How Heat Generation in Stencil Printing Affects Solder Joint Quality


Reading time ( words)

The challenge of stencil printing in the midst of ongoing miniaturization is achieving repeatable solder paste deposition from print to print and pad to pad; this requires an understanding of paste rheology characteristics and the influences of temperature. "Rheograms" demonstrate the need to control heat generation in the paste roll to ensure consistent depositions at flip chip dimensions.

By M.H.A. Riedlin and N.N. Ekere

Solder paste deposition on printed circuit boards (PCB) at acceptable quality standards is essential for successful reflow soldering of surface mounted assemblies, i.e., forming good thermal, mechanical and electrical solder joints. Printing is a complex process that can be summarized as the extrusion of a paste roll, formed by the action of a moving squeegee blade, through the apertures of an extremely thin metal stencil and onto a board. Having passed over the openings, the stencil and PCB are separated mechanically (Figure 1).1 Components are then placed on the deposits and the entire assembly is heated to the eutectic tin/lead temperature of the solder paste to complete formation of the joints.

69371-th_71074.gifFigure 1. The solder paste stencil printing process. Several variables are critical: paste quality, stencil design, and the heat generated by the squeegee and stencil.

Several variables are critical for successful printing: quality of the solder paste, stencil design and manufacture, and the printing equipment. The aim here is to investigate the effect of heat generated by the printing equipment, particularly by the squeegee and the stencil, on solder paste quality.

Solder Paste RheologySolder paste consists of solder particles suspended in a flux/vehicle system. Particle diameter and size distribution, typically 25 to 45 μm, can vary depending on the application. The volume fraction, or metal content, can also vary; typically it is 88 to 91 percent metal content. The flux/vehicle system, which dictates how well paste prints and how susceptible it is to environmental variations, can consist of resins, solvents, activators and rheological additives. The formulation embodies much experience and expertise, and is usually regarded by paste manufacturers as proprietary information.

The base of most systems is a resin. It is dissolved in a rosin-soluble solvent whose composition partly determines paste viscosity. One or more surface agents are normally included to ensure proper dispersion of the particles and to provide adequate wetting of the substrate. Additives can also be included to give the paste the desired rheological performance. The viscosity of the paste is important for two reasons: It should be low enough so that the squeegee can successfully force the material through the stencil apertures, yet sufficiently high to prevent the paste from smearing or slumping after separation.

Effect of Temperature on ViscosityA major difficulty in achieving consistent printing performance and quality is that the properties of most solder pastes are extremely responsive to changes in ambient temperature and humidity. As with most suspensions, two competing effects occur when temperature rises: the falling viscosity of the flux/vehicle system obeys Arrhenius' equation, and as the system becomes more fluid and the dispersed particles acquire more energy, the rate of structure formation increases. In other words, the fall in flux/vehicle viscosity may be more than compensated for by an increase in suspension viscosity. Hence, a possibility exists for viscosity to rise, remain unchanged or fall with an increase in temperature.

A clear trend appears to relate solder-paste viscosity to temperature, i.e., an increase in temperature leads to a decrease in viscosity. One report presents a schematic of the variation of viscosity against a small range of ambient temperatures.2 However, it does not appear to be substantiated with experimental evidence, making it pointless to estimate the change in viscosity. Other researchers report that a solder paste's viscosity can change by as much as 40 Pas from a 1°C change.3 Again this information is meaningless because the parameters at which it was measured are not given. (It must be remembered that solder paste is a non-Newtonian fluid; its viscosity is not only a function of temperature but also of shear rate. Providing data for one is useless without the other.)

Research that does provide enough information to estimate the effect of temperature and shear rate on viscosity describes experiments performed over a range of temperatures 25°, 50°, 75°C for a variety of solder pastes.4 Although its purpose is to investigate the "printability" of solder pastes and their ability to resist slumping at increasing temperatures, the data did allow for the estimation of the pastes' performance at ambient temperatures. Conclusions:

  • Solder pastes exhibit a general trend of decreasing viscosity with increasing temperature.
  • The rate of change is different for each paste.
  • It can be estimated from the data for one solder paste that an increase of 1°C results in a decrease in viscosity of approximately 50 Pas at a shear rate of 1.0 s-1 between 20° and 30°C.

Because such a small change in temperature can affect the printing performance of solder paste, several stencil-printer manufacturers have developed "climate control" systems around their equipment. If the environment of the paste is kept at a constant, the material can be expected to print in a consistent manner. However, during printing, energy transferred to the solder paste is converted via viscous motion into heat that, if built up in the paste roll, might affect viscosity and, consequently, printing performance.

