skip to main content

Automation Helps Overcome Productivity Challenges in Wafer Level Packaging

Shekar Krishnaswamy

Return on investment can be fast, too. One user realized a payback in just a few months from improved throughput and MES transaction automation.

Manufacturing complexity in the outsourced assembly and test (OSAT) industry is exploding, driven both by demands for higher functionality, thinner form factors, and longer battery life in handheld devices, and by competitive dynamics (see figure 1).

On the technology front, OSAT factories are moving into more complex packaging technologies that blur the line between where wafer processing ends and packaging begins. In order to meet the challenges of 2.5D and 3D wafer-level architectures, they are becoming more like modern wafer fabs.

This means traditional packaging methods are no longer necessarily relevant, and that fab-like characteristics, life cycles, and activities will have to be learned and deployed in the OSAT environment. Fab automation technologies are a key enabler for this transition.

Figure 1. Packaging technology is rapidly changing to accommodate demand for improvements in battery life, performance, and functionality in thinner form factors.

In fact, it’s no longer possible to neatly divide the manufacturing of handheld devices such as smartphones and tablets into the traditional categories of front-end-of-the line (FEOL), where transistors are fabricated, and back-end-of-the-line (BEOL) where interconnections, packaging, and assembly take place. A so-called mid-end-of-the-line (MEOL) approach (figure 2) is evolving that has both FEOL and BEOL characteristics. It came about because in vertical architectures the interconnections between layