{insert
in-page links}
By
K.J. Whitlaw, LeaRonal,
Buxton Derbyshire
UK
Abstract
The
use of pulse periodic reverse current for improving the
metal distribution of copper on high aspect ratio printed
circuit boards will to measure. Data will be presented
comparing tests carried out using both direct current and
pulse periodic reversal to measure the effect of current
density on metal distribution for high aspect ratio printed
circuit boards (e.g. 8:1). A theoretical basis for
the improvement obtained by using pulse periodic reversal
will also be presented. The stability of the electrolyte
under pulse periodic reversal conditions will be investigated.
Details will also be given concerning the power supply/current
switching unit requirements.
Introduction
Over the
last 10 years the complexity of printed circuit board
design has increased dramatically, the board thickness
has increased significantly and hole diameters reduced
significantly compared to the relatively simple boards
which were produced during the early 1980s. Aspect
ratios (board thickness divided by hole diameter) of 10
to 20:1 or even greater are now commonplace. High
aspect ratio boards present particular problems for electrolytic
copper plating since with traditional direct current techniques,
unless very low current densities are used, it is impossible
to obtain uniform coatings with even thickness on both
the surface of the board and in the through-hole.
Operation at low current density is not compatible with
today's industry needs for high production output.
The use of pulse periodic reversal is proposed as a means
of achieving the desired uniformity of metal distribution
at sufficiently high current densities to maximize productivity.
Theoretical
Considerations
The simplest
way of considering an electrolytic plating system is by
considering the anode and cathode as capacitors with resistors
in parallel connected via solution bulk resistance.
This is shown schematically in Figure 1. The uniformity
of current distribution is dependent upon the uniformity
of the potential distribution, which in turn is dependent
on the potential drop due to the ohmic resistance of the
solution. The key to uniformity of secondary current
distribution (non-geometric) is polarization resistance.
If we consider 2 points on a printed circuit board, Ps
being on the surface and Ph being in
the hole, then the ratio of the current density on the
surface to the hole (Js/Jh) is given
by the following:
Js/Jh
= (Rh + Rp)/(Rs +Rp)
- (1) This is shown diagrammatically in Figure 2.
The
greater the polarization resistance the more even the metal
distribution, this can be demonstrated by substituting simple
numbers into equation (1). For example:
If
Rh = 6, Rs = 1 and Rp =
0 then Js/Jh = 6/1=6.
If
Rp = 1 then Js/Jh = (6+1)/(1+1)
= 7/2 = 3.5.
If
Rp = 4, Js/Jp = (6+4)/(`1+4)
= 10/5=2.
Polarization
curves showing variation in current with potential are used
to describe the behavior of an electrochemical system.
A typical curve is described in Figure 3. It is generally
accepted that the reaction Cu2+ + 2e-
Cu takes place in two steps, each involving a 1 electron
exchange with the cuprous ion as an intermediate.
Cu2+
+ e- Cu+
Cu+
+ e- Cu
The
first step is rate determining and controls the overall
kinetics.
In
electrolytic copper plating for printed circuit boards the
range where i/iL = 0.2 - 0.4 is most important,
ie region 3 of Figure 3. This is known
as the Tafel region. The important characteristic
for uniformity of current distribution is the slope of the
polarization curve and this is dependent on the charge transfer
co-efficient.
Figure
4 shows three polarization curves where has values of 1,
0.5 and 0.25. In order to determine the Tafel slope
of these curves, it is necessary to compensate for the transport
contribution to the current and plot the log of the current
density versus applied potential. This is depicted
in Figure 5, where Tafel slopes of 60, 120 and 240 mV per
decade are obtained for values of 1, 0.5 and 0.25 respectively.
In pulse periodic reverse plating, Tafel slopes as high
as 500 mV per decade are not uncommon and differences between
the Tafel plots using direct and pulse periodic reverse
current are shown in Figure 6.
So
what are the practical implications of this for uniformity
of copper electrodeposition on through hole printed circuit
boards? Considering the diagram of Figure 2, assuming
that the over potential at Ps is 250 mV and the
over potential at Ph is 150 mV, then using Figure
6 the current density at Ps in the case of DC
is 3.98 amps/dm2 whereas in the case of pulse
periodic reverse Ps is 2.81 amp/dm2
and Ph is 1.78 amp/dm2. This
results in a surface to hole ratio in the case of direct
current of 6.3:1 and pulse periodic reverse current of 1.58:1.
