POLITECNICO DI MILANO

Faculty of Materials Engineering

Masters of Science in Materials Engineering
(Academic Year 2012– 2013)

Electrodeposition of Gold and Gold Alloys for bonding

In MEMS

Advisor : Prof. Luca MAGAGNIN

Master thesis by : Fevzi ERSEN 751668

1. Gold and Gold Alloy in Microelectromechanical System

The development of microelectromechanical systems (MEMS) industry gave inputs to investigate

new materials due to importance of thin films. Most of the electronic devices fabricated in MEMS
use materials such as silicon and many other thin films, which are not well characterized regarding
their electrical properties relevant to MEMS application. Little research has been conducted on
microscale gold, because it has not been considered earlier as an engineering material[1].

Thin gold films are used in many MEMS devices because of high electrical conductivity. Some of
the applications include radio-frequency (RF) MEMS which can operate at gigahertz frequencies,
allowing for large bandwidth and extremely high signal-to-noise ratios and in inertial MEMS to
increase mass of proof-mass for achieving higher sensitivity in MEMS accelerometers. These films
can also be used in variable capacitor, chemical and biological sensors, optical detectors etc.

Some basic requirements for Au films used in MEMS applications are as follows[2]:

1. High conductivity.
2. Film–substrate adhesion should be strong enough to prevent cracking, delamination and
spallation.
3. Residual stress developed should be minimum to prevent deformation of MEMS structure.
4. Films should be stable within a wide range of temperature (from below zero to several hundred
degree Celsius).
5. Films should be resistant to surface wears, corrosion and oxidation.

1.1 Electrodeposition of Gold and Gold Alloys in MEMS.

Electrochemically deposited gold has satisfied many of the demands of the electronics industry.
Gold has the third best electrical and thermal conductivity of all metals at room temperature. It has
high ductility and excellent wear resistance, which are important for electrical contacts. The
inertness of gold prevents the formation of insulating surface oxides (as compared to metals like
aluminum), so insulating surface oxide does not form on its surface. Gold is an excellent metal for
wire bonding integrated circuits (ICs). Gold wires can be bonded to pure, soft gold pads by
thermocompression bonding (300–400˚C at high pressure to form a weld) or thermosonic bonding
(150–200˚C with ultrasonic energy to form a weld)[3].
Electroplating is an electrodeposition process for producing a dense, uniform , and adherent coating,
usually of metal or alloys, upon a surface by the act of electric current. The coating produced is
usually for decorative and/or protective purposes, or enhancing specific properties of the surface
[4]. The surface can be conductors, such as metal, or nonconductors, such as plastics. Electroplating
products are widely used for many industries, such as automobile, ship, air space, machinery,
electronics, jewelry, defense, and toy industries. The core part of the electroplating process is the
electrolytic cell (electroplating unit). In the electrolytic cell (electroplating unit) a current is passed
through a bath containing electrolyte, the anode, and the cathode. In industrial production,
pretreatment and post-treatment steps are usually needed as well [5].

The electrodeposition of pure gold is an enabling technology for wafer bump and wirebonding
interconnect applications. Most conventional pure gold electroplating processes used in the
electronics industry are cyanide based [6]. Due to the toxicity of the free cyanide formed during
electrolysis, as well as cyanide’s incompatibility with positive photoresists, sulfite-based gold
processes have been the traditional alternative for wafer applications. However, traditional sulfite-
based processes have suffered from issues with solution stability and the necessity for annealing to
achieve the desired deposit hardness.

Soft gold electroplating is mainly used for wire bond and surface mount applications, which involve
a large number of bump bonds on a chip. The gold should be soft enough so that the bumps can be
easily deformed to accommodate small variations in thickness. Cyanide based electrolytes are the
most extensively used in the industry. The gold (I) cyanide complex Au(CN)-2 has a stability
constant of 1039, which makes the solution very stable, and shifts the standard reduction potential of
gold (I) from 1.71V to –0.61V (SHE). However, in recent years, the health and safety concerns
regarding the use of cyanide have overwhelmed the benefits it brings. In addition, free cyanide
containing solutions can be incompatible with positive photoresists used in the microelectronics
industry because they attack the interface between the resist film and the substrate, lifting the resist
and depositing extraneous gold under the resist. Therefore, the demand for cyanide-free gold
electroplating processes has increased significantly [6].

The workpiece to be plated is the cathode (negative terminal). The anode, however, can be one of
the two types: sacrificial anode (dissolvable anode) and permanent anode (inert anode).The
sacrificial anodes are made of the metal that is to be deposited. The permanent anodes can only
complete the electrical circuit, but cannot provide a source of fresh metal to replace what has been
removed from the solution by deposition at the cathode. Platinum and carbon are usually used as
inert anodes. Electrolyte is the electrical conductor in which current is carried by ions rather than by
free electrons (as in a metal). Electrolyte completes an electric circuit between two electrodes [7].
Upon application of electric current, the positive ions in the electrolyte will move toward the
cathode and the negatively charged ions toward the anode. This migration of ions through the
electrolyte constitutes the electric current in that part of the circuit. The migration of electrons into
the anode through the wiring and an electric generator and then back to the cathode constitutes the
current in the external circuit. The metallic ions of the salt in the electrolyte carry a positive charge
and are thus attracted to the cathode. When they reach the negatively charged workpiece, it provides
electrons to reduce those positively charged ions to metallic form, and then the metal atoms will be
deposited onto the surface of the negatively charged workpiece.

Electrodeposition or electrochemical deposition (of metals or alloys) involves the reduction of metal
ions from electrolytes. At the cathode, electrons are supplied to cations, which migrate to the anode.
In its simplest form, the reaction in aqueous medium at the cathode follows the equation [5]:

Mn+ + ne-  M

with a corresponding anode reaction. At the anode, electrons are supplied to the anions, which
migrate to the anode. The anode material can be either a sacrificial anode or an inert anode. For the
sacrificial anode, the anode reaction is [5]:

M  Mn+ + ne-

Gold exists primarily in the +1 and +3 oxidation states. The standard potential versus the normal
hydrogen electrode (NH.E) for a variety of Au(I) and Au(III) complexes is given in Table 1.1 [3].
The most important ion for electrodeposition is [Au(CN)2]-. The stability of the Au(I) cyanide
complex is reflected in the shift of the reduction potential for Au(I) from 1.71V (aqua complex) to -
0.611V (cyanide complex). The stability constant for [Au(CN)2]- is 1039. Two other gold complexes
are of interest for electrodeposition, gold sulfite (K=1010) and gold thiosulfate (K=1028) [3].

Au+ + 2CN-  [Au(CN)2]- K = 1039

Au+ + 2S2O3-2  [Au(S2O3)2]-3 K=1028
Au+ + 2SO3-2 [Au(SO3)2]-3 K=1010
The potentials for the reduction of Au(III) to Au and Au(III) to Au(I) are important because Au(III)
can also be used as the source of gold in baths in place of Au(I), and Au(III) can also be formed at
the anode during plating through the oxidation of Au [3].

Table 1.1: Standard Reduction Potentials for Gold Ions (V vs. NHE)

Gold film is a polycrystalline film. The microstructural characteristics of polycrystalline films

control their properties, performance, and reliability in applications in wide variety of electronic,
magnetic, photonic, chemical, and micromechanical devices and systems. Gold films have
columnar structure. The knowledge of diffusion process operating in gold film is important for
application and design of microelectronic devices, fabrication and reliability of the thin-film
package. Adhesion of gold film on silicon is poor, so a layer of chromium (Cr) is used before gold
deposition, which enhances the adhesion of the film with substrate. Owing to this discrete layer of
Cr, gold film has poor thermal stability because of the natural tendency of Cr to diffuse to the film
surface at higher temperatures [1]. Besides Cr diffusion the other main factor controlling the
resistance of the film is its grain structure, which in turn is controlled by the deposition parameters
like current density, electrolyte temperature, thickness of deposited gold film and the annealing
temperature. The variation of stress and resistivity of electroplated gold film with various deposition
parameters such as current density, deposition temperature of the electrolyte, thickness of the gold
film and annealing temperature.

Gold-tin eutectic alloy solders are commonly used for microelectronic device packaging, because of
their excellent combination of mechanical and thermal properties. The industry standard contains 70
at% gold and has a eutectic temperature of 280°C [8]. A possible alternative could be the Sn-rich
eutectic composition that is present in the gold-tin system (Figure 1.1). The second eutectic exists at
6.8 at% Au and a temperature of 217°C. This composition may also be a candidate Pb-free solder
for microelectronic applications.

Figure 1.1: Au-Sn binary phase diagram [9]

According to Lee and Chuang [10], the Sn-rich eutectic has suitable properties to be used for
packaging applications, and its use would significantly reduce material costs. Department of
Chemical and Materials Engineering in University of Alberta has developed a simple solution for
the electrodeposition of Sn; this solution consists of tin chloride and tri-ammonium citrate. Gold
particles (submicron size) can be added to this solution to form a suspension, which can be used to
electrodeposit Sn-rich, Au-Sn eutectic solder films, with uniform thickness and composition [8].
The films in this study were deposited onto Pt metalized silicon wafers from the aforementioned Sn
solution with suspended Au particles, using a pulse plating power supply.

The Au particles are incorporated between the Sn grains. Solder joint integrity is dependent on the
formation and distribution of brittle intermetallic phases. As such, Sn rich solder films were aged
and/or reflowed to study intermetallic formation. After aging or reflow, the two eutectic phases
form in the film [8]. The study will focus on the effects of plating conditions on the morphology
and composition of Sn rich, Sn-Au solder films and intermetallic formation during aging and
reflow.

The residual stress of gold films deposited on silicon wafers is important for microsystems and
microelectromechanical systems (MEMSs). Gold is an excellent structural and contact material for
MEMSs because it is difficult to oxidize. The residual stress of gold films annealed at different
temperatures plated from a neutral-pH cyanide bath was measured. Recrystallization of the gold
occurs at high temperature, which relieves the stress due to mismatch of the coefficient of thermal
expansion (CTE) between the gold and the silicon substrate. Normally, the gold achieves a stress
neutral point at the highest temperature. The CTE mismatch between the metal and the substrate
caused tensile stress in the gold during cooldown. Stresses on the order of 50, 200, and 350 MPa
were observed at anneal temperatures of 100, 250, and 400˚C [11].

Residual stress is a particular problem for free-standing gold structures where flatness is desired,
such as in X-ray lithography masks. The stress films deposited from a neutral pH cyanide bath was
found to be 50 MPa tensile whereas the addition of 25 ppm arsenic changes the stress to 30 MPa
compressive. Pulsed current reduced the stress to <5 MPa compressive. In other applications, the
stress gradient in gold films deposited at different conditions was used to build precise-curvature
cantilever beams for microswitches [12].

It has also been found that the residual stress of the plated gold can be controlled in non-cyanide
baths. Low-stress deposits are of particular interest for X-ray lithography masks because they must
be thin, flat, and made from high-atomic-number elements [3].
2. TYPICAL DIRECT CURRENT (DC) PLATING BATHS:

The cyanide-based baths are divided into three classifications [3];

1. The Alkaline Gold Cyanide Bath (pH>8.5)

2. Acidic, buffered baths (pH between 1.8 and 6)
3. Neutral, buffered gold cyanide bath (pH between 6 and 8.5)

- The fourth group of baths, noncyanide plating baths.

Gold readily forms alloys with other metals during the deposition process substrate onto substrate
which it was plated. Gold or its alloy deposits become hard or soft, dull or bright, depending on the
impurities and deposition conditions.

The materials of plated gold in use for this purpose can be classified into two categories: soft gold
and hard gold. Soft gold is used for metalizing bonding pads and fabricating micro bumps on silicon
IC chips and ceramic packaging boards, while hard gold is used as a contact material on electrical
connectors, printed circuit boards, and mechanical relays. Electrolytic methods are available for
plating both soft gold and hard gold.

Electroplated hard gold is the most commonly used contact material for high reliability electrical
connectors and relays. The major factor determining the hardness of this material is its small grain
size. On the other hand, with the recent advances in miniaturization of electronic components and
devices, dimensions of electrical contact areas are expected to reach, and may become even smaller
than, the grain size of the hard gold that is 20–30 nm. When this ultimate miniaturization is realized,
the bulk hardness and the resistance to mechanical wear of the hard gold will no longer be retained.
To meet the requirement of this trend, we considered it meaningful to have available a technology
of electroplating an amorphous gold or gold alloy, because physical properties of amorphous
materials in general are known to be independent of their size or physical dimensions.

The properties of hard-gold films of such small dimensions are expected to deviate significantly
from those of bulk hard gold. On the other hand, mechanical properties of amorphous metals and
alloys in general are known to be independent of their size because of the absence of crystal grains
in such materials. Thus, to be prepared for the future need of an amorphous electrical contact
material. Investigation on the possibility of electroplating amorphous Au–Ni alloy from a bath
prepared by adding a gold salt into a bath that is already known to deposit an amorphous alloy such
as Ni–W. This approach was successful in developing a process to electroplate amorphous Au–Ni
alloy with a hardness value greater than twice that of the conventional hard gold without adversely
affecting the electrical contact resistance. Using a similar approach, we also developed a process for
electroplating amorphous Au–Co alloy films.

At pH<8, the predominate form of cyanide is HCN(g), which evolves as a gaseous product. HCN
gas is highly toxic, and its evolution from plating baths is a health concern. Appropriate ventilation
is required for all cyanide baths. The presence of significant concentrations of free cyanide in the
bath, such as in the case of alkaline baths, is also a health risk because accidental cyanide ingestion
or injection can occur [3].

The presence of free cyanide in the bath is a major consideration in the selection of the anode and
metal ion replenishment method. In the presence of free cyanide, gold metal can be used as a
consumable anode because it can be electrochemically oxidized forming the gold cyanide complex.
In the absence of free cyanide, gold is not oxidized to any appreciable extent, so it acts like an inert
electrode. The ramifications of this will be included for each bath [3].

2.1. Alkaline Cyanide Baths

Alkaline cyanide baths operate at high pH and with an excess of free cyanide. Potassium cyanide is
generally used as a source of the free cyanide in these baths. Since there is appreciable free cyanide
present, the cyanide released during the deposition reaction does not significantly alter the cyanide
concentration in the bath. Thus the Nernstian shifts in the reduction potential of Au(I) cyanide due
to accumulation of free cyanide (even in the vicinity of the electrode surface) do not occur [3].

Cathode: [Au(CN)2]- + e-  Au + 2CN-

Anode: Au + 2CN-  [Au(CN)2]- + e-
Full cell: No change to plating solution

The main changes are appeared in the bath due to oxidation or reduction side reactions, drag-out of
plating salts, and drag-in of water or impurities. The free cyanide promotes the corrosion of the gold
anode (forming [Au(CN)2]-), increases the throwing power, and improves the conductivity. Free
cyanide also retards the codeposition of some metals because the stability of their cyanide
complexes shifts their reduction potential to values more negative than those used during deposition.
The absorption of atmospheric carbon dioxide increases the concentration of carbonate in the
solution as a result of the reaction between cyanide and CO2. The carbonate can help to stabilize the
pH and improve throwing power slightly [3].

Nonsoluble, dimensionally stable anodes, like platinized titanium or stainless steel, are sometimes
used. Oxygen gas and hydrogen ions are the dominant products at the insoluble, dimensionally
stable anodes,

Cathode: 4[Au(CN)2]- + 4e-  4Au + 8 CN-

Anode: 2H2O  4H+ +O2+ 4e-
Full cell: 4[Au(CN)2]- + 2H2O  4Au + 8 CN- + 4H+ +O2

The hydrogen ions produced in anode equation and shown in equation of full cell for the full cell
can either be neutralized by hydroxide (lowering the pH of the solution) or form HCN (g) and
evolve from the bath as a gas depending on the operating pH. Thus the pH and the gold cyanide ion
concentration of the bath will drop as current is passed, as shown by the two half reactions. Both
these effects must be reversed through the addition of plating salts and KOH. The addition of KOH
and KAu(CN)2 as a replenishment for the right-hand side of full cell equation will result in the
accumulation of KCN (in the form of K+ and CN- until the KCN solubility product is reached) in
the bath. The buildup of KCN is somewhat mitigated by drag-out. An insoluble anode is convenient
because the cathode current distribution is reproducible and consistent over long periods of time.
Also the anode does not have to be replaced as it corrodes. Thus, anode maintenance is replaced by
bath maintenance [KOH and KAu(CN)2 additions]. The cathodic current efficiency for the
deposition of the metal from alkaline baths can be as high as 90–100% with adequate gold content
in the bath and sufficient agitation [13].

2.2. Acid Cyanide Baths

The nonalkaline, cyanide-based baths contain other metals and normally use a citrate buffer. They
are operated around pH 4. Low-pH buffered baths were consumed in wide application throughout
the electronics industry for contact surfaces, corrosion protection, bonding surfaces, and special
electroforming. The pH of the baths allows the use of photoresist and other polymers.
Anodes in these baths are usually platinized titanium, gold, or gold-plated platinized titanium [3].
The gold-plated electrodes are found not to dissolve readily at low current densities because of the
low free-cyanide concentration. This, however, is not the case at high current densities and low
solution volume where appreciable transient concentrations of free cyanide can build up and lead to
the oxidation and dissolution of a pure gold anode [3]. Stainless steel and carbon anodes are found
to introduce contamination, and their use is discouraged. The use of insoluble anodes requires that
plating salts be added to the bath to replenish the gold content [e.g., KAu(CN)2] and maintain pH
control. Although K+ removal and pH control can be accomplished by an ion exchange method,
drag-out and salt addition are the usual method of bath maintenance [14].

The codeposition of other metals with gold is easily compatible with acid cyanide baths, resulting in
marked changes in the physical properties (especially hardness) of the deposits. Hard gold is a great
milestone for electronic components where low contact resistance, pore-free deposits, wear
resistance, and chemical inertness are functional requirements. Such applications include the
multiple-insertion electrical contacts found on printed wiring board contacts or spring contacts [3].

Bright, hard deposits can be produced from alkaline and acid cyanide baths containing cobalt,
nickel, indium, silver, arsenic, or cadmium.

A typical range of concentrations is given below [13]:

KAu(CN)2 12- 15 gL -1
Citric Acid 90-115 gL -1
Cobalt (added as acetate or sulfate) 0.07- 0.1 gL -1
pH (adjust with KOH) 3.6-4.7
Temperature 40-65˚ C

-1
To establish the proper pH, KOH is added to the bath. A concentration of 50 gL KOH will
produce a bath of about pH 4.0. Baths that produce pure gold deposits often operate at close to
100% current efficiency [13]. The high current density used in high-speed baths can produce a
temporary buildup of free cyanide in the bath that may lead to the corrosion of a gold anode or even
the slow oxidation of a platinum anode. It has also been found that [Au(CN)4]- can be formed at the
anode, especially under high current and high cobalt concentrations at a platinum anode.
One of the limiting factors for the current density is the rise in pH in the vicinity of the electrode
during the reduction process, through either gold deposition or hydrogen gas production [3].

At higher pH or higher temperature, the cobalt and carbon contents of the deposit decrease. An
increase in pH above 5.5 virtually eliminates the cobalt from the deposit. This increase
demonstrates the ability of acid cyanide baths for codeposit metals with gold. Lower current
densities (smaller rate of consumption of protons at the cathode) or higher mass transport (higher
rate of delivery of protons from the solution) will mitigate the rise in pH in the electrode boundary
layer during deposition. Thus the higher the solution agitation, the greater is the allowable current
density. The cobalt or nickel ions are known to codeposit in the gold in at least two different forms,
the cyano complex and as a substitutional alloy. The increase in hardness is due to the grain refining
effect of the codeposited cyanocomplex. The porosity of the deposited gold also decreases with
higher mass transport and lower current density.

2.3. Neutral Cyanide Baths

Nonalkaline baths exhibit a wider range of physical and chemical properties than possible at higher
pH [3]. The hardness is the result of a reduction in the grain size of the gold caused by the codeposit
increasing the rate of grain nucleation. It was also noted that the amount of codeposited metal (e.g.,
cobalt or nickel) decreased to nearly zero at pH>5.5. Although the metals used to harden gold are
not readily codeposited from neutral pH baths, it has been found that hard gold (as well as soft gold)
can be formed in neutral baths by the careful selection of the plating parameters without the use of
additives. In the neutral baths the gold salt used is the same as in alkaline and acid baths,
KAu(CN)2. The neutral pH and absence of free cyanide make these formulations the preferred ones
for use with photoresist and other polymeric materials.

The neutral baths produce pure, soft gold at relatively low current densities, typically 2–5 mA cm-2
and 60–70˚C. The phosphate salts serve only as the supporting electrolyte and pH control, so that
strict control of their concentration is not critical. The phosphate buffer can be used for acid,
neutral, or basic conditions with pH adjustment carried out by the addition of KOH or H3PO4.
Control of pH, temperature, and gold concentration is critically important.
Additive-free hard gold (AFHG) can be produced from neutral baths using phosphate. The
advantages of AFHG baths over metal hardened baths are [15];
1. No need to control additive concentration
2. High current efficiency
3. Greater tolerance for heavy metal impurities due to the insolubility of many metal phosphates
4. High ductility and thermally stable contact resistance for electronic components.

However, AFHG baths are highly susceptible to bath contamination. Small amounts of impurities,
such as nickel, cobalt, or iron ions, can cause large changes to the physical properties deposited
from the bath. It can be difficult to clean the plating bath of the unexpected impurities. The AFHG
bath was shown to have superior tensile strength, total elongation (high ductility), and more stable
contact resistance than comparable deposits from a cobalt-hardened acid bath. However, a small
increase in contact resistance was found when the deposit was heated to>100˚C due to the surface
segregation of AuCN [16].

The use of a gold anode is usually an unacceptable solution to the Au(III) problem because of the
anode maintenance and changing cathodic current distribution as the anode corrodes and changes
shape. The buildup of Au(III) is somewhat mitigated by the dilution effects due to drag-out. Other
possible solutions to the Au(III) problem are to use a RuO2/TiO2 (often called dimensionally stable
anode, DSA), adjust the current density to correct for the lower current efficiency, or chemically
treat the Au(III) by reduction (e.g., hydrazine) or removal (activated carbon).

2.4. Noncyanide Baths

The Au(I) cyanide complex is by far the most important for electrodeposition. However, there are
several approaches for the use of [Au(CN)2]- which have stimulated the investigation and
commercialization of other gold complexes in plating baths. The stability of the gold cyanide
complex causes the reduction potential to occur at very negative potentials, resulting in the
coreduction of hydrogen ions, which lowers the plating efficiency and makes the development of
electroless plating baths difficult. The release of free cyanide during the reduction of [Au(CN)2]-
can be incompatible with positive photoresists used in the microelectronics industry. The operating
parameters of commercial gold sulfite baths have been presented with special attention given to the
effect of plating temperature and current density on residual stress in the deposited metal. Lastly,
the health and safety of workers and the environmental impact of the wide-scale use of cyanide are
a concern. But, Alkaline sulfite baths can also be used without the use of thallium, which itself is a
health hazard [3].

Although Au(I) sulfite has been used in commercial baths, the complex is susceptible to
disproportionation, forming Au(III) and metallic gold. This spontaneous decomposition of the bath
has led commercial baths to use proprietary stabilizing additives. The gold sulfite complex
Au(SO3)2-3 is the most common alternative to cyanide. This complex has a stability constant of
1010,much less than that of the gold cyanide complex. Gold (I) sulfite is came across some problems
particularly at pH<7 where sulfite transforms to bisulfite. Therefore, stability is always an issue for
sulfite-based gold electroplating solutions, and stabilizing additives are usually necessary. Gold (I)
thiosulfate has a stability constant which is between that of Au(I) sulfite and Au(I) cyanide. Gold (I)
thiosulfate complex Au(S2O3)2-3is another alternative to cyanide. Its stability constant (1028) is
between cyanide and sulfite complexes and it is stable in weakly acidic solutions. However,
thiosulfate ions themselves are not stable without the presence of free sulfite. They undergo
disproportionation and produce sulfite ions and free sulfur, which precipitates.

The Au(I) thiosulfate complex is stable in weakly acidic solutions because of the low bath stability
for thiosulfate. The dominant species in mixed-complex baths, S2O2-3 and S2O2-3, at neutral or basic
conditions have been reported to be Au(S2O3)(SO3)-3 [17]. The thiosulfate-only complex is
dominant at low pH. The existence of the mixed ion complex was suggested in previous studies
based on the deposition potential. It was also reported that the results from the sulfide-only bath are
consistent with a higher stability constant (1027), rather than the commonly assumed value of 1010
[16].

The effect of mixed thiosulfate–sulfite electrolytes produce soft gold deposits suitable for forming
gold “bumps” on ICs and in electronic interconnections. The Vickers hardness of the
electrodeposited gold was 80 kg mm-2 in the as-deposited state and 50 kg mm-2 after annealing. The
optimum bath with thallium was composed of the following formulation [16]:

NaAuCl4.2H2O 0.06 M
Na2SO3 0.42 M
Na2S2O3.5H2O 0.42 M
Na2HPO4 0.03 M
+
TI (added as TI2SO4) 0.03 M
pH 6
Temperature 60˚C
Current density 5mAcm-2

The mixed sulfide–thiosulfate plating bath was used to produce soft gold deposits for gold bumps
and wire-bonding applications. Bright gold deposits were obtained at a cathodic deposition
efficiency (current efficiency) of greater than 99%. The purity of the gold deposits was greater than
99.99% and had a hardness between 60 and 80 Knoop. The bath composition is as follows [3]:

Gold as metal 10 g L-1

Sulfide 70 g L-1
Thiosulfate 70 g L-1
Additives 10 g L-1
pH 5
Temperature 50˚C

The deposition of gold from the iodide–thiosulfate bath suggests that the chemical reaction
preceding the electron transfer step is the dissociation of the [Au(S2O3)2] -3
to form [Au(S2O3)2] - and S2O3-2.

In each of the studies involving a mixed-thiosulfate bath (sulfite–thiosulfate and iodide–thiosulfate),

it was observed that the mixed-salt gold complex is more stable and harder to reduce than either of
the single-salt complexes. That is, the Au(I) sulfite–thiosulfate complex is reduced at potentials
more negative than either the Au(I) sulfite or Au(I) thiosulfate complex.

Electroplating of soft gold is generally carried out with a bath containing KAu(CN)2 in a phosphate
buffer of pH 7 at a mildly elevated temperature. It does not contain free cyanide ions initially, but
they are generated at the cathode surface as a result of the gold deposition reaction. The cyanide
ions are partly converted into HCN, which escape into the atmosphere, and partly remain in the
bath. The presence of free cyanide is undesirable not only for its toxicity but also for its
incompatibility with photoresists used to delineate circuit patterns through which the gold is to be
plated. Cyanide attacks the interface between the substrate and the photoresist, lifting the latter and
depositing extraneous gold under the photoresist. To avoid these problems, an Au(I) sulfite bath is
used instead. However, the sulfite bath easily undergoes a disproportionation reaction to form
Au(III) and metallic Au because of the relatively low stability of the Au(I) sulfite complex, causing
the bath to decompose spontaneously on standing, unless a suitable stabilizer is added to the bath.
All commercially available Au(I) sulfite baths contain proprietary stabilizing additives.

In contrast, the Au(I) thiosulfate–sulfite mixed ligand bath we developed is highly stable and
requires no stabilizing additive. We selected this system in view of the fact that it is used
successfully for formulating non-cyanide gold plating baths yielding good bath stability and deposit
properties.

3. ELECTROLYTIC METAL DEPOSITION

There are three types of electrolytic metal deposition processes: direct current electrodeposition,
pulse plating, and laser-induced metal deposition.

3.1. Direct Current Electrodeposition

In the direct current (DC) electrodeposition, the current source is. A power source in the form of a
battery or rectifier (which converts alternating current electricity to regulated low-voltage DC
current) provides the necessary current [7].

Electroplating is performed in a plating unit. Electrodes, immersed in the electroplating bath

(electrolyte), are connected to the output of a DC current source. The workpiece is to be plated acts
as a negatively charged cathode. The positively charged anode(s) completes the electric circuit. This
type of circuit arrangement directs electrons (negative charge carriers) into a path from the power
supply (rectifier) to the cathode (the workpiece to be plated).

The geometric shape and contour of a workpiece to be plated affect the thickness of the deposited
layer. In general, workpieces with sharp corners and features will tend to have thicker deposits on
the outside corners and thinner ones in the recessed areas. The cause of this difference in the
resulting layer thickness is that the DC current flows more densely to sharp edges than to the less
accessible recessed areas. In other words, the current distribution is not uniform [5].
3.2. Pulse Plating

Electrodeposition using pulsed currents is known as pulse plating. The pulsed currents can be
unipolar (on–off) or bipolar (current reversal). Pulses can be used along or be superimposed on a
DC feed. By using the bipolar pulse, metal deposition occurs in the cathodic pulse period, with a
limited amount of metal being redissolved in the anodic period. This repeated deposition and partial
redissolution could improve the morphology and the physical properties of the deposit. The
replacement of a direct by a pulsed current in the electrodeposition of gold has a marked effect in
improving the mechanical properties of the deposits and in reducing their internal stresses. For
many applications in the electronics industry pulse plating therefore offers considerable advantages
[5].

Au (SO3)2-2 + e-1  Au + 2SO2-2 (Cathode)

SO3-2 + 2OH-1  SO4-2 + H2O + 2e (Anode)
SO3-2 + ½ O2  SO4-2 (Air Oxidation)
The structure and surface roughness of gold and gold alloy deposits can be influenced by
periodically reversed or pulsed current. Apparently the need for a high conductivity gold
electrodeposit for integrated circuitry initiated the development of a pulse plating unit which
supplied an on-off D.C. pulse with on- and off-times in the millisecond range and operated with
current peaks up to 8 A. In general the on- and off-times were in the range of 1:10. The advantages
claimed for the pulse plating of gold were [19]:

1. A substantial increase in rate of deposition.

2. Less danger of hydrogen embrittlement.
3. A reduction in the need for additives.
4. A lower level of impurities from the electrolyte.

Pulse plating of gold in the production of semiconductor devices seems to be a procedure which is
comparable to vacuum deposition but much more convenient and rapid, especially for experimental
purposes.

Cheh has done theoretical calculations and experimental measurements which clearly show that for
phosphate, citrate and cyanide gold electrolytes the limiting current density is higher in pulse than
in D.C. plating because of the higher concentration of gold ions in the diffusion layer, although the
limiting overall plating rate is lower. According to Cheh this increase in the limiting current density
might be the reason for the improved physical properties by yielding a different type of structure in
the deposit [20]. Pulse plating markedly affects the properties of the deposit and its fine gold
deposits produce a dense fine-grained structure with about half the resistivity of normal D.C. plated
layers. The fact that pulse plated deposits have the lowest porosity of all gold films deposited by
various methods was attributed by Rehrig to the densely packed continuous fine grained structure
with barely visible grain boundaries. It is shown at Figure 3.1. [21].

Figure 3.1: Relationship between porosity and thickness of various types of gold films according to
Rehrig [21].

The very small grain size and the high purity also serve to explain the difference in tensile strength
and elongation observed by Lendvay and Raub and shown in Figure 3.2 [19]. In the as-deposited
condition fine-grained deposits gave a tensile strength about 25 percent higher than with D.C. plated
deposits. D.C. plated deposits show a slight increase in tensile strength after heat treatment at
100°C, which is followed by a continuous decrease at higher temperatures, while elongation rose
more or less steadily. The increase at 100°C is very likely due to the beginning of decomposition of
incorporated impurities. Pulse plated deposits with lower impurity concentrations do not show any
hardening effect. As one might expect from a purer metal, its tensile strength also decreases much
more rapidly during heating.

Figure 3.2: Tensile strength and elongation of pulsed current and direct current plated deposits from
a phosphate gold solution [19].
3.3. Laser-Induced Metal Deposition

In laser-induced metal deposition, a focused laser beam is used to accelerate the metal deposition.
Experiments have shown that the deposition rate can be increased by 1000 times [7]. The plating
equipment mainly consists of a laser head with focusing optics and the electrolytic cell. The focused
laser beam can pass through a hole in the anode through the electrolyte and impinge on the cathode
surfaces.

4. ELECTROLYTE

Different metals may need different types of electrolyte. Types of electrolytes include water
solutions of acids, bases, or metal salts, certain pure liquids, and molten salts. Gases may act as
electrolytes under conditions of high temperature or low pressure. In addition to metal salts,
electrodeposition electrolytes usually contain a number of additives for various purposes. Some
agents are used to increase electrolyte conductivity (supporting electrolytes). Others may be used
for increasing bath stability (stabilizers), activating the surface (surfactants or wetting agents),
improving leveling or metal distribution (leveling agents), or optimizing the chemical, physical, or
technology properties of the coating. These coating properties include corrosion resistance,
brightness or reflectivity, hardness, mechanical strength, ductility, internal stress, wear resistance,
or solderability [7].

The properties of electrolyte are usually characterized by electrolytic conductance, covering power,
macrothrowing power, and microthrowing power [18]. Inorganic and organic salts, acids or alkalis
can be used to increase the electrolytic conductance. The conductivity of an electrolyte is a function
of the degree of dissociation, the mobility of the individual ions, the temperature and viscosity, and
the electrolyte composition.

Covering power describes that how electrodeposition electrolyte can cover the entire surface of a
workpiece being plated, with desired uniform thickness. Covering power is influenced by the nature
of the substrate surface, the electrolyte composition, the temperature and viscosity, and the current
density. Macrothrowing power predicts the ability of an electrolyte to lay down as nearly as
possible a uniformly thick deposit across the surface of a workpiece [5]. A good covering power is a
prerequisite for good macrothrowing power. Other factors that affect macrothrowing power include
the current distribution and current density, electrolyte composition, electrolytic conductance, and
electrolyte agitation.

Microthrowing power indicates that how metal electrodeposition occurs at the outer plane of the
substrate or at the base of valleys or cracks. Microthrowing power can be improved by activating
the surfaces at the base of valleys or cracks to promote electrodeposition there, while inhibiting the
outer surfaces by using inhibitors preferentially. In many cases, microthrowing power is inversely
related to macrothrowing power [5].

4.1. ELECTROLYTE ADDITIVES

Electrodeposition electrolytes usually contain various other species. Thus both inorganic and
organic salts, acids or alkalis will be added to increase electrolyte conductivity. Other species, both
organic or inorganic, may be added for specific purposes, for example, to increase bath stability, or
improve leveling or metal distribution, or to optimize the chemical, physical or technological
properties of the deposited metal (corrosion resistance, brightness or reflectivity, hardness,
mechanical strength, ductility, internal stress, wear-resistance or solderability) [7].

In working with such additives, it should be recognised that they are often effective over only a
defined concentration range, and should their concentrations fall outside this range, all kinds of
problem can arise either in the deposition processor the deposit properties.

There are three principal types of additives associated with high purity gold electrolytes:
Complexing agents, grain refiners and hardening agents. Complexing agents such as pyrophosphate
ion, organophosphorus compounds and polyphosphates are added to reduce the activity of metallic
impurities in the solution by forming stable complexes and hence minimizing codeposition. Organic
chelating agents such as EDTA and related compounds are also used. Relatively small amounts of
base metals are used for providing grain refinement. These additives also provide smoothing and
semi-brightening of the deposit, while not being codeposited to a significant extent. In neutral
solutions, arsenic and thallium have been used. Some additives such as alums and hydrazine sulfate
have been claimed to harden the electrodeposit without being codeposited.

Use of heavy metal ions in trace quantities (parts per million) in gold electroplating solutions,
induces a marked cathodic depolarization which extends the range of current densities over which
smooth, fine grained deposits can be obtained. In slightly alkaline phosphate electrolytes the most
effective additives comprise the family of elements Hg, Ti, Pb and Bi, which lie immediately
adjacent to gold in the periodic table.

They exhibit a strong tendency to form an adsorbed monolayer on gold and platinum electrodes.
This is done at potentials positive to those at which their cathodic deposition as bulk metals would
begin, i.e. at under potentials. Deposits obtained with these additives have a very fine and highly
uniform grain size. The various heavy metal ions have a brightening effectiveness which is in the
order "I >Pb> Bi > Hg. This is the inverse order of their electron work functions (the amount of
energy required to lift an electron out of a lattice). The postulated mechanism for this performance
is that the elements form an adsorbed monolayer on the surface of the gold. This lowers its work
function and thereby lowers its deposition potential so that deposition then occurs at under
potentials.

Carriers’ are compounds of large molecular weight which, in the presence of small numbers of
chloride ions, will move the start of the curve towards the right, increasing the voltage needed to
produce the desired current. This gives a deposit with a tighter grain structure, and improves plating
distribution and throwing power.

‘Brighteners’ are disulphide compounds with a low molecular weight that reduce the effect of the
carrier by increasing the current. They have a key role in determining the physical properties of the
deposit, in particular tensile strength and percentage elongation.

‘Levellers’ are a class of aromatic compound that are absorbed preferentially at high current areas,
suppressing the plating locally, and allowing the thickness in areas of lower current density to
‘catch up’.

The electrodeposition of metal coatings in industrial metal finishing is usually based on aqueous
electrolytes, known in the industry as electrodeposition baths or simply, baths. Their primary
constituent is the metal salt of the metal to be deposited, then in most cases an acid or alkali to
promote conduction. In some cases the bath is buffered. In addition, there will be additives to
promote the electrodeposition process or optimize the deposit properties [7].

4.1.1. Brightening agents

In metal finishing, bright or highly reflective surfaces (the terms are usually synonymous)can be of
both decorative and functional importance, if for no other reason that bright surfaces require less
mechanical finishing, after plating, if indeed any at all.
Defining brightness or reflectivity of a surface as the extent to which incident light is reflected, it
follows that surface roughness and unevenness are antithetical to good reflectance. Reflective
surfaces are characterized by a crystallite size at the surface of less than 0.3 µm2. Thus, highly
reflective deposits are found under conditions where fine-grained deposits are formed. To achieve
this, it is long-established practice to add metallic or organic compounds, known as brighteners or
brightening agents. Their use is based on empiricism, there being no certain means of predicting
their effects. In the same way, selecting the most appropriate such compound(s) will usually be
based on trial-and-error.

Two different types of brightener are recognized, the first being brightener carriers, the second,
brightener additives [7].

Addition of brightener carriers, also known as primary brighteners or Class I brighteners, leads to a
significant grain refining action in the deposited metal. While this greatly increases the brightness
of the metal, it will not impart a mirror-finish or anything close to this. Among the best-known
brighteners of this type are;

- sulfonamides

- sulfonimides

- benzenesulfonic acids (mono-, di- or trisulfonic acids)

- napthalenesulfonic acids (mono-, di- or trisulfonic acids)

- alkylsulfonic acids

- sulfinic acids

- arylsulfonesulfonates.

These primary brighteners are essential components of a complete brightening system, in that they
reinforce the efficacy of Class II or secondary brighteners, the latter being rarely used without the
former. Class II brighteners, even at very low concentrations, impart a near-mirror finish to the
deposited metal. The most important of these are listed in Table 4.1.
 Table 4.1: Some commonly used brighteners.

Out of a very large number of potential Class II brighteners, only a few are used in practice, largely
because many such species are highly effective in their brightening action, they can adversely affect
process or deposit properties in other ways. In order to minimize this, secondary brighteners are
usually employed with a primary brightener in a combination known for the compatibility and
synergy of both.

A number of electrolytes form bright deposits only in a comparatively narrow temperature and
current density operating window. Figure 4.1 shows this for chromium deposition.

As Figure 4.1 shows at any given temperature, a range of current densities exists, within which
bright deposits can be formed. Alternatively, at any given current density, there is range of
temperatures in which the same is true. Figure 4.1 is valid only for a particular chromium
electrolyte system, and depending on concentrations of species used, and the catalyst, similar
diagrams can be drawn, though the general shape of these is similar. At a given temperature, the
widest possible range of operating current densities is desirable, especially when items of complex
geometry or with deep profiles are involved, as in such cases, there will be marked variations in
current density at different locations on the work, An approach such as that shown in Figure 4.2 is
essential in ensuring that bright deposits are formed over the entire surface, however complex this
might be.

Figure 4.2: Operating window for bright chromium deposition as function of deposition temperature
and current density.

There are many theories to explain the brightening action of this class of additive. Most probably,
the effect is due to a controlling effect, and perhaps a rate-limiting action, on the
electrocrystallization step. Such theories link with the accepted mechanism for metal
electrodeposition, based on a nucleation and growth at a series of centers. From the growth nuclei,
grains develop with a range of sizes and orientations.

It is assumed that the brighteners are preferentially adsorbed on the developing growth peaks,
thereby inhibiting further growth at these points. Having no alternative, the electrocrystallisation is
forced to continue in the valleys rather than the peaks, whereby a levelling action results. The end
result is a smooth and shiny surface. It follows that brightening is associated with a levelling action
as well.
4.1.2. Levelling agents

Either before or after electroplating, it may be necessary for the work to be mechanically finished,
using grinding, then polishing stages, and these can significantly add to the overall cost, especially
when the deposited metal is hard, for example, hard chromium. Such kind of these additives is not
necessary for the gold due to its mechanical and physical properties. Levelling agents can act to
smooth pre-existing irregularities in the surface, such as pits or scratches and thus, they compensate
for poor surface quality of incoming work.

The levelling action of an electrolyte, that is to say its ability to form a smoother surface than that of
the original substrate, is defined as 'levelling power'.

By levelling action is implied the deposition of a greater thickness of metal in the depths of a
groove than at its outer edge. This can be quantitatively expressed as:

Figure 4.3: Illustration of levelling action: (a) geometric levelling; (b) true levelling; (c) levelling as
function of deposit thickness [7].
E is the levelling effect, dt the deposit thickness at the base of the groove, d the overall deposit
thickness and Rz the roughness of the substrate.

