Throwing Power Measurement: The Assaf Cell

D.R. Gabe, M. Ward & Y. Assaf

A new throwing power cell is described in detail; it has

the virtue of being usable as a small fixture that can
be positioned in an electroplating tank. This so-called
Assaf Cell is particularly useful for simulating small
recesses, typically through-holes found in printed circuit
boards (PCBs).
This paper demonstrates how a Throwing Power Index,
given by the ratio of metal thickness on the front and
rear faces of the Assaf-Cell test panel, may be used to
evaluate/optimize electrodeposition conditions. Data are
presented from work using various electrolytes, including
high-throw acid copper solutions, thereby illustrating its
usefulness and application.
Fig. 1—Schematic of
The ability to ‘throw’ metal into a recess, hole or cavity the Assaf Cell design.
during electrodeposition is an important characteristic of an
electroplating process and is related primarily to the type of while most recently the use of personal computers has
solution and its constituents. Additional factors include the enabled other measures to be used, e.g., Luke.10 Table 1
cell design, the agitation employed and the disposition of the summarizes some formulae and Throwing Power Index
anodes. This characteristic was recognized during the last ranges that arise from their use.
century, but it was in the 1920s that Haring and Blum1,2 in the There is no shortage of ideas, but practicality is important!
U.S., and Field3 in the UK attempted to measure throwing Consequently, a number of other suggestions11 and patents12,13
power and to define a Throwing Power Index. remain to be exploited to general acceptability.
The Haring cell is well known (but not actually well used!)
as a box in which a channel of electrolyte is contained, The Assaf Cell
with a central anode and cathodes on either side at movable The Assaf Cell14 was proposed to overcome three shortcomings
distances away. The ratio of distance (x1 / x2) is then related of the Haring Cell:
to the weight (w1 ,w2) respectively or thickness of cathode
deposit when current is passed. The difference in thickness 1. Difficulty of providing uniform agitation.
is attributed to the solution’s resistance (iR drop), the 2. Differing anode current densities on each side with
electrode reaction efficiency and the electrode overpotentials. consequent differing voltage drops.
Field3 first proposed a formula for the Throwing Power 3. Need for specific relevance to PCB through-holes.
Index, so that:
An obvious advantage of the Assaf Cell is that it is, in effect,
(1) a small fixture that can be placed in the corner of a tank
or even clipped on a plating rack. A further advantage to
where L = x1 / x2 and M = w2 / w1. practical users is that it offers a narrow recess which can
be adjusted and used as a PCB through-hole simulator
Since that time, developments have been such that both other where the aspect ratio (i.e., board thickness-to-hole diameter
cells and other formulae have been employed, e.g., Gabe.4 ratio) is large.15
The variety of cells studied is substantial but none have The cathode fixture comprises a Perspex backing board
become established as standard, the Hull Cell being often and a 4 mm thick spacer onto which a 40 x 40 mm cathode
used as the basis for novel design: sheet is fixed, as shown in Fig. 1. In an alternative format, a
circular cathode may be used (40 mm diameter ≅ 0.25 dm2)
• Hull cell used in throwing power mode5 and different thickness spacers may be preferred to simulate
• Modified Hull Cells for agitation, heating and scale specific recess geometries. Whatever the choice of shape
linearity (based on convenience or applicability), it is important
• Gornall cell for PCB applications6 to maintain the desired anode:cathode area ratio for that
• Rotating electrode “Hull-type” cells7,8,9 particular electrolyte.
Although no anode-to-cathode spacing and size ratio
The formulae used have depended upon two issues: first, the are prescribed, a spacing of 150 mm has been reported
range of values; second the sensitivity of the value obtained. previously.16 A 100 mm spacing, 1.5:1 size ratio and 40
For example, the Field formula extends from +100 (very x 40 mm cathode have been preferred in this work.15
good) through 0 to -100 (very poor). For many years this has Electrolyte agitation is achieved by either attaching the
been found to be convenient (see British Standard 205, Part cathode assembly to a reciprocating rod, by a magnetic
5) but ranges involving ∞ and of 0 to 100 have protagonists, stirrer, or by conventional air agitation. Following deposition,
127 PLATING & SURFACE FINISHING
 Table 1
Throwing Power Index Formulae

