Chapter 2 Literature Review
2.5. Motivations
UBM is commonly manufactured by PVD methods such as e-beam evaporation or sputtering method. Recently, EN layer is applied on UBM manufacture. As mentioned above, electroless plating possesses the advantages such as simple equipments and low material cost, and is benefited by its selective and uniform deposition. Electroless plating hence becomes one of the competitive processes for
UBM manufacture in FC-related technologies.
O’Sullivan et al. [7] reported a superior barrier capability of Co(P) to inhibit the
interdiffusion between Cu and ILD in Cu-IC in comparison with EN. Our previous work confirmed that the electroless Co(P) film might also inhibit interdiffusion between eutectic PbSn solder and Cu [8]. In addition, enhancement of thermal property and diffusion barrier capability by incorporating W in Co(P) has been
reported previously [9-11]. This thesis work hence investigated the feasibility of electroless Co(W,P) layer, both amorphous and polycrystalline, as the diffusion
barriers to various solders including eutectic PbSn, SnBi, and SAC. The samples were treated by liquid-state aging at 250C up to 5 hrs and the solid-state aging at 120C or 150C for 1000 hrs. Afterward, microstructure and composition characterizations
were performed in order to analyze the alloy reactions and morphology evolution at the solder/Co(W,P) interfaces. The values of Ea for IMC growth and shear strength of the PbSn/Co(W,P) and SAC/Co(W,P) systems were also investigated. Analytical results illustrated the good diffusion barrier characteristics of electroless Co(W,P) to various solders and thus are feasible to the UBM for FC bonding.
Chapter 3
Experimental Methods
This work presents a study on the diffusion barrier characteristics of electroless Co(W,P) to PbSn, SnBi and SAC solders. Ea’s of IMC growth and bonding strengths are investigated in PbSn/Co(W,P) and SAC/Co(W,P) systems regardless of crystallinities of Co(W,P) layers. Si wafers sequentially coated with 50-nm thick Ti and 100 nm-thick Cu were chosen as the substrate to simulate the Cu interconnects.
After pretreatments, about 6-8 m electroless Co(W,P) layers with various crystallinities were deposited on Cu/Ti/Si substrates. The structure of samples prepared in this study is schematically illustrated in Fig. 3-1. The methods of sample preparation, thermal treatments and microstructure/composition characterizations are described in the following sections.
3.1. Sample Preparation
3.1.1. Substrate Preparation
First, Si wafers were cleaned by RCA process. After forming a thin SiO2 layer on Si by wet oxidation, the wafers were coated with Ti (50 nm)/Cu (100 nm) layer by e-beam evaporation. Ti layer serves as the adhesion layer whereas Cu layer simulates
the Cu interconnects. The wafer was cut into pieces for further sample preparation.
Before the electroless deposition, the Si/Ti/Cu substrates were cleaned by de-ionized water, acetone, and de-ionized water in sequence to eliminate the contamination.
Figure 3-1. Structure of samples prepared in this study.
3.1.2. Pretreatment
Prior to the electroless plating, a pretreatment including roughening, sensitization and activation was performed to obtain a catalytic surface. The chemical formulation of pretreatment is listed in Table 3-1 and the purposes of each step are described as follows [103,105].
(1) Roughening: a treatment to eliminate the oxidation on Cu surface and to enhance Ti/Cu (50 nm/100 nm)
Si substrate Co(W,P) (6-8 μm)
Solders
the roughness of substrate surface for improving the adhesion between substrate and deposited layer. With the increase in the substrate roughness, the surface concentration of the adsorbed Sn and Pd ions increases so that the sensitization becomes less essential in this case. Roughening is carried out with 5 wt.% H2SO4
for 30 sec.
(2) Sensitization: a treatment to escalate the adsorption of Pd ions on the surface.
Besides, the sensitization reduces the induction period of the electroless metal deposition reaction, and reduces the size of deposited particles.
(3) Activation: a treatment to form catalytic nuclei for electroless plating thereon before immersing in an electroless bath. The conditions of activation such as pH of palladium chloride (PdCl) solution concentration, temperature, and surface roughness determine whether the sensitization is necessary or not.
Table 3-1. Chemicals and processing conditions of pretreatment.