A distinction is made between geometric and true levelling. In the former case, the key factor is that
the deposit is thicker at the bottom of a groove than at the edge (Figure 4.3a). The effect is
measured in terms of the geometry and shape of the groove. The hallmark of true levelling is a
higher metal deposition rate in the groove than at the adjacent areas (Figure 4.3b) and this results
when the so-called levelling agents or levellers are added to the electrolyte. The manner in which
these additives work is by restricting the rate of metal deposition, but only at those locations where
they are freely supplied, specifically at peaks and edges. The levelling action of such species can
vary quite widely, apart from which it is usually temperature dependant and can often be enhanced
by work movement or solution agitation. Levelling action is a function of deposit thickness and
increases as thickness increases (Figure 4.3c).

4.1.3. Surfactants (wetting agents)

Plating baths will usually contain wetting agents, also known as surfactants. These are characterized
by their asymmetry and dipolar nature. They have a hydrophilic ('water loving') 'head' and a
hydrophobic ('water-repellent') 'tail'-. Surfactants can be 'anionic', 'cationic' or 'non-ionic' depending
on the charge at the hydrophilic end. The polar hydrophilic radicals are typically-COO- or-SO3
groupings, while the hydrophobic part of the molecule might be a non-polar aliphatic long-chain
hydrocarbon. Non-ionic surfactants are themselves uncharged, acquiring their charge from ions
present in aqueous solution [7].

Wetting agents are surface-active species, that is, they act by accumulating at a surface or interface.
At an air-electrolyte interface, they are oriented with hydrophilic head into solution, the
hydrophobic tail to the air-side. In this way they reduce the surface tension of a liquid. Anionic or
non-ionic surfactants are most often used. The lowering of surface tension facilitates the release of
gas bubbles adhering to the work surface, whether these be cathodically evolved hydrogen or
trapped air. Where such bubbles adhere to the surface, their shielding action results in little
electrodeposition occurring and they can be the cause of pitting or deposit porosity. In addition to
the prevention of pits or pores in the deposit is the main purpose of wetting agents or surfactants.

Addition of surface-active agents also results in the following:

  certain organic compounds, insoluble or poorly soluble in water can be solubilised
or their distribution in solution made more uniform;
 coagulation and sedimentation of suspended solid matter as dirt in solution
which would otherwise lead to rough or porous deposits, is reduced;
 crystal growth is inhibited and a brightening action of the deposit results;
 draining of adherent electrolyte, on withdrawing the work from solution is
improved, drag-out losses are reduced.

In aqueous electrodeposition solutions using surface additives is extremely important owing to

significant effects on the growth and structure of deposits. The striking effects of additive agents on
electrocrystallization in low concentrations (ranging from a few milligrams per liter to a few
percent) point to their adsorption on high energy surface areas where it causes an inhibiting effect
on the most active growth sites. There are three principal types of additives associated with high
purity gold electrolytes: grain refiners, complexing agents, and hardening agents. Additives act as
grain refiners and levelers because of their effects on the structure of the electrical double layer on
the surface and electrode kinetics. A large number of mechanisms have suggested to explain
behavior of additives such as blocking the surface, changes in Helmholtz potential, complex
formation including induced adsorption and ion bridging, ion pairing, changes in interfacial tension
and filming of the electrode, hydrogen evolution effects, hydrogen absorption, anomalous
codeposition and the effect on intermediates .

It has long been known that the formation of crystals depends sensitively on the presence of
additives and impurities. It was found that soluble, as well as insoluble additives could affect
crystallization. Moreover, the foreign of foreign atoms can influence the mutual binding energy and
thus the nucleation probability. Addition of foreign elements is one of the methods for the control of
nucleation.

Proper additives can accelerate or decelerate crystal growth and affect the size, shape and the
robustness of the crystals. The influence of additives on the growth of gold single crystals and also
the possible growth of complex crystals was investigated. Additives were mixed with the silica gel
before the gelling process; their concentration, about 20 mM, was 10-times lower than that of the
HAuCl4 component [23].
 Figure 4.4: Digital camera photo of the hexagonal gold microcrystals preferential growth after

addition of Pd(CH3COO)2 to the silica gel [23].

Because of the specific gold reaction that occurs on the anode which leads to the formation of the
microcrystals, complex crystals were not observed. However at the same time it was found that the
presence of additives such a copper sulfate, palladium acetate or chromium oxide in the gel affects
the crystal growth. For example, in the presence of palladium acetate, the crystals preferentially
grew into hexagonal shapes with similar sizes of about 20µm (Figure 5.4). Hexagonal crystals make
up about 10-15% of all crystals when no additives are present; this increased in the presence of
palladium acetate to 85-90%.

Acetic acid reversibly absorbs on the Au crystal surface, and acts as an inhibiting agent to reduce
crystal growth. The surface free energy of the equilibrium crystal must have a minimum value. Due
to the absorption of additive molecules, the surface tension and the ledge tension change and affect
the equilibrium form of the crystal. It is assumed this leads to the growth of relatively small
hexagonal crystals of similar size.

On the other hand, addition of copper sulfate to the gel produced a harvest of big crystals with
diameters in the range of hundreds of microns and with high-quality surfaces [23]. A possible
explanation for this behavior is that the ledge tension decreases as a result of sulfate adsorption on
the surface of gold. Adsorption would lead to a change in the shape of the nuclei and also the mean
step separation for spiral growth, resulting in the formation of more kink positions and an increase
in crystal growth rate.

Other additives such as chromium oxide caused the growth of gold crystals in vertical lines in the
space between the anode and the cathode. We attribute this to an increase in the stability of the
[AuCl2]- complex which slowly disproportionates further from the anode and forms crystals in the
central region of the cell.

Cysteine (2-amino3-mercaptopropanoic acid) is an organic material which can make complex with
gold ions. These complexes can deposit Au particles on the cathode surface separately and cause
grain refining. The oxygen reduction on gold nanoparticles which are deposited from 1 mM
solution with cysteine additive is eight times higher than that on the gold bulk deposit at −0.8 V
potential. This behavior can be attributed to increase of effective surface area by decreasing
nanoparticles size. Cysteine was found as an appropriate agent to refine gold nanoparticles due to
increase of nucleation sites and improve kinetic parameters without considerable change in the
purity of deposits. Particle size considerations in the presence of cysteine indicated that the gold
nanoparticles are uniformly distributed in the range of 10–40 nm. Improved catalytic properties of
gold nanoparticles can be obtained when their effective surface area is increased. It is established
that oxygen reduction on refined gold nanoparticle deposits is eight times greater than that on
conventional gold deposits [22].

Chloride solution with cysteine additive was used as electrolyte in gold nanoparticles
electrodeposition. Atomic Force Microscope was used to evaluate the effect of cysteine on the
morphology and topography of gold nanoparticles.

Further investigation of additives on the nucleation and growth of gold single microcrystals will
provide valuable information for the preparation of a variety of single crystals.

The electroplated Au films were obtained from the bath consisting of a mixture:

0.05M KAu(CN)2 + 0.25M (NH4)H2HPO4 + 0.40M (NH4)2HHPO4 without and with additives of
surfactants, namely 10 mg/l oxyethylendiphosphoric acid (H4OEtDP) + 0.75 mg/l
cyanethilendiamine (CyEn) at 60-70 °C and pH 5.6. The deposition was performed at the cathodic
current densities 0.2 to 0.7 A/dm2 on the brass substrate using Pt anode [24].

Table 4.1. Comparative properties of the gold coatings produced at different

condition of electrolysis

In order to reduce the porosity of film the effect of some surfactants was studied. The positive effect
of composition additives (H4OEtDP and CyEn) is revealed. The optimal composition of surfactants
is: 10 mg/l H4OEtDP and 0.75 mg/lCyEn. The gold films obtained from phosphoric acid bath with
additives are light, fine-crystalline, well-adherent to the brass substrate. Adding of these specific
surfactants almost excludes the pores in electroplated film. Simultaneously the gold electrode-
position rate increases at higher current densities 0.3 – 0.5 A/dm2 due to reducing the hydrogen
evolution rate (Table 4.1). Moreover, the adding of these surfactants enable to obtain Au films (at
0.5A/dm2) that possess the excellent microhardness. In our opinion, the increase of microhardness is
resulted by the grain size decrease with increasing cathodic current density.

So, the adding of H4OEtDP and CyEn to the phosphoric gilding bath expands the range of current
densities up to 0.6 A/dm2. Properties of the obtained films are reasonable for microelectronics
application.

The gold sulfite complex Au(SO3)2-3 is the most common alternative to cyanide. This complex has a
stability constant of 1010, much less than that of the gold cyanide complex. Consequently, it is
susceptible to disproportionation resulting in the formation of Au (III) and metallic gold, especially
at pH less than 7, where sulfite protonates to form bisulfite. Therefore, stability is always an issue
for sulfite-based gold electroplating solutions, and stabilizing additives are usually necessary.
The effect of temperature, pH, sulfite concentration, gold concentration and additive concentration
play on the achievable role for current density range. Temperature has the strongest effect on the
current density range, which increases with increasing temperature. The current density range can
also be increased by raising the pH and gold concentration, or decreasing the sulfite concentration,
however is not significantly affected by the additive concentration [25].

Figure 4.5: The effects of gold and additive concentrations on the plating current density range [25].

The presence of ammonia in the copper–ammonia–thiosulfate system is essential as it stabilizes

copper (II) and reduces its rate of reaction with thiosulfate. The gold oxidation polarisation curve
obtained for ammonium, potassium, and sodium thiosulfate solutions using a polished pure solid
gold electrode are shown in Figure 4.6. It is immediately obvious that gold oxidises much more
readily in ammonium and potassium thiosulfate solutions than sodium thiosulfate solutions [26].
Figure 4.6: Linear sweep voltammograms showing the gold oxidation half reaction in either 0.2 M
(NH4)2S2O3, K2S2O3, or Na2S2O3. Also shown is the polarization curve obtained for 0.2 M
(NH4)2S2O3 with 0.05 M CsCl added [26].

The cations such as potassium, which are higher in atomic mass and less hydrated compared with
sodium, are more likely to be adsorbed onto the gold surface. It is known that the ion pair
dissociation constants for the alkali metal ions with thiosulfate ions decrease in the sequence of
Na>K>Rb>Cs. Therefore, one would expect that K+ is more likely to form an ion pair with gold
thiosulfate or thiosulfate than Na+. An additional experiment was also carried out where 0.05 M
caesium chloride was added to an ammonium thiosulfate solution, and it can be seen from Figure
4.6 that the most active polarisation curve was obtained in this instance. This result is consistent
with the presence of the heavy alkali metals giving rise to a higher rate of gold oxidation in
thiosulfate solutions. In practical terms, it should be apparent that the gold leaching reaction would
occur more readily in ammonium thiosulfate solutions than in sodium thiosulfate solutions.

It has been shown that the presence of ammonium ions dramatically enhance the oxidation of gold
in thiosulfate solution (Figure 4.6). Therefore, the effect of alkyl ammonium ions on the oxidation
of gold was also studied. The polarisation curve for the oxidation of gold in sodium thiosulfate
solutions in the presence of 0.05 M alkyl ammonium chloride was obtained using the pure solid
gold disc, and these are shown in Fig. 6. It is clear that the addition of mono-methyl and di-methyl
ammonium ions also enhance the gold oxidation half reaction, although the effect is not as
pronounced as for ammonium ions. It can also be seen from Figure 4.7 that the gold oxidation
reaction is actually hindered in the presence of tri-methyl ammonium ions. A similar result was
observed for the tetra- methyl ammonium ion. It would appear that in the case where tri-methyl and
tetra-methyl ammonium ions are present in the thiosulfate solutions, the decrease in the gold
oxidation rate could be attributed to the coverage of the gold surface with the adsorbed ions, since
the degree of adsorption of alkyl ammonium ions increases as the number of alkyl groups increase.

In order to achieve higher gold leach rates in thiosulfate solutions using oxidants other than copper
(II), the gold oxidation half reaction needs to be improved. Therefore, the effect of additives, such
as potassium ethyl xanthate (PEX) and thiourea (TU) on the gold oxidation reaction was studied.
Using the standard electrochemical techniques, it is difficult to interpret the data in the presence of
either xanthate or thiourea, as both these species can be oxidized.

Figure 4.7: Linear sweep voltammograms showing the oxidation of gold in 0.2 M sodium
thiosulfate in the presence of 0.05 M alkyl ammonium chloride [26].

The polarisation curves are shown in Figure 4.8 for ammonium thiosulfate solutions containing 5
mM of either xanthate or thiourea. It should be clear that the addition of 5 mM of thiourea is
extremely beneficial to oxidation of gold in thiosulfate solutions, with the current rapidly increasing
as the potential is increased above 125 mV. This type of polarization curve is typical of active
dissolution processes, such as the oxidation of copper or silver in thiosulfate solutions. It is not
possible that the gold thiourea complex is being formed in this instance, since an experiment
conducted with 10 mM thiourea and 0.2 M ammonium sulfate did not result in any gold oxidation
in the potential region 0–300 mV. In comparison, the addition of potassium ethyl xanthate is shown
in Figure 4.8 to completely block the oxidation of gold, with no gold oxidation being observed
within a wide potential range. It is worth noting that both xanthate and thiourea are used as surface
modifiers in many different processes, and hence one could attribute the effects shown in Figure 4.8
to the adsorption of these species onto the gold surface in the thiosulfate solutions.

An important point to note is that thiourea is considered a potential carcinogen, and hence its use as
an additive may be restricted. A qualitative description of the results is shown in Table 4.2. These
additives were chosen because they are either known to adsorb readily on metal surfaces, and/or
they have characteristics which are similar to thiourea. It can be seen that most of the additives
actually resulted in passivation of the gold surface. However, the presence of thioacetamide was
found to have a similar effect to thiourea in dramatically improving the gold oxidation half reaction.

Figure 4.8: Linear sweep voltammograms showing the effect of thiourea and potassium ethyl
xanthane on the gold oxidation half reaction in 0.2 M ammonium thiosulfate [26].