Author Formula Range of values

Haring-Blum 100 (L - M) / L +80 to -∞
Heatley 100 (L - M) / (L - 1) +100 to -∞
Field 100 (L - M) / L + M -2) +100 to –100
Subramanian 100 (L - M) / M (L - 1) +100 to –25
Luke 100 L / (L + M - 1) +100 to 0

Table 2
Variation in Throwing Power
Using Different Electrolyte Conditions
Fig. 2—Comparison of throwing power ratio between “standard” and
Electrolyte Back:front proprietary PPR electrolytes using direct current.
thickness ratio
Well-formulated silver bath, DC 2:3 to 3:4
Pulsed current 1:1 and compare the “standard” electrolyte containing no
Aged bath with contaminants 1:17 additives and its proprietary counterpart * under direct current
Aged bath after filtration treatments 2:3 conditions. This throwing power reduction is attributed to
the change in slope (dη/di) of the polarization curve as
current density rises:
the cathode is rinsed and dried and the minimum deposit
thickness values (located around the central region of the • At lower current densities when dη/di is high, the current
cathode surfaces) are recorded. Throwing power is calculated tends toward a secondary-type distribution, producing a
by expressing the minimum metal thickness values as a more even metal coverage.
percentage ratio: • As current density increases and dη/di falls, the
current assumes a more primary-type distribution, and
consequently, throwing power falls.
(2)
The aim in analyzing such data is to identify high throwing
Experimental Results & Discussion power values at as high a current density as possible, and
Previous Work thereby upwards shifts of the trendlines in Fig. 2 are sought.
In his original paper, Assaf14 pointed out that in a conventional It should be recognized that the proprietary solution contains
silver plating bath with a typical thickness of 8 to 10 µm the additives designed for periodic pulse reverse (PPR) operation.
throwing power ratio was 2:3 to 3:4, but using pulsed current However, it is interesting to note that for the duration of
it could be reduced to 1:1. This was the demonstration of its this initial experiment, bright and lustrous electrodeposits
value. It is also true that the method can identify solution were produced. By comparison, those produced from the
contamination as shown by the values in Table 2. “standard” solution were level but relatively dull. Longer-
In a more detailed study of throwing power in silver term operation of the former electrolyte under DC produced
cyanide solutions, Leisner, et al.17 reported a series of results progressively duller deposits, presumably as additive
for various pulse current schedules and showed that the degradation occurred. It should also be noted that at higher
values were within the range 0.64 to 0.95, with DC plating current densities, the throwing power of both electrolytes
giving values as low as 0.49 or 0.62. They also showed that tended towards similar values.
graphs of throwing power enabled favorable pulsed current Figure 3 illustrates the effect of pulsed current on throwing
conditions to be identified and in one case, to establish an power using pulse waveforms as follows:
optimal current density for the system. Rasmussen18 has used
the Assaf Cell to study pulse plating of Sn-Zn alloys from an • Cathodic:anodic cycle time 10,1 and 20,1 ms
acid solution. When his values are converted to back:front • Current density ratio, ia:ic 2.8:1
thickness ratios, values of 50 to 65% have been obtained,
with a slight decrease as current density was increased. A It was found that at mean current densities up to 2.3
small change in alloy composition was also noted, but the A/dm2, improved throwing power may be obtained (20,1 ms
increasing current density could be at least partly responsible waveform) when compared to direct current; beyond this
for such a change. limit, DC throwing power remains the most effective. This
transition may be explained by the fact that a pulse reverse
IPTME Research waveform requires a higher cathodic current density than
In our work,15 high-conductivity acid-copper solutions for the equivalent DC condition in order to maintain the same
a PCB electroplating process were investigated and the deposition rate. Since it has been reported19 that the overall
influence of pulsed current measured. This investigation current distribution tends toward a primary distribution
was also concerned with the use of eductors for electrolyte under pulse conditions, this would explain the reduction in
agitation. Initial trials (Fig. 2) illustrate the reduction in throwing power between the pulse and DC data. The higher
throwing power that occurs with increasing current density throwing power achieved by the 20,1 ms pulse versus 10,1
ms is attributed to the same phenomenon. Occurrence of
___________________________ the throwing power transition at a particular magnitude of
*
Copper Gleam PPR©, Shipley-Ronal, Marlborough, MA. current density is understood to be related to the type of
May 2001 128
 fact that 25,1 ms was the optimum in this study. It
may be that there was a synergistic effect between
this particular timing and the additives (i.e., the
pulse-sensitivity of the additive is quantified), which
becomes suppressed at 30,1 and 40,1 ms timings.
If so, there must be scope for optimizing additives
further if the 25,1 ms timing was proven to be
applicable in a commercial process. It should be noted
that while it is advantageous to increase throwing
power, operation at higher current densities must
not compromise the physical/mechanical properties
of the electrodeposit.
The applicability of the Assaf test method to high
aspect ratio through-hole PCBs may be of concern
when considering the results obtained. This is a
Fig. 4—Comparison of throwing power ratio between DC and PPR consequence of the cell geometry and also the annular
with modified cathodic:anodic pulse times in the proprietary PPR gap behind the cathode surface (normally 4 mm),
electrolyte. which is considerably larger than a through-hole.
Electrolyte flow across the cathode surface is also
somewhat different in this cell compared to full-scale
process. However, Assaf Cell studies have been
successfully carried out elsewhere as a means of
analyzing PPR behavior in copper electrodeposition16.
In terms of determining trends, therefore, this method
is simple and allows a wide range of parameters to
be studied relatively quickly.
Finally, it should be noted that in this study
no attempt has been made to present a critical
experimental comparison of the available methods
of measuring throwing power. Instead, it has been
considered more useful to illustrate the way in which
the Assaf Cell test is performed and to show how the
data obtained can be constructively used.
Fig. 3—Comparison of throwing power ratio between DC and PPR current
in the proprietary PPR electrolyte. General Discussion
Industrial users of commercial plating solutions have
shown the virtues of the Assaf Cell as an analysis
additives used and would be expected to vary with additive tool. Practical experience with a number of common plating
concentration. Therefore, certain additives can be expected solutions including chromium, zinc, copper and tin, plus less
to be preferred in pulse-sensitive terms. common solutions such as bismuth has shown that in situ
Figure 4 illustrates the improvement in throwing power analysis of problematic baths can indicate:
that was achieved beyond the 2.3 A/dm2 threshold through
the use of other cathodic:anodic cycle times but still with • Incorrect current density
an ia:ic ratio of 2.8:1. The results show that throwing power • Metal ion depletion (including the effects of ineffective
may be significantly increased up to mean current densities agitation)
of 3.5 A/dm2 using a (cathodic:anodic) cycle time of 25,1 • Additive and/or brightener depletion
ms. This experiment was repeated to validate the accuracy
of these particular results. Furthermore, there is a point of One example cited is for data obtained from a commercial
inflection in the throwing power behavior that occurred in plating shop, in which process adjustments were made using
the range 2.7 to 4.0 A/dm2 using the 25,1, 30,1 and 40,1 ms Assaf Cell analysis. An acid-copper plating solution had
conditions. It is reasonable, therefore, to state that: become depleted of brightener content, achieving 75 percent
throwing power. With brightener additions, throwing power
• Changes sometimes occur at mean current densities fell to 35–50 percent, depending upon current density (noting
of 3 A/dm2 or above. that that throwing power usually falls with increased current
• Through-hole throwing power using 20,1 ms pulsed density). Results also indicated that the bath producing
current with additives at “high” current densities is samples with best back-surface appearance achieved a
comparable with that from direct current at “low” throwing power of 47 percent. This represented the optimum
current densities. brightener content. This case, therefore, serves as a reminder
that metal distribution is not always related to visual
As mentioned previously, improved throwing power at longer appearance!
pulse on-times can be expected since a lower peak current In practice, it is clear that a throwing power value of
density is required, and therefore, the current distribution 100% represents “perfect” thickness distribution between
would tend toward DC conditions. While this may explain panel faces, which can only be achieved with ideal solutions
the improved throwing power recorded for the 25,1, 30,1 (probably complexed) and conditions. In practice, a range
and 40,1 ms timings, it cannot, however, account for the of 30 to 80 percent has been found normal for acid copper

