7-1 Equivalence Factors for ROSE Testing Equipment
Instrument Equivalence Value
Manual Method 10.06
Omegameter 14.00
Ionograph 20.00
Zero Ion 37.00
ICOM 5000 22.00

There were two basic approaches for these machines: (1) the test assembly would sit in a static
bath of extract solution and the solution conductivity monitored over time; and (2) the test
assembly would sit in the extract solution which would circulate and the conductivity of the
solution monitored as it exited the cell and the result integrated over time. The first method
became the static method and was covered as IPC-TM-650, Method 2.3.26. The second method
became the dynamic method and was covered as IPC-TM-650, Method 2.3.26.1. Both of these
test methods have since been superseded by IPC-TM-650, Method 2.3.25.

Until about 1985, all of the test units used a room temperature solution and a 10 minute run time
(usually). Then, in an effort to move more contaminants into solution, test units began
incorporating both heated solutions 46 C [115 F] and spray agitation. No equivalence factors
were determined for such instruments, or have been since, to the IPCs knowledge. It has been
suggested that heated solutions and spray agitation are inadvisable for ROSE testing of bare
boards as this increases the likelihood of extracting ionic materials from the materials of
construction.

Over the last two decades, many customers and suppliers have struggled with comparisons
between ionic cleanliness testers. Bergendahl and Dunn, in 1984, argued that the "equivalency
factors" in the military and commercial specifications were not valid (see Evaluation of Test
Equipment for the Detection of Contamination on Electronic Circuits, European Space Agency
Technical Memorandum ESA STM-234, Paris (1984).). A 1992-93 round robin study of ionic
cleanliness testers was performed by an IPC task group at the Electronics Manufacturing
Productivity Facility (EMPF), then located in Indianapolis, Indiana. That study was documented
in IPC-TR-583, and largely agreed with Bergendahl and Dunn, indicating that the equivalency
factors were not valid and the ionic cleanliness testers suffered from a lack of repeatability and
reproducibility. That study concluded that these kinds of ionic cleanliness testers could be used
as process control tools, but not for product acceptance testing. Unfortunately, because the
industry could not provide scientifically sound alternatives, these values have continued to be

Page 28 of 38
used, flawed as they are. IPC technical committees are working on providing better tools and
guidance on this matter.

It is important for the reader to understand that all of this developmental and validation work was
done for ionic cleanliness testers used to examine the cleanliness of printed board assemblies, not
bare printed boards. The residues found in bare printed boards are different from those found on
printed board assemblies.

7.1.1 Limitations The ionic cleanliness testers use an extraction-based test. The idea is to
dissolve the contaminants on a substrate and bring them into the extract solution so that you can
examine them. An ionic cleanliness tester examines the solution conductivity. In theory, as
more ionic material, such as flux residue, is dissolved into solution, the conductivity grows and
registers on the meter. Because one can only see the results of the meter, it is critical to
remember the following limitations. All extractions are a time-temperature-solubility equation.
1. If the contaminant will not dissolve in either isopropyl alcohol or water, then it does not
get into the solution and the meter does not measure it. If the residues of interest are not
soluble in the extract solution, then the test results may be erroneously interpreted as
clean. Many modern flux residues are based on synthetic resins and are not soluble in
isopropanol and water, but the residues for these fluxes may not be detrimental.
2. If the contaminant dissolves slowly, then one may only see a fraction of the contaminant
present in the typical 10 minute ionic cleanliness test. If the conductivity values continue
to rise following the 10 minute minimum test time, then the test should be run until the
conductivity reading stabilizes. However, caution must be exercised as prolonged
exposure to the test solution can encourage leaching effects that might otherwise be
passive.
a. If the conductivity continues to rise with time, this may indicate a problem with
solder mask cure.
b. If the test runs for very long times (hours), the exposure to the alcohol/water
solution will begin to chemically attack the bare board, freeing ionic material.
3. If a contaminant does not dissolve at the temperature of the extract solution, then the
meter will not detect the contaminant.
4. If the extract solution cannot get to the contaminant, such as for very low standoff parts,
then the residues cannot be moved into the extract solution and so cannot be detected.
5. Many materials can cause an increase in solution conductivity, but are not detrimental to
printed board reliability. Carbon dioxide from the air is a good example.
6. If a contaminant is not electrically conductive in solution, it cannot be measured by this
test.
7. It is important to note that calibration of a machine (ROSE tester) takes into account the
Specific Gravity of the test solution at a defined Temperature at a known Conductivity.
When one considers that the present acceptance values for both manufactured assemblies and
printed boards were developed from high solids rosin fluxes with solvent cleaning, and that the
chemical nature of modern fluxes is dramatically different from that of fluxes of the past, it
becomes readily apparent that the use of ROSE for product acceptance is not a valid practice,
even though it is commonly done in the industry. Many of todays fluxes are based on synthetic
resins and have no solubility in isopropanol or water. In addition, the chemical residues from

Page 29 of 38
printed board fabrication are also dramatically different from those found in printed board
assembly operations and so ROSE values for printed board acceptance are likewise not
recommended.

This begs the question of why the instruments are still used to the extent that they are. There are
several reasons for this:
1. ROSE testing is still a valuable process control tool. If ones result was 5 last week and 5
yesterday, and today, the result is 10, then something in the process changed and
investigations must be made.
2. This test is often contractually required, as are the acceptance values. Because of this,
almost every OEM and every printed board fabricators has one such instrument in house.
3. ROSE testing is a frequently misunderstood / misapplied test, or at least not well
understood, and the industry inertia in this area is considerable.
4. The ROSE test, even as flawed as it is, is better than no cleanliness data at all. Most
other forms of ionic residue detection, such as ion chromatography, are not as amenable
to in-line process control and are considerably more expensive.
An example of #4 is that in one manufacturing process, based on saponified in-line aqueous
cleaning, the ROSE values (contractually required) jumped dramatically during the course of an
operating shift. It was found that the DI water rinsing in the cleaner had stopped (due to pump
failure) and as such, the conductive saponifier chemical was detected by ROSE. The problem
was corrected and all suspect hardware was re-cleaned and ROSE values returned to baseline.
An important consideration for #3 is that ROSE is still a valuable tool, but the numbers have to
be understood: What does a value of 8 g/in
2
NaCl equivalence mean for the product being
manufactured? A recommended approach is to use a better method of ionic analysis, such as ion
chromatography, to determine acceptable or unacceptable levels of ionic residues for the product,
while also doing ROSE testing to set process control limits.
It should be noted that no cleanliness standards exist for components, ROSE or otherwise.
7.1.2 ROSE and Printed Boards Reference [2] is an article on bare printed board cleanliness
requirements in IPC and military printed board specifications, circa 2000. All of the
specifications of that time recognized several fundamental cleanliness considerations:

Cleanliness of printed boards is important, even if the industry cannot all agree on how it
should be measured or what pass-fail levels are needed. Failure to pay attention to cleanliness
considerations will have a negative impact.
A surface needs to be clean before one puts down solder mask or before layers are laminated.
If it is not clean before these steps, there will be no opportunity to go back and remedy the
problem later.
Printed boards need to be clean for delivery to the assembler. Process fluids encountered
after solder mask application, such as fluxes and fusing fluids, need to be adequately
removed/reduced. Because different end-item applications will have differing printed board
cleanliness needs, it is left to the assembler/user to state in their procurement document how
clean the printed boards need to be and by what measure cleanliness is determined.

