COPPER PLATING

Plating

HIGH ASPECT RATIO PCBs

Achieving proper copper deposition in holes on thicker boards is no easy task, even with reverse pulse plating. Understanding the ins and outs of electrolyte agitation is critical.
by STEPHEN KENNY, BERT REENTS and DR. JENS ZOSEL

Achieving quality electrolytic copper metallization during the production of high aspect ratio PCBs, such as back panels, isnt easy. These panels can be from 3 mm to 10 mm thick, with typical aspect ratios of 10:1, and the current trend is for thicker panels with aspect ratio up to 15:1. One of the limiting factors in copper deposition is the mass transport of ions into high aspect ratio holes. Achieving the required copper thickness in the hole without overplating the surface (causing resist overplating with pattern plate or poor line definition with panel plate) is the biggest problem in the production of high aspect ratio panels. Also, component mounting with the press-fit technique is difficult when copper deposit distribution is poor. To overcome throwing power problems, low electroplating current densities have been used, but this has a negative impact on productivity. As a solution to these problems, reverse pulse plating can allow the use of higher current densities with improved surface distribution and throwing

power in the through holes. However, pulse plating does not eliminate mass transport limitations; to achieve good plating results an adequate electrolyte agitation must be ensured. As part of an Atotech/Kurt Schwabe Institute joint project, the flow dynamics of copper deposition were investigated1, and the influencing factors on copper deposition in blind microvias were documented2. Further experiments have been carried out to investigate the influences on through-hole plating, particularly in high aspect ratio holes. TABLE 1 gives a summary of the electrolyte exchange mechanism considered and also the influencing factors. The influencing parameters were held constant as far as possible and the artificial convection by means of forced flooding was investigated. A specially designed multilayer PCB with an electrochemical flow sensor was used as part of these investigations. A schematic of one hole on the test board is shown in FIGURE 1. The test board was placed in a test chamber, allowing the variation of key parameters such as nozzle diameter, angle, distance, lateral flow, pressure/flow, density and pulse pumping. FIGURE 2 shows a microsection through one test coupon showing the inner layer electrode connections.
TABLE 1. Electrolyte exchange mechanisms and

influencing factors
ELECTROLYTE EXCHANGE BY Diffusion Migration Natural convection Artificial convection FIGURE 1. Schematic of test hole in circuit board showing microelectrode array.
22

INFLUENCING FACTORS Concentration Temperature Surface tension Viscosity Density

PRINTED CIRCUIT DESIGN & MANUFACTURE

JANUARY 2005

COPPER PLATING
TABLE 2. Test conditions for Inpulse 2 trials for 3.2 mm thick panels TEST I AVERAGE/ PULSE I REVERSE PARAMETERS (ms) (A/dm2) FORWARD/REVERSE 4 DC 4/4 4 / 16 4 / 16 4 / 16 4 / 16 DC 240 / 20 80 / 4 80 / 6 250 / 25 240 / 12 PHASE SHIFT () 180 180 180 180 180 PULSE RISE TIME FACTOR 1.2 1.2 1.2 1.2 1.2 SURFACE FINISH Good Good Good Rough Rough Good THROWING POWER MIN. (%) 32 53 49 55 54 71 FIGURE 2. Microsection through test coupon with a hole diameter of 0.2 mm showing inner layer electrode connections.

1 2 3 4 5 6

TABLE 3. Test conditions for Inpulse 2 trials for 3.2 mm thick panels, 0.3 mm holes TEST I AVERAGE/ PULSE I REVERSE PARAMETERS (ms) (A/dm2) FORWARD/REVERSE 1 2 3 4 5 4 4 / 16 4 / 16 4 / 16 4 / 16 DC 40 / 2 80 / 4 160 / 8 240 / 12 PHASE SHIFT () 180 180 180 180 PULSE RISE TIME FACTOR 1.2 1.2 1.2 1.2 FREQUENCY AVERAGE (Hz) MIN. THROWING POWER (%) 23.8 11.9 6.0 4.0 30 40 50 60 70

velocimetry (PIV) was used to image the flow of electrolyte through a high aspect ratio panel. Here, a dynamic system is illuminated by two laser beams and the resulting interference pattern information is recorded on a camera. The data from one of the experiments of flow through a high aspect ratio panel are shown in FIGURE 3.

