Press Releases

 
 
 
 
 
 

【News】Using Plasma Treatment for the Underfill Process in Flip Chip Packaging

 

2015/09/02

 
 

Introduction

Flip chip packaging technology has become a sophisticated and highly demanded packaging technique due to its high performance, heat dissipation properties, speed and I/O density. The underfill process plays an important role as it can considerably increase the yield and reliability in flip chip assembly.

Such a process can reduce the relative displacement between the substrate and die, thus reducing the stress associated in the solder interconnection that occurs from the structure’s mechanical loading and thermal cycling. It can also protect the device against environmental hazards, thus enhancing the mechanical reliability of flip chip packages.

However, the underfill process is marked with a number of difficulties, such as potential delamination, unbalanced and lower fillet height and slow wicking speed. Slow wicking speed reduces production throughput and increases product cost, while unbalanced and lower underfill height will reduce the tolerance of a flip chip device to delamination as well as mechanical and thermal shock. These factors force the solder balls to tolerate most of the deformation of the package assembly throughout the cycling loading process, which in turn promotes premature shear failure.

These complications are related to the characteristics of the flip chip device, underfill fluid, and substrate, such as bump density, gap, die size, substrate’s surface property and varying die passivation materials. Although the underfill fluids, geometries and packaging materials are specified, underfill performance only relies on the surface characteristics of the die and the substrate.

Modification of plasma surfaces is important to realize a successful underfill process in flip chip packaging. Such modifications are carried out to improve both surface bondability and surface adhesion and also customize surface energy.

The plasma process is used in a number of applications, such as increasing height and uniformity of fillet; enhancing the pull strength of wire bonds and the underfill adhesion of flip chip devices; and modifying surfaces for improved adhesion in mold, lamination and encapsulation processes.

Plasma Technology

Plasma is a partially ionized gas that contains an electrically neutral mixture of chemically and physically active gas phase species, such as photons, electrons, ions, neutral particles and free radicals. The molecular species can do physical work through sputtering and the reactive radical species can do chemical work by chemical reaction. Hence, plasma is capable of performing various surface modification processes, such as crosslinking, surface activation, contamination removal, and etch by physical bombardment and chemical reaction.

The Argon plasma process is a typical example of surface cleaning by physical bombardment. Oxygen plasma is capable of both chemical and physical work on surfaces due to the presence of free radicals and ions. Plasma applications can use either direct mode or downstream mode. In the former mode, the substrate is exposed directly in the glow discharge zone. Since this process involves both free radicals and ions, direct plasma has been shown to be fast, aggressive and effective but it tends to affect certain devices that are sensitive to photon emission, UV light, and/or charge damage.

In the latter mode, including IFP or ion free plasma, the substrate is located beyond the plasma glow discharge zone, often downstream of the gas flow. Downstream plasma process is believed to be a mild process. This is because most of the UV light and ions are filtered before the activated species can reach the surface of the substrate. This process can be applied in devices that are susceptible to ion bombardment and UV exposure. When compared to the direct plasma process, downstream plasma process is slower. On a theoretical basis, the IFP process is believed to be the pure chemical process because mostly free radicals take part in the surface modification.

Study performed in the past has shown that plasma active species are capable of entering the gap between the substrate and die and activating the surface under the die. This activation of surface is critical for using plasma for underfill process. The results of the experiment suggest that the surface wettability beneath the flip chip die relies on the flip chip geometry, plasma chemistry, and the flip chip package materials. The size of the chip impactsaffects the effectiveness of surface cleaning in flip chip packaging.

When the die size is reduced, the surface contact angle located on the middle of the substrate and die also reduces. This means, with decreasing die size, the effectiveness of plasma cleaning increases. Following plasma treatment, the surface material also impacts the surface contact angle. Within the same plasma condition, it was observed that the surface contact angle located on the die surface is comparatively lower than the substrate surface.

Additional experimental results suggested that the surface contact angle, both on the substrate below the flip chip and at the center of the die, relies rely on the plasma source gas. When compared to argon and nitrogen, the oxygen-based plasma has more impact on the contact angle. The trend of the contact angle is O2 plasma >Nplasma >Ar plasma.

It is possible to obtain higher surface energy or lower contact angle through oxygen plasma treatment. Under the same plasma conditions, the surface contact angles on the die and substrate reduce with increased ratio of oxygen in O2/N2 and O2/Ar gas mixtures. Moreover, plasma treatment changes the surface composition below the die with a passivation layer of polyimide. Upon comparing polyimide’s oxygen content with treated and untreated samples, the oxygen composition on the surface of the die increases about 36% following oxygen IFP plasma treatment.

Analysis of the surface functional group suggested that the overall oxy-functional group increases approximately 19%. This increase in oxy-functional groups on the die surface can chemically link the underfill materials during the course of the underfill process. If the concentration of oxy-functional groups is greater, the potential for delamination would be lower. This interface delamination can be reduced through removal of contaminants from the interfaces, followed by chemical activation of the surface.

Plasma for the Underfill Process

High performance during the underfill dispensing procedure includes high fillet, high wicking speed, and homogeneous fillet height. It is often preferred that the underfill fluid dispenses rapidly, while creating an adequate fillet height with excellent uniformity.

Underfill’s wicking speed is inversely proportional to the flow-out time, which in turn is proportional to the cosine of contact angle:

T=3µL2/hγcosθ.

Where, µ is = fluid viscosity, T is = flow-out time in seconds, L is = flow distance, γ is = surface tension of liquid vapor interface, h is = bump or bump height, and θ is = the contact and wetting angle.

In some flip chip packages, the wicking speed will be faster if the contact angle is lower. Hence, the faster the wicking speed, the higher the manufacturing line throughput. A case in point from the earlier work demonstrated that the contact angle on the middle of the die was 40-degrees and 20-degrees prior to plasma treatment and following plasma treatment, respectively. The reduced contact angle changed into a reduced flow-out time from 60 to 22s, which is an improvement of 270%. This means the production throughput improves considerably when plasma is utilized in the underfill dispensing process of flip chip packaging.

Generally, a fillet heights that is are more uniform leads to a more reliably packaged device because , whereas fillet heights that are not uniform result in irregular stresses on the chip that. This may may eventually result in fillet cracking and package failure. Both height and uniformity of fillet are affected by the low surface energies of the die and substrate. The earlier experiment suggested that following oxygen IFP plasma treatment, imbalance of fillet height decreased from 44.2% to 12.60%, and the height of the fillet on the opposite side increased from 12.4% to 27.70%. This meets the parameter of fillet uniformity and height in flip chip packaging industries.

New Experimental Data and Discussion

Plasma for Underfill Dispensing Performance

To demonstrate how the underfill dispensing performance is improved through plasma, three samples A, B, and C with varied surface finishes and die sizes are taken for assessment. Table 1 shows the results. From this data, it can be observed that plasma treatment improves the wicking speed, which is seen in the reduction of the time to flow-out during the course of the underfill dispensing process.

Table 1. All samples were treated in an AP-1000, batch plasma machine from Nordson MARCH

  Sample Information
Sample A Sample B Sample C
Die Size 5 x 5 mm 5 x 5 mm 7 x 7 mm
Surface Finishes Cu/Ni/Au Cu/OSP Cu/Ni/Au
Pitch Between Joints 200 µm 200 µm 12.5 µm
Joint 100 µm 100 µm 350 µm
Plasma Parameters
O2 100 sccm 400 W 200 m Torr 3 min
Result
  Time to Flow-Out (sec)
Average St. dev Improvement
Sample A Untreated 9.07 1.61 17%
Sample A Treated 7.56 1.58 17%
Sample B Untreated 9.64 1.34 11%
Sample B Treated 8.6 1.77 11%
Sample C Untreated 24.23 4.15 37%
Sample C Treated 15.26 1.75 37%

This result matches with the earlier results; however, the scale of enhancement was lower owing to the differences in materials. Plasma treatment, in particular, increases the surface energy, enabling increased wicking speed and flow. This increased wicking speed relies on the die size and surface finish. The gold finished-flip chip displays better wicking speed following plasma treatment, while the flip chip with larger die size exhibits much more improvement.