Page 30 of 38

At that time, the only ionic cleanliness test required on printed boards was the ROSE test prior to
solder mask application. The required metric was the 10.06 g/in
2
NaCl equivalence (1.56
g/cm
2
NaCl equivalence) found in printed board assembly levels specifications. This meant
that an OEM could get a printed board with acceptable levels of cleanliness per printed board
standards, but which left the OEM with no margin for printed board assembly processes. This
was true for most or all printed board standards for the military at the time.

Many companies developed their own printed board cleanliness standards based on ROSE
testing, some scientifically based and others with no discernible scientific base. One OEM, for
instance, was known to require boards to have 0.75 g/cm
2
NaCl equivalence because they
wanted boards twice as clean as the standards allowed.

The danger in using a ROSE pass/fail level of 1.56 ug/cm2 sodium chloride equivalence leads to
a false sense of security in many cases. A "pass-fail" metric leads one to believe that ionic test
results up to 1.56 ug/cm2 NaCl equivalence leads to "safe or acceptable" levels of ionic residues,
and values above 1.56 are "bad or unacceptable". Modern assemblies can be very sensitive to
ionic residues and so this historic pass-fail level is often inadequate or unreliable. Assemblers
often have to meet the same 1.56 metric, and so having a bare board at the 1.56 level leaves them
no margin. Modern fabricators usually work to levels down around 0.2 ug/cm2 NaCl
equivalence.

7.1.2.1 IPC-6012 The predominant bare printed board specification for rigid printed boards in
use today is IPC-6012.

1. This specification requires the use of ROSE testing at the time of publication of this
document. Alternative methods are allowed, but the user has to demonstrate equal or
better sensitivity with the ability to dissolve flux residues or other contaminants
dissolvable by isopropanol and water.
2. Cleanliness testing by ROSE is required prior to solder mask application with an
acceptance value of 1.56 g/cm
2
NaCl equivalence .
3. Cleanliness testing by ROSE on inner layers, prior to stack up and lamination, is an as
agreed between user and supplier (AABUS) item. This refers to both method and
acceptance criteria.
4. Cleanliness testing by ROSE on finished printed boards is done when specified in the
procurement document. The procurement document must also call out the methodology
and the acceptance values.
5. Any special requirements are AABUS items and must be in the procurement document.

It is common for procurement documents to call out the same ionic cleanliness acceptance values
(1.56 g/cm
2
NaCl equivalence) on finished printed boards, but the wording of the specification
is such that more modern approaches can still be used.

7.1.2.2 Cautions on ROSE and Bare Printed Board Cleanliness


Page 31 of 38
Keeping in mind the limitations of ROSE testing, and the different chemical residues in
printed board fabrication, it is possible that a board may meet all IPC-6012 cleanliness
criteria, yet still contain high levels of residues harmful at the OEM level.
Consumer-based detergents, commonly used for dishware and glassware (e.g. Cascade, J oy)
often contain high levels of organic surfactants which may not be detected by ROSE testing,
but can affect adhesion of materials at the OEM level. Organic residues, not detectable by
ROSE, often need alternative methods of analysis, such as Fourier Transform Infrared
(FTIR) spectroscopy, water drop testing (contact angle) or simple smell (lemon/orange
scent).
In production, ROSE tests are seldom run longer than contractually required as the
conductivity levels can continue to rise over time for a variety of reasons, such as carbon
dioxide from the air mixing in or as materials are leached from the board. If the ROSE
values continually rise after the recommended test time, then other test should be run to
determine whether it is a solder mask cure issue. The shortest possible test time is preferred
and there are test systems that are capable of shorter test times by the use of "curve-fitting
algorithms.
One method of differentiating rising levels due to CO
2
and those due to another source is to
run the ROSE tester with no sample, mimicking a test. This should show how much of the
increase is due to CO
2
absorption over the course of the test. This is especially true for
ROSE testers which incorporate agitated spray in the method, which increases the amount of
mixing from the air.
If the CO
2
contribution is shown to be a minimal contributor to a continuing rise in ionic
levels in a ROSE test, then the next likely ionic source is improperly cured solder mask, as
discussed previously. A steadily rising ROSE result can be either ionic material coming
from the solder mask itself, or process fluids that were absorbed into the mask and
consequently leached out in the extraction process.

7.2 Ion Chromatography If the industry cannot rely on ROSE testing to yield accurate
information on ionic cleanliness, what methodology and criteria should be used?
A more modern and more precise method of measuring ionic cleanliness of printed assembly
boards and materials is the use of Ion Chromatography (IC). The IPC Test method for ion
chromatography is IPC-TM-650, Method 2.3.28. A variation on that test method, which uses a
different extract solvent, can be found in IPC-TM-650, Method 2.3.28.2.
Both of these methods rely on a longer duration (10 min vs. 60 min) , higher temperature (room
temp vs. 60 C [140 F] vs. 80 C [176 F]) extraction than does the ROSE testing. IC also
quantifies what ionic species are present and in what amounts.
IPC-WP-008 is a white paper on what is involved in setting up a laboratory for ion
chromatography analyses.
7.2.1 Why IC data is better and Preferred The results from total ionic methods (e.g. ROSE) do
not directly correlate to the individual ionic contamination on the printed board, which can result
in misleading information. IC data is preferred and better because it provides speciation
information that is useful in identifying contamination reactivity, thereby making it easier to
determine whether the contamination will cause failures. Speciation provides specific compound

Page 32 of 38
information needed during failure analysis investigations. IC is also more sensitive than other
methods so it can be used to quantify and identify contaminants at the failure site on the printed
board or for small spot analysis.

7.2.2 Why the Industry is moving this Way The industry is expected to move this way because
the printed board industry increasingly uses solder fluxes with organic acids and esters which are
not easily identifiable by total ionic methods, which more readily identify strongly ionic
compounds. The printed boards are more susceptible to failures at lower contaminant
concentrations because the devices have lower clearances and can easily trap contamination, and
the boards are more heavily populated.

