IPC-7095C-2013.pdf - 第98页

to the solder proﬁle due to nonuniform heating (wide band width). Note too that consistent problems may also be related to paste quality and land pattern design. Once the program is found to give the desired results (ass…

should be maintained between 230-245°C for lead-free.

The time above liquidus (TAL) should be 60 to 90 seconds.

Extended duration above the solder melting point or TAL

will damage temperature sensitive components. It also

results in excessive intermetallic growth which makes the

solder joint brittle and reduces solder joint fatigue resis-

tance.

7.1.4.8 Cooling Zone The typical cooling rate for most

assemblies has been 4-6°C per second, driven primarily by

throughput and SnPb intermetallic thickness concerns.

With the transition to Pb-free solders, the pad cratering

defect has become more common due to the increased stiff-

ness of the SAC solders and the reduction in the ﬂexure

resistance of laminates. Pad cratering has been identiﬁed

directly after the reﬂow process leading to several experi-

ments designed to understand the impact of cooling rate.

During the cooling phase, the various materials in the

y-axis of a solder joint will cool at different rates. Typically

the BGA package will cool faster than the solder joint and

much faster than the PCB. This differential cooling can

create mechanical strain on the weakest spot in the

interconnect, the laminate below the BGA pad. By slowing

the cooling rate signiﬁcantly, to as low as 1.5°C per sec-

ond, all of the materials in the y-axis will cool more slowly

and will reduce the strain placed on the laminate. Experi-

mental work performed by consortia has shown that this

slowed cooling does not appear to negatively impact the

solder joint intermetallics nor the grain structure of the

joint. If pad cratering is found immediately after reﬂow or

if the assembly is determined to be at risk for cratering,

then the cooling rate of the PCA should be slowed to

reduce strain.

7.1.4.9 Thermal Profile for Backward Compatibility

Developing a reﬂow proﬁle is made difficult when dealing

with backward compatibility issues where some lead-free

components are used on a primarily tin/lead board. Back-

ward compatibility is a scenario where some components

are available only with lead-free surface ﬁnishes. Such a

scenario arises since it may not be economical for many

component suppliers to supply both tin/lead and lead-free

versions of the same component. It is not an issue when

using leaded components such as small outline integrated

circuit (SOIC), plastic leaded chip carrier (PLCC) or ﬁne

pitch with lead-free surface ﬁnishes. Most tin/lead compo-

nents primarily have 85% tin surface ﬁnish with about 15%

lead. When component manufacturers eliminated the lead

from parts they plated pure tin; however, the solderability

suffered. To correct this condition, up to 5% bismuth was

added to the plating alloy to improve the wetting or solder-

ability to the part.

The real problem arises when using lead-free BGAs on a

primarily tin/lead board. If the tin/lead proﬁle with maxi-

mum peak temperature of 220°C is used, the lead-free

BGA balls will not reﬂow at all or will partially reﬂow,

creating a serious solder joint reliability problem.

If tin/lead components are soldered along with some lead-

free BGAs in the same oven (since tin/lead versions were

not available), a peak temperature must be used that is not

damaging to all the tin/lead components, but is also suffi-

cient to reﬂow the lead-free BGAs. Using tin/lead solder

paste is appropriate, since most of the components on the

board are tin/lead. As shown in Table 7-3, peak temperature

of 210-220°C will be ﬁne for tin/lead but inadequate for

lead-free BGA balls with a melting point of 217-221°C.

But a peak temperature of 228-232°C with 60 to 90 sec-

onds time above liquidus or TAL will be sufficient to

reﬂow lead-free BGAs without seriously damaging all the

tin/lead components on the same board.

If the tight reﬂow temperature band of 228-232°C is diffi-

cult to achieve, in order to solder both tin lead and lead-

free BGAs in backward compatibility scenario, consider

selective laser soldering to solder lead-free BGAs after

other tin lead components have been soldered in a convec-

tion reﬂow oven, or ﬁnd an alternative source for BGAs

with tin/lead balls.