129 PLATING & SURFACE FINISHING

using DC, noting that care must be taken when measuring About the Authors
deposit thickness. Experience has shown that this will vary Dr. David R. Gabe is professor of
slightly between the cathode edge and middle regions, even Materials Engineering at Loughborough
when using a small surface area in experiments. University of Technology, Leics., LE11
3TU, England, and formerly director of
Conclusions the Institute for Polymer Technology and
The Assaf Cell and fixture have been used to measure Materials Engineering. He is a graduate
throwing power in a number of solutions and for small of the University of Wales (Cardiff) and
recesses, typically as represented by through-holes on PCBs. received a PhD from the University of
Useful data has been generated which has enabled optimal Sheffield. He has published more than
pulsed-current conditions to be identified. The graphical 250 papers on electrodeposition and
representation of data (% throwing power vs. mean current related coating processes, and was recipient of the AESF
density) has been shown to give clear indications of preferred Scientific Achievement Award in 1995. He also served
pulse conditions. a two-year term as president of the Institute of Metal
This approach to throwing power measurement has been Finishing.
claimed to be of greatest use in the context of through-holes Dr. Matthew Ward is a research technologist at Mars UK
on PCBs. For this reason, it has not so far gained wide usage but Ltd. Prior to this, he was a research associate at the Institute
appears to have been limited to the aforementioned research of Polymer Technology and Materials and Engineering
exercises. In commercial solutions the requirements have (IPTME), Loughborough University in the UK, where he
been adequately met by traditional methods. Nevertheless, was involved in researching properties of hard anodic films
use of the Assaf Cell in these fields could offer a degree of as part of a collaborative program with the European Union
convenience, making further studies worthwhile. and three anodizing companies. Dr. Ward holds a B.Eng.
in mechanical and materials engineering, and a PhD for
Acknowledgments research into electrodeposition of copper for printed circuit
The authors acknowledge supply of chemicals by Shipley- board manufacture. He is a member of the Institute of
Ronal (U.K.) and encouraging discussions with Dr. J.N. Metal Finishing, and to date, has published 15 papers on
Crosby and Mr. K. Whitlaw during the early part of his research interests, which include low-friction-anodized
this work. surfaces, agitation-system design, and pulsed-current
technology for electrochemical processing.
References Yair Assaf has owned and operated Bohak Plating Co.
1. H.E. Haring & W. Blum, Trans. Amer. Electrochem. Ltd., Israel, since 1974. He also spent several years in the
Soc., 43, 303 (1923). research and development of fishing gear for the government
2. H.E. Haring, ibid., 46, 107 (1924). of Israel, including underwater research. He worked for
3. S. Field, Metal Ind., 44, 614 (1934). seven years at the Corrosion and Surfinishing Lab, Technion
4. D.R. Gabe, Metal Finishing Guidebook & Directory, Haifa Israel, and two years as director of the plating shop
Elsevier Science, New York, NY (2001). of Tademon Company.
5. T.Q. Mei, Plating & Surface Finishing., 83, 79 (December,
1998).
6. Gornall Cell, McGean Rohco Inc., Cleveland OH.
7. A.F.S. Afshar, D.R. Gabe & B. Sewell, Trans IMF,
69, 37 (1983).
8. C. Madore, D. Landolt, C. Hassenpflug & J.A. Herman,
Plating & Surface Finishing, 80, 36 (August, 1995).
9. I. Kadija, J.A. Abys, V. Chinchankar & H.K. Straschil,
Plating & Surface Finishing, 78, 60 (July, 1991).
10. D. A. Luke, Trans IMF, 61, 64 (1983).
11. S. Shawki, F. Hanna & Z. A. Hamid, Metal Finishing,
85, 59 (December, 1987).
12. R. F. Bernards, U. S. patent 4,932,518 (1990).
13. B. Stein, U. S. Patent applied for. (1993).
14. Y. Assaf, Plating & Surface Finishing, 67, 12 (October,
1980).
15. M. Ward & D. R. Gabe, Proc. Europ. PCB Conference,
Munich, Germany (1999).
16. P. Leisner, PhD Thesis, Lyngby Technical University,
Denmark (1992).
17. P. Leisner, P.T. Tang & G. Bech-Nielson, Proc. AESF
SURFIN® 1994, Indianapolis, IN., AESF, Orlando,
FL (1994).
18. J. Rasmussen, Proc. AESF SURFIN® 1993, Anaheim,
CA, AESF, Orlando, FL (1993).
19. T. Pearson, J.K. Dennis & J.F. Houlston, Trans. IMF,
69, 9 (1991).

May 2001 130