IC instrumentation has made many recent technology advancements:
Suppressed conductivity which lowers baseline noise and provides higher signal-to-noise,
which increases sensitivity.
Inline eluent generation is another technology advancement that generates precise and
accurate eluent inline, eliminating separation effects from variations or errors in eluent
preparation. Inline eluent generation and new column technologies allow the use of
hydroxide eluents which result in faster separations of strongly ionic compounds and greater
separations of similar weakly ionized compounds, such as organic acids, that are not easily
separated by carbonate-based eluents. These improved separations result in more accurate
quantifications and increase reliability and accuracy of data reporting.
In capillary-based IC systems, the sample injection volume requirement can be as small as
0.4 uL. This may lend itself well to localized contaminant investigations (i.e. gathering a
drop from the surface of a circuit board). This allows for the injection and quantification of
extremely small sample volumes.
There is a wide array of column resin chemistries available, allowing for custom separation
of unique/niche blends of ionic material elements.

IC is more desirable as a tool because modern assemblies are much denser and have much higher
sensitivities to residues than did assemblies from the 1990s. Bulk ionic cleanliness testing is not
sensitive enough to detect the subtle differences in ionic residues that spell the difference
between acceptable and unacceptable circuit performance. Some ions are very harmful (e.g.
chloride) and some are benign (e.g. nitrate).

7.2.3 When to Test There are three critical points where ionic cleanliness should be measured,
very similar to those discussed in the ROSE section:

1. Before mass lamination. If inner layers are not clean prior to lamination, the ionic material
becomes entrapped in the inner layers and cannot be removed.
2. Prior to solder mask application. Solder mask is a polymeric coating that seals off the
surface of most of the printed board. If the printed board is not clean prior to solder mask
application, there is no way to clean the residues without stripping the mask. Mask stripping
is usually a very chemically aggressive process and it is seldom cost effective to do.
3. Final board cleanliness prior to packaging. The final plating operations, as outlined before,
can leave a variety of residues on the printed board surface. These are the residues that most
ROSE or IC tests will focus upon. It should be noted that cleanliness as determined by ion

Page 33 of 38
chromatography would only appear in an IPC-6012 order if it was written into the purchase
contract.

7.3 IPC-5704 Realizing that ROSE was not an adequate measure of ionic cleanliness for modern
designs, the IPC 5-32c Bare Board Cleanliness Assessment Task Group worked to develop new
standards for ionic cleanliness as determined by ion chromatography. The task group could have
either done massive amounts of round robin testing, or adopt a commercial specification with a
proven track history. The latter was pursued.

Delphi Electronics, a high volume automotive electronics OEM headquartered in Kokomo,
Indiana, volunteered their Q-7000 printed board specification as a starting point. This
specification is considered as public domain and can be obtained from Delphi Electronics.

In this Q7000 specification, when a printed board fabricator is qualified, the ionic residues are
determined using IC and acceptance based on the levels of the ionic species determined. At the
same time, ROSE testing is performed to develop process control criteria for the fabricator.

Delphi implemented this approach in the mid 1990s with good success. Using this method,
process problems rooted in printed board residues were dramatically reduced or eliminated.
OEMs that reported using the Delphi criteria saw similar reduction in printed board cleanliness
related issues. Independent test labs, while each having their own recommended cleanliness
levels for IC data, generally agree with these published values.

The IC method used by Delphi was slightly different than the IPC standard IC method (IPC-TM-
650, Method 2.3.28). The difference was the composition of the extract solution, where the
Delphi method used 10% isopropanol and 90% DI water, while the IPC method used 75%
isopropanol / 25% DI water. The Delphi method was written up as a standard IPC test method
and is now IPC-TM-650, Method 2.3.28.2. It is recognized that the two different methods yield
different results. The IPC committees are currently working to develop a correlation between the
two methods.

The result of the Delphi/IPC effort resulted in the publication of IPC-5704 in December 2009.

When a printed board is extracted per IPC-TM-650, Method 2.3.28.2, IPC-5704 requires the
following:

From IPC-5704
4.1.1 Analysis of Extract Solution For all bare PBs that were tested per 4.1, three
milliliters of extract solution shall be reserved for IC analysis. The remaining extract
solution shall be tested using the modified Resistivity of Solvent Extract (ROSE) method
of IPC-TM-650, Method 2.3.25.1.

Table 4-1 Bare PB Ionic Contamination Maximum Limits
Ions Non-OSP OSP
Chloride (Cl) 0.75 g/cm
2
0.75 g/cm
2
Bromide (Br) 1.0 1.0

Page 34 of 38
Sodium (Na) + Potassium (K) 2.0 4.0
Total Inorganic 3.8 5.9

The units in Table 4-1 are in micrograms (g) of the ion per square centimeter (cm
2
)
based on the extraction volume and the calculated sample surface area.

For additional information, the IPC has recordings of a webinar on the IPC-5704 specification,
presented in J une 2011.

It is important to remember to match the IPC-5704 criteria to IPC-TM-650, Method 2.3.28.2
generating the numbers and not to numbers generated with IPC-TM-650, Method 2.3.28. It is
also important to note that the units used here are different than those in ROSE testing.

The modified ROSE test, outlined in IPC-TM-650, Method 2.3.25.1, referenced in both the
Delphi standard and the IPC-5704 specification, was also developed by Delphi and adopted by
the IPC. Essentially, this modified-ROSE method combines the longer term, higher temperature
extraction method of IC and the use of the ionic cleanliness tester as the meter. The extract
solution, after reserving a small amount for IC analysis, is poured into the ionic cleanliness tester
sample cell and a short test is run. The use of modified ROSE can also be used as a process
control metric that is an improvement on the standard ROSE method.

7.3.1 Why These Ions? In most published failure analysis studies on electrochemical failure
mechanisms, certain ions are often the common root cause of the failure. Chloride and bromide
are frequently cited as problem ions. Consequently, these two ions have maximum allowed
amounts. Research by Delphi Electronics, supplemented by the findings of other test labs,
showed that when printed boards had high amounts of sodium or potassium, it often meant there
was a problem with solder mask cure and development. Sodium and potassium are common
cations found in solder mask processing. Consequently, there is a maximum allowed amount
for combined sodium and potassium. The final requirement is for total inorganics. IC can often
find additional ionic species, such as methane sulfonic acid, nitrates, phosphates, etc. If a printed
board meets the requirements for chloride, bromide, and combined sodium and potassium, yet
has high amounts of other ionic material, then it indicates a cleanliness condition that is suspect
and may indicate a fabrication operation that is not in control. Table 7-2 shows the most
common anions and cations included in the Inorganic classification.

Table 7-2 Inorganic Ions
Cations Anions
Sodium (Na) Fluoride (F)
Potassium (K) Chloride (Cl)
Lithium (Li) Bromide (Br)
Ammonium (NH
4
) Nitrite (NO
2
)
Calcium (Ca) Nitrate (NO
3
)
Magnesium (Mg) Sulfate (SO
4
)
Phosphate (PO
4
)


Page 35 of 38
At the time of publication of this document, the IPC-5704 specification only differentiates
between a printed board with an organic solderability preservative (OSP) and that with a non-
OSP finish, such as immersion tin or ENIG. Non-OSP also covers HASL final finishes. The
requirements for chloride and bromide are the same between these two classes of finish. The
values for sodium and potassium are different for an OSP coated printed board assembly because
this surface finish is naturally higher in these ions. For this reason, the total allowable inorganic
amount is higher for OSP coated printed board assemblies.