No clear variation is seen between untreated substrates and plasma treated substrates in X-ray images, but the variation of underfill performance can be seen after completing the underfill dispensing process. Figure 1 shows the comparison of optical images between untreated and plasma treated flip chips following the underfill curing process. If plasma treatment is not done, the underfill bleed-out and imbalance on the dispensing side is are comparatively larger than the opposite side. This variation is markedly reduced in the flip chip receiving plasma treatment. It is thus obvious that plasma treatment helps in enhancing the underfill uniformity along the flip chip edge.

Figure 1. Top view images of flip chip after underfill curing process.

The underfill performance from the side view is illustrated in Figure 2. The fillet height variation between untreated and plasma treated samples can be seen at the device’s corners. The underfill material takes up more area in case of the plasma treated flip chips when compared to the untreated flip chips at the corners. This suggests that plasma improves the surface energy and leads to improved fillet height along the edge of the flip chip, allowing underfill fluid to surround the chip’s corners during the course of the underfill dispensing process.

Figure 2. Side view images of flip chip after the underfill curing process

The optical images of CSP packages following underfill curing are shown in Figures 3 and 4. This is the same as seen in Figures 1 and 2. The plasma treatment before underfill enhances the underfill performance by balancing fillet height, increasing wicking speed, and enhancing fillet uniformity.

Figure 3. Top view images of CSP after the underfill curing process

Figure 4. Side view images of CSP after underfill curing process

Plasma for Adhesion

If there is poor adhesion between the underfill and the metal finishes, between the underfill and die passivation layer, and between the substrate and underfill, it will promote the interface to delaminate in flip chip packages. Often, delamination of a flip chip package is observed at the interface between the die passivation layer and the underfill material around the die edges. Various factors influence underfill adhesion such as chemical nature and surface contamination of the passivation layer. Plasma is capable of removing impurity and chemically altering the surface for improved adhesion. The temperature cycling test is often utilized for interface delamination assessment, and confocal scanning acoustic microscopy or C-SAM is utilized for assessing whether the devices go through the temperature cycling test.

The results of the C-SAM analysis following the temperature cycling test are shown in Table 2. All samples utilized for this assessment were gold finished-flip chips. The temperature cycling range is -40 to 125°C. From the results, the O2 plasma-treated samples had no device failure following 4000 temperature cycles, whereas the two untreated samples exhibited a failure rate of 11.3% and 7.5%, respectively.

Table 2. Result of Temperature Cycling Test

Substrate Finish   No. of Cycles
1000 2000 3000 4000
PCB AU Plasma 0/58 0/58 0/58 0/58
No Plasma 0/62 0/62 1/62 6/62
PCB AU Plasma 0/50 0/50 0/50 0/50
No Plasma 0/40 0/40 1/40 2/40

The adhesion force was quantized with T-peel and Lap Shear tests so as to determine the performance of plasma performance for improved adhesion (Figure 5). A Henkel underfill material and FR-4 coated with solder mask 125µm-thick polyimide film were used for testing. The thickness of the sample was 10mm, while the adhesion area was 10mm x 25mm.

For this assessment, two plasma modes were utilized. IFP plasma was operated in the XTRAK-IFP system utilizing 45sccm of oxygen gas, 200W RF power and 200mT pressure for 120s. Direct plasma was operated in an AP-1000 plasma treatment system utilizing 120sccm of oxygen or argon gas, 400W RF power and 200mT pressure for 120 seconds.

The strength of the interface adhesion enhances following plasma treatment, as shown in Figure 5, increasing the shear and peeling forces by 216% and 435%, respectively. . The shear and peeling force increase 216% and 435%, in that order following plasma treatment. Also, the failure mode is transformed to polyimide film cohesion failure from polyimide- underfill adhesion failure during the shear test. These results suggest that a surface treated by plasma can enhance the adhesion and eliminate or decrease the delamination in flip chip packages.

Figure 5. Shear and peel force comparison between the plasma treated and untreated samples.

Conclusion

This article has shown how plasma treatment can be effectively used in capillary underfill processes. Plasma treatment markedly enhances underfill dispensing adhesion and performance at the interface, and thus eliminates or decreases the delamination at the interface. Visual analysis shows the improvement in the height and uniformity of the fillet following the plasma treatment. Oxygen plasma treatment increases the reliability of devices, thanks to reduction or elimination of device failure.

 
  相關連結: http://www.azom.com/article.aspx?ArticleID=12306