The Driving Force Behind Whisker Growth
by Reima Lahtinen and Tom E. Gustafsson

Whisker growth on tin is a well-known phenomenon. But whiskers grow on other metals, too. Whisker growths on zinc, silver, cadmium, indium, aluminum, lead, and even gold have been reported¹.

Figure 1: A hot-dip galvanized support structure with extensive whisker growth on the surface.

Whisker growth on tin plating has been linked to the formation of copper-tin intermetallics in the plating process². Growth on silver has been reported to be linked to a very low concentration of hydrogen sulphide in surrounding areas³; growth on gold plating was related to the presence of rubidium in the film⁴; growth on zinc was reported by Lindborg in 1975⁵. In his paper, Lindborg suggested that to initiate zinc whisker growth, a minimum stress must be present in the zinc coating. Any other irregularities causing micro-stresses in the coating, dislocations, interstitials, or vacancies have no effect on growth.

In a paper published by Sugiarto et al.⁶ in 1984, the zinc whisker growth was linked to micro-stresses caused by brightener residues left in the electrodeposited coating. Films produced without brighteners showed no whisker growth. Passivation treatment slowed down whisker growth but did not prevent it. To grow whiskers, the specimens were placed in an oven for 24 hours at a temperature of 170°C.

In their presentation at the IPC/JEDEC Conference in 2004, Reynolds and Hilty⁷ showed that zinc whisker and the plated film had similar fine grained structure. The zinc was presumed to have diffused over long distances so no thinning of the film was observed under the whisker. Intermetallic compounds appeared to have no role in whisker growth.

It has been assumed that hot-dip galvanized (zinc) coatings (HDG) are immune to whisker growth. In their paper, Brusse and Sampson⁸ mention that they had seen one report on presumed whisker growth on HDG coating.
In this report, SEM (scanning electron microscope) investigations on whisker growth on hot-dip galvanized zinc coatings are presented.

The Specimens
Two specimens were investigated. Figure 1 shows a hot-dip galvanized support structure. Extensive whisker growth is seen on the surface. The surface appearance varied from a shiny spangled to a dull gray appearance.
The support structure is a U profile with dimensions of 30 x 30 x 30 mm. The longest whiskers on the surface are more than 3mm in length. The support structure had been in use for several years in a power plant.

Figure 2: An EDS-spectrum showing heavy whisker growth.

The second specimen investigated was a bend of pipe. The surface appearance was dull gray. This specimen showed extensive whisker growth inside the pipe. The bend of pipe had been on a warehouse shelf for several years. The whiskers inside the pipe were also about 3mm in length.

SEM/EDS investigations of the surfaces: A sample from the support structure for SEM/EDS analysis was cut at the point were both shiny and dull surfaces appeared next to each other.

This area is seen on the top of the photograph in Figure 1. A sample from the pipe was cut from the middle of the bend for SEM/EDS analysis of the inside surface.

In EDS-analysis (energy dispersive spectrum) of the inside surface of the pipe, beside normal elements associated with HDG zinc coating, both chlorine (Cl) and sulfur (S) were detected.

In the analysis of the support structure, both chlorine and sulfur were detected also. In the area near the whisker root and in areas where heavy whisker growth was observed, the chlorine peak was higher than the sulfur peak.

In areas where whisker growth is not as extensive, the chlorine peak was lower than the sulfur peak. In the dense whisker area, small amounts of potassium were detected at the root of a long whisker. The EDS-spectrum of this spot is shown in Figure 2.

Figure 3: An SEM photograph of a whisker that has just started to grow.

Figure 3 is an SEM photograph of a whisker that has just started to grow on the support structure. These kinds of whiskers were found in between the shiny and dull surfaces. The whisker has a "hat," which is the original surface of the zinc coating. In EDS-analysis, sulphur andchlorine were found on the hat. In the areas around these small whiskers, sulfur was not found and chlorine existed only in minor quantities.

In our investigations on whisker growth on electroplated zinc coating, these same elements were detected. This information will be detailed in an upcoming article.

Figures 4 and 5 show the difference in cross sections of the shiny and dull surfaces of the support structure. Some large crystals can be seen on the cross section of the shiny area. In the dull area there are no distinguishable features. The area of heavy whisker growth contains less "voids," also.

Factors That Cause Whiskers Growth
The effect of crystal size and the coefficient of thermal expansion: The support structure in Figure 1 has a surface where there are (bright) large zinc crystals and (dull) small zinc crystals. The dull areas grow long whiskers, but the bright areas do not. Incompatibility between coefficients of thermal expansion (CTE) has been suggested as one factor behind whisker growth.

The International Zinc Association⁹ provided the CTEs listed in Table I. The figures show that thermal expansion along the c-axis is about four times that of the a-axis.

In the sample from the support structure, the dull areas are made of small crystals and the bright areas of large crystals. The thermal stresses caused by fluctuation of temperature in the dense environment of small crystals have no room for relaxation and the probability for whisker growth is much greater. The anisotropic CTE of the zinc crystal has the possibility to create large pressures between individual crystals in the dull areas.

For this phenomenon, we propose the name "intergranular fretting." In the pipe sample, the coating is made altogether from small dark crystals, and the whisker growth has happened inside where compressive stresses are more concentrated.

The HDG coatings contained small amounts of aluminum as an alloying element, as seen in the EDS spectrum in Figure 2. Aluminum decreases the crystal size. In the past, hot-dip galvanizers used a layer of molten lead in the tank bottom and the zinc was saturated with lead. Lead increases the crystal size. From the past to the present, the average crystal size of HDG zinc coatings decreased and HDG zinc coatings started to grow long whiskers.

The sources of temperature differences: In the case of the bend of pipe, natural temperature fluctuations (summer to winter) create large temperature differences. But also inside air-conditioned spaces, where the support structure had been, temperature fluctuations were equally expected.

Figures 4 and 5: SEM photographs show the difference in cross sections of the shiny and dull surfaces of the support structure. Some large crystals can be seen on the cross section of the shiny area. In the dull area, there are no distinguishable features.

The normal variation around the air conditioning set point is ±2°C. The anisotropic nature of the CTE of zinc crystals creates, with the above temperature fluctuation, along the c-axis a pressure that effectively corresponds temperature fluctuation of ±8° C. A temperature difference of this magnitude is enough to push a whisker out. However, the rate of temperature change in air-conditioned spaces is usually slow. Much faster temperature change is possible if the source of energy is different.

Photochemical reactions as a possible heat source: Zinc compounds have photoluminescence properties. Both zinc sulfide and zinc oxide absorb and emit photons. Impurities hinder photoluminescence. The absorbed energy is then dissipated as heat. This kind of local heating in combination with the anisotropic CTEs can create high local stresses that lead to whisker growth.

The fluorescence of zinc oxide is hindered if impurities are present. Omichi et al.¹⁰ have made zinc oxide films using zinc chloride as the starting material. The zinc oxide films showed fluorescence below 180°K but not above. The small amount of chloride left in the film prevented the fluorescence. Various cations are used to activate ZnS, the most common being copper. We did not find copper in the X-ray analysis, but the whisker root contained potassium. Potassium is the element of group Ia and copper is an element in group Ib. A possible light source to excite the compounds is always present. In the investigated cases, fluorescent lighting that furthermore is tuned towards daylight (blue end) radiation, can be the light source and power generator for whisker growth. In the case of the bend of pipe, however, the inside is unlikely to be greatly illuminated.

The absorption of photons takes place mainly in the UV region of the spectrum, and the fluorescent lighting spectrum begins just outside this. There is some overlapping with zinc sulfide, though.

The photon absorption can obviously not be the prime source for the energy supporting whisker growth in the investigated cases, but it may offer extra energy to boost whisker growth, especially when zinc sulfide is present.

Summary
In our investigations, the following preconditions, in the growth of long zinc whiskers on hot dip galvanized zinc coatings, were found:

• Small crystal size, corresponding to dull HDG surface;
• chlorine and sulfur must co-exist, and the amount of chlorine must be greater;
• temperature fluctuations are required, but they can be small and slow (intergranular fretting);
• and compressive stresses in the coating.

The driving force behind whisker growth is related to the anisotropic coefficient of thermal expansion of zinc crystals. The energy for the whisker growth may, in some cases, rely on the photochemical properties of zinc compounds.

Acknowledgement
The authors express their gratitude to Teollisuuden Voima Oy (Finland) for providing the samples and financial support to this investigation.

References

1. http://nepp.nasa.gov/whisker/other_whisker
2. Galyon, G.T., "A History of Tin Whisker Theory: 1946 to 2004," IBM eSG Group, Poughkeepsie, New York.
3. http://nepp.nasa.gov/whisker/other_whisker/silver/index.htm
4. http://nepp.nasa.gov/whisker/other_whisker/gold/ index.htm
5. Lindborg, U., "Observations on the Growth of Whisker Crystals from Zinc Electroplate," Metallurgical Transactions A–Physical Metallurgy and Material Science, 6(8):1581-1586; 1975.
6. Sugiarto, H. et al., "Studies of Zinc Whiskers Formation and Trowth from Bright Zinc Electrodeposites," GEC Research Laboratories, Hirst Research Centre; 1984.
7. Reynolds, H.L. and Hilty, R., "Investigations of Zinc Whiskers Using FIB Technology," presented at IPC/JEDEC Lead-Free North American Conference Boston; December 3, 2004.
8. Brusse, J. and Sampson, M., "Zinc Whiskers: Hidden Cause of Equipment Failure," IT Pro; November/December 2004.
9. International Zinc Association: Frequently Asked Questions About Zinc, http://www.iza.com.
10. Omichi, K. et al., "AP-HVPE Growth of ZnO with Room Temperature Ultraviolet Emission," J. Mater. Chem., 11:3158-3160; 2001.

For more information, contact Reima Lahtinen at (e-mail) reima.lahtinen@vtt.fi.