7.3.2 IPC-5704 Samples

From IPC-5704:
3 TEST SPECIMENS

The test sample shall be a standard qualification printed board and/or test coupon, or a bare
production printed board having a defined material set (laminate, solder mask, surface
metallization) and fabrication process.

The specification invoking the use of this document shall define the sample size. If this document
is used as a stand-alone document, the sample size shall be as agreed upon between user and
supplier (AABUS).

If the sample size is not provided, a recommended sample size is 11, however the sample size
should be in accordance with the sampling plan defined in IPC-6012, taking into account the lot
size and statistical significance required.

This indicates that the test sample can be a production printed board or quality conformance
coupon (better than using entire panels).
Sample size is always a hotly debated topic, which is why this specification does not require
a mandatory sample size that may conflict with a sample size from the dominant
specification (e.g. IPC-6012).
The suggested sample size of 11 comes from Delphi and represents a sound starting sample
size, if one is not defined elsewhere (e.g. IPC-6012). In practice, the sample size is adjusted
on a case by case basis, depending on the history with the supplier.
The frequency of testing is another hotly debated point. For a supplier that demonstrates
good process control, the requirements may be less than that for a vendor with no history, or
a history of questionable process control.

7.3.3 5704 Flexibility

FROM IPC-5704:
4.1.4 Qualification by Similarity Bare printed boards having the same combination of materials
and processes can be qualified by similarity to bare printed boards that have been tested and
qualified in accordance with the requirements of this document. Bare printed boards with the
same laminate, same solder mask, same surface finish, same manufacturing process, same
manufacturing line and same manufacturing facility, could be qualified by similarity, as an
example. Deviations or exceptions to the qualification by similarity process shall be AABUS.


Page 36 of 38
4.2 Ionic Cleanliness Testing for Process Control The excess extract solution as noted in 4.1.1
from the bare printed board extraction process shall be tested per IPC-TM-650, Method 2.3.25.1
(modified ROSE method). This data shall be used to set process control parameters.

IC shall be used to disposition the non-conformance when an out of control point is determined.

Section 4.1.4 of IPC-5704 was added to allow some latitude for similarity in materials,
however:

Many of the OEMs have seen dramatic differences in cleanliness between line A and line
B, in the same building, or Building 1 vs. Building 2 in the same fabrication site.
Consequently, a fabricator can expect resistance from OEMs when trying to apply a
single qualification to multiple sites.

7.3.4 Acceptance of IPC-5704 At the time of publication of this standard, the methods and
requirements found in IPC-5704 have not been adopted in major bare printed board
specifications such as IPC-6012. However, those committees have been looking for better ionic
cleanliness criteria for many years and the eventual adoption of these values is expected.

7.4 SIR Testing ROSE, modified-ROSE and IC are all chemical analytical methods designed to
determine what chemical ions are present and in what amounts. The question often remains on
what effect those ionic materials have on electrical performance and how to best demonstrate
acceptable or unacceptable electrical performance.

Once such method is the use of SIR testing. From the standpoint of printed board cleanliness,
fabricators who have supplied boards to the telecommunications industry have probably been
required to meet SIR criteria developed by Bell Telecommunications Research (Bellcore), found
in Bellcore specification GR-78-CORE (Chapter 13). This periodic SIR testing of inner layers
was done prior to lamination, prior to solder mask application, and on outer layers of finished
boards. In such testing, a number of standardized patterns (e.g. Bellcore pattern) are tested at 35
C [95 F]/90% relative humidity for 24 hours and must meet certain very high minimums. If
the test patterns are not ionically clean, the high values required cannot be attained.

The test method can also be found in IPC-TM-650, Method 2.6.14.1.

In terms of the time frame, this SIR test requires 24 hours of time, a controlled temperature-
humidity chamber, and suitable apparatus to measure resistances in the tera-ohms. The test also
allows that if the SIR values fall below the minimums, the test can be extended to 96 hours for a
retrial. If the SIR minimums are below the specification values as the second measurement, then
the result is a failed test.

7.5 Other Tests

7.5.1 Other Tests - Water Break Testing There are additional tests that are sometimes used to
check the cleanliness condition of printed boards. One such test is a water break test and can
be found as ASTM F21 and ASTM F22. Distilled or DI water is sprayed onto a cleaned surface.
If the surface is clean, and generally free of oils or greases, the water will wet a surface evenly

Page 37 of 38
in a sheeting action. If the surface is not clean, then the water will bead up. This is generally a
quick and easy test to run prior to plating operations.

7.5.2 Other Tests Surface Energy Alternatively, the surface cleanliness or surface energy can
be characterized by measuring the contact angle of a drop of DI water placed onto the surface in
question. Devices called goniometers are used to make such a deposition and to measure the
shape of the droplet in terms of its contact angles. Low contact angles, where the water drop
wets out along the surface, generally represents a clean surface. A high contact angle, where the
water drop forms beads, may indicate an unclean surface. It should be noted that the surface
material plays a major role in this approach. A fluorinated surface (e.g. PTFE) will bead up no
matter what its cleanliness condition is.

Surface energy measurements can often show if there is a problem with the solder mask cure.
From an OEM perspective, printed board surfaces with low surface energy will experience
problems with adhesion for adhesives and for conformal coatings. Surface energy can be
quickly measured using devices called dyne pens.

The following general rules of thumb for surface energy may be helpful:

Under 30 dynes/cm problems are likely for adhesion and surface wetting
30-35 dynes/cm intermittent problems, but more frequently
35-40 dynes/cm intermittent problems, but more rarely
Above 40 dynes/cm generally good results with rare exceptions.

Such measurements can often show the effects of residual surfactants which do not show up in
ROSE or IC testing.

7.5.3 Other Tests - Water Drop Testing (Comparative Tracking Index) Water drop testing is
a way of observing qualitative surface insulation resistance without having to purchase expensive
equipment. A voltage is applied across a "Y" pattern, or across any pads or conductors that are in
the area of interest. A drop of water is placed on the printed board so that it bridges the two
conductors. The water drop is viewed under an optical microscope, and the time it takes for a
short to form between the conductors is measured. The greater the ionic cleanliness, the longer it
will take for a dendrite to bridge the conductors. This is a very old method and is seldom used
anymore.

8 REFERENCES

The following reference documents are available upon request through answers@ipc.org:

1. Ionic Cleanliness in IPC and Military Bare Board Standards IPC Review, October
2000
2. What is Considered Clean? - IPC Review, November 2000.

9 ACKNOWLEDGEMENTS

Page 38 of 38
A special recognition goes to Mr. Dale Murry and Mr. David Sullivan of Collins Printed
Circuits, Cedar Rapids, Iowa. Their combined 50+years of fabrication knowledge, and what can
go wrong in a fabrication process, was the foundation of this document.