The results show that a maximum diffusion current is achieved at a flow angle of 90, with the highest impingement velocity, of course. In larger-scale tests, the technique of particle image

Horizontal Application
The high aspect ratio throwing power in Atotechs Uniplate systems has restricted their use in the production of thicker panels. Even in panels thicker than 1.6 mm, the copper throwing power has not been fully acceptable, depending on the aspect ratio. The reason for this is the emphasis on the production of thinner material at higher current densities with blind microvias. The high current density, on the order of 10 A/dm2 average, and the requirement to produce blind microvias under such conditions, have required the use of relatively high copper concentrations at above 35 g/l. These factors have not enabled the best throwing power in high aspect ratio panels. Attempts have been made to improve the throwing power in standard pulse plating equipment but these have

FIGURE 3. PIV measurement of vertical solution flow through high aspect ratio panel.
24

PRINTED CIRCUIT DESIGN & MANUFACTURE

JANUARY 2005

only provided marginal improvement. These trials were limited by the pulse parameters available with the standard system. The standard Inpulse module has a spray bar-to-cathode separation of 95 mm and an anode-to-cathode separation of 75 mm. In the Inpulse 2, both the spray bar and the anode are set much closer to the cathode at 15 mm and 8 mm for the anode. The spray system itself has been modified to give a more directed agitation towards the panel, these changes were made primarily to enable the more efficient flooding of blind microvias. Using this system, experiments were conducted to investigate the optimal electrolyte composition and pulse plating parameters to achieve best throwing power in 3.2 mm thick panels with an aspect ratio of 10:1. The results have shown that primarily, the pulse waveform setup and electrolyte adjustment are critical to throwing power improvement. The following was found to be the best electrolyte composition: Copper 20 g/l Sulphuric acid 270 g/l Chloride ions 40 mg/l Iron(II) 7 g/l Iron(III) 1 g/l Leveller Inpulse H6 1.7-2.0 ml/l Brightener Inpulse 4.0-5.5 ml/l The additive concentrations are more typical of electrolytes adjusted to produce high aspect ratio panels; in particular, the copper concentration is 15-20 g/l lower than in a standard electrolyte. The pulse plating parameters were varied from DC plating conditions at 4 A/dm2 to pulse plating, with a forward of 250 ms and a reverse of 25 ms. A selection of the parameters used and the throwing power achieved is shown in TABLE 2. Due to the weakness of corner flattening caused by high reverse conditions and surface roughness, the best throwing power results were achieved with a forward of 240 ms at an average current density of 4 A/dm2 and a reverse of 10 ms at current density 16 A/dm2, and not with 25 ms in reverse time. A general tendency can be seen that throwing power is increased with a lower frequency, as in TABLE 3. In all tests a phase shift in pulse parameter of 180 was used, meaning
JANUARY 2005

PRINTED CIRCUIT DESIGN & MANUFACTURE

25

COPPER PLATING
TABLE 4. Pulse plating batch sequence for thick panel trials SEQUENCE 1 2 3 EXPOSURE TIME (MIN) 100 15 5 A/dm2 RATIO FORWARD TO REVERSE 1:31 1:2.5 1:1.1 PULSE TIMING (ms) 30:1.5 20:1 20:1

that the reverse pulse was applied to the anodes on one side of the test panel while the forward pulse was applied to the anodes on the other side. See FIGURE 4. The pulse rise time factor was set at 1.2 in these experiments, and its effects on plating results are currently being investigated. As a comparison with similar panels, a throwing power of only 30% would be achieved at 3 A/dm2 with horizontal DC and at 2 A/dm2 with a throwing power of 55% under vertical conditions in DC. Only with pulse plating under standard vertical conditions with air agitation is a throwing power of 90% achieved, but this is at a current density of 2 A/dm2. With forced agitation, improved throwing power is possible, but even this is not at such a high current density.

Vertical Application
In vertical systems the electrolyte agitation is usually produced with a combination of air agitation in the electrolyte itself and a mechanical agitation of the PCB being plated. This mechanical agitation must ensure that the panels are moved evenly and remain vertical in the electrolyte, or the solution flow will not be uniform through all the holes in the

panel. To ensure this cathode movement, systems are used which clamp the panel securely and supply current to the panel. These agitation systems, air in the electrolyte and movement of the panel can lead to uneven fluid transport due to the non-defined air agitation and the movement of the panel through the agitation bubbles. To overcome these problems, eductors are becoming more common. Eductors use the Venturi principle (the kinetic energy of one solution will cause the flow of another) to allow small pumps to circulate larger volumes of liquid. Typically the use of eductors can produce a 4 to 6X increase in volume of solution movement compared to the volume pumped, but this increased volume is at a lower pressure than the directly pumped solution. A smaller eductor will pump a lower volume but will allow more eductors to be placed on one pipe, thus providing a more even flow. The current method of eductor installation is in a vertical plating tank on the floor underneath the cathode. Two pipes are placed below the cathode, one on each side, with the eductors pointing upwards toward or away from the cathode. There are similar installations with the eductors

