IPC-D-279 EN.pdf - 第27页

somewhat higher or lower fatigue reliability and are, on the whole, signiﬁcantly less well characterized from a reliabil- ity point of view . 3.6 Plated-Through Hole and Via Reliability In surface mounting the plated-thr…

the optimum solution (see A-2.3 and A-3.6 for explana-

tions).

3.5.1.8 Cyclic Temperature Swing The cyclic tempera-

ture swing (∆T) of components and substrate is the differ-

ence in the maximum and minimum steady-state tempera-

tures experienced during either externally (daily) imposed

temperature variations or operationally (on/off, load ﬂuc-

tuations) imposed variations. It needs to be realized that the

temperature swing of the components is typically not the

same as the temperature swing of the substrate due to the

power dissipated in active devices. Smaller ∆Ts result in

improved reliability. It needs to be noted that for some

applications the temperature swings during transport and

storage prior to operation can be more severe and a bigger

threat to reliability than the operating conditions.

3.5.1.9 Cyclic Expansion Mismatch The cyclic expan-

sion mismatch, ∆(α∆T) results from the difference in the

thermal expansion of components and substrate which are

determined by the respective thermal expansion coeffi-

cients (CTE) and cyclic temperature swings (∆T). Smaller

expansion mismatches result in improved reliability.

3.5.2 Secondary Design Parameters While the effects

of secondary design parameters are, by themselves, of

second-order importance, their additional contribution to

the effects of the ﬁrst-order parameters can be signiﬁcant.

The effect of some of these second-order parameters might

be different in accelerated testing and actual operational

use. The effect of these secondary parameters is indirectly

included in the reliability predictions of Equations A-3 and

A-4 in Appendix A, Solder Joint Reliability, is the ‘‘non-

ideal’’ factor, F. This F-factor is empirically determined.

Design parameters having second-order effects on solder

joint reliability are as follows:

3.5.2.1 Solder/Base-Material CTE-Mismatch The large

CTE-mismatch (∆α) between the solder and some base

materials (ceramic, Alloy 42, Kovar, silicon) can make sub-

stantial contributions to the cyclic fatigue damage (see

A-2.3 and A-3.6).

3.5.2.2 Solder Joint Shape/Fillet/Volume Experimental

evidence indicates that solder joint shape/ﬁllet/volume

have only secondary importance for reliability. In some

highly accelerated tests cyclic life improvements of about

a factor of two have been achieved, but it is not clear

whether even these small beneﬁts would result for the

slower conditions prevalent in most product operations.

The improvements result from the time necessary for crack

propagation through the ﬁllet.

Stress concentrations, e.g. from solder-mask-deﬁned sol-

dering lands for ball grid arrays (BGAs), can reduce the

solder joint fatigue life by as much as a factor of three

depending upon the severity of the loading conditions.

3.5.2.3 Solder Joint Uniformity Some experiments in

which solder joints were loaded primarily in a stress-driven

mode (high cyclic frequencies, no hold times, very large

temperature swings with fast transitions) showed the need

for extreme uniformity of all the solder joints of a compo-

nent to avoid unequal stressing; accelerated tests utilizing

test conditions more closely resembling product use condi-

tions did not reveal a need for extraordinary solder joint

uniformity.

3.5.2.4 Initial Solder Joint Grain Structure Aﬁne initial

grain structure in solder joints results in cyclic life

improvements of about a factor of two in highly acceler-

ated tests. The grain structure of solder is inherently

unstable and will grow with time. Higher temperatures and

cyclic loading accelerate the grain growth. Thus, for most

product applications a ﬁne initial grain structure will not

result in a signiﬁcant improvement of fatigue life; the sol-

der joints of accelerated test vehicles should be artiﬁcially

aged to start the tests with more product-related grain

structures.

3.5.2.5 Conformal Coating Conformal coating can have

different effects on solder joint life during thermal cycling

depending on the type of material, thickness, and location.

The advantage of conformal coating is that it slows the

absorption of water and oxygen into surface cracks. The

presence of oxides on the cracked surfaces may accelerate

the crack propagation. Oxidation layers prevent ‘‘re-

welding’’ of the solder during crack closure.

On the negative side, conformal coating may add another

material with a very high thermal coefficient of expansion

which may inﬂuence reliability. This addition can be sig-

niﬁcant especially if the coating wicks under components,

ﬁlling the gap between the printed board and component.

In addition, conformal coating can become rigid below the

glass transition temperature. This condition can exert con-

siderable stress on the components and solder joints during

the thermal cycles.

Because of the large variation in conformal coating mate-

rial properties, thickness applied, methods of application,

etc., the effect of conformal coating, in general, needs to be

evaluated empirically for each application.

3.5.2.6 Compliant Substrate Surface Layers Compliant

layers at substrate surfaces can provide additional reliabil-

ity margins, but by themselves are not adequate to counter-

act the effects of large expansion mismatches.