Page 1
IPC Compilation of Comments
IPC-5703 Final Draft for Industry Review August 2012

No. Commentor Paragraph/
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Recommendation Reason for Recommendation IPC Committee
Disposition
on each comment
submitted
(Accept/Not Accept/Withdrawn/
OBE (Overcome by Events)
1 Todd
MacFadden,
Bose
Figure 1-1 T I suggest photos showing dendrites growing on actual
PCBs, rather than on capacitors. Here are a couple
examples:
Better relevance. Accepted. Will replace
1-1 with these two
photos.
2 M. Paddack
Boeing
#1.3 E Delete reference to paragraphs 1.3.1 thru 1.3.3: The
definition of all terms used herein shall be in
accordance with
IPC-T-50 and as defined in 1.3.1 through 1.3.3.
Definitions in already define in IPC-T-50. Accepted. Reference
made to IPC-T-50
3 M. Paddack
Boeing
1.3.1 thru 1.3.3 E Delete paragraphs 1.3.1, 1.3.2, 1.3.3 Definitions are already defined in IPC-T-50 Accepted. Reference
made to IPC-T-50
2
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
4 J ohn
Rohlfing,
Delphi
Section 1.2, 4
th

paragraph
E Change . . . depends on a wide number of factors:, to
. . . depends on a number of factors:
Incorrect usage (numbers are either large or small
not wide)
Accepted.
5 J ohn
Rohlfing,
Delphi
Section 1.2, 4
th

paragraph, 1
st

bullet
E Change e.g. to i.e. Incorrect usage (e.g., meaning for example,
would imply that something more than just smaller
lines and spaces is directly resulting a higher
voltage gradient; i.e., meaning in other words,
more adequately describes the relationship
between smaller lines and spaces and higher
voltage gradient)
Accepted.
6 M. Paddack
Boeing
2.1 E Delete IPC-TR-583 Document is not referenced anywhere within IPC-
5703
Accepted.
7 M. Paddack
Boeing
3, 1
st
paragraph T Rewrite Water quality varies season to season.
Water quality may vary by season or water source.
If the water is from a river it may very well change
from season to season but if from a well it might
not. Likewise we see water quality change
depending if water is from the aquaduct, ground
wells and even which holding tank the water is
stored in.
Accepted.
8 M. Paddack
Boeing
3, 2
nd
paragraph,
6
th
bullet
E Delete A thorough cleaning process, incorporating
cleaning agents and rinsing, is better than a mild or
low cost cleaning process.
This sounds like motherhood. Should remove of
be more specific.
Not Accepted.
This was specifically
included to indicate low
cost is not the way to go.
9 J ohn
Rohlfing,
Delphi
Section 3, 3
rd

paragraph
E Change . . . cost per liter . . . to . . . cost per unit
volume . . .
While everyone pays for their water by some
measure of volume, that unit is not necessarily
expressed in liters
Accepted.
10 M. Paddack
Boeing
3, 3
rd
paragraph,
1
st
bullet
E Revise: Unfiltered tap water: to Tap water: Simple is better. Not Accepted.
There is a difference
between the two. Some
places filter tap water,
but only for particulates,
nothing else.
11 M. Paddack
Boeing
3, 3
rd
paragraph,
2
nd
bullet
E Revise: Tap water that has been pH altered with acids
or alkaline materials: to Tap water that has been pH
altered:
Too wordy of a title Accepted.
12 M. Paddack
Boeing
3, 3
rd
paragraph,
3
rd
bullet
T Question on this paragraph. Filtered tap water: Water can be filtered without being carbon treating.
Is the intent to just discuss carbon treating and not
filtration.
Not Accepted.
Present wording talks
about carbon filtration.
13 M. Paddack
Boeing
3, 3
rd
paragraph,
4
th
bullet
T Question on this paragraph. Softened water: Isnt
water often passed through a water softner when
carbon treated?
Seems like the paragraphs are out of order. Accept with modification.
Added information on
water softening.
3
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
14 Todd
MacFadden,
Bose
3, 3
rd
paragraph,
last bullet
E Suggest add phrase in italics:

DI water is the most desirable of water grades for
purposes of cleaning.
For clarity and precision of intent. Accepted
15 M. Paddack
Boeing
3, 4
th
paragraph E & T Revise: An industrial aqueous cleaning operation will
typically have tap water plumbed first into a particulate
trap and then into a carbon filtration column. The
output of this column will feed into a series of
deionizing columns which make the water
progressively cleaner. In-line resistivity
monitors can show either a numeric readout of the
water resistivity or a simple red light / green light
arrangement (aka dummy lights) if the resistivity falls
below a set value, usually 2 megohm-cm. Most often,
these lights are located at the DI source. It is highly
recommend that
these lights monitoring also be incorporated at the
destination since as well as DI water quality can
degrade over long piping runs.
Improve text.

We dont care where the lights are located. What
is important is where the monitoring is done.

By-the-way, this is always done with in-line
monitors. May time the monitoring is done by
sampling. The sampling method works quite well.
Accepted with
Modification
16 J ohn
Rohlfing,
Delphi
Section 3, end of
6
th
paragraph
E/T . . . copper, unpolished stainless steel, garden hoses,
or Tygon tubing (bacteria, green fuzzy stuff)
Not clear what the writer is intending to say:
- Is it that Tygon tubing is itself bacteria and/or
green fuzzy stuff?
- Or that Tygon tubing encourages bacterial growth
that manifests as green fuzzy stuff?
- Or that copper, unpolished stainless steel, garden
hoses and Tygon tubing all may encourage
bacterial growth that manifests as green fuzzy
stuff?
Accepted with
Modification
17 M. Paddack
Boeing
3, 6
th
paragraph E Remove paragraph: Highly polished deionized water
green fuzzy stuff).
Earlier we said not to use the highly polished DI
water then here we put in a large paragraph
discussing it. If we arent using it then this
paraghaph is not relavent and should be removed.
OBE see disposition to
comment #16
18 Todd
MacFadden,
Bose
Section 3, 7
th

paragraph
E and since silica is not ionic, deionizing columns will
not remove it.

Is there any significance/consequence to this? Can
silica pose a reliability risk or interfere with PCB
fabrication? I dont suggest removing this, but should
at least indicate whether its something to be
concerned with.
Clarification Accepted with
Modification.
4
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
19 Todd
MacFadden,
Bose
Section 3, 8
th

paragraph
E Suggest move the paragraph that starts,

A RO and DI water setup should incorporate
ultraviolet (UV) irradiation

so that it follows the sentence about green fuzzy stuff.

Does UV irradiation allow use of copper, unpolished
stainless steel, garden hoses, or Tygon tubing.?
Clarification Accepted.
Paragraph moved up.
20 Todd
MacFadden,
Bose
Section 3, 8
th

paragraph
E Clarify:

An RO/DI system involves membrane
technology which requires active maintenance.

