Note: Descriptions are shown in the official language in which they were submitted.
~ 3 ~ 2 ~
Background o the Invention
Ceramic chip capacitors are widely used in hybrid and
printed circuit modules because theyoffer ruggedness,
volumetric efficiency J wide available range and attractive
cost. Generally, the chips are fabricated by interleaving
rectangular electrode plates and dielectric layers, alter-
nately attaching the plates to two termination bands on
the opposing ends of the chip. A substantial percentage of
cermic chip capacitors are fabricated using barium titanate
ceramic as the dielectric material with a typical thickness
of 1 mil. Palladium silver is frequently used or the plates
and thicknesses in the order of 0.1 mil are typical. It is
common practice to use solder joints to connect the termination
bands to the circuit of the module.
The primary means of soldering chip capacitors and chip
components in general, is with a hand soldering iron which can
damage not only the component but also the module. For example,
if heat is permitted to conduct into the component, internal
pressure may develop from trapped gas molecules within the
component that are not able to escape as rapidly as they expand.
Also, if the adjoining layers of the component have a mismatch
in coeficients of thermal expansion, shear stresses develop
within the component. Further~ even if the component materials
are selected to have similar coe~ficients o thermal expansion,
they likely will exhibit different thermal conductivities such
that a temperature gradient will exist between the layers and
cause shear stresses of expansion. Also, stress can be created
in the solder joints substantially caused by two conditions.
First, during the heating process, the chip is free to expand
but the module area is hindered from expansion by the mechanical
-1-
~ ~ ~ 3 ~ Z ~
restraint provided by the cooler surrounding material of the
module; compr0ssive stress occurs in the local area of the
module but no expansion. Then, upon cooling to room temperature,
the chîp and solder joints are in a state of tensile stress
as the chip is partially restrained from decreasing to its
original length by the solidified solder. SecondJ if the joints
are soldered one at a time, substantial stresses occur in the
joints and the chip caused by even minimal contact with the chip
during the formation of the second solder joint.
The internal component stresses heretofore described can
cause component failures and thereby substantially reduce the
reliability of a module. Examples of structural defects caused
by stress are delaminations, material crumpling, voids, and
cracks. The effects may, for example, be short and open circuits
or changes in capacitive properties. Also, stress on the
soldering joints m~y result in a poor connection. The prior
art includes various techniques for soldering chip capacitors,
the most common of which is the use of hand soldering iron.
Other methods utilize flame, hot air, or focused radlant energy.
However, a need still exists for a means and method for soldering
chip capacitors to modules so as to minimize the internal
component stresses and the stress on the solder joints.
3~
Summary of the Invention
Electromagnetic radiation having a wavelength in the
range from one millimeter to one nanometer is separated into
a plurality of beams which are directed at spaced impingement
areas on a support means for supporting materials to be bonded.
More specifically, the radiation preferably comprises pulsed
coherent laser energy having pulse wid-ths less than 10 milli-
seconds, a pulse repetition rate less than 10 pulses per second,
and a pulse train duration of less than 1 second. It is also
preferahle that these pulse parameters be selectable so as to
be optimized for various applications.
The apparatus so described may be used for soldering.
More specifically, it is preferable to use the apparatus for
the soldering of chip components onto circuit modules. In this
application, it is preferable that the radiation have a wave-
length of 1.06 micrometers and pulse widths of 4 to 6 milli-
seconds. It may also be preferable to support the module on a
X-Y table which may be controlled by a microprocessor and which
is used to precisely position the module while soldering chip
components to it.
In accordance with the present invention, there is
provided a method of soldering an electronic chip component
having solder on opposing encls to two supporting conductor pads
connected to a substrate, comprising the steps oE: providing
pulsed coherent laser radiation having pulse widths of less than
10 milliseconds with a repetition rate of less than 10 pulses
per second, the pulse train duration of said radiation being
less than one second; separating said radiation inko two simul-
taneous spatially separate heams having substantially equal
energy; directing said beams inward toward said opposing ends
of said electronic chip component to reflow said solder; and
-- 3
.. ~J
-
~ ~Z3~2C~
removing said beams to substantially simultaneously form
solder joints bet~een said component and said pads.
- 3a -
~39;~
Description of the Drawings
The invention will be described in more detail by
refe~ence to the drawings in which: .
FIG. 1 is a mechanical ~chema~ic diagram depicting the
operation of the apparatus ~mbodying the invention,
FIG. 2 is a pictoral view of the outward appearance of
a laser soldering apparatus;
FIG, 3 is an elevation view with partial section
showing a chip capacitor soldered to a module circuit;
FIGS. 4A and 4B are diagrams demonstrating the relation-
ship of the creep ang~le for two different solder thicknesses
tl and t2, for a given creep distance ~; and
FIG. 5 shows a graph of chip tensile stress plotted as
a function of temperature for two different solder thicknesses.
~ 3~ Z ~
Descri~tion of the Preferred Embodiment
Referring to Figure 1, an illustration is shown which
demonstrates the operation of the apparatus embodying the
invention. Two laser beams 2 and 4 are formed by directing
a coherent beam 6 of laser radiation at a double slotted
aperture disc 8. Preferably, the disc may be selected from
a plurality of discs so that the length and width of the
slots can be varied for different applications. Any con-
ventlonal pulse laser such as, for example, laser model SS-380
supplied by Raytheon Company may be used as the source of
laser radiation. Preferably, laser radiation with a wavelength
of approximately 1.06 micrometers is used for reasons herein-
after described. Also, the preferable pulse widths, pulse
repetition rate, and total energy requirements will be
described later. The total energy of the two formed beams
is monitored by directing the beams through a beam splitting
mirror 10, well known in the art, that reflects 1~ o the
incident energy to a conventional energy monitor 12 as shown
in Figure 1. The remaining 99~ of the incident energy of
two beams is transmitted through the beam splitting mirror
to reflecting prism 14, the f~mction of which is to provide
a spacing between the two beams which is sufficien~ly large
so that the beams may be simultaneously directed at two
opposing ends of a component 16 as shown in Figure 1. Next,
each beam passes through a conventional cylinder lens 18 and
20 to ocus the energy to a size and shape, the preerential
specifications o~ which will hereinater be described. Each
focused beam then impinges on a beam power splitting mirror 22
and 24 which has a characteristic property at the frequency of the
energy of reflecting 99~ of the beam toward ~he soldering plane 32
~2392~
where solder is to be reflowed. The mirrors 22 and 24 transmit
the remaining 1% of the energy through to energy monitors 26
and 28. The position of mirror 24 may be adjusted as illus-
trated in Figure 1 so that the beams may be directed to be
incident on the soldering plane at different distances apart
to facilitate the simultaneous soldering of both ends of dif-
ferent sized components. Additional adjustment is urther
attained through step changes in -the position of mirror 22
in a like manner to the adjustment of mirror 24. Also, as
shown in Figure 1, 100~ reflecting mirrors 34 and 36 direct
the images of visible light 38 and 40 from the soldering plane
to te~evision cameras 4? and 44 for aligning the position of
a component 16 relative to the sight lines where the two laser
beams are directed when activated. Visible light may be trans-
mitted to the soldering area by the use of fiber optics ~not
shown). The hardware heretofore ~.escribed is well-known to
one skilled in the art.
In operation, referring to Figure 2, a protective shroud
46 is opened by the operator to expose a conventional X-Y
table (not shown). Typically, a printed circuit module 70, as
shown in Figure 3, is rigidly affixed to the X-Y table by any
means and a chip component 16 is placed across the conductor
pads 74 which are part of the printed circuit module and to
which the chip component is to be soldered. After the pro-
tective shroud is closed, the X-Y table control panel 48 which
has both coarse and fine adjustments for both axes is used to
position the module so that when the laser 50 is activated,
the two beams as described with reference to Pigure 1 will
simultaneously impinge at the bases of the opposing ends of
the chip component. The proper positioning of the X-Y table
--6--
~ ~ ~ 3~ ~ ~
may be accomplished with the television monitors 52 and 54
connected to the television cameras in Figure 1 and which
have transparent overlays 56 and 57 to indicate the area and
location of laser radiation when activated. After the laser
firing parameters, l~hich are discussed hereinafter, are
selected using the laser control panel 58, the laser lS acti-
vated. The laser radiation energy absorbed by the solder on
the pad and the solder on the chip component is converted to
thermal energy which reflo~s the solder. When the energy is
removed and the solder cools, a solder joint is formed at each
end of the chip component connecting it to the circuit of the
module. The energy ln joules at the total energy monitor and
the individual beam monitors is displayed on the laser control
panel.
In an alternate embodiment, a microprocessor ~not shown)
is used to control the position of the X-Y table and the exci-
tation of the laser such that a plurality of chip components
can be sequentially soldered to a single module without operator
intervention. Whether in automatic or manual mode, the X and Y
coordinates of the X-Y table are displayed on digital readouts
59 on the X-Y control panel.
The soldering apparatus heretofore described provides for
a soldering technique that has substantial improvements over
the prior art. The improvements substantially result from the
simultaneous reflow of solder at both ends of the chip and the
operator o the apparatus havin~ very precise control over
the soldering parameters.
With re~erence to the background and Figure 3, chip capa-
citors are fabricated by interleaving rectangular electrode
plates 60 and dielectric layers 62 alternately attaching the
~ ~ ~ 3 ~ 2 ~
pla~es to two termination bands 64 at opposing ends of the chip.
For example, the elec~rode plates may be fabrica~ed of palladium
silver while the dielectric is a ceramic material. A chip may
be dipped into solder 66 during fabrication. It has been found
that when two strips of materials having different coeficients
o thermal expansion are bonded along an interface and heated
isothermally, the composite will bow concave to the material
having the lower coefficient of expansion. ~ neutral region
is established within each material which comprises a locus o
unstressed points which define a boundary between compressive
a~d tensile stresses; the maximum compressive and tensile
stresses generally exist on the surfaces of the materials. In
the case of a chip capaci~or where the bending moments are
counterpoised by successi~e layers of ceramic and electrode
plates such that the interfaces are maintained planar, the
stresses are at a minimum~at the center of each layer and
maximum at the boundaries. Also, in the case of a chip capa-
citor, even though the adjacent layer may have a similar
coefficient of thermal expansion, the materials typicalIy have
different thermal diffusivities such that stresses result
from d~fferent expansions caused by different temperatures
rather than different coefficients of expansion. Thermal
diffusivity of a material is defined as its thermal conduc-
tivity divided by the product of its specific heat and density.
Accordingly, to prevent damage to a ch:ip capacitor or a chip
component, it is, in general, important to minimize the heat
permitted to transfer into the chip during the soldering process.
The laser solder apparatus h~retofore described provides
very precise control over the amount and profile of energy
transerred to the solder and thereby provides a means of
~L~23~3~
minimizing the heat transferred to the chip. Por example,
the basic equation for the energy required to melt a substanee
is given by the equation
E~ /V = ~C~T + Hs)
where EM is the required melting energy in joules, V is the
volume to be melted in cubic centimeters, ~ is the material
density in grams per cubic centimeter, C is the material
specific heat in calories per gram per degree centigrade,
~T is the melting tempera~ure minus the initial material
temperature in degrees centigrade, and Hs is the heat of
fusion in calories per gram. Generally, solder should
be approximately 30 to 50 above its mel~ing point to
achieve good wetting. Therefore, the total energy required
to reflow is approximately given by the equation
Er/V ~(C~Tl + Hs) ~ 6(C~T2)
where ~T2 is the desired reflow temperature minus the melting
temperature
Assuming an "A" size (.035 x .055 x .65 inches} chip
capacitor including an end termination of 0.2 mils of silver
and 1.5 mils of copper followed by a 1.5 mil 63 Sn/37 Pb solder
dip, and all three materials reachlng the temperature of
215C or 32C above the 63/37 solder eutectic temperature,
- a total energy of .434 joules is required to re~low each solder
joint. Further~ assuming a reflection factor at the solder of
50% and a 90% optics transmission factor, approximately 0.964
joules would preferably be delivered in each beam for the
specified conditions. Similarly~ for example, 3.7 joules would
preferably be d~livered in each beam to solder an "F" size
~.235 x .075 x .210 inches) chip capacitor. A series of tests
~ ~ 2 ~
were made with the apparatus and energy monitor data was taken
for ~arious combinations of double slotted discs and laser
parameters. The precise amount of energy as theoretically
calculated above with any deviations arising from practical
considerations can be delivered by the apparatus with a high
degree of repeatability and consistency.
The laser model SS-380 is a Neodymium Yttrium Aluminum
Garnet (Nd-YAG) type that operates at a wavelength of ~.06
micrometers. The pulse width is variable from two to ten
milliseconds with the pulse repetition rate selectable to
6, 4, Z and 1 pulse per second and manual single shot~ There
is a maximum power of 40 watts. In the preferred embodiment,
the SS-380 laser system was modified to provide even greater
flexibili~y. Pulse widths to 30 milliseconds in 0.1 milli
second increments are a~ailable. l~lso, 30 joules per pulse
protected at 50 watts may be delivered. Further, a burst
mode is provided which operates above 10 pulses per second.
The use of 1.06 micrometer wavelength minimizes light
reflection from solder as compared to a 10.6 micrometer Carbon
Dioxide laser. I~ith the increased energy absorption by the
solder, a smaller percentage re-flects ~oward the chip to result
in increased stress within the chip. Also, because the energy is
coherent, ~he cylindrlcal lens can be desi~ned to precisely
focus substantially all the energy so that the incidence of
energy directly onto the chip is substantially eliminated.
Furthermore, the preferred laser energy profile comprises a
train of pulses with relatively narrow pulse widths. l~or example,
to solder an "~" size chip capacitor, it has been determined
to be preferable to deliver six pulses at a pulse repe~ition
-10-
~ ~ ~ 3 ~ 2 ~
rate of six pulses per second, each pulse having a pulse width
of four to six milliseconds. The double slotted aperture disc
used is one which provides a beam having a depth dimension
approximately equal to the depth (.0~5 inches) dimension as
viewed on Figure 3. Also, the laser amplitude is set at 20%
of maximum. It is preferable not to vary the amplitude when
changing from one component size to another; rather, the
double slotted aperture is changed and the pulse repetition
rate, pulse width, and duration are varied to optimize for the
particular application.
The narrow pulse ~rain energy profile described in the
previous paragraph provides significant protection against
component damage during and immediately af~er the soldering
process. First, the emissivity of liquid solder is substantially
greater than in the solid state. Accordingly, a relatively
large percentage of heat added to the solder af~er it becomes
liquid is conducted to the component. It follows that if the
energy that is added to the solder is precisely controlled in
small increments so that the solder passes into the liquid
state for only a very brief period of time such as milliseconds,
the amount of heat conducted to the component is minimized.
Second, for any given point within a chip capacitor, the stress
is maximized when the diference in temperature between the
electrodes and dielectric layers is also maximum. Further,
studies have shownthat a maximum diference occurs at approximately
three time constants where one time constant is defined as the
time for a point to rise to 63% of its total temperature rise
caused by the addition of heat to the material. This is
Iogical because if the temperature is raised very gradually
over a relatively long period of time, ~he different diffusivities
~23~32~3
of the layers have very little e~fect and the dierence in
temperature between the two layers is very small. Provided
their coefficients of thermal expansion are well matched~ the
stress t Up to a limit, will become smaller as the number of
time constants is increased. Also, if ~Lhe number of time
constants is reduced from three time constants, the material
has insuicient time to react. In most industrial applications
the production rate is important such that it is not easible
to expend considerable time to raise the ~emperature o
components slowly. Therefore, in order ~o minimize stress,
it is important to trans~er heat to the solder very rapidly.
The four to six millisecond pulses provide an introduction
of heat into the sol~er in a time subs~antially less than
one time constant.
The simultaneous soldering of both ends of a chip
capacitor by the use of two beams of laser radiation provides
substantial improvements in stress relief over prior art
methods. More specifically, after coollng down there is less
stress in the solder joints and accordingly less stress applied
to the chip. The simultaneous re1Ow at both ends of the
component without external contact permits the component to
float on the liquid solder. This provides a sufficient thickness
of solder between the component and the printed circuit board.
More speciically, it is preferable that not less than .0025
inches of solder solidiy under the chip to limit stress in
the chip and plastic atigue in the 5O1der.
Reerring to Figures 4A and 4B, it is illustrated that
given ~, the creep distance at each end of the component
required to compensate or the component expansion during
soldering being greater than the printed circuit module, the
-12-
~ 3 ~ 2 ~
creep angle ~ gets smaller as the thickness of the solder
is increased. More specifically, 32 is smaller than 01 because
t2 is greater than tl. Figure 5 plots emperical data that
demonstrates that the tensile stress applied to the component
after cooling down is reduced when the thickness is increased
from .003 inches to .OlO inches. Also, the simultaneous reflow
substantially eliminates the vertical stress applied to the
chip when soldering one end after the other end has already
been soldered.
Although the in~ention has been described with a preferred
embodiment, it will be appreciated to those skilled in the art
that various modifications can be made without departing
from its scope. For example, although the discussion is
primarily directed at chip components and chip capacitor in
particular, the general soldering technique would be applicable
to a variety of applications. It is therefore intendad that
the invention not be limited except as defined by the claims.
2~
-13-