7.1.4.10 Unique Profile for Each Printed Board Assem-

bly

A reﬂow program is not the same as a reﬂow proﬁle.

A program is the combination of machine settings and con-

veyor speed, while a proﬁle is the visual representation of

the time and temperature that a thermocouple sees during a

PCAs travel through the reﬂow oven. Every unique PCA

needs to be proﬁled to show that all locations on the board

meet the various requirements for creating acceptable sol-

der joints. A single program will produce very different

proﬁles for different, unique PCAs. There is some misun-

derstanding that one oven proﬁle will work for all boards

and hence there is no need for developing unique proﬁles

for each board. This is simply not true because each board

has a unique thermal mass or different loading patterns

(distance between boards as they are loaded in the oven).

A double-sided board, depending upon component place-

ment and distribution of copper planes, will require a dif-

ferent proﬁle for each side. Proﬁles may look the same for

many PCAs but will generally require different machine

programs to create these similar proﬁles. It is common to

have a small number of standard machine programs, but it

must be shown that a speciﬁc program produces an accept-

able proﬁle.

Once the program has been optimized to create the desired

proﬁle, it is recommended that an actual production board

with solder paste and components for reﬂow be created.

After reﬂow, inspect the quality of the solder joints to

verify that the solder joints across all of the various com-

ponents meet the requirements of the IPC-A-610 and any

customer speciﬁc requirements. A random problem only in

a certain section of the board may be related to solderabil-

ity; a consistent problem in a given section may be related

January 2013 IPC-7095C

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

--`,,```,,,,````-`-`,,`,,`,`,,`---

to the solder proﬁle due to nonuniform heating (wide band

width). Note too that consistent problems may also be

related to paste quality and land pattern design.

Once the program is found to give the desired results

(assuming design and other material variables have been

optimized), document the program. After this point, no

changes should be allowed in the program and the result-

ing proﬁle.

7.1.5 Material Issues Flux has two key attributes. First,

it must remove contamination and, second, it must protect

the solderable surfaces after contamination removal. A

common mistake is to use a time/temperature proﬁle that

consumes the ﬂux before the solder melts. Ideally, the ﬂux

would be consumed just as the solder begins to melt. Acti-

vation time should range from 90 to 120 seconds. Flux

usually becomes active at around 130°C for tin lead solder

pastes. Typically, solder paste activation for lead-free sol-

der will be higher, in the 150°C range; however, it is rec-

ommended to work with your solder paste supplier for rec-

ommendation on that speciﬁc solder paste.

Components can be damaged by the incorrect application

of heat. All components have a heat exposure limit. Most

tin/lead surface mount components should tolerate a peak

temperature of 220°C for up to 60 seconds. Lead-free

BGAs will be rated to a higher temperature which is

approximately 240-260°C. Thermal shock, caused by the

rapid application of heat, can crack certain components.

However, since the peak temperature of reﬂow ovens var-

ies, the intent is to heat the solder in a controlled estab-

lished proﬁle to a solder joint temperature of 210-220°C

for tin lead products and for 235-245°C for lead-free

products.

Component lead ﬁnish will affect solderability. There are a

number of lead ﬁnishes being used today, including tin/

lead, gold, tin and palladium. It is important to select a ﬂux

and solder alloy that works well with the lead ﬁnish being

used.

7.1.6 Vapor Phase Vapor phase reﬂow can be operated

as a single ﬂuid system or a two ﬂuid system, utilizing a

primary and a secondary ﬂuid. The process was developed

using the two ﬂuid approach in batch equipment; but mod-

ern in-line systems are normally operated with only one

ﬂuid. Whichever system is used, the maximum temperature

reached by assemblies in vapor phase (VP) reﬂow depends

on the choice of the primary ﬂuid. Primary ﬂuids are avail-

able in a number of temperature ranges, with 218-222°C

being common with tin lead products and 235-245°C for

lead-free products. While all the primary ﬂuids can be

classed as perﬂuorocarbons, the basic structure (amine,

cyclic or ether) will determine the key properties of in-use

stability, solder paste chemicals solubility, and overall

process economics. The choice of a ﬂuid is normally based

on the melting point of the solder alloy to be reﬂowed.