Why? What is meant by active maintenance? Is the
implication that membrane systems are inherently
vulnerable to bacteria contamination?
Clarification Accepted.
Clarified.
21 M. Paddack
Boeing
3.1 T Add: Some cleaning agents can be corrosive and may
pose a risk to some board finishes if not completely
removed during the rinsing process. Care must be
taken in selecting the cleaning agents for the particular
application.
Accepted.
22 M. Paddack
Boeing
4 T Delete: The newer systems have numerous self-
checks built in.
What systes are we referring to and is this actually
true? I dont think so.
Accepted with
Modification
23 Todd
MacFadden,
Bose
Section 4 E Suggest add a few other items for PM:

Pumps in cascade rinse systems (to ensure proper
and adequate flow rates), periodically calibrate DI
resistivity monitors
Additional info needed Accepted.
Included.
24 M. Paddack
Boeing
5 T Delete: and components in 1
st
sentence This guideline is about PWBs not components. Accepted.
25 M. Paddack
Boeing
5 T 3
rd
Paragraph add: On bare copper and on laminate
surfaces between copper conductors, they reduce
insulation resistance. Fingerprints can also reduce
solderability. Fingerprints on component
We have lost sight that contamination of
termination surfaces can affect solderability and in
term affect reliability. I think this issue needs to be
expanded on within this document in several
places.
Accepted with
Modification.
26 J ohn
Rohlfing,
Delphi
Section 5, 1
st

paragraph, 3
rd

bullet
E Change Assorted dust and dirt to Dust and dirt Extraneous word (while assorted does lend a
touch of linguistic color, it doesnt add to the
meaning; why add assorted to dirt and dust, but
not to the equally deserving food particles, hand
creams and cosmetics, lint or debris from
gloves, or bodily fluids?)
Accepted.
5
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
27 J ohn
Rohlfing,
Delphi
Section 5, 1
st

paragraph, after
bullets
E Change Such contamination may be water-soluble,
solvent soluble, or insoluble in either., to Such
contamination may be water-soluble, solvent-soluble,
or insoluble in either water or solvent.
Consistency (adding hyphen) and clarity (adding
water or solvent at the end of the sentence)
Accepted.
28 M. Paddack
Boeing
6.1 T Paragraph 3 need to expand on or remove. Some
resin systems contain brominated
It there a point to laminates with brominated fire
retardants? Do they cause corrosion, dendrite
groth, affect ROSE results? Shuld we recommend
not using brominated fire retardants. In other
workds what is the guidance?
Accepted with
Modification.
29 M. Paddack
Boeing
6.1 E Paragraph 3 Separate the additives from the absorbed
chemicals.
Mixing subjects here. Suggest moving the
additives into paragraph 2 or at least separate the
two sources of chemicals into their own paragraph.
Accepted. Moved
paragraph up and
clarified.
30 M. Paddack
Boeing
6.1 E Paragraph 4 Souldnt this paragraph be combined with
paragraph 2. Extracted items can
So what if the materials are extracted. Does that
casue an issue?
Accepted with
Modification.
Added clarification.
31 M. Paddack
Boeing
6.1 E Paragraph 5 Expand What is the guidance here? Is there some testing
needed? How do different laminate compare for
absorption?
Accepted.
Added clarification.
32 J ohn
Rohlfing,
Delphi
Section 6.2, 1
st

paragraph, 5
th

bullet
E Add Air Handling Systems at start of bullet Consistency (this is a list of residue sources in the
storage environment and most of the bullet items
start with the item itself, followed by a long hyphen)
Accepted.
33 J ohn
Rohlfing,
Delphi

Section 6.2, 1
st

paragraph, 9
th

bullet
E Change Fibers from packing materials, dust from
packing peanuts, and threads from clothing . . ., to
Fibers and dust from packing materials and threads
from clothing . . .
Colloquialism: packing peanuts (may not
translate well); packing materials produce both
fibers and dust as contaminants, so they can be
combined when the colloquialism is removed.
Accepted.
34 J ohn
Rohlfing,
Delphi
Section 6.2, 1
st

paragraph, 11
th

bullet
E Add Conveyors at start of bullet Consistency (this is a list of residue sources in the
Accepted.
35 J ohn
Rohlfing,
Delphi
Section 6.2, 1
st

paragraph, 12
th

bullet
E Add Separation Sheets at start of bullet Consistency (this is a list of residue sources in the
Accepted.
36 M. Paddack
Boeing
6.4 T 1st bullet expanded on or removed Photoresist is a
coating of photo sensitive material
Is there a point to this bullet? So what if a
photoresist is used? Probably should make this
the fist sentence of 6.4.
Accepted for deletion.
37 M. Paddack
Boeing
6.4 T 2
nd
bullet Cleaning can be What is being cleaned? Accepted with
Modification.
Possibly make this
6.3.1.
38 M. Paddack
Boeing
6.4 E Combine third bullet with the 2
nd
bullet. Keep the chemical cleaning bullets together. Accepted.
6
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
39 Todd
MacFadden,
Bose
6.4, 5
th
bullet T Suggest to remove:

However, in operations such as tab cleaning, where
phosphoric acid is used, the minerals in tap water can
be beneficial in neutralizing the acid.

This conflicts with aforementioned recommendation to
use DI water throughout, and may or may not be true
anyway depending on the content of the city water
which is highly variable in certain regions of the world.
Sentence is more confusing that value-adding.
Accepted.
40 M. Paddack
Boeing
6.4 E 5
th
bullet Need to clarify If the water used for cleaning or rinsing? Not clear
to me.
OBE
See disposition to
comment #39
41 M. Paddack
Boeing
6.5 E Move 1
st
bullet in first paragraph of 6.5 This bullet is not about contamination but is part of
the process description.
Accepted with
modification.
42 M. Paddack
Boeing
6.5 E Move 1
st
bullet of the second grouping to the beginning
of the group introduction: In the developing process,
the unexposed portion of the resist is chemically
removed. For the developing process, there are a
number of different chemistries used, depending on
what resist is being used.
Seems like this is describing the process rather
than define a contamination source.
Accepted with
modification.
43 M. Paddack
Boeing
6.5 E 2
nd
bullet of 2
nd
grouping, remove A common
developing solution
So what if tap water is used. What is the
contamination isse? If there isnt one the bullet
should be removed.
Accepted with
modification.
44 M. Paddack
Boeing
6.5 T 3
rd
bullet of 2
nd
grouping, expand In developer
seldom works well.
What is meant by seldom works well? Is there
some cross contamination issue or something?
Need to elaborate.
Accepted with
modification.
45 M. Paddack
Boeing
6.6 E Move 1st bullet to intoroduction paragraph. The
copper protected ..
This is part of the process definition not a
contamination source.
Accepted.
46 M. Paddack
Boeing
6.6 E Combine bullet 2 and 3 This etch process is often
done with cupric chloride etchant and hydrochloric
acid, usually with a pH of around 1, and often with a
sodium chlorate oxidizer. It is important that these
etching chemicals be fully neutralized or they continue
to attack the
copper.
The two bullets are one related subject. Also isnt
there an issue with the acid being absorbed and
later cause isses?
Accepted.
47 M. Paddack
Boeing
6.6 T Expand on or remove 5
th
bullet In general, etching
inner layers
So what if the resist is thicker than on the outer
layers. Is there a contamination issue here? If so,
what is it? If not, remove the bullet.
Accepted incorporate
new Table 6-1.
48 M. Paddack
Boeing
6.6 T Expand or remove the 6
th
bullet: Inner layers of a
printed microetch.
What is the contamination issue? Accepted.
7
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
49 M. Paddack
Boeing
6.6 T Delete 7
th
bullet: Laminate material should be age
affect stackup.
This is not related to contamination and does not
belong here.
Accepted
50 M. Paddack
Boeing
6.6 T Delete 8
th
bullet: Brown oxide still being used in
Asia.
Not relavent to contamination and does not belong
here.
Accepted
51 Todd
MacFadden,
Bose
6.6, last bullet T Develop this point:

Brown oxide still being used in Asia.

Are there residues from this process that require special treatment to remove? Something special about
brown oxide; what about black and red oxides? Other lamination pretreatments that should be considered?
Accepted
52 J ohn
Rohlfing,
Delphi
Section 6.7, 6
th

bullet
E/T Treatment transfer remaining copper after etch
deep remaining tooth
Editorial Incomplete sentence (needs a verb)
Technical Unclear (what is the writer trying to
say?)
Accepted bullet
rewritten.
53 M. Paddack
Boeing
6.7 T Delete 5
th
bullet: The pH in the etching process
copper hydroxide can precipitate in plating tanks,
affecting plating quality.
here.
Accepted
54 M. Paddack
Boeing
6.7 E Expand on 6
th
bullet: Treatment transfer remaining
copper after etch deep remaining tooth.
Need to explan what the comtaminate is and its
impact. Add guidance on wha to do about it.
Accepted bullet
rewritten.
55 M. Paddack
Boeing
6.8 T Delete bullets 2, 4 and 5. Bullets are not related to contamination. Accepted
56 M. Paddack
Boeing
6.9 T Add to third bullet: Mixed resin chemistry dark
spots or charring of the epoxy resin. Cross
contamination can also result in delamination.
Epoxy dust trapped in polyimide has often be the
cause of delaminated boards.
Accepted
57 M. Paddack
Boeing
6.10 T Delete first bullet: Overcooking . be cured. Not relavent to contamination and does not belong
here.
Accepted with
modification added
Such degradation
cannot be cured and
may result in
undesirable chemical
products breakdown on
the surface of the
laminate.
58 M. Paddack
Boeing
6.10 T Delete 4
th
bullet: Under curing resins may also not
achieve for the polymeric structure, affecting
thermal and mechanical properties.
here.
Accepted with
modification Under
curing resins may allow
for migration of chemical
constituents in the resin
to migrate out as a
contaminant.
8
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
59 J ohn
Rohlfing,
Delphi
Section 6.10, 1
st

bullet
T Change Such degradation cannot be cured., to Such
degradation cannot be reversed.
Word choice (Typically, the word cured in the
context of electronics manufacturing refers to a
desired chemical crosslinking of a thermoset
material; however, the usage in this context is
intended to refer fixing a problem. Reversed is a
better choice, since the degradation is not a
medical condition that can be cured.)
Accepted
60 Todd
MacFadden,
Bose
6.10, last bullet T Make the point that anti-static gloves are not
necessary in PCB manufacturing:

Gloves advertised as antistatic or electrostatic
discharge (ESD) compliant
Accepted with
modification change
last sentence in this
bullet to read ESD or
treated gloves are
typically not used in
fabrication. If they are
used, gloves should be
investigated for this prior
to use.
61 Todd
MacFadden,
Bose
6.11, 1
st
bullet T Suggest clarify first bullet:

This is a very dusty process generating large amounts of particulate matter, even when
vacuum systems are employed.

So any steps that can be taken beyond the vacuum? Do the drill areas need ful-room air
filtration/purification systems?

I suggest modifying this, or adding new bullet:

Drill vacuum is critical to ensure complete removal of drill debris from holes, which could otherwise obstruct
or inhibit plating
Accepted last sentence
in recommendation as
2
nd
bullet in 6.11.
62 Todd
MacFadden,
Bose
6.12, 1
st
bullet T I suggest removing this bullet. Make the point in the
previous section (per my suggestion above). Also, it is
not always true that the deburr/desmear process will
remove particles/residue left in the holes, so I wouldnt
want to encourage operators to rely on it.
Accept with modification:
The particulates that
remain are often but not
always removed during
the subsequent
deburring process.
63 J ohn
Rohlfing,
Delphi
Section 6.12, 3
rd

bullet
E Change If this process is not well controlled, the resin
can be swelled and might not be removed in
subsequent rinsing operations., to If this process is
not well controlled, the resin can be swelled and the
potassium permanganate might not be removed in
subsequent rinsing operations.
Clarity (otherwise, it sounds like the swollen resin
is supposed to be removed)
Accepted.
64 M. Paddack
Boeing
6.13 E Move bullets 1, 2 and 3 to the introduction paragraph
of 6.13.
There are part of the process description and not
listing for contamination.
Accepted.
9
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
65 M. Paddack
Boeing
6.16 T Delete 6.16 There is nothing in 6.16 about contamination and
should be removed.
Accepted with
modification - revised to
indicate how these affect
contamination.
66 M. Paddack
Boeing
6.17 T Delete 6.17 There is nothing in 6.17 about contamination and
should be removed.
Accepted with
modification - revised to
indicate how these affect
contamination.
67 Todd
MacFadden,
Bose
Section 6.18 T Suggest the first bullet describe the importance of good cleaning prior to soldermask (city water can be a
source of contamination that presents reliability risks that sometimes do not manifest themselves until a
product is in the field):

One of the most important operations in the soldermask process is the pre-clean step. It is essential that
PCB shops use DI water in this step and not city water. City water quality is highly variable and often
contains very high levels of chlorine and other contaminants which can affect soldermask adhesion to the
PCB, as well as contributing to corrosion problems on the PCB in the field.