It should be apparent that the leaching of gold in solutions containing thiourea is likely to be much
more rapid than leaching in the absence of thiourea. The effect of thiourea concentration on the gold
oxidation half reaction was investigated and Figure 5.9 shows the current density for gold oxidation
at various thiourea concentrations. At any potential, the gold oxidation half reaction is slower at
lower thiourea concentrations. However, the improvement of the gold oxidation observed is
significant even at the lower thiourea concentrations. In general, it is recommended that thiourea
concentration greater than 5 mM should be maintained in order to achieve appreciable gold
oxidation rates.
 Table 4.2. Effect of additives on the gold oxidation half reaction [26].

The presence of thiourea significantly improves the oxidation of gold in ammonium thiosulfate
solutions. So can the ammonium ions be removed when thiourea is present? Thus, experiments
were performed to determine the current density for gold oxidation within the potential range from
0 to 300 mV. The polarization curves are shown in Figure 4.9. It is immediately obvious that the
presence of thiourea also enhances the gold oxidation in sodium thiosulfate solutions.

Figure 4.9: Linear sweep voltammograms showing the effect of thioureaconcentration on the gold
oxidation half reaction in 0.2 M ammonium thiosulfate [26].
The addition of caesium ions also increased the reactivity of the gold. The effect of additives such
as thiourea and potassium ethyl xanthate was also investigated, and it was found that the addition of
xanthate resulted in the complete passivation of the gold surface. However, the presence of small
amounts of thiourea substantially improves the oxidation of gold in thiosulfate solutions. This effect
was more pronounced in ammonium thiosulfate solutions than in sodium thiosulfate solutions.

5. SURFACE PREPARATION

Gold plating is often the last step in a long series of processing steps, making it immediately suspect
when difficulties are encountered. However, in many cases the problems that appear at the gold
plating step are due to the cleaning and preparation processes preceding gold deposition. Poor,
incomplete, or inappropriate cleaning can result in inadequate gold deposits. While problems do
occur in the gold plating process, only a good engineering evaluation of the defective deposits can
elucidate the failure mechanism.

Workpieces to be plated may be put through a variety of pretreating processes, including surface
cleaning, surface modification, and rinsing [18]. A schematic data flow of a typical electroplating
plant, including surface treatment and waste treatment, is depicted in Figure 5.1.

The purpose of surface pretreatment is to remove contaminants, such as dust and films, from the
substrate surface. The surface contamination can be extrinsic, composed of organic debris and
mineral dust from the environment or preceding processes. It can also be intrinsic, such as a native
oxide layer. Contaminants and films interfere with bonding, which can cause poor adhesion and
even prevent deposition. Therefore, surface pretreatment is important to ensure plating quality.
Most (metal) surface treatment operations have three basic steps: surface cleaning, surface
treatment, and rinsing.
 Figure 5.1: Process data flow of a typical electroplating plant [18].

5.1. Surface Cleaning

Cleaning methods should be able to minimize substrate damage while removing the contaminants,
dust, film, and/or debris. Cleaning processes are based on two approaches: Chemical approach and
Mechanical approach.

5.1.1. Chemical Approaches

A chemical approach usually includes solvent degreasing, alkaline cleaning, (soak cleaning), and
acid cleaning (acid pickling).
5.1.1.1. Solvent Degreasing: Contaminants consist of oils and grease of various types, waxes, and
miscellaneous organic materials. These contaminants can be removed by appropriate organic
solvents, either by dipping the workpieces in the solvent or by vapor decreasing.

5.1.1.2. Alkaline Cleaning: Workpieces are immersed in tanks of hot alkaline cleaning solutions to
remove dirt and solid soil. A special type of alkaline cleaning is electrocleaning. In electrocleaning,
the workpiece can be either the cathode (namely direct cleaning) or the anode (reverse cleaning).
Electrocleaning adds to the chemical action of the cleaner the mechanical action caused by plentiful
gas evolution at the surface of the workpiece.

5.1.1.3. Acid Cleaning: Acid cleaning can move heavy scale, heat-treat scale, oxide, and the like.
The most commonly used acids include sulfuric and hydrochloric. Pickling can also be combined
with current to be more effective.

5.1.2. Mechanical Approaches

Mechanical preparations include polishing, buffing, and some variations. Polishing is to remove
small amounts of metal by means of abrasives. It produces a surface that is free of the larger
imperfections left by grinding, and is a preliminary to buffing. Buffing is similar to polishing, but
uses finer abrasives to remove very little metal. Buffing can produce an extremely smooth surface.

5.2. Surface Modification

Surface modification includes change in surface attributes, such as application of (metal) layer(s)
and/or hardening.

5.2.1. Rinsing

In wet plating, when workpieces are transferred from one treating solution to another, or when they
leave the final treating solution, they carry some of the solution in which it has been immersed. This
solution is called drag-out. In most cases, this residue solution should be removed from the
workpieces surface by rinsing before the workpieces enter the next step in sequence, or come out of
the final processing solution. The dirty rinse water will be sent to the wastewater treatment facilities
before being discharged to a public sewage system [5].
Copper, brass, and silver parts are easily plated once cleaned. The normal practice may require
degreasing in an alkaline spray or soak or by a solvent vapor degreaser. A copper part may require a
bright dip (10% sulfuric acid in water) to remove the native oxide. Beryllium copper spring material
or castings should be carefully cleaned by a 15-min soak in boiling NaCN and NaOH solution (38 g
NaCN, 38 g NaOH, and 1 L H2O) followed by a water rinse and bright dip and final rinse. Other
procedures use a dip in 60% by volume H2SO4 at 50˚C for 30 s followed by a water rinse.

Nickel parts must be activated before gold plating or the plate may peel. Activation can be
accomplished by a dip in warm 1 : 1 HCl in water. It is very important not to allow the nickel parts
to dry after activation before gold plating. Nickel alloys like Kovar can be cleaned by soaking for
many minutes in hot 18% HCl. The oxide-free surface is prepared for plating by bright dipping for
several seconds in 1 : 3 nitric acid–acetic acid with 15 mL HCl per liter of solution at 70˚C. Kovar
can be first plated with Woods or Watts nickel before gold plating or to gold strike the nickel alloy
directly at high current in a low-gold-content bath followed by normal gold plating [3].

Gold plating onto aluminum and aluminum alloys has become more common, particularly for
electronics and space applications. The surface is first degreased, such as in trichloroethylene
vapors, followed by alkaline cleaning in trisodium orthophosphate (Na3PO4, 12 gL-1 ; H2O, 25 gL-1 ;
and Na2CO3, 20 gL-1) at 60˚C for 2–3 min. After water rinsing, the aluminum surface is cleaned in a
mixed-acid solution (sulfuric acid, 10mL L-1; hydrofluoric acid (40%), 12.5mL L-1, and nitric acid,
255mL L-1) for 2–3 min. A zinc displacement deposition process can precede the electroless nickel
underlayer (nickel sulfate, 30 g L-1; zinc sulfate, 40 gL-1; sodium hydroxide, 106 g L-1; potassium
cyanide, 10 gL-1 ; potassium bitartrate, 40 gL-1; copper sulfate, 5 g L-1; and ferric chloride, 2 g L-1)
for 1 min. The first zinc layer is stripped in 50% nitric acid followed by a second zinc layer by
immersion for 30 s. Extreme caution should be practiced when handling hydrofluoric acid and
potassium cyanide, especially when the cyanide is in the same vicinity as acids.

Gold films are widely used in MEMS applications because of its excellent physicochemical
stability, low resistivity, and high reflectivity for infrared radiation. However, adhesion onto silicon
has been problematic. A two-step electrodeposition process was found to give growth directly onto
a silicon surface. After stripping the surface oxide from the silicon, gold is plated and annealed and
then a final coating of gold is plated. The current density of the initial gold plate is critical.
The adhesion of gold directly to glass is very difficult. An ion beam mixing process has been used
to produce gold films directly bonded to glass. An Au–C bilayer film was deposited on glass
followed by ion beam “mixing” caused by the Xe+ ion implantation. The ion implantation process
provides for atomic mixing of the thin layer with the glass. A thicker film of gold could then be
deposited. Excellent adhesion of the gold film on glass was obtained.

Thin electroplated gold wires were bonded to hydrogen silsesquioxane glass by a transfer process
using imprint lithography. The transfer process first produced thin gold wires on silicon from
AuCN-4 immersion plating onto silicon. Silicon serves as the reducing agent for gold deposition.
The plating bath contained free fluoride to dissolve the oxidized silicon. The lack of adhesion of the
gold wire onto the silicon surface allows easy release and transfer onto the glass.

The direct thermocompression bonding of plated gold surfaces was used to achieve MEMS
packaging. Two silicon wafers were bonded together at a pressure of 2.5 MPa and temperature of
320˚C. The gold film produced on the side walls of via holes was used for bonding to two parts
together, providing a hermetic seal for the MEMS package.

Service requirements may need an underplate between the base metal and the gold electroplate.
Specifications for gold plated copper parts exposed to temperatures > 60˚C often require an
intervening nickel-plated barrier to prevent copper–gold interdiffusion and subsequent copper bleed
through to the surface. The barrier may be 5–10 µm of low-stress nickel, like that plated from a
sulfamate bath. Gold plating on zinc-based alloys should have at least 8 µm of nickel over 8 µm of
copper. Tin or lead should have 8 µm of nickel over 2 µm of copper.

The use of a gold strike is common before the deposition of a thick gold plate. The thickness of a
gold strike may be 0.02–0.25 µm, depending on the thickness and requirements of the full gold
thickness. A gold strike bath is usually a dilute version of a normal gold plating bath. A high current
density is used to overcome activation control of the plating process. A large fraction of the gold
cyanide ions at the electrode surface are reduced to gold at a high nucleation rate. Thus a uniform
deposit is formed over the electrode surface compensating for inhomogeneities across it, which may
have created problems under normal gold plating conditions. The gold strike surface has increased
adhesion and porosity, and less contamination is dragged into the main plating bath. A typical strike
bath would have 0.8 g L-1 gold versus 8 g L-1 and current densities 2–10 times higher than the
conventional bath.
5.2.2 High-Speed Strip Plating

There has been extensive use of high-speed plating as well as selective plating in the semiconductor
industry since around 1970 [3]. In particular, reducing cost and materials in the reel-to-reel
electroplating of electrical connectors has led to many innovations in high-speed plating. Reel-to-
reel plating of connectors includes facilities for cleaning and preparation of parts, strike plating, and
transport of parts and chemicals. The sequence of the processes is similar to those in batch plating;
however, size, space, and time limitations have stimulated a fine tuning of the processes, which
makes their review worthwhile. The strip plating process for electrical connectors is composed of
the following processes: (1) ultrasonic clean, (2) electropolish, (3) acid dip, (4) nickel plate, (5) gold
strike, (6) hard gold plate, (7) hot rinse, and (8) dry. Cleaning is limited to non-foaming solutions
because of the need for rapid recirculation of the liquid from a reservoir tank. Ultrasonic agitation
speeds up the cleaning process by dislodging the contaminants from the surface through cavitation.
Electropolishing is an anodic process where metal dissolves (cleaning and deburring) in proportion
to the anode current density. Copper alloys used for electrical connectors electropolish well in 1 : 1
phosphoric acid in water. Current densities as high as 4000 mAcm-2 can be used to accommodate
high-speed processing with small electroplating cells. A major problem in electropolishing is the
deposition of copper powder at the cathode, which must be removed periodically to avoid cell
shorting. While it is desirable to have the electrodes in close proximity to reduce solution heating
and voltage, wider spacing (e.g., 1.5 cm) is necessary to mitigate cell shorting from powdery copper
loosely adherent to the cathode. The copper phosphate residue can be removed by a hot-water rinse
followed by a rinse in 10% sulfuric acid (by volume) at 60˚C. The nickel underplate (0.14 mm) can
be from a nickel sulfate or nickel sulfamate bath.

Figure 5.2: The AFHG baths used in reel-to-reel gold strike.

The current density for the gold strike is 30 mA cm-2. The hardness of the final deposit is achieved
by producing small grains, which is the result of high current density at low temperature. The
hardness of the gold plated from the high speed bath given above is 180 KHN25 for current densities
between 150 and 250 mAcm-2. The hardness drops off at lower current densities, to 150 KHN25 for
100 mAcm-2, and 125 KHN25 at 50mAcm-2. Others have used even higher current densities (800–
2000 mA cm-2), although the conditions are different. A final hot rinse and drying complete the strip
plating process. The plating process can be accomplished in relatively short-length plating cells (18
in. long) with insoluble anodes that allow precise placement of the hardware in an attempt to
selectively plate the gold only on the functional area of the connectors. Although a majority of the
gold can be deposited on the contact area of connectors, it is extremely difficult to prevent some
plating on nonfunctional areas immersed in the plating bath. Laser-assisted plating has been shown
to assist in the selective plating process. Laser processes have been developed to improve the
selectivity of the deposition process. The direct irradiation by a pulsed laser has been shown to
cause extremely rapid surface heating, increasing the agitation and deposition rate. It also increases
the interdiffusion of the deposited atoms. Lasers have also been used in a patterning technique
where the whole component was coated with an insulator, such as glycol phthalate, paraffin wax, or
photoresist, and an yttrium–aluminum–garnet (YAG) laser was used to burn off the material in
small areas which were then plated [27]. One problem is the precise alignment of the laser to the
part either in a moving or a step-and-repeat process.

6. GOLD AND GOLD ALLOY BONDING IN MEMS

Several practical micromachming techniques in silicon have emerged from recent research.
However, the complexity of the micromechanical systems that would potentially benefit from these
developments often exceeds the performance limits of these technologies when applied to a single
wafer. Silicon wafer-to-wafer bonding offers the designer of such a system an extra degree of
flexibility to trade-off single-wafer concentrated complexity for multi-wafer solutions. Wafer-to-
wafer bonding has been used in applications such as power devices and has been particularly
successful in integrated silicon sensors.

The main advantage of wafer-to-wafer bonding in the application is that it enables the separate
fabrication of the sensor wafer and the wafer in which the active readout electronics is integrated,
until the very last processing step. The sensor wafer can, therefore, be designed for maximum
performance of the sensing element, without jeopardizing the performance of integrated active
devices; the compatibility is not impaired. Moreover, the separation also allows for the use of
different specialized foundries for processing of the sensor wafer and the readout wafer. Assembly
can take place afterwards using relatively simple equipment, so that the batch fabrication advantage
of the silicon technology is maintained. This approach, however, makes low-temperature wafer-to-
wafer bonding and through-wafer interconnect mandatory [28].