JANUARY 2005

PRINTED CIRCUIT DESIGN & MANUFACTURE

27

COPPER PLATING
tions. The system consists of a mounted on a single pipe runspray head that moves regularning directly below the cathly over the surface of the cathode, with the eductors mountode and produces an intensive ed at a fixed angle pointing forced flooding of the panel alternately away from the and through-holes at the point panel. The disadvantage with of spraying. The head moves this setup is that the electrolyte in a plane between anode and flow uniformity depends on cathode, but is so dimensioned the positioning of the eductors that it does not interfere with and also the distance between the electrodeposition process. the nozzle and the panel. Initial results with high aspect To give a more uniform ratio panels show a significant flow, eductors can be posiimprovement in throwing tioned between the anodes in FIGURE 4. Pulse waveform schematic showing phase shift power when compared to the plating cell pointing between top and bottom anodes. standard air agitation, and a directly towards the cathode. more uniform deposit in comparison to This setup has the advantage of giving a eductor agitated equipment on the same more direct flow of electrolyte towards scale. FIGURE 6 shows plating results the panel (FIGURE 5). The disadvantage of all eductor installations is that the from a 3.2 mm panel with 0.3 mm hole. solution flow can never be completely Investigations are ongoing in the use uniform over the panel surface. A comof so-called batch plating parameters to promise must be made between the improve throwing power, particularly in number of eductors installed and the panels thicker than 5 mm. During the flow uniformity. plating cycle, the pulse parameters are To overcome eductors limitations varied. Normally at the start of the on flow uniformity, a moving spray sys- FIGURE 5. Eductors installed with flow cycle, a strong reverse charge is used to tem is being tested in laboratory condi- directly towards the cathode. give good throwing power, followed by

28

PRINTED CIRCUIT DESIGN & MANUFACTURE

JANUARY 2005

COPPER PLATING
FIGURE 6. Microsection of 10:1 aspect ratio panel plated with moveable spray system. Notes: panel thickness 3.0 mm; hole diameter 0.3 mm; average current density 2 A/dm2; throwing power 90-95%; reinforcement plated by CD horizontal at 5 A/dm2.

FIGURE 7. Microsection of 10:1 aspect ratio panel plated with modified pulse parameters. Notes: panel thickness 5.0 mm; hole diameter 0.5 mm; average current density 1.7 A/dm2; throwing power 95100%.

a lower reverse charge at the end of the plating cycle for good surface finish. An example of such a plating sequence is given in TABLE 4. FIGURE 7 shows plating results from a 5.0 mm panel with a 0.5 mm hole using a pulse plating sequence together with the moveable spray system to give optimal electrolyte exchange. The use of both optimized pulse parameters and electrolyte agitation gives significant improvements in throwing power in trial line experiments. Experiments in basic electrochemistry have shown a strong influence of electrolyte agitation on copper electroplating characteristics. Modifications to horizontal pulse plating equipment, together with optimized plating parameters, show improved throwing power under experimental conditions. However, the influence of improved solution flow appears to be limited in first investigations. In new vertical equipment, the use of eductors to improve agitation is becoming a standard. However, agitation systems used in vertical systems still do not match the quality of horizontal systems.

The use of a moving spray flood system shows advantages in the trial line scale, but this must be confirmed in full-scale production equipment. PCD&M
STEPHEN KENNY is manager, vertical systems panel and pattern plating at Atotech, Berlin. He can be reached at Stephen.Kenny@atotech.com. BERT REENTS is manager, horizontal systems panel and pattern plating at Atotech, Berlin. He can be reached at bert.reents@atotech.com. DR. JENS ZOSEL has been with the Kurt Schwabe Insitut, Meinsberg, Germany, since 1992. He can be reached at zosel@ ksi-meinsberg.de.

REFERENCES
1. B. Reents, A. Thies, P Langheinrich, J. Zosel, . Online measurement of flow and mass transfer in micro-holes with PIV and an electrochemical sensor array. ISE Sympo sium 2002 proceedings. 2. B. Reents and S. Kenny, The influence of fluid dynamics on plating electrolyte for the successful production of blind microvias, IPC Expo 2002 proceedings.

JANUARY 2005

PRINTED CIRCUIT DESIGN & MANUFACTURE

29