The following four images illustrate soldermask adhesion problems caused by inadequate PCB rinse before
soldermask process

Accept 2
nd
paragraph as
new first bullet and
incorporate the first
three images provided
(excluding the image
with solder mask
bubbles).
Todd
MacFadden,
Bose
Section 6.18
(contd)
T This cross section shows soldermask lifting from the trace due to inadequate pre-rinse. Accepted (see above
comment)
10
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
Todd
MacFadden,
Bose
Section 6.18
(contd)
T Another example of soldermask performance problems caused by poor pre-rinse operation:

Corrosion on copper traces under soldermask due to the presence of chloride under the soldermask
because the PCB shop used city water instead of DI water in the soldermask pre-clean operation. City
water can have extremely high levels of Cl and other contaminants.
Accepted section
revised along with bullet
relating to corrosion on
copper traces along with
the last photograph
displaying corrosion.
68 M. Paddack
Boeing
6.18 T Delete last sentence of 3
rd
bullet: If solder masks have
a low surface energy solder mask surface.
The material is not related to contamination. Accepted.
69 M. Paddack
Boeing
6.18 T Delete bullets 6, 7, 8, and 9 Not relavent to contamination and does not belong
here.
Accepted with
modification bullets
revised based on a
cross-contamination
issue.
11
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
70 J ohn
Rohlfing,
Delphi
Section 6.18,
10
th
bullet
E Change . . . a nightmare condition . . ., to . . . an
undesirable condition . . .
Colloquialism (may not translate well partially
cured solder mask will certainly lead to
undetectable contamination issues; but whether or
not the user experiences bad dreams as a result is
not for the writer to say)
Accepted.
71 J ohn
Rohlfing,
Delphi
Section 6.21, 8
th

bullet
E Change . . . they are a quite spectacular failure in a
250V test., to . . . they typically fail in a 250V test.
Colloquialism (may not translate well failure or
not, saying it is spectacular might encourage one
to try it anyway, just to see what it does)
Accepted with
modification.
72 J ohn
Rohlfing,
Delphi
Section 6.23, 5
th

bullet
E Change These are notorious for accumulating finger
salt, . . ., to These have a high tendency toward
accumulating finger salts, . . .
Existing word choice more op-ed than technical
(may not translate well)
Accepted with
modification.
73 Brian
Madsen,
Continental
AG
#7.1, page 25 E I think one of these was supposed to be static and
one dynamic, but right now it seems repetitive.

The first method became the static method and was
covered as IPC-TM-650, Method 2.3.26. The second
method became the static method and was covered as
IPC-TM-650, Method 2.3.26.1.
Did these both really become the static method, or
was one dynamic or something different?

Accepted 2
nd
paragraph
in recommendation for
inclusion in paragraph 8
of 7.1.
74 Graham
Naisbitt
Section 7 T/E This entire section needs re-writing. I consider that the sole focus should be on what
should be done today and drop all historic
reference. The equivalency factors are a
nonsense. The essence of the test is based upon a
sound and widely used measurement of ionics.
ALL instruments can be calibrated by means of
dosing a specified amount of a specified saline
solution into a known tank volume at a known level
of conductivity / resistivity. If that value is 1.00 then
they ALL can measure the same way. If they
dontthen the Guage R&R results will show this.
I have already proposed a new ionic contamination
test method to replace ROSE and a Guage R&R
can be run to prove this point.
Not accepted
comments here apply to
a previous Working draft
version of the document.
75 J ohn
Rohlfing,
Delphi
Section 7.1, 2
nd

paragraph
E Question: should (e.g. RMA) instead be (i.e. RMA)? Clarify Is RMA supposed to be an example of
high solids rosin fluxes (in which case, e.g. is
correct)? Or is RMA the principle type of of high
solids rosin fluxes (in which case, it should be
rendered, i.e.)? It seems to me the latter is the
case, but Ill defer to those who are more in-the-
know.
Accepted with
modification.
12
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
76 Brian
Madsen,
Continental
AG
#7.1.1, point #2 T Text:
If the contaminant dissolves slowly, then one may
only see a fraction of the contaminant present in the
typical 10 minute ionic cleanliness test. When ROSE is
being used for product acceptance, the ROSE test is
seldom run longer than the minimum 10 minutes of the
specification because longer run times increase the
chance of a failing result.

Recommend: to include some informative text that a
slow-dissolving contaminant should create an
increasing or non-baseline conductivity so the ROSE
test needs to be run to completion regardless of the
amount of time required.

Comment: Within the last months we have received
ionic conductivity test results showing a continuously
rising conductivity over the time of the test (20 min, I
think) that were considered a Pass because they did
not exceed our ionic limit (barely) at the time limit. It
was clear that if the test would have been run to
completion (per rules in 2.3.25 at right), the boards
would have failed the test.

I did not find the 10 minute limit in the 2.3.25 test
method. I did find

And
.

The current text seems to accept the 10 minute
limit without clearly pointing out it is in violation of
the test method procedure.

Accepted with
Modificaiton
77 J ohn
Rohlfing,
Delphi
Section 7.1.2,
3
rd
paragraph
E Change . . . some scientifically based, and some
involving Tarot-card-reading psychics and various
practices of voodoo., to . . . some scientifically based,
and others with no discernible scientific basis.
Word choice. Accepted.
78 J ohn
Rohlfing,
Delphi
Section 7.1.2.2,
2
nd
bullet
E Change . . . often need alternative methods of
analysis . . . or simple smell (lemony freshness) can be
used., to . . . often need alternative methods of
analysis . . . or simple smell (lemon, orange, etc.).
- Extra verb at end
- Advertising-speak, will not translate well
(freshness is vague and subjective, if not
outright indiscernible; whereas lemon and
orange, which are typical oils in some
commercial and industrial detergents, are
generally recognizable smells by most
members of any given society)
Accepted.
13
Figure/Table
Type of
comment
(T=Technical
or
E=Editorial)
Disposition
on each comment
submitted
79 Peter
Bodsky,
Thermo
Fisher
Scientific
Section 7.2.2,
2
nd
paragraph
E Add the following after the first sentence in this
paragraph:

In addition, with Capillary based IC systems, the
sample injection volume requirement is as small as 0.4
uL. In sample limited scenarios (ie; gathering a drop
from the surface of a circuit board), this allows for the
injection and quantification of extremely small sample
volumes.
With the availability of Capillary IC systems and
smaller and smaller boards, this addition might
provide for the utility of CapIC in future applications
without having to modify the method or prove
equivalency.
Accepted with
modification.
80 J ohn
Rohlfing,
Delphi
Section 7.3.1,
2
nd
paragraph
E Change For this reason, the total inorganic amount is
higher . . ., to For this reason, the total allowable
inorganic amount is higher . . .
Clarity Accepted.
81 J ohn
Rohlfing,
Delphi
Section 7.5.2,
last sentence
E Change . . . of residual surfactants (e.g. Cascade,
J oy) that do not show up . . ., to . . . of residual
surfactants that do not show up . . .
Inappropriate usage / over-specificity / used as
examples earlier, but referring to a certain practice
Accepted.

5703

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

5703

Transféré par

Droits d'auteur :

Formats disponibles

DRAFT DOCUMENT FOR INDUSTRY CONSENSUS REVIEW ONLY

Vous aimerez peut-être aussi