The
increase in polarization resistance (reduction of transfer
coefficient or increase of Tafel slope) is postulated to
be due to the adsorption of an insulating film containing
the long chain organic carrier component of the electrolytic
copper additive system. copper (I) and chloride. This
is depicted in Figure 7. The degree of absorption
of this species is reduced as applied potential increases,
as is shown in Figure 8. A high degree of surface
coverage results in lower charge transfer co-efficient.
It is therefore proposed that the absorption of this material
is enhanced by the strong positive potential during the
anodic phase of the pulse reverse cycle.
There
are two other aspects which need to be considered from a
mechanistic standpoint. Firstly, if the organic suppressing
agent is added to the bath without chloride, no suppression
results, adding weight to the theory that the absorbing
compound contains chloride and organic carrier. Secondly,
the addition of a sulphur based additive to the system causes
acceleration in DC mode, which is overcome by the application
of the pulse reverse technique, indicating desorption of
this species during the anodic phase. All these mechanisms
are in agreement with those already reported.
The
Wagner number is a dimensionless quantity which is an electrochemical
value used to define uniformity of current distribution.
The use of pulse current electrodeposition for enhancing
Wagner numbers has been previously reported.
The Wagner number is calculated as follows:
Wa
= d/dj x k/L = Ra/Re
ie
this is charge transfer resistance divided by the ohmic
resistance.
d/dj
= (2.303RT/Zf)/j
k
= solution conductivity
L
= characteristic length
For
high aspect ratio holes the characteristic length is the
length of the hole (thickness of the board) squared divided
by the hole diameter. This characteristic length results
from a mathematical analysis of a plated through hole system
which is very similar to that of a tubular electrode where
the ratio of ohmic resistance to charge transfer resistance
is given by
k = (4zFl2/RTd)io
where l = length of hole, d = diameter of hole
It
can be seen from the above that increasing the charge transfer
resistance, increasing the temperature of solution or conductivity
(reducing the ohmic resistance), reducing the current density
all tend to make the current distribution more uniform.
Increasing the thickness of the board or reducing the diameter
of the hole make the current distribution less uniform and
since the thickness of the board is a squared term its effect
is more pronounced.
Figure
9 shows the variation of Wagner number with applied current
density for values of 0.12 and 0.5 and L values of 1.92
cm and 4.17 cm. More uniform distribution is obtained
when = 0.12 compared to 0.5 (NB this is the most pronounced
effect) and at each value more uniform distribution is obtained
at the lower L value as predicted by the above equations.
Practical
Considerations
A
standard test board of 8:1 aspect ratio (0.3 mm diameter
holes drilled in 2.4 mm thick laminate) was plated in a
specifically developed acid copper electrolyte and the effect
of average applied current density on metal distribution
was measured for both direct current and pulse periodic
reverse. Surface to hole ratio was determined by standard
micro section technique and typical pulse periodic reverse
waveform is shown in Figure 10. The results are given
in Table 1 and expressed graphically in Figure 11.
It can be clearly seen that with direct current the metal
distribution deteriorates as the average current density
is increased, whilst the surface to hole ratio is maintained
close to 1 with pulse periodic reverse irrespective of the
applied current density. Photomicrographs of the actual
boards are shown in Figure 12.
Table 1 - Comparison of pulse periodic
reverse and direct current
Ratio
board surface: hole centre
Hole
diameter 0.3 mm
Board
thickness 2.4 mm
| Average current density |
PPR * |
DC |
| 1.6 amp/dm2 |
1.14:1 |
1.34:1 |
| 2.6 amp/dm2 |
0.90:1 |
1.49:1 |
| 3.6 amp/dm2 |
1.05:1 |
2.08:1 |
*
Reverse
current 0.4 amp/dm2 (mean)
10
msec forward, 0.5 msec reverse
Many
types of difficult boards have now been successfully processed
in full scale production, boards which proved very difficult
to plate using traditional DC techniques. Two typical
examples will now be considered.
1.
0.3 mm diameter hole in 3.2 mm thick laminate
2.
0.4/1 mm diameter holes in 5 mm thick laminate
The
following pulse conditions were used for each type of board:
| peak forward current density |
4 amp/dm2 |
| Forward: reverse ratio |
3.67:1 |
| Forward time: reverse time |
20 msec: 1 msec |
| Plating time |
1 hour |
The
thickness values obtained are shown in Table 2. A
typical photomicrograph of the 1 mm hole in the 5 mm board
is shown in Figure 13. The 3.2 mm board was a British
Standard test panel and the 5 mm board had large ground
plane areas (holes in low current density regions) and isolated
areas (holes in high current density regions) making it
extremely difficult to process. Much more uniform
distribution is obtained using the pulse periodic reverse
technique.
Solution
Stability
During
the development of the process for use with pulse periodic
reverse, additives generally tended to fall into two categories.
One which resulted in an enhancement of metal distribution
but unfortunately was unstable in the solution or secondly
those which were stable towards the high current density
reverse pulse but did not produce the desired improvement
in metal distribution. A new generation of additives
has been developed which are both stable in the solution
and produce the required metal distribution characteristics.
This solution stability can be demonstrated by continuous
electrolysis tests under laboratory conditions. This
was carried out in a 3 litre beaker at a current of 0.33
am/litre with the additive concentrations being maintained
by addition after analysis using cyclic voltammetric stripping
techniques. The variation in additive consumption
with electrolysis time is shown in Figure 14.
Table 2 Overall Averages Computed from all
the Holes Sectioned for Each Board
| Hole diameter |
Board thickness |
12/d |
Surface |
Hole entry |
Hole centre |
a:b |
a:c |
b:c |
Typical DC result
at 3 A/dm2a:c |
| mm |
mm |
mm |
(a) |
(b) |
(c) |
|
|
|
|
| 0.4 |
5.0 |
62.5 |
33.2 |
42.3 |
22.3 |
0.78 |
1.49 |
1.89 |
7.0 |
| 1.0 |
5.0 |
25.0 |
27.0 |
34.0 |
31.4 |
0.79 |
1.16 |
1.08 |
2.4 |
| 0.3 |
3.2 |
34.1 |
34.3 |
43.2 |
26.1 |
0.79 |
1.31 |
1.65 |
3.0 |
3.2mm
board - 3 sections taken from 18 x 12"panel
5.0
mm board - 14 sections taken from 18 x 24" panel,
2 holes measured per section
Average surface
= 1+ 2+3+4
4
Average hole
entry = 5+ 6+ 8+10
4
Average hole
centre = 7 + 8
2
This
low rate of replenishment and the stability of the solution
have been confirmed by production operation.
Recommended
Operating Conditions
The
recommended solution composition is shown in Table 3.
As mentioned above, the additive and carrier components
can be determined by CVS and combined into a replenisher
solution for convenience of dosing.
Table 3 Copper Gleam PPR
| Operating
Conditions |
| Copper sulphate |
75 g/l |
| Sulphuric acid |
120 ml/l |
| Chloride |
50 ppm |
| Temperature |
25°C |
| Copper Gleam PPR
Carrier |
15 ml/l |
| Copper Gleam PPR
Additive |
0.5 ml/l |
|
|
| Replenishment (10,000
ampere hours) |
| Copper Gleam PPR
Replenisher |
2.5 litres |
Deposit
Properties
The deposits
obtained from the Copper Gleam PPR process are comparable
with the deposits obtained from the highest performance
DC electrolyte. Elongation is approximately 20%
and ultimate tensile strength 300 N/mm2. Deposits
are fine grained equiaxed and matt in appearance and will
withstand the most rigorous thermal shock testing without
cracking, eg IPC, BS 9760. These characteristics
are maintained throughout the lifetime of the electrolyte.
Additional
Benefits
Since
pulse periodic reverse tends to equalize potentials across
a surface where in DC strong differences in potentials exist,
pulse periodic reverse produces more uniform surface distribution
in pattern plated boards with unfavorable geometries and
reduces overplating of isolated tracks. This can significantly
reduce the quantity of solder mask required to cover these
tracks with the attendant cost savings.
On
a specific test board with isolated tracks in high CD areas
and other tracks in low CD areas, using DC plating a thickness
ratio of 2.3:1 was achieved whereas with pulse periodic
reverse this was reduced to 1.2:1. Using a base copper
thickness of 17 microns, this gives rise to the following
track heights.
1.
DC Low CD region 17 + 25 = 42 microns
High CD region 17 + 58 = 75 microns
2.
Pulse periodic reverse
Low CD regions 17 + 25 = 42 microns
High CD regions 17 + 30 = 47 microns
To
cover a 75 micron track requires a dry solder mask film
coating of 50 microns, which is equivalent to a wet film
weight of ca. 105 g/m2 . For a 45 to 50
micron track a dry film thickness of 40 microns is required,
equivalent to 90 g/m2. Therefore when using
pulse plating a solder mask saving of (105 - 90)/90 = 16.7%
is achievable. This is demonstrated diagrammatically
in Figure 15.
Also
as a result of the improved metal distribution, both surface
to hole and across the surface, significant copper metal
savings can also be achieved by using pulse periodic reverse
e.g. for a board with a surface hole ration of 1.40: 1 then
to achieve a thickness of 25 µm in the hole an average surface
thickness of 35 µm is required. If by use of pulse
periodic reverse plating a 1:1 surface to hole ratio can
be produced i.e. 25 µm on both surface and hole, the resultant
savings are (35 - 25)/25 = 40%.
Instrumentation
For
a plating window of say 3 m x 0.6 m and pattern plating
with 50% copper exposed, a unit capable of generating a
forward current of 400 amps per side and a total reverse
current of 2500 amps is required. It is important
that the pulse generating system contains two separate rectifiers,
one for the forward current and one for reverse current,
since they are of different magnitudes. A range of
water-cooled, switch mode rectifiers which generate this
pulsed output incorporating the high energy reverse pulse
is now available.
Technical
Description
Each
system comprises a pulse engine and a separate control unit
which sets and controls the pulse output in terms of
forward pulse
time ( 1 - 50 milliseconds)
reverse pulse
time (0.2 - 3 milliseconds)
forward peak
current (50 - 800 amps)
reverse peak
current (50 - 2500 amps)
A
single unit linked to its individual control unit can be
used as a stand alone system but for larger multiple installations
all control units are linked using an RS485 data link.
Pulse timing is then set and controlled from one controller
which is configured as the master. Alternatively,
a personal computer or plant computer can control all pulse
units in the installation.
The
forward and reverse current densities are preset by the
operator or process engineer and the reverse current
is automatically calculated and set by the control unit.
The equations for calculating the currents are as follows:
a
= average current density
b
= peak forward current density
c
= peak reverse current density
d
= forward plating time
e
= reverse plating time
f
= ratio of peak reverse current density: peak forward
current density
i)
To calculate average current density from known peak current
densities and plating time:
a = [(b x d) - (c x e)]/(d + e)
-1
ii)
To calculate forward and reverse peak current densities
from known average current density peak reverse: peak
forward and plating time:
c = f x b substitute in (i)
-2
a = [(b + d) - (f x b x e)]/(d + e) rearranging
-3
b = (d - f x e)/a(d + e) from (2)
-4
c = f x b
The
pulse engines should be situated as near as possible to
the actual plating tank to minimize inductance effects and
the units can be configured to provide either single anode
or dual anode capabilities. If the system is configured
for local control and not computer control, all the pulse
parameters and plating current may be entered using the
keypad and display.
Conclusions
1.
Pulse periodic reverse current can be used to give improved
metal distribution of copper on high aspect ratio printed
circuit boards when compared to direct current plating.
This improved distribution can be achieved at high plating
rates. There is no need to reduce current density
for difficult boards to achieve good surface to hole ratios.
When using pulse periodic reverse with pattern plating build
up on isolated tracks is minimized. This results in
increased productivity and lower manufacturing costs.
2.
An acid copper electrolyte has been developed for use in
combination with pulse period reversal which is capable
of delivering the benefits and a reliable pulse plating
rectifier is now also available.
3.
The use of this technology will increase in the future as
the demand for more complex boards at high production rates
increases.
References
1.
M McCormick, DeMontfort University, private communication
2.
M Goodenough, K J Whitlaw, Trans Inst Metal Finishing,
67 (1989), 57 - 62
3.
A Hubin, D Segers, J Vereeken, unpublished data, part
of a LeaRonal/Siemens project
4.
T Pearson, J K Dennis, Journal of Applied Electrochemistry,
20 (1990), 196- 208
5.
M R Kalantary, D R Gabe, M R Goodenough, Journal of Applied
Electrochemistry 23 (1993), 231 - 240
6.
T Kessler, R Alkier, Journal of Electrochemical Society
123, (1976), No 7, 990-999
7.
LeaRonal Copper Gleam PPR
8.
Chemring Plating Systems