3.5.2.7 Solder Composition The most widely used sol-

der compositions are eutectic (63/37) and 60/40 tin-lead

solder. Solder compositions other than these can have

July 1996 IPC-D-279

somewhat higher or lower fatigue reliability and are, on the

whole, signiﬁcantly less well characterized from a reliabil-

ity point of view.

3.6 Plated-Through Hole and Via Reliability In surface

mounting the plated-through holes (PTHs) have been

reduced to the single function of providing electrical con-

nections through the circuit board. Because the PTH-vias

(PTVs) no longer need to be able to accept component

leads, there is no need for the traditional large PTH diam-

eters. The drive towards higher circuit board densities also

has put pressure on designers to reduce PTV diameters. At

the same time, the number of layers and, thus, the circuit

board thicknesses have been increasing. Therefore, the

PTV aspect ratio (the ratio of circuit board thickness and

drilled PTV diameter) has been increasing. The results of

an IPC round robin study (IPC-TR-579, Round Robin Reli-

ability Evaluation of Small Diameter Plated Through Holes

in PWBs) show that for PTVs with aspect ratios (AR) > 3,

special care is necessary to produce high quality PTVs.

Copper plating quality in the barrel was found to be a sig-

niﬁcant parameter; nickel plating in the barrel increases the

robustness of the PTV to temperature cycling.

The PTVs are most stressed during temperature excursions

into the solder reﬂow range. For high quality PTVs ﬁve (5)

excursions to solder reﬂow temperatures consume about

1/6 of the available fatigue life of the PTH-via copper bar-

rels; for low quality PTH-vias the solder reﬂow operations

can cause PTH-via barrel fracture during manufacture. For

high quality PTH-vias the 1/6 loss of available life is not

signiﬁcant; however, for applications with more severe use

conditions (see Table 3-1) this 1/6 loss of life could be a

sizeable portion of the design life.

The most important aspect of high quality PTVs is the

quality of the copper deposit in the via barrels. As the

PTV’s aspect ratio increases, it becomes more difficult to

plate high quality, uniform copper deposits inside the holes.

Special electrolytic plating formulations or electroless plat-

ing may be required.

3.7 DfR of SM Solder Attachments The material in

Appendix A gives a detailed treatment of DfR of solder

attachments.

3.8 DfR of Insulation Resistance The material in

Appendix C gives a detailed treatment of DfR with regard

to Insulation Resistance.

4.0 SUBSTRATES

This section addresses the materials related issues of sub-

strates. For printed board design and layout, see section 3.3

above.

See also IPC-D-275 for more details on rigid boards.

4.1 General Substrate Categories Interconnect sub-

strate technologies can be divided into the following cat-

egories:

1. Organic based printed wiring board

2. Discrete wiring printed board

3. Ceramic based printed circuit board (thick ﬁlm,

co-ﬁred)

4. Ceramic based printed circuit board (thin ﬁlm)

Categories 1-3 above can be further classiﬁed:

Type 1—Single sided

Type 2—Double sided

Type 3—Multilayer with blind or buried vias

Type 4—Multilayer with blind and/or buried vias

Type 5—Multilayer metal-core board without blind or

buried vias

Type 6—Multilayer metal-core with blind and/or buried

vias

Type 7—Rigid-ﬂex multilayer without blind or buried

vias (organic only)

Type 8—Rigid-ﬂex multilayer with blind and/or buried

vias (organic only)

See Table 4-1 for the advantages and disadvantages of

common substrates.

Surface mount components are held on these substrates

with solder. The components normally have a different

CTE than the substrate; consequently, there will be

mechanical stresses on solder joints as the ambient tem-

perature changes. The cycling stress is a potential reliabil-

ity problem. To minimize this problem, packages can use

compliant leads and have the CTE matched to that of the

substrate. Substrates are also used for removal of heat.

Selected substrates must maintain their function in adverse

environmental conditions and be manufactured/assembled

at reasonable cost.

Substrates for surface mount technologies include the most

commonly used ﬁber glass reinforced epoxy system with

ﬂame retardant, FR-4. The next commonly used material is

polyimide glass because of its higher temperature. Each of

the materials has its own particular characteristics and

properties and behaves differently under varying conditions

of temperature, humidity, and other stresses.

4.2 Substrates and Their Functions:

a. Power distribution;

b. Signal distribution, interconnecting lands to each

other and to the next level with acceptable signal

integrity;

c. Structural, providing a stable platform with a CTE

close enough to that of the components that sufficient

reliability under use conditions is obtained.

IPC-D-279 July 1996

For the assembly to function properly, the electrical,

mechanical, and thermal requirements of each material

used in the substrate must be considered for the operating

conditions and use environments.

Table 4-1 provides guidance in the material choices. See

IPC-D-275 for other material discussions.

4.3 Moisture and its Effects on Polymer Substrates

Polymers commonly used in SMT printed boards absorb

and adsorb water when exposed to moist atmospheres (high

relative humidity) for durations ranging from several days

to several weeks; the equilibration time depends upon the

thickness of the laminates and the geometry of the conduc-

tor pattern. The relative permittivity of water is 80 and that

of common substrates ranges from 3 to 5; the absorption of

1-3% by weight of water can signiﬁcantly (but reversibly)

increase the dielectric constant between conductors and

hence the capacitive coupling between conductors, over

time. The absorption and adsorption of water also

decreases the insulation resistance between conductors at

the surface (SIR). Together with ionizable contaminants

and DC bias, condensed moisture can lead to electro-

chemical corrosion and dendrites on the surface of the sub-

strates; conductive anodic ﬁlament (CAF) formation at the

glass ﬁber-resin interface; and electrochemical corrosion

and dendrites at delaminations and voids such as occur

between conductors on inner layers and between barrels of

PTVs and PTHs. Moisture effects are more signiﬁcant in

SM printed boards because the spaces between conductors

and the interbarrel distances are much less than the corre-

sponding dimensions in through hole printed boards; in

addition, solder masks may not be easily applied between

land patterns in the ﬁne and extra ﬁne pitch SM patterns.

See Appendix C for DfR information on SIR. See also

IPC-TR-476.

Some materials with higher glass transition temperatures

) such as bismaleimides and polyimides, appear to

absorb more water than the lower T

materials, such as the

epoxies. Drying of the higher T

materials (as well as

thicker buildups of the epoxy systems) prior to SM reﬂow

exposure or rework/repair is recommended to minimize

delamination or separation, for instance, of the conductor

from the resin or the glass ﬁber from the resin.

The laminate surface is porous when treated by etching to

enhance adhesion of conductors; this surface porosity can

retain hydrolyzable and ionizable contaminants and water,

as well as hydrophilic materials such as polyglycols which

are used in the formulation of some water soluble SM sol-

der pastes. Solder mask and conformal coating materials

cover and seal the porous surface and help to retain SIR

values and reduce the risk of corrosion.

Common solder masks (and conformal coatings) are per-

meable to water vapor; the presence of water soluble con-

taminants between solder mask or conformal coating and

the underlying surface can result in vesication or mealing

and in electrochemical corrosion/migration.

Chemisorption of water into polymers appears to reduce

the T

slightly, reduces the adhesion of the polymer to

other materials and reduces the strength of the polymer.

See also the bibliography of IPC-SM-786.

4.4 Coefficient of Thermal Expansion (CTE) of Polymer

Systems

Polymer systems expand with increasing tem-

perature, demonstrating a glassy phase response below T

with a CTE or α

and a rubbery phase response above T

with a much higher α

, typically 3 times α

. The transition

from glassy phase to rubbery phase is gradual, but for most

polymer substrates may be characterized by T

, the glass

transition temperature.

Glass ﬁber reinforced substrates exhibit signiﬁcantly differ-

ent CTE in the z (out of plane) axis compared to the CTE

in the x and y axes; for example, below its T

, Quartz/

Bismaleimide material with 35% resin by weight exhibits a

CTE(x-y) of 6 ppm/°C and a CTE(z) of 41. Woven glass

ﬁber reinforcement exhibits an additional difference

between x and y axes on the order of 1-5 ppm/°C; this dif-

ference may be signiﬁcant where the CTE of a large SM

component package is to be matched to the CTE of the

substrate to enhance cyclic life of the solder attachments.

The CTE(z) is particularly signiﬁcant in determining the

cyclic life of PTH and PTVs in SM PWAs because the

aspect ratio (ratio of substrate thickness to ﬁnished hole

diameter) is generally much larger than the corresponding

aspect ratio achieved in printed boards manufactured for

through hole technologies. Higher CTE(z) values result in

higher cyclic tensile stress in the barrel of the PTH or PTV

during temperature excursion during SM reﬂow, or SM

component removal/rework/repair as well as during printed

board fabrication, solder dipping, hot air leveling, or wave

solder. See IPC-TR-579 and IPC-SM-782.

The thermal cycle reliability, vibration robustness, and the

thermal management of high performance Surface Mount

(SM) products are heavily dependent upon the constraining

core such as copper-molybdenum-copper (CMC), copper-

Invar-copper (CIC) and molybdenum-graphite-

molybdenum (MGM) composite material systems.

The ratios of the various materials in those composite sys-

tems can be adjusted to tailor the effective CTE to the opti-

mum value. The tradeoffs include weight and cost. See

IPC-MC-324.

4.5 Constraining Cores in Substrates A constraining

core is an internal supporting plane in a packaging and

interconnecting structure, used to alter the coefficient of

thermal expansion of printed boards.

July 1996 IPC-D-279