Heat Generation During PrintingTheoretical attempts have been made to develop conceptual models of the stencil printing process.5 To calculate temperature rise in solder paste, it is important to determine where heat is generated; Riemer's model of a Newtonian fluid roll can be used as a basis.6 (This will only provide an approximation for the non-Newtonian solder paste. Nevertheless, the results should give an indication of heat generation.) In the model, the components of paste velocity can be found using the Navier-Stokes equations of fluid flow at low Reynold's numbers. The resulting expressions for the stress tensor in the fluid are evaluated in the usual way, and the heat generation rate, Q, is given as:

69371-th_71069.gif

where σij are the components of the stress tensor, α is the contact angle of the squeegee, l is the contact length of the paste bead on the stencil, d is the solder particle diameter, Θ and r are polar coordinates where the squeegee tip is the origin, V is the velocity of the squeegee, and f(α) is given by:

69371-th_71070.gif

and a being given by:

69371-th_71072.jpg

In equation 1, the particle diameter, d, is introduced as a lower cut-off in the integration over the volume of the paste roll. If it is not introduced, the integrand becomes infinite at the origin, which is acceptable because physically, the suspension cannot flow below d. This consideration obeys the Navier-Stokes equations of motion, i.e., below d, the solution becomes invalid. Similarly, the original solution involves an infinite volume of fluid, which is not the case for the finite volume of the solder paste bead. Equation 1 is only an approximation to the total heat generated.

Thus, from this approximation, the temperature rise δT in the paste is given by:

69371-th_71073.jpg

where ρ is solder-paste density, Cp is the specific heat capacity of the paste at constant pressure and W is the print stroke length.

By substituting typical values for the variables, an estimate of the temperature rise can be determined, e.g.:

l = 0.01 m, W = 0.5 m, V = 0.02 ms-1, μ = 103 Pas, d = 30 x 10-6 m, r = 5 x 103 kgm-3 and Cp = 350 Jkg-1K-1.

69371-th_71075.gifFigure 2. Schematic of the heat generated during stencil printing shows 90 percent is created at the 1 mm-thick outer layer of the paste roll.

The temperature rise, δT, is 2.2 K for a squeegee angle of 60°. The heat generated that prompts the rise is near the edge of the roll, which contacts the stencil surface and the squeegee tip. In fact, from this adaptation of Riemer's model, it can be estimated that 90 percent of the total heat generated is at the 1 mm-thick outermost layer of the paste roll (Figure 2). Because such a temperature rise can have a significant effect on the solder paste's viscosity, it is important to estimate the time scale required for this heat to dissipate. The relaxation time, t, for thermal conduction gives an indication of the heat dissipation:

69371-th_71076.jpg

where k is the thermal conductivity (typically 0.65 Wm-1 K-1) and L is the length over which the heat is dissipated. The relaxation time, when L is 1.0 mm, is approximately 2.9 seconds. The heat generated in the paste roll during a print stroke lasting 25.0 seconds has plenty of time to dissipate into the stencil, squeegee and the ambient air.7

Experimental MeasurementsTo investigate the possibility of a temperature rise (and the validity of Reiner's model), a range of experiments is performed. A brief summary of the printing follows:

  • A semiautomatic printer with a polyurethane squeegee blade* (95 durometer, 60° angle of attack) is set at a speed of 20 mm per second with a load of 6 kg. These parameters are found to produce optimum deposits for a nickel-coated, stainless-steel electroformed stencil (150 μm thick) having a standard quad-flat-pack pattern.
  • The solder paste is a no-clean flux/vehicle system suspending Sn/Pb/Ag-alloy particles of a size range of 25 to 45 μm. The metal content is 90 percent, re-sulting in a viscosity of 750 to 1,000 Pas (based on IPC test methods). Before testing, the solder paste is permitted to reach room temperature over a 24-hour period.
  • To monitor temperature rise in the paste, three K-type thermocouples are attached near the tip of the squeegee. After printing one stroke, no temperature rise is observed; after five strokes in quick succession plus further repetitions at higher speeds, it is concluded that no significant rise in paste temperature occurs during printing. This is in stark contrast to stirring a 500 g jar of the paste for 30 seconds, where a 5°C temperature rise is recorded using the same thermocouples. It must be remembered that the objective is not to accurately measure the temperature of the solder paste, but rather the change in temperature. Thus, the choice of equipment is determined not by accuracy but by resolution.
  • Bearing this in mind (and the poor results from the thermocouples), a thermal imaging camera** is used. The temperature measurement range is -20° to 450°C, with a sensitivity of less than 0.1°C at 30°C. To determine the heat generated in the paste roll, the camera is focused in front of the paste roll to monitor the complete stroke and record both the environmental temperature and the temperature of the paste. After taking the pictures, the data are downloaded so they can be viewed. Points of interest are labeled using the accompanying software. Figure 3 shows the paste separating from the squeegee as it lifts at the end of a print stroke. The top third depicts the squeegee mount mechanism, the middle section is the solder paste bead and the bottom third is the stencil.
  • By monitoring the solder paste just after separation, an indication of the temperature at the squeegee tip where a large portion of the heat is generated is gained. The results indicate that the solder paste could be experiencing a temperature rise during stencil printing. Its temperature at the squeegee tip is slightly higher than the paste on the stencil. There is a more marked contrast between the solder paste and the surrounding environment where the temperature difference is approximately 1.5°C. This thermal rise matches the predicted value from the model and implies that the paste's rheological properties could be changing significantly during printing. In fact, the viscosity could drop approximately 10 percent owing to such a temperature rise. What measures can be taken to counter this effect?