For the range cited, the lower temperatures are suitable for

the typical tin/lead or tin/lead-silver alloys used for stan-

dard attachment processes. The upper end of the range will

permit reﬂow of high lead alloys, which are used to attach

pins to PGA packages. Users faced with reﬂow of a spe-

cialty alloy have been successful in mixing two primary

ﬂuids to tailor a vapor phase system for a speciﬁc stable

boiling point. Higher temperatures will permit shorter

times, which may be advantageous with some solder

pastes.

The primary vapor phase should be inert and not introduce

contaminants that must be removed later. Solder paste

chemicals that dissolve in the ﬂuid are carried in the high

boiling vapor then deposited on the surface of the boards.

Such residues tend to be difficult to remove. Minimizing

solder paste residue in the primary ﬂuid will maximize the

lifetime of the ﬂuid, prevent boiling point elevation due to

dissolved paste ingredients, and simplify cleaning.

The secondary vapor blanket was originally CFC-113, a

lower boiling ﬂuorinated material, which formed a low cost

sacriﬁcial ‘‘lid’’ over the more costly primary ﬂuid. The

constant exposure to the high boiling primary ﬂuid at the

interface of the two ﬂuids could cause the secondary ﬂuid

to undergo thermal decomposition at the interface, generat-

ing HCl (hydrochloric) and HF (hydroﬂuoric) acid vapors.

These corrosive vapors often attacked the soldering equip-

ment over time. While in theory the vapors could be

absorbed in ﬂux residues and cause problems for high reli-

ability products, this was rare in comparison to the attack

on the equipment. With the phase out of CFC-113, a low

boiling perﬂuorocarbon was introduced to replace it. This

second generation secondary blanket ﬂuid was more stable

than CFC-113 for prolonged exposure to the high boiling

vapor phase ﬂuids.

As surface mount technology grew, most users converted

to the higher throughput in-line machines, which used the

single ﬂuid approach. Deﬂuxing after vapor phase reﬂow

should be done with either a bipolar solvent formulation or

include an aqueous cleaning formulation that can ensure

removal of all the solder paste residues, with the choice of

cleaning process based on the composition of the solder

paste. Secondary factors inﬂuencing the decision would be

compatibility, and the component to PWB surface spacing.

In addition, most companies gave serious thought to con-

sidering the potential chemical loss from using this type of

equipment since many perﬂuoro compounds are very long-

lived global warming compounds.

7.1.7 Cleaning vs. No-Clean The selection of the solder

paste/ﬂux technology has many implications. The two

basic types are materials where the residues require clean-

ing or can be left on the PCA without causing damage.

IPC-7095C January 2013

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

--`,,```,,,,````-`-`,,`,,`,`,,`---

7.1.7.1 Flux Residues that Require Cleaning (Clean)

Those residues that require cleaning can be separated by

the chemicals necessary to remove the residue, typically

either solvent/surfactant, basic water, or deionized water

cleanable. Selection of a cleaning solvent should follow a

careful evaluation of technical, economic, and environmen-

tal considerations.

In general, cleaning under BGAs is difficult because the

tight gap between the board and components may entrap

ﬂux which may be difficult to remove during cleaning.

Entrapped ﬂux residues that require cleaning are typically

highly corrosive so this situation can lead to serious reli-

ability concerns. However, if proper care is taken in select-

ing the cleaning processes and equipment, and if the sol-

dering and cleaning processes are properly controlled,

cleaning under BGAs can be successful. In addition, if

no-clean pastes are used, the stencils are required to be

cleaned to ensure good printing. It does need to be empha-

sized, however, that good washer process control is essen-

tial when using aggressive water-soluble ﬂuxes.

When establishing a cleaning process for water-cleanable

solder pastes, it is necessary to verify cleanliness speciﬁ-

cally with low proﬁle BGAs or CSPs, and BGA connec-

tors. It is common to utilize a method such as surface insu-

lation resistance (SIR) to verify that ﬂux residues are

completely removed between the BGA balls and to estab-

lish your cleanliness process settings. Other methods such

as ion chromatography or ionography are recommended to

be used to monitor the process after qualiﬁcation.

7.1.7.2 Flux Residues that Do Not Require Cleaning (No-

Clean)

The use of no-clean ﬂuxes has increased due to

the environmental concerns of using ﬂuxes that require

cleaning, the disposal of used solvents and the cost of the

cleaning equipment. However, no-clean is not a drop-in

process. No-clean ﬂuxes are generally not as active as other

types of ﬂux and hence the soldering results may be less

than desired unless adequate steps are taken. Some

no-clean ﬂuxes require nitrogen reﬂow environments to

produce acceptable results; however, most currently avail-

able no-clean ﬂuxes can be reﬂowed in air.

No-clean ﬂux residues must be sufficiently inert so that

they will not damage the PCB or components. These resi-

dues are often mildly ionic and/or acidic and can poten-

tially cause product failure. The possible failure modes

caused by those residues include:

1. Corrosion of the PC board.

2. Shorting between metal traces due to dendritic growth.

3. Functional degradation of the ICs due to ionic contami-

nation diffusing into the active junctions of the die.

The ﬁrst two failure modes are surface phenomena. The

effects can be detected on the board surface, and the incu-

bation time is relatively short. These failure mechanisms

are detected with the surface insulation resistance (SIR)

test or the electrochemical migration test (ECM). The third

failure mode depends on how long it takes for the external

ions, if any, to diffuse through the package encapsulation

into the die area. Once these foreign ions reach the silicon

die, they may impact its function. The diffusion time is

usually on the order of years in a normal operating envi-

ronment. Typically, some form of acceleration testing, such

as SIR or ECM, is required in order to measure this effect

within a reasonable test period. In addition, all no-clean

ﬂuxes used on a PCA such as solder paste, wave solder ﬂux

and even rework ﬂuxes can all remain on the PCA and can

become mixed. It is recommended that these ﬂuxes be

tested in combination, as well as individually, to under-

stand the complete cleanliness picture.

7.1.8 Package Standoff The package standoff is one of

the prime parameters determining the reliability of the

BGA solder joints. Package standoff for a BGA is deﬁned

as the distance between the land on the bottom of the pack-

age substrate and the land on the top of the board surface.

This distance varies depending on the type of solder ball:

the high lead ball type stays a fairly consistent size; the

eutectic solder ball reduces the package standoff height.

This is also known as ball collapse height. For example,

when 1.27 mm tin/lead BGA is soldered on the package,

it collapses 0.1 mm and when it is soldered on the board,

the balls ‘‘collapse’’ additional 0.2 mm. So overall, for

1.27 mm ball you can expect 0.25 mm to 0.30 mm reduc-

tion from the pre-reﬂowed value. The high lead and copper

balls do not collapse because they do not melt. The stand-

off height on lead-free BGAs should be veriﬁed during the

set up of your process. It is recommended to establish the

process to reproduce the standoff height for speciﬁc com-

ponents.

Factors that determine the post-reﬂow BGA package stand-

off from the board include the BGA package weight, the

ball size, the ball material, the land size and land conﬁgu-

ration (solder mask deﬁned or nonsolder mask deﬁned).

Standoff height decreases with increased package weight.

However, for packages with a large ball count, the package

weight may have less effect on the standoff height. One

study on the relationship between these two parameters

discovered that increasing the package weight by 5X

decreased the standoff height by only about 0.05 mm for a

615 ball, 1.27 mm ball pitch BGA package.

Larger ball sizes will lead to larger package standoff

heights due to the larger volume of solder in each ball.

Standoff heights are inversely proportional to the land

diameters. For nonsoldermask deﬁned (NSMD) lands, a

solder mask relief around the land may reduce the standoff

height, because the solder will wet out along the conduc-

tors as well as along the edges of the land. This is shown

in Figure 7-10.

January 2013 IPC-7095C

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

--`,,```,,,,````-`-`,,`,,`,`,,`---