Wafer-to-wafer bonding has the potential to overcome many of problems that are not overcame by
single wafer technologies. Individual wafers can be subjected to relatively simple bulk and/or
surface micromachining steps and can subsequently be combined to realize a complex
micromechanical function. The electronic circuits can, in principle, be integrated in any of the
wafers, which further improves the design flexibility. As dicing and packaging take place after
bonding, the micromechanical devices can be designed to be less prone to breaking during
fabrication that would otherwise result from the reduced integrity of the individual bulk-
micromachined wafer(s).

Wafer-to-wafer bonding processes for microstructure fabrication processes have an impact in

packaging and structure design. Processes are categorized into direct bonds, anodic bonds, and
bonds with intermediate layers. Micromachining encompasses a broad range of technologies
anchored in the core technology of microlithographic pattern transfer. A large fraction of the
micromachining technologies are specific to the silicon material system principally due to the
origins of the field, namely, the silicon integrated circuit (IC) industry. In the silicon
micromachining field, there have been two dominant fabrication methods, broadly classified as bulk
micromachining (etching deep features into a wafer) and surface micromachining (depositing,
patterning, and selective etching of films on a wafer) [29]. Fundamentally, both of these techniques
rely on some form of etching or material removal. A complementary, additive, micromachining
technology is the wafer-to-wafer bonding of patterned substrates, often simultaneously involving
alignment of the substrates.

These wafer-to-wafer bonds were performed at low temperatures (less than 450˚C) and involved
either field-assisted silicon-to-glass bonding (anodic bonding) or a eutectic bond between silicon
wafers using a gold thin film. The advantage of the low temperatures of these bonds was that wafers
with common IC metals such as Al could withstand this temperature without degradation. These
bonding methods offer the advantage in some circumstances of being high-temperature stable
bonds, thus permitting a wide range of subsequent processes. The emergence of these techniques
has catalyzed a great deal of activity in wafer-to-wafer bonding for a range of applications.
The wafer-to-wafer bond illustrated in Figure 6.1., is a thin-film glass frit bond. The pressure sensor
has been fabricated using bulk micromachining and thin-film processes for the piezoresistors. In the
case of a gauge pressure sensor, the lower wafer is etched to form an inlet port, and thus the role of
the lower wafer bond reverts from that of a vacuum seal to a first-level pressure inlet manifold. In
either case, the primary function of the wafer bonding is to effect a package function.

Particularly in the sensors industry, that packaging can be a dominant component in the final
manufactured cost of a device. By achieving package function at the wafer level, it is possible to
realize tremendous overall savings in cost since this enables the packaging of a multitude of sensors
or actuators simultaneously, eliminating costly individual chip-packaging steps.

Figure 6.1: A Motorola pressure sensor using glass frit wafer bonding for packaging.

Another example of the use of wafer bonding in the manufacture of pressure sensors is shown in
Figure 6.2. Here, the pressure sensor is formed by a high-temperature silicon direct bond, followed
by thinning of a wafer and fabrication of the piezoresistors. In comparison with the conventional
bulk micromachined pressure sensor of Figure 6.1, this pressure sensor has several interesting
features that are enabled by the wafer bonding. The wafer bonding produces a reference cavity for
the absolute pressure sensor as in the previous case; however, the cavity depth can be made
considerably smaller (1–10 µm) as compared to the bulk micromachined devices (~500 µm). The
benefit of this is that the lower surface of the sealed cavity can function as an overpressure stop. In
the event of a large pressure, the sensing membrane can be designed to “bottom out” against the
cavity lower surface, preventing catastrophic failure of the pressure sensor. There are several other
noteworthy advantages of this device.

Figure 6.2: A silicon wafer-bonded pressure sensor. Die photo (Left). Cross section of process
(Right).

A multitude of other interesting examples of sensors built using wafer bonding exist. One
noteworthy early example is that of an accelerometer in which the bonding not only defines the
critical mechanical flexures in the device but also is used to implement overacceleration stops for
the proof mass. This greatly enhances the manufacturability and packaging of the device.

6.1. Categories of Wafer-to-Wafer Bonding

The types of wafer bonding that are most commonly employed in microstructure fabrication can be
placed in three categories:
6.1.1. Direct Bonds

Wafers are directly contacted without the assistance of significant pressure or any intermediate
layers or fields. These bonding schemes rely on the tendency for smooth surfaces to adhere, and
always utilize some form of thermal cycling after the contacting to increase the bond strength. The
maximum temperature of the anneal defines the bonding as being low temperature (450 ˚C) or high
temperature ( 800 ˚C) [29].

6.1.2. Anodic Bonds

The wafers are contacted and the bond is completed by application of an electric field at
temperatures in the range of 300–450 C. The bond is typically performed between a sodium-baring
glass wafer and a silicon wafer [29].

6.1.3. Intermediate-Layer Bonds

This category includes all bonding mechanisms that require an intermediate layer to promote the
wafer bond. This could include eutectic bonds, polymers, solders, or thermocompression bonds.
Many of these wafer-level bonds are a scale-up of bonds that have been routinely performed at the
die level in the packaging of IC’s.

Intermediate bonding is the bonding of two silicon wafers using an intermediate layer; a polymer
glue, a soft glass or gold beyond the eutectic temperature of the Au-Si binary system. The special
property of the silicide-based bond is the good electrical contact between the two bonded wafers.
There is the gold-silicon hard solder that has already been frequently used for silicon die bonding to
a substrate. The silicon-gold binary system reveals the most dramatic reduction of the melting
temperature. As shown in Figure 6.3, the melting temperature is reduced from 1063 ˚C at pure gold
to 363 ˚C, whereby 19 at.% silicon is dissolved in the eutectic silicon-gold compound. Eutectic die
bonding is often used in industry. An Au/Si compound with 19 at.% Si is used as a substrate and
heated up to a temperature slightly above the eutectic temperature (dashed line in Figure 6.3). This
Au/Si substrate acts as a solder and consumes silicon from the die after it is brought into direct
contact. Silicon is dissolved until the saturation composition is reached (X, in the Figure 6.3). Upon
cooling a reliable bond was obtained. An extension of this technique is the micron bump bonding,
in which multi-electrode contacts can be made between a die and the Si/Au substrate [28]. Although
this technique has so far only been used in die bonding, it can in principle also be employed for
silicon wafer-to-wafer bonding with gold as an intermediate layer.
 Figure 6.3: Silicon-gold phase diagram [28].

Problems associated with gold eutectic bonding are the long-term drift in sealed-cavity devices and
the possible trap formation halfway to the bandgap. Contamination of silicon with gold would result
in a severe reduction of the minority carrier lifetime in integrated active devices. Under normal
conditions silicon dissolves in the flowing gold and not vice versa, so there is, in principle, no gold
doping of the wafer. Experiments have demonstrated the suitability of the eutectic low-temperature
wafer-to-wafer bonding for multi-wafer smart sensors, the compatibility with techniques for
through-wafer interconnect and the functional integrity of the bonded device. The sealing
performance needs to be demonstrated using an endurance test on an integrated pressure sensor.
Smart micromechanical sensors and actuators can be fabricated with electronic and/or
micromechanical functions distributed over various wafers, due to the low-temperature processing
and the implementation of through-wafer interconnect. The area consumption required for bonding
(sealing) and etching of through holes makes the technique less-suitable for high-density wafer-
scale integration, but it is of great promise in smart sensors, where up to 10 leads is usually
sufficient.
Silicon microstructures can be sealed together by Eutectic Bonding. For instance, it is possible to
bond bare silicon against Au covered silicon, or Au covered silicon against Au covered silicon.
Another alternative is to use a Au/Si preform with a composition close to the eutectic concentration.
The eutectic temperature for Au/Si is 363 degrees Celsius, which is well below the critical
temperature for Al metallized components. In addition Au is a quite commonly used thin film
material. These are reasons to prefer the Au/Si choice, but the eutectic sealing can also be
performed with other materials [30].

The major problem with the eutectic bonding is to obtain complete bonding of large areas. Even
native oxides prevent the bonding to take place. To reduce this problem the native oxide on the
silicon should be removed before Au deposition, for instance by backsputtering. In addition some
mechanical motion could be required during bonding to enhance the alloying process.

Eutectic Bonding by a Au/Sn preform has been reported to introduce substantial mounting stress in
piezo-resistive sensors causing long term drift due to relaxation of the built-in stress. The materials
usable for eutectic bonding of metallized silicon components are quite ductile, and this is a probable
disadvantage resulting in long term mechanical drift.

Silicon wafer bonding has been used to fabricate surface-rnicromachining-like structures for a
variety of sensors and actuators. Figure 6.4 illustrates the process. We begin by joining two silicon
wafers, a device wafer containing an etch-stop layer and a handle wafer containing etched cavities,
using wafer bonding. The bulk of the device wafer is then thinned to produce a wafer with sealed
cavities which looks virtually identical to an unprocessed wafer. At this point the bonded and
thinned structure can be returned to a standard IC facility for electronics fabrication. Following
circuit processing, low-temperature back-end processes are performed to functionalize the sensors.
The process has two significant features. First, the mechanical layer which results from this process
has reliable, reproduceable mechanical properties. Second, the electronics integration strategy is
compatible with standard IC processes and equipment [31].

Modification to the thinning process is to achieve electrical isolation of the mechanical layer from
the substrate. presented in the future. Thinning methods are critically important in wafer bonded
micromachining processes since the thickness tolerance of the mechanical parts will be defined by
this step. We described an electrochemical thinning process that produces a n-type mechanical layer
on a n-type substrate. However, electrical isolation is important for a class of structures built in this
technology (accelerometers and gyroscopes, for example).

Figure 6.4: MIT wafer-bonded integrated sensor process [31].

Two technical approaches are either to use a buried chemical etch-stop (p++ doping or oxide) and
bonding to an oxide coated substrate wafer (dielectric-isolation) or to use a junction electrochemical
etch-stop and a p-type substrate wafer (junction-isolation) as illustrated in Figure 6.5.
Electrochemical etch-stop methods are known to produce high quality mechanical structures, but
the layers are most conveniently formed by silicon-silicon bonding, where the lower wafer provides
an accessible electrical connection during the etch-stop process. This process thus precludes
electrical isolation of the mechanical parts from the substrate.
 Figure 6.5: Sealed cavity processes for electrical isolation of the mechanical parts from substrate.

A second important feature of the integrated process is the ability to fabricate these parts using a
completely dry release process. A resist-mask is used for device patterning followed by a Si plasma
etch to pattern the mechanical structure and a plasma etch to remove the -resist-With this dry
release method, large suspended plates have been successfully fabricated as tuning fork tines for the
gyroscopes and the proof masses for the accelerometers.

In the process of piezo-resistive pressure sensor packaging, a simple thermo-compression bonding

setup has been fabricated to achieve the wire bonding interconnection of a silicon chip with printed
circuit board. An annealed gold wire is joined onto a pad surface with a needle-like chisel under a
force of 0.5–1.5 N/point. The temperature of the substrate was maintained in the range of 150–
200˚C and the temperature of the chisel was fixed at around 150˚C during wire bonding operation.
The tensile strength of the wire bonding was measured with a bonding tester by the destructive-
pulling experiment and was found to be at the average of 132 mN/mm2 [32]. The microstructure of
the bonding point was examined by scanning electron microscopy. The interface of the thermo-
compression boning was shown to possess an acceptable level of reliability for a micro-
electromechanical system (MEMS)-based device. The results showed that this setup can be easily
operated for fabrication and is suitable for fabricating not only low-cost pressure sensors, but also
other

6.2. Wirebonding

Wirebonding is an electrical interconnection technique using thin wire and a combination of heat,
pressure and/or ultrasonic energy. Wire bonding is a solid phase welding process, where the two
metallic materials (wire and pad surface) are brought into intimate contact. Once the surfaces are in
intimate contact, electron sharing or interdiffusion of atoms takes place, resulting in the formation
of wirebond. In wirebonding process, bonding force can lead to material deformation, breaking up
contamination layer and smoothing out surface asperity, which can be enhanced by the application
of ultrasonic energy. Heat can accelerate Interatomic diffusion, thus the bond formation [33].

Wirebonding process begins by firmly attaching the backside of a chip to a chip carrier using either
an organic conductive adhesive or a solder (Die Attach). The wires then are welded using a special
bonding tool (capillary or wedge). Depending on bonding agent (heat and ultrasonic energy), the
bonding process can be defined to three major processes: thermocompression bonding (T/C),
ultrasonic bonding (U/S), and thermosonic bonding (T/S), as shown in Table 6.1.

Wirebonding Pressure Temparature Ultrasonic energy Wire

Thermocompression High 300-500 ˚C No Au
Ultrasonic Low 25 ˚C Yes Au, Al
Thermosonic Low 100-150 ˚C Yes Au

Table 6.1. Three wirebonding processes [33].

There are two basic forms of wire bond: ball bond and wedge bond, the corresponding bonding
technique, bonding tool and materials are listed in Table 6.2. Currently, thermosonic gold ball
bonding is the most widely used bonding technique, primarily because it is faster than ultrasonic
aluminum bonding. Once the ball bond is made on the device, the wire may be moved in any
direction without stress on the wire, which greatly facilitates automatic wire bonding, as the
movement need only be in the x and y directions.

Ball Bond Wedge Bond

Figure 6.6: Two basic forms of wire bond [33].

Wirebond Bonding technique Bonding tool Wire Pad Speed

Ball bond T/C Capillary Au Al,Au 10 wires/sec (T/S)
Wedge bond T/S, U/S Wedge Au, Al Al,Au 4 wires/sec

Table 6.2. Wire bond formation [33].

In this technique, wire is passed through a hollow capillary, and an electronic-flame-off system
(EFO) is used to melt a small portion of the wire extending beneath the capillary. The surface
tension of the molten metal from a spherical shape, or ball, as the wire material solidifies. The ball
is pressed to the bonding pad on the die with sufficient force to cause plastic deformation and
atomic interdiffusion of the wire and the underlying metallization, which ensure the intimate contact
between the two metal surfaces and form the first bond (ball bond). The capillary is then raised and
repositioned over the second bond site on the substrate, a precisely shaped wire connection called a
wire loop is thus created as the wire goes. Deforming the wire against the bonding pad makes the
second bond (wedge bond or stitch bond), having a crescent or fishtail shape made by the imprint of
the capillary’s outer geometry. Then the wire clamp is closed, and the capillary ascends once again,
breaking the wire just above the wedge, an exact wire length is left for EFO to form a new ball to
begin bonding the next wire. Ball bonding is generally used in thermocompression (T/C) or
thermosonic bonding (T/S) process. This technique requires a high temperature raging from 100˚C
to 500˚C depending on bonding process. Heat is generated during the manufacturing process either
by a heated capillary feeding the wire or by a heated pedestal on which the assembly is placed or by
both depending on the bonding purpose and materials. Relatively small gold wire (< 75 mm) is
mostly used in this technique because of its easy deformation under pressure at elevated
temperature, its resistance to oxide formation, and its ball formability during a flame-off or
electronic discharge cutting process. Ball bonding is generally used in application where the pad
pitch is greater than 100 mm. However, the application of the pitches with 50 mm has been
reported.

Wedge bonding is named based on the shape of its bonding tool. In this technique, the wire is fed at
an angle usually 30-60˚ from the horizontal bonding surface through a hole in the back of a bonding
wedge. Normally, forward bonding is preferred, i.e. the first bond is made to the die and the second
is made to the substrate. The reason is that it can be far less susceptible to edge shorts between the
wire and die. By descending the wedge onto the IC bond pad, the wire is pinned against the pad
surface and an U/S or T/S bond is performed. Next, the wedge rises and executes a motion to create
a desired loop shape. At the second bond location, the wedge descends, making a second bond.
During the loop formation, the movement of the axis of the bonding wedge feed hole must be
aligned with the center line of the first bond, so that the wire can be fed freely through the hole in
the wedge. Several methods can be used to end the wire after the second bond. For small wires (<75
µm), clamps can be used to break the wire while machine bonding force is maintained on the
second bond (clamp tear), or the clamps remains stationary and the bonding tool raises off the
second bond to tear the wire (table tear). The clamp tear process offers a slightly higher yield and
reliability than the table tear process due to the force maintained on the second bond during the
clamp tear motion. The clamp tear process also offers a light speed advantage over the table tear
process due to fewer required table motions. However, the table tear process, with a higher wire
feed angle capability and stationary clamp, has the potential to provide slightly more clearance from
package obstructions such as a bond shelf or pin grid. For large bonding wires (>75 µm), other
methods can be used such as a cutting blade or the placement of the wire into a channel in the
wedge for wire termination. As the wedge ascends, the clamped wire is fed under it to begin
bonding the next wire.

Wedge bonding technique can be used for both aluminum wire and gold wire bonding applications.
The principle difference between the two processes is that the aluminum wire is bonded in an
ultrasonic bonding process at room temperature, whereas gold wire wedge bonding is performed
through a thermosonic bonding process with heating up to 150˚C. A considerable advantage of the
wedge bonding is that it can be designed and manufactured to very small dimensions, down to 50
mm pitch. However, factors based on machine rotational movements make the overall speed of the
process less than thermosonic ball bonding. Aluminum ultrasonic bonding is the most common
wedge bonding process because of the low cost and the low working temperature. The main
advantage for gold wire wedge bonding is the possibility to avoid the need of hermetic packaging
after bonding due to the inert properties of the gold. In addition, a wedge bond will give a smaller
footprint than a ball bond, which specially benefits the microwave devices with small pads that
require a gold wire junction.

A Multi Chip Modules, abbreviated “MCM”, is described as a package combining multiple ICs into
a single system-level unit. The resulting module is capable of handling an entire function. An MCM
can in many ways be looked upon as a single component containing several components connected
to do some function. The components are normally mounted un-encapsulated on a substrate where
the bare dies are connected to the surface by wire bonding, tape bonding or flip-chip. The module is
then personated by some kind of plastic moulding. The module is then mounted on the PCB in the
same way as any other QFP or BGA component.

Deposited MCMs are MCMs where the conducting layers are deposited on some sort of carrier,
usually ceramic but also on silicon. The metal deposition is normally done by sputtering or
evaporation in vacuum. After the deposition process the metallized surface is patterned by applying
a photosensitive resist that after exposing and developing function as an etch resist. All metal not
covered by the resist is removed. After the patterning of one layer, a dielectric coating can be
applied. For organic dielectrics this are often done by spinning or spraying, whereas non-organic
dielectrics are deposited by Chemical Vapour Deposition (CVD). Holes in the dielectric are opened
by a similar photolithographic process and new layers can be added.

6.2.1. Wirebonding on SBU

The MCM-L and MCM-Flex have significantly softer dielectrics than most other MCM
technologies. This means that a careful design of the bond pad and careful specification of the
metallization system. Also the bonding process is important.

- Gold thickness on the substrate is defined by the type of bond process used and the bond strength.
- Secondary metal finishes may be required to provide a rigid bonding surface at high temperature
- The substrate Tg is critical in determining the bond strength and reliability. If the wire bonding
process takes the board above the Tg, then localized deformation and sinking of the bond pad area
will occur.
Wire bonding typically requires between 100 and 1250 nm of soft gold on the bond pad. For thin
layers (100–200 nm) immersion gold can be used. This process is self-limiting, and gold is plated
on all exposed surfaces, after the solder mask has been applied. Once a complete gold coverage of
the surface is achieved, the plating process stops. Immersion gold is the least expensive gold finish
available. This coating is commonly used for ordinary surface mount assembly, and provides
excellent solderability and coplanarity without a significant cost increase compared to other
finishes. This means that the same finish can be used for both wirebonding and soldering.

If higher gold thickness is needed (500–1000 nm) for gold wire bonding, electroless gold can be
used. The electroless process means that all exposed metal surfaces will be plated. The process will
continue as long as it immersed into the plating bath. However, this amount of gold will have a
negative effect on the solder, causing a brittle joint.

If thicker gold is needed, electroplating has to be used. Electroplated gold is selectively added onto
the metal surfaces that are included in the external electrical circuit. This means that all wire
bonding pads must be connected together in a bus. The electroplating is a more costly process, and
the required bussing can cause some design (layout) problems.

The gold thickness influences on the wire bond pull strength. A study by McDermott is shown in
Table 6.3 below. In this study a 25 µm gold wire has been bonded onto different gold surfaces. In
all cases a 5 µm nickel layer underneath the gold.

Gold finish Thickness [nm] Average Bond Strength Standart Deviation [g]
[g]
Immersion Gold 110 8,01 1,31
Immersion Gold 125 9,60 1,30
Electroless Gold 530 9,77 1,66
Electroless Gold 900 9,93 1,24

Table 6.3. A study by McDermott

6.3. Classic Thin Films

Classic thin-film hybrid technology is an example of a technique that fall into this technology [34].
This technology is based on one thin gold signal layer on top of an alumina substrate. Crossovers
are made by wire bonding. The backside of the substrate can be gold coated for the use as a ground
plane.

Advantages
- Narrow lines possible.
- Low dielectric loss tangent (» 0.0001)
- Resistors and small capacitors can be integrated in the wiring
- Trimming of resistors is possible
- Very good thermal performance, depending on the substrate (λ~ 30-200 W/mK)

Disadvantages
- One layer may cause routing problems.
- Long bond wires for crossovers may be too inductive for some applications.
- 50 ohm lines will be as wide as the thickness of the substrate.
- High dielectric constant (~7-10)
- Medium cost

6.4. Gold wire:

Gold wire is used extensively for thermocompression bonding and thermosonic bonding. In
producing the gold bonding wires, surface finish and surface cleanliness are the critical issues to
ensure the formation of a strong bond and to prevent clogging of bonding capillaries. Pure gold can
usually be drawn to produce an adequate breaking strength (ultimate tensile strength of the wire)
and proper elongation (ratio of the increase in wire length at rupture to the initial wire length given
as a percentage) for use as bond wire.

Ultra pure gold is very soft, therefore small amounts of impurities such as 5-10 ppm by weight of
Be or 30-100 ppm by weight of Cu are added to make the gold wire workable [34]. Be-doped wire
is stronger than Cu-doped wire by about 10-20% under most conditions, thus advantageous for
automated thermosonic bonding where high-speed capillary movements generate higher stresses
than in slow or manual bonders.

6.5. Metallurgical systems

In wirebonding process, different pad metallizations are used, depending to the production
requirements. Therefore, different metallurgical systems can be formed with different reliability
behaviours. The typical metallurgical systems are [34]:

Au-Au system
Gold wire bonded to a gold bond pad is extremely reliable because the bond is not subject to
interface corrosion, intermetallic formation, or other bond-degrading conditions. Even a poorly
welded gold-gold bond will increase in strength with time and temperature. Gold wire welds best
with heat although cold ultrasonic Au-Au wire bonds can be made. Either thermocompression or
thermosonic bonds are easily and reliably made. Thermocompression bondability, however, is
strongly affected by surface contamination.

Au-Al system
Au-Al welding system is the most commonly used in wirebonding process. However, this bonding
system can easily lead to formation of Au-Al intermetallic compounds and associated Kirkendall
voids. The formation can be accelerated with the temperature and time of the operational life. There
are five intermetallic compounds that are all colored: Au5Al2 (tan), Au4Al (tan), Au2Al (metallic
gray), AuAl (white), and AuAl2 (deep purple). AuAl2 can initially form in the interface between
gold and aluminum during bonding process even at room temperature and could transform to other
Au-Al compounds depending on the temperature, time and bonding configurations. Therefore, this
system often presents a problem in reliability of the bonds.

Au-Cu system
Bonding gold wires to bare copper lead frames can cause the formation of three ductile intermetallic
phases (Cu3Au, AuCu, and Au3Cu) with overall activation energies of 0.8 to 1 eV. The formation of
these intermetallic compounds can decrease the bond strength at higher temperatures (200-325˚C)
as a result of Kirkendall voiding. The degradation is apparently dependent on the microstructure,
weld quality, and impurity content of the copper. Cleanliness of the bonding surface is extremely
important to ensure good bondability and reliability in Cu-Au systems. In addition, if polymer
material is used for die attach, the polymer must be cured in an inert atmosphere to prevent
oxidation.

Au-Ag system
The Au-Ag wire bond-system is very reliable for very long times at high temperatures. This bond
system does not form intermetallic compounds and does not exhibit interface corrosion. Gold-wire
bonds to silverplated lead frames have been successfully used in high production for many years.
Bondability problems can be caused by contaminants like sulfur. Thermosonic Au-Ag bonding is
usually performed at high temperature (approximately 250˚C) which dissociates thin silver-sulfide
films thus increases bondability of silver.

7. EXPERIMENTS

Soft gold for electronics or other applications can be obtained by controlling the structure by metal
additions into the electrolyte and also by electrodeposition with very short current pulses and long
off times. Soft gold is obtained from baths containing a small quantity of thallium ions as an
additive, while hard gold is obtained from similar baths with nickel ions as an alternative additive.
Thallium is not deposited, but affects strongly the adsorption phenomena at the gold cathodic
surface. Normal metal ions addition gives soft gold deposits, while inert metal addition give hard
gold layers. The main influence of normal metal addition is a decrease in the cathodic voltage
during deposition, with decrease in inhibition condition coming from complexant interaction with
the cathodic surface. A similar effect can be obtained with pulsed current, when the voltage increase
is limited, favouring the capacitance charging with respect to the faradaic discharge. In these
conditions, lateral growth prevails with respect to outgrowth, which becomes predominant in
inhibition conditions. Cyanide and sulphite gold baths show a similar behaviour with respect to
inhibition phenomena by complexant at surfaces.

8.1 Experimental Section

The silicon substrates wafers coated by sputtering with a gold thin layer on titanium. Before
applying electrodeposition, all substrates were degreased with acetone and rinsed with distilled
water. The electrolyte was prepared from some certain quantity chemicals and distilled water. The
electrolyte had the following composition: for the study of additives, KAu(CN)2 0,04M, K2C2O4
0,5M, KH2PO4 0,123M and K2HPO4 0,057M. To this electrolyte, were added Ni+2 15 ppm from a
nickel sulphamate, or Tl+ 40 ppm from a thallium chloride solution and the results were compared
with those obtained from the bath without additions. In some tests, a commercial gold sulphite
electrolyte was used. Platinised titanium mesh was used as the anode.

The solution composition and operating conditions are shown in Table 7.1. The bath electrokinetic
behavior was controlled by cyclic voltammetry on a gold plated electrode. So as to obtain the
voltage transients during steady state electrodeposition, apply the secondary current pulse (SCP)
technique.

Current was supplied to the cell as well as power in galvanostatic mode (AMEL 550). The current
density was 1 mA cm-2. All electrochemical tests were performed in prismatic Pyrex cells including
about 300 mL of solution; hydrodynamic conditions and current distribution were fully
characterised. Electrochemical measurements were performed in a three-electrode configuration,
utilising a measuring cell with an SCE reference electrode in connection to the working electrode
with a lateral channel Piontelli probe (mean distance of lateral slit from surface 30 µm). No ohmic
corrections were made, because of the Piontelli probe. All substrates were electrodeposited thick
enough to obtain a crystallographic orientation imposed by the electrolyte. Potentiodynamic runs
were performed on electrodes previously coated (about 3 µm) in the same electrodeposition bath,
with scan rate 0.1 mV s-1.

Electrolyte Au Electrolyte AuNi Electrolyte AuTI Operating Parameters

KAu(CN)2 0.04M Electrolyte Au + Electrolyte Au + pH 6.8
K2C2O4 0.4M 15 ppm Ni 40 ppm TI Deposition time: 100 s or 4h
KH2PO4 0.123M Current density: 1 mA cm-2
K2HPO4 0.057M Substrate: Au/Ti sputtered
Si wafer
Anode: Platinum (Pt)

Table 7.1. Composition and operation conditions for Au, Au-Ni and Au-TI electroplating solutions

The screened coulomb potential method is related to a voltage relaxation technique, as well as short
current pulses superimposed on the metal cathode during electrodeposition. Four effects are
considered for the evaluation of the voltage increase:
 1. If ohmic drop happens at less than one nanosecond, this quantity of the drop is not important
for the voltage η (t) increase.
2. The capacitance current related to the nature and amount of electroactive species adsorbed at
the electrode.
ic = Cads exp (η /Bt) dη/dt
Cads, adsorption pseudocapacitance, determines the slope of the voltage.
Bt: Transient Tafel slope
3. Current is applied for the faradaic reaction
iF= iD exp (η /Bt)
iD: Deposition current density and it reaches a steady state within a few milliseconds.
4. At the adsorbed layer, the change of concentrations occurs in the second run and introduced
with τ as a transition time constant respect to a Sand type behavior. The pulse current
density;
iP=iC +iF

The equation of the overvoltage transient;

η (t) = Bt In [iP / iD - (iP / iD-1) exp (-iDt / Bt/ CADS)]+ (RT / z / F ) t1/2 /τ

The composition was measured by EDS analysis. Thickness and morphology were investigated by
means of laser profilometry (UBM Microfocus) and SEM (Cambridge Stereoscan 360). The surface
Dimension of the surface features was obtained by image analysis software. The phase structure and
the texture of the deposits were determined by X-ray diffractometry with Cu Kα radiation and a
powder goniometer (Philips PW1830). Gold deposit structure was measured by SEM and AFM also
with using XRD and we obtain a thickness around 10 µm. Gold deposition was compared respect to
thermal analysis records and we obtain very different grain size and defects presence considering
metal additives (Ni or Tl) or without metal additives. We obtain different curves and they identify
different mechanical behavior.

7.2. Experimental Results

7.2.1. Effect of Tl+ and Ni+2 on electrokinetics and structure of gold coating from cyanide
baths

Figure 7.1 depicts the relation between voltage relation and time that is duration of gold or gold
alloy electrodepositon in the cyanide-oxalate-phosphate bath. If we compare to different baths,
different cathodic voltage increase are obtained during the electrodeposition of gold. In Tl-added
electrolyte, cathodic voltage increase occurs more than 100 mV with respect to the other
electrolytes which is additive-free or Ni added. Deposition voltage is more stable in the case of no
additives or Ni-added bath. Voltage changes during electrodeposition as a result of uncontrolled
inhibition phenomena.

Figure 7.1. Cathodic voltage variation with time during electrodeposition of gold and gold
alloys, Au-Ni and Au-Ti

If we see the cyclic voltammetry graph (Figure 7.2), we obtain very different behavior for the Tl
additive bath, no additive and Ni additive. The maximum peak reaches to 0.75 V which is related to
gold deposition, but other two solutions one of them occurs at about 1.08 V and the other occurs at
0.33 V.

SCP curves related with these three solutions are shown in Figure 7.3 give the transient parameters
and reported in Table 7.2. The transient Tafel parameter is very low for the AuTl bath, intermediate
for the Au bath and >RT/F, high for the AuNi bath and in the range of 2RT/F (typical of hydrogen
discharge). The adsorption pseudocapacitance is very high for AuTl and much less for Au or AuNi.
We can obtain that adsorbed layer’s modification dedicates to diffusion of discharcing species
during the pulse.
 Figure 7.2: Cyclic voltammetry obtained from Au, Au-Ni and Au-Tl electroplating solutions.

Figure 7.3: Secondary current pulses obtained from Au, Au-Ni and Au-Tl surfaces after 1 hour of
deposition
The diffusion coefficient is higher for AuTl rather than pure Au and AuNi at Figure 7.4. After 100s
gold deposition the surface layer surfaces behave differently with respect to pure gold bath, Ni ions
additives bath and Tl ions additives bath and this behavior is shown in Figure 7.5. Additives affect
the nucleation of the gold layers; Ni+2 ions additive decreases the grain size, while Tl+ ions additive
increases the grain size.

Figure 7.4: Variation of transition time as function of current density in Au, Au-Ni and Au-Tl
electroplating solutions

Solution Bt/mV dec-1 CADS/µF cm-2 RT/(αzF) τ/s

Au 75 45 80-250 0.25-0.55
AuNi 125 46 140-220 0.10-0.22
AuTl 25 140 200-350 0.4-2.0

Table 7.2 SCP parameters obtained after deposition of 1h with Au, Au-Ni and Au-Tl electroplating
solutions

After 4 hour deposition time, 10 µm thickness is obtained with applying a current 1 mA cm -2.
Figure 7.6 shows the surface after experiments. Table 7.3 presents the parameters of thickness,
microhardness, elastic modulus and roughness.
 Solution Thickness/µm HV E/GPa Ra/nm
Au 6.0 128 76 86
Au+Ni 8.2 140 76 34
Au+Tl 9.0 86 77 129

Table 7.3 Thickness, hardness, elasticity modulus and surface roughness for the Au, Au-Ni and
Au-Tl coating surfaces after 4h of deposition at 1 Ma cm-2.

The hardness of soft AuTl is lower than Au and hard AuNi. Each elastic modulus is nearly constant
and the roughness of AuTl is greater than values of Au and AuNi. XRD patterns show a decrease in
the peak enlargement at middle height for AuTl with respect to Au and AuNi, and an increase in the
preferred orientation (111) for AuTl, while AuNi shows a strong (110) preferred orientation.

Figure 7.5: a)Au, b)Au-Ni and c) Au-Tl electrodeposited surfaces after 100 s of deposition at 1 mA
cm-2 on Au/Ti plated Si wafer

Figure 7.6: a)Au, b)Au-Ni and c) Au-Tl electrodeposited surfaces after 4h of deposition at 1 mA
cm-2
7.2.2. Effect of very short pulse time of deposition on structure of gold layers from sulphite
bath:

A commercial gold suphite based electrolyte was used for direct and pulse plating on sputtered gold
substrates and it contains 10 g L-1 gold and the other supplementary compounds Au(S2O3)2-3 0.05M,
Na2SO3 0.4M and Na2S2O3.5H2O. During the direct plating, current density was applied 4 mA cm-2
for 35 minutes. Besides, pulse plating was performed for 60 minutes with tON = 0.1 ms at iP=400
mA cm-2 and tOFF=100 ms at zero current. Other plating conditions: room temperature and pH 6.8.

Figure 7.7: XRD patterns of Au deposited with direct or pulsed current from sulphite bath and
comparison of (111) peak

Figure 7.7 shows the XRD patterns of the two samples, with a higher half width full maximum for
directly electroplated gold and an almost half value for the pulse plated deposits. Pulse plating also
induces a strong preferential orientation along the (111) plane, while directly plated gold is
characterised by a more intense XRD peak for the (220) plane. Figure 7.8 shows the surface of the
electrodeposited gold layer obtained with direct or continuous current.
 Figure 7.8: Surface of Au deposited with a) direct or b) pulsed current from sulphite bath.

8. CONCLUSION:

During my study, I tried to understand behavior of gold and gold alloy electrodeposition in micro
electromechanical systems (MEMS) with searching previous experiments and sources. First, the
advantages and disadvantages of this plating process were investigated. Pure gold is so expensive
therefore, reduction of the gold quantity on the substrate was aimed for the commercial demands
with respect to its desired properties. Au, Ag, Ni, Cr generally were used in the replace of huge
amount of gold.

Electrolytes that are used in the gold electrodeposition process, were investigated. Due to the good
compability with the photoresist materials and its bath stability, sulfite baths were used. Also in the
condition of the laboratory and in the terms of human health, non-cyanide baths were chosen during
the experiments.

Gold electrodeposition occurs in strongly complex solutions of cyanides or thiosulphates–sulphites.

The effects of temperature and laboratory conditions must be thought and experiments should be
obtained in strongly inhibited conditions. Applied current densities were evaluated in terms of
physical and chemical features of deposited material. Generally for DC plating, current density, in
the window chosen between 0,5 mA cm-2 and 5 mA cm-2, were applied to observe the its effects on
the stress, adsorption, growth mechanism, constituents distribution of deposited gold alloy.
Additives were studied and the effect of metal ion additives on the structure and morphology of
gold deposited from cyanide or sulphite baths was confirmed by the present results: Ni+2 decreases
grain size, increases hardness and favours outgrowth; Tl+ increases grain size, decreases hardness
and favours lateral growth. One of the most important results in the study is the decrease in the
cathodic overvoltage during the deposition in the presence of Tl+.

Some important results have been achieved: it is possible to obtain soft gold deposits without
additives, but by controlling the current pulses. If the current pulse is in the range of tenths of
milliseconds, the capacitance current prevails and the voltage increase at the cathode is limited. As
a consequence, the grain size of the gold deposit becomes larger, with fewer defects and smoother
morphology.
REFERENCES:

1. S.Kal, A. Bagolini, B. Margesin, M. Zen, Stress and resistivity analysis of electrodeposited gold films for MEMS application, Microelectronics
Journal 37 (2006) 1329–1334
2. Y. Fu, H. Du, W. Huang, S. Zhang, M. Hu, TiNi-based thin films in MEMS applications: a review, Sensors Actuators A 112 (2004)
395–408.
3. P.A. KOHL, Electrodeposition of Gold, Section 4, Modern Electroplating, Fifth Edition Edited by Mordechay Schlesinger and Milan Paunovic
Copyright _ 2010 John Wiley & Sons, Inc.
4. ASTM International. In B374-96 (2003) Standard Terminology Relating to Electroplating; ASTM International: West Conshohocken, PA, 2003.
5. H.H.Lou, Y.Huang, Electroplating, Encyclopedia of Chemical Processing DOI: 10.1081/E-ECHP-120007747 Copyright # 2006 by Taylor &
Francis.
6. K.Wang, R.Beica and N.Brown, Soft Gold Electroplating from a Non-cyanide Bath for Electronic Applications, 2004 IEEE/SEMI Int'l Electronics
Manufacturing Technology Symposium.
7. Kanani, N. Electroplating: Basic Principles, Processes and Practice; Elsevier Advanced Technology: Oxford, U.K., 2004.
8. C. Watt, Q.Liu, and D. G. Ivey, Electrophoretic deposition of tin-rich, tin-gold eutectic solders for micro/optoelectronic applications, 214th ECS
Meeting, Abstract #2499, © The Electrochemical Society
9. H. Baker, ASM Handbook. Alloy Phase Diagrams Vol. 3, ASM International, (1992) 2-76.
10. C.C. Lee and R.W. Chuang, IEEE Transactions on Components and Packaging Technologies, 26 (2003) 416-422.
11. S. Kal, A. Bagolini, B. Margesin, and M. Zen, Microelectron. J., 37, 1329 (2006).
12. J. J. Kelly, N. Yang, T. Headley, and J. Hachman, J. Electrochem. Soc., 150, C445 (2003).
13. E. D. Winters, Plating, 213 (1972).
14. L. Silverman, B. Bernauer, and F. Pettinger, Met. Finish., 68, 48 (Aug. 1970).
15. H. A. Reinheimer, J. Electrochem. Soc. (1974).
16. S. Nakahara and Y. Okinaka, J. Electrochem. Soc., 128, 284 (1981).
17. T.A.Green, Gold Electrodeposition for Microelectronic, Optoelectronic and Microsystem Applications, Gold Bulletin, 2007
18. Lowenheim, F.A. Electroplating; State Mutual Book & Periodical Service: New York, 1997.
19. Ch. J. Raub, A. Knödler, The Electrodeposition of Gold by Pulse Plating
20. C. C. Wan, H. Y. Cheh and H. B. Linford, Plating, 1974, 61, (6), 559
21. D. L. Rehrig, Plating. 1974, 61, (1), 43
22. A. Dolati∗ , I. Imanieh, F. Salehi, M. Farahani, The effect of cysteine on electrodeposition of gold nanoparticle, Materials Science and
Engineering B 176 (2011) 1307– 1312
23. M. Muzikar, V. Komanicky, W.R. Fawcett, A detailed study of gold single crystal growth in a silica gel, Journal of Crystal Growth 290 (2006)
615–620
24. O.BERSIROVA, V. KUBLANOVSKY, The Electrodeposition Conditions of Functional Gold Films for Microelectronics, ISSN 1392–1320
MATERIALS SCIENCE (MEDŽIAGOTYRA). Vol. 10, No. 4. 2004
25. K.Wang, R. Beica, N. Brown, Soft Gold Electroplating from a Non-cyanide Bath for Electronic Applications, IEEE/SEMI Int'l Electronics
Manufacturing Technology Symposium 2004
26. I. Chandra, M.I. Jeffrey, An electrochemical study of the effect of additives and electrolyte on the dissolution of gold in thiosulfate solutions,
Hydrometallurgy 73 (2004) 305–312
27. I.R. Christie, Trans. Inst. Met. Finish., 47(1986)
28. R.F. Wolffenbuttel, K.D. Wise, Low-temperature silicon wafer-to-wafer bonding using gold at eutectic temperature, Sensors and Achcators A, 43
(1994) 22X29
29. M.A. Schmidt, Wafer-to-Wafer Bonding for Mirostructure Formation, Proceedings of the IEEE, VOL. 86, NO.8, AUGUST 1998
30. Anders Hanneborg, SILICON WAFER BONDING TECHNIQUES FOR ASSEMBLY OF MICROMECHANICAL ELEMENTS, CH2957-
9/9110000-0092, IEEE 1991.
31. C.H. Hsu, M.A. Schmidt, Micromachined Structures Fabricated Using a Wafer-Bonded Sealed Cavity Process, 0-9640024-0-X, TRF 1994
32. Q.Wang , X. Yang ,Y, Zhang , J. Ding, A simple thermo-compression bonding setup for wire bonding interconnection in pressure sensor silicon
chip packaging, Microsyst Technol (2011) 17:1629–1633
33. The Nordic Electronics Packaging Guideline: Wire bonding: Introduction of Wirebonding technology, Chapter A: Wirebonding
34. The Nordic Electronics Packaging Guideline: Multi Chip Module Center