69371-th_71077.gifFigure 3. A thermal image of the stencil printing process.

Minimizing the EffectThere are two options to minimize the effect of heat generation on solder paste viscosity. One is to control the temperature at the source; the second option is to alter the formulation of the solder paste so that it is less susceptible to small changes in temperature.

Research reports the benefits of a heated squeegee.8 By modifying the squeegee mounting system to accommodate a heating element, the user can raise the temperature across the back during the print stroke. The reported benefits are that bead rolling action is improved due to a lubrication layer formed along the blade, and that sticking, stencil snap-back and whiskering are eliminated as a result. Similar benefits are also reported for vibrating squeegees.9

The second option does not require modification of standard printers. The solder paste manufacturers instead have responded to the effect of heat generation by modifying the formulation of the flux/vehicle system. The changes most likely are made to the rheological additives.

To measure the effect of temperature on solder paste viscosity, a rheometer*** with a 40 mm-diameter parallel plate at 0.5 mm gap is used. Several shear rates are investigated, but for comparison purposes, only the 1 s-1 is discussed.4 The temperatures investigated were 10°, 15°, 20°, 25° and 30°C.

69371-th_71078.gifFigure 4. A rheogram of viscosity vs. temperature for a no-clean solder paste. Although it is clear that as the temperature rises viscosity decreases, the rate of change has negligible effect on performance.

Figure 4 displays the results. It is clear that as temperature increases, viscosity decreases. What is reassuring is the rate at which viscosity changes with the temperature rise. The viscosity changes by only 20 Pas per 1°C over the range of 20° to 30°C (under typical printing conditions).

The paste is less susceptible to changes in temperature and is a significant improvement over that in which several shear rates are investigated.4 This would suggest (although not conclusively prove) that solder paste manufacturers have improved their formulations over the last five years so that viscosity is no longer as susceptible to changes in the ambient temperature.

ConclusionA process model for solder paste temperature rise during stencil printing predicts that the paste will experience an increase of approximately 2°C. The model is experimentally validated using a thermal imaging camera system. The temperature rise can seriously affect solder paste viscosity and, in turn, damage printing performance. However, it is shown under rheometric conditions that a typical modern solder paste is not as susceptible to changes in ambient temperature as those developed earlier in this decade. Nevertheless, these slight variations in viscosity may cause significant deterioration in print performance at flip chip dimensions.

  • DEK 260.

** ThermaCAM Infrared Imaging Radiometer from Inframetrics.

*** Reologica StressTech.

ACKNOWLEDGMENTSThis work was funded by EPSRC under grant GR/L 59320. The authors thank Samjid Mannan (for his work under grant GR/H 39239), Mark Currie and Thay Ainal for conducting the experiments and Multicore Solders Ltd. and Inframetrics for providing samples and apparatus.

REFERENCES1 N.N. Ekere, S.H. Mannan and M.A. Currie, "Solder Paste Printing Process Modeling Map," Proceedings of Japan International Electronic Manufacturing Technology Symposium, December 4 through 6, 1995.

2 C.A. MacKay, "Solder Creams and How to Use Them," Electronic Packaging and Production, February 1981, p. 116-133.

3 P.S. Maiso and B. Bauer, "Statistical Process Control in Solder Paste Manufacture and Use," IPC Technical Review, September 1990, p. 20-26.

4 M.J. Mindel, "Solder Paste Rheology as a Function of Temperature," Proceedings of SMI '91, San Jose, Calif., p. 490-495.

5 N.N. Ekere, E.K. Lo, S.H. Mannan and I. Ismail, "Application of Ink Screen Models to Solder Paste Printing in SMT Assembly," Journal of Electronics Manufacturing, Vol 3, 1993, p. 113-120.

6 D.E. Riemer, "Analytical Engineering Model of the Screen Printing Process: Part 1," Solid State Technology, August 1988, p. 107-111.

7 N.N. Ekere, S.H. Mannan, M.A. Currie and I. Ismail, "Flow Processes of Solder Paste During Stencil Printing for SMT Assembly," Journal of Materials Science: Materials in Electronics, Vol 6, 1994, p. 34-42.

8 M. Curtin, "Getting Better Print Quality," SMT, February 1994, p. 34-37.

9 H. Da and N.N. Ekere, "The Study of Solder Paste Response Under Sinusoidal Vibration," 21st IEMT Symposium, U.S.A., 1997, p. 37-43.

10 M.H.A. Riedlin and N.N. Ekere, "Rheological Techniques for Measuring Normal Stress Differences of Solder Paste," 21st IEMT Symposium, U.S.A., 1997, p. 178-84.

M.H.A. RIEDLIN and N.N. EKERE may be contacted at the Electronics Manufacture and Assembly Group Department of Aeronautical, Mechanical and Manufacturing Engineering, University of Salford, Salford M5 4WT, UK; 0161 295 3128; Fax: 0161 295 5575; E-mail: n.n.ekere@amme.salford.ac.uk.

Share

Print



Copyright © 2019 I-Connect007. All rights reserved.