Step Component Concentration Immersion Time
Roughening H2SO4 5 wt.% 30 sec
Sensitization SnCl22H2O 10 g/L
5 min
HCl 40 ml/L
Activation PdCl22H2O 0.1 g/L
1 min
HCl 8 ml/L
The existence of the adsorbed Sn ions on the substrate in sensitization process provides both a greater number of Pd ions on the substrate and a greater bonding
strength of Pd ions on the surface. The mechanism of sensitization and activation is known to involve the concept of an equilibrium shift towards formation of complex Pd anions and predominance of the number of Pd ions over Sn ions on the surfaces.
3.1.3. Deposition of Electroless Co(W,P)
Afterward, about 6- to 8-m thick electroless Co(W,P) layers were deposited on
the Cu/Ti/Si substrates. The chemical formulation for electroless plating bath is listed in Table 3-2. During electroless plating, the pH value of plating bath was monitored by a pH meter and adjusted with KOH solution so as to achieve the Co(W,P) layers with desired crystallinities. The amorphous and polycrystalline Co(W,P) films (term
as -Co(W,P) and poly-Co(W,P) hereafter) were separately obtained at pH = 8.6 and 7.6, respectively. The -Co(W,P) contains about 9 to 10 at.% of P and 0.4 to 0.9 at.%
of W, whereas the poly-Co(W,P) contains about 4 at.% of P and 8 at.% of W.
Table 3-2. Chemicals and processing conditions of electroless Co(W,P) plating.
Component Concentration (g/L)
Co Source CoSO47H2O 23
Reducing Agent NaH2PO2H2O 18
Complex Agent Na3 Citrate 144
Buffer Agent H3BO3 31
W Source Na2WO42H2O 10
pH value 7.6 or 8.6
Temperature 90C
After the electroless deposition, appropriate amounts of eutectic PbSn (Shenmao Technology Inc., SH-6309 RMA, Lot No.:7736; Tm = 183C), SnBi (Shenmao Technology Inc., PF602-P, Lot No.:0802002; Tm = 138C) and SAC (Sn-3.0 wt.%
Ag-0.5 wt.% Cu, Shenmao Technology Inc., PF606-P, Lot no.:D0806083; Tm = 217 to 219C) solder pastes were immediately applied on the Co(W,P) and a brief reflow at 250C for 30 sec in forming gas (95% N2-10%H2) was performed to solidify the
solder.
3.2. Thermal Treatment Methods
High-temperature storage test is typically used to determine the effect of time and temperature for thermally activated failure mechanisms of electronic devices. During the test, elevated temperatures (accelerated test conditions) are used without the
applied electrical bias. JESD22-A103C [106] is the standard for high temperature storage test of electronic devices and it lists 7 test conditions: (a) +125 (0/+10)C, (b) +150 (0/+10)C, (c) +175 (0/+10)C, (d) +200 (0/+10)C, (e) +250 (0/+10)C, (f) +300 (0/+10)C, and (g) +85 (0/+10)C as shown in Table 3-3. Test condition
should be selected in accord with the applicability of electronic device, material types and physical properties, packaging styles, customer requisitions, etc. Note that Condition B listed in Table 3-3 with 1000-hr test duration is most commonly adopted
for evaluating the quality and reliability of electronic devices.
Table 3-3. High-temperature storage test conditions [106].
In order to study the diffusion barrier characteristics and the interfacial reactions
of electroless Co(W,P) to PbSn, SnBi and SAC solders, test conditions in JESD22-A103C standard were chosen, e.g., Condition E (liquid-state aging at 250C) and Condition B (solid-state aging at 150C) although the JESD22-A103C is primary
for testing the electronic devices. Notably, the melting point of SnBi solder is about 138C, so that the test temperature was reduced to 120C for solid-state aging test.
The Ea of IMC growth and the bonding strength of Co(W,P)/PbSn and Co(W,P)/SAC systems were also evaluated so as to gain an in-depth understanding on the applicability of electroless Co(W,P) as the diffusion barrier layer in UBM structure.
The details of the aging tests are described as follows.
3.2.1. Liquid-state Aging
Liquid-state aging test is to study the interfacial reactions between electroless Co(W,P) and liquid solders and the diffusion barrier mechanism of electroless Co(W,P)
layer. The testing samples were placed on a hot plate, which is connected to a controller. The temperature of hot plate was set at 250C, and a glass funnel with two
pipes was used as a covering to ensure the testing samples were fully exposed to the forming gas (95% N2-10%H2) for up to 5 hrs regardless the solder types in the samples. After annealing, testing samples were quenched to room temperature and encapsulated with Acrylic cold mounting materials, afterwards, the cross-sectional surface of testing samples were grinded and polished for further micro-structural analyses.
3.2.2. Solid-state Aging
Solid-state aging test is to study diffusion barrier characteristics of electroless Co(W,P) layer, the interfacial reactions between solder and electroless Co(W,P), and Ea’s of IMC growth of testing samples. Before solid-state aging test, the solder/Co(W,P) samples were encapsulated with Acrylic cold mounting materials, grinded, polished, and observed by SEM to investigated the initial IMC thickness Afterwards, the testing samples were vacuum-sealed and sent to the furnace for
solid-state aging test at 120C or 150C for up to 1000 hrs. Samples annealed for 50, 200, 500, 800, and 1000 hrs at 120C or 150C were encapsulated, grinded, polished,
and observed by SEM to investigate the IMC growth rate and the types of IMC phases.
For the determination of the Ea’s of IMC growth in PbSn/Co(W,P) and SAC/Co(W,P) systems, solid-state aging at 130C and 170C for at least 500 hrs were also
performed and, for the purpose of comparison, the PbSn/pure Co samples were prepared by applying the PbSn solder paste on pure Co foil and tested at the same aging condition.
3.3. Ball Shear Test
As to samples for ball shear test, an array of circular bond pad pattern with 200
m in diameter was formed on the substrates deposited with -Co(W,P) or
poly-Co(W,P) by the photolithography process using an SU-8 permanent photoresist
(supplier: MicroChem Corp.; 5%-weight-loss thermal decomposition temperature = 279C in air) as the mask layer. After attaching the 300-m-diameter PbSn or SAC
solder balls on the pads, a reflow treatment in forming gas (95% N2-10%H2) at 250C for 1, 10, 20, 30 and 60 min was then carried out to form the solder ball joints. The shear test was performed in accord with the JSDEC Standard, JESD22-B117A [107], by using a Dage 4000 multipurpose bond tester supported by Schmidt Scientific
Taiwan Ltd. at Hsinchu, Taiwan, R.O.C. The shear tool standoff = 30 m above the sample surface and shear speed = 100 m/sec. The average shear force was deduced
from the test results of at least 25 bumps for each sample preparation condition.
According to the shear tests, the influence of P and W contents on the bonding joint strength was also investigated.
3.4. Microstructure and Composition Characterizations
3.4.1. Scanning Electron Microscopy (SEM)
After polishing, the testing samples were taken out from Acrylic cold mounting materials and fixed on a sample holder. Afterwards, samples were etched by etchant solution (98% methyl alcohol and 2% HCl) and coated with a thin Pt layer by Pt coater immediately to prevent from oxidation. The cross-sectional views of the samples were examined by the SEM (Jeol JSM-6500F or Hitachi S-4700) operated at 10 or 15 kV within the secondary electron image (SEI) mode so that high-resolution surface morphology can be obtained. Backscattered electron image (BEI) SEM analysis was performed for distinguishing the differences in atomic types involved in the sample. It was done on the samples with relatively smooth surfaces so that the etching and Pt coating can be omitted.
3.4.2. Composition Analysis
SEM analyses in conjunction with Energy Dispersive Spectrometer (EDS, Oxford Inca Energy 7557 or Genesis) and line scanning analyses were adopted to examine the microstructure and composition changes in the samples. IMCs were tested with at least 10 spots to ensure the phases. The computer’s Dead Time (DT) is set to run between 30% and 40% of the X-ray capture time to obtain the optimum EDS signal for EDS line scan analysis.
3.4.3. Transmission Electron Microscopy (TEM)
The cross-sectional views of the samples were examined by TEM (Philips Tecnai F-20) operated at 200 kV. Since the SEM samples mentioned above were prepared, the interfacial portions of the samples were cut by using the focused-ion-beam (FIB, FEI-201) as the cross-sectional TEM (XTEM) samples and were put on TEM copper grid for TEM analyses. The FIB and TEM were supported by Materials Analysis Technology, Inc. at Chupei, Taiwan, R.O.C.
3.4.4. X-ray Diffraction
The crystal structures of electroless Co(W,P) films were characterized with an x-ray diffractometer (XRD, M18 XHF, MacScience) operated at 200 mA and 50 kV.
The x-ray source was Cu-K radiation ( = 0.154 nm) and the signal scanning rate was 4°/min.
Chapter 4
Results and Discussion
Electroless Co(W,P) to PbSn Solder 4.1.1. PbSn/-Co(W,P) Samples
4.1.1.1. Liquid-state Aging for Long Times
Figures 4-1(a)-(f) present the cross-sectional SEM micrographs of PbSn/-Co(W,P) couples subjected to liquid-state aging for as-reflow, 20 min, 30 min,
1 hr, 3 hrs and 5 hrs, respectively. We note the prolonged liquid-state aging is for examining a complete microstructure evolution in such sample types. As shown in Fig.
4-1(a), granular CoSn2 intermetallic compound (IMC) coated with CoSn3 by the peritectic reaction [12] or CoSn3 IMCs emerged at the solder/Co(W,P) interface at the early stage of aging. Voids were occasionally observed and they might be caused by the evaporation of organic additives in the solder paste. In the sample subjected to
20-min aging, the IMCs coarsen and detachment of IMCs away from the reacting interface occurs as shown in Fig. 4-1(b). In addition, an about 1-m thick continuous
layer neighboring to unreacted Co(W,P) can be observed and the EDS analysis reveals it contains about 20 at.% of P and about 2.3 at.% of W. As a result of the accumulation of P elements at the reacting interface, such a P-rich layer is in fact a mixture of
Figure 4-1. Cross-sectional SEM micrographs of PbSn/α-Co(W,P) samples subjected
SEM images shown in Figs. 4-1(e) and 4-1(f) indicate that there is no dramatic change in interfacial morphology in the samples subjected to prolonged aging in which the thickness of P-rich layer remains the same at about 1 m. The IMC spallation from PbSn/Co(W,P) interface into solder during prolonged liquid-state aging is ascribed to the P accumulation which reduces the adhesion the IMCs.
Figure 4-2. Cross-sectional SEM micrograph of PbSn/α-Co(W,P) sample subjected to 250C/30-min liquid-state aging followed by 150C/200-hr solid-state
aging. (SEI mode; accelerating voltage = 15 kV)
Above results seem to indicate that the formation of continuous P-rich layer may
deter subsequent Co-Sn reactions. In order to verify this phenomenon, a 250C/30-min aged sample was further aged at 150C for 200 hrs. As shown in Fig.
4-2, the growth of secondary IMCs from the P-rich layer into solder was observed.
A supply of Co and Sn elements would be required for forming such bush-like
IMCs, implying the P-rich layer cannot block the Co-Sn interdiffusion and the IMC spallation is in fact an unceasing process at the reacting interface during liquid-state aging. This also indicates the P accumulation at the reacting interface plays a key role in the evolution of interfacial morphology. When the P content at reacting interface remains low at the early stage of aging, the IMCs form and coarsen accordingly at the solder/Co(W,P) interface. With the increase of aging time, a continuous P-rich layer forms and the sufficiently high amount of P deteriorates the adherence of IMCs to P-rich layer and consequently leads to the IMC spallation. During the prolonged aging, the Sn elements continuously diffuse across the P-rich layer to react with Co to form the IMCs at the interface neighboring to Co(W,P) while the IMCs neighboring to the molten solder unceasingly spall away. In the meantime, the highly accumulated P elements interrupts the coarsening of IMCs and results in an ultrafine IMC mixture in the P-rich layer as mentioned in previous study [12]. Since the spallation of nano-scale IMCs could not be visibly detected by SEM in an in-situ manner, a reacting interface containing the P-rich layer with fixed thickness was hence observed in the samples subjected to prolonged aging.
4.1.1.2. Solid-state Aging
Analytical results regarding to the PbSn/-Co(W,P) subject to solids-state aging
have been reported in detail previously [12]. In conjunction with the results of liquid-state aging presented above, the formation of IMCs apparently implies the sacrificial-type barrier feature of -Co(W,P) to eutectic PbSn solder. Further, our
previous TEM characterization revealed a finely dispersed Co2P precipitates in the
-Co(W,P) layer and the electroless layer tends to recrystallize during the aging
treatment [12]. The formation of phosphide compounds and supersaturated P elements in -Co(W,P) may block the diffusion of Cu into Co and, hence, the -Co(W,P) may
also possess the stuffed-type barrier capability since EDS has revealed a negligible interdiffusion between Co and underlying Cu layer. The -Co(W,P) is hence a combined-type, i.e., sacrificial- plus stuffed-type, diffusion barrier to PbSn solder.
4.1.2. PbSn/poly-Co(W,P) Samples
4.1.2.1. Liquid-state Aging
Figure 4-3(a) depicts the cross-sectional SEM image of PbSn/poly-Co(W,P) sample subjected to liquid-state aging for 1 hr and the corresponding EDS line
scanning profiles is shown in Fig. 4-3(b). Unlike the sample containing -Co(W,P), about 5-m thick, scallop-type CoSn3 IMCs form at the solder/poly-Co(W,P) interface
without spallation. In addition, the XTEM/EDS analysis of PbSn/poly-Co(W,P) sample revealed an about 500-nm thick, amorphous layer with relatively high W
0 4 8 12 16 20
Figure 4-3. (a) Cross-sectional SEM micrograph of PbSn/poly-Co(W,P) sample subjected to 250C /1-hr liquid-state aging and (b) corresponding EDS
line scanning profiles. (SEI mode; accelerating voltage = 15 kV). (c) XTEM micrograph of the PbSn/poly-Co(W,P) sample subjected to
0.5 μm
250C/1-hr liquid-state aging. The dotted circle denotes the area where
the selected area electron diffraction (SAED) pattern was taken. (d) TEM/EDS spectrum of CoSn3 in (c). (e) TEM/EDS spectrum of W-rich layer in (c).
content (about 15 at.%) emerging in between the CoSn3 IMCs and unreacted Co(W,P) as shown in Fig. 4-3(c). Since the W content is high in comparison with that i n previous -Co(W,P) system, we hence term it as the W-rich layer. Similar to the accumulation of P, the W-rich layer should result from the accumulation of W at the interface when Co reacts with Sn to form the IMCs due to the low solubility of W elements in the samples.
Since the amorphism has been categorized as a plausible barrier mechanism [29], a liquid-state aging up to 5 hrs was hence performed to verify whether the W-rich
layer may retard subsequent interdiffusion. Nevertheless, the IMC layer was found to thicken to about 7 m after the 5-hr aging in the presence of such an amorphous layer.
Hence the barrier capability should ascribe to the nature of chemical bonds [108], rather than to the structure amorphism as proposed by previous classification scheme [29]. The decoupling of structure amorphism with the barrier capability will be discussed in Section 4.2 [85].
4.1.2.2. Solid-state Aging
Figures 4-4(a) and 4-4(b) separately present the cross-sectional SEM images of PbSn/poly-Co(W,P) samples subjected to solid-state aging for 1000 hrs and the corresponding EDS line scanning profiles. An about 3 m in thickness, layer-like IMC formed at the solder/poly-Co(W,P) interface and, according to EDS analysis, it is mainly the CoSn3 type. The formation of IMCs iterates the sacrificial type barrier
Figure 4-4. (a) Cross-sectional SEM micrograph of PbSn/poly-Co(W,P) sample subjected to 150C/1000-hr solid-state aging and (b) corresponding EDS
line scanning profiles. (SEI mode, accelerating voltage = 15 kV)
XTEM images of the 150C/1000-hr aged sample are presented in Figs. 4-5(a) and 4-5(b). An about 250-nm thick amorphous W-rich layer (W content 10 at.%) in
5 m
between IMCs and unreacted Co(W,P) was similarly observed. The thinner W-rich layer in such a sample is attributed to the lower temperature of solid-state aging.
Figure 4-5. (a) XTEM micrograph of PbSn/poly-Co(W,P) sample subjected to 150C/1000-hr solid-state aging. (b) Enlarged picture of W-rich layer and
corresponding SAED pattern.
We note that TEM and EDS analyses detect negligible amount of Co2P precipitates and/or the Co-W alloy phase in unreacted poly-Co(W,P) and the presence of grain boundaries in poly-Co(W,P) might serve as the fast diffusion paths. Although supersaturated P and W contents are present in poly-Co(W,P), their effects on stuffed-type barrier seem comparatively less than those in -Co(W,P). Besides, subsequent kinetic analysis indicated the poly-Co(W,P) exhibits a lower Ea of IMC growth. This weakens the stuffed-type barrier capability in poly-Co(W,P) and, hence,
W-rich layer CoSn3
Co(W,P)
(b)
100 nm CoSn3
Co(W,P) W-rich layer PbSn
1 m
(a)
the poly-Co(W,P) is termed mainly as a sacrificial-type barrier.
4.1.3. Determination of Ea of IMC Growth
Figure 4-6 presents the thickness consumption of Co(W,P) layer against the square root of aging time in the samples containing various electroless Co(W,P) layers
deduced from the SEM characterizations. It can be readily seen that at the same aging
deduced from the SEM characterizations. It can be readily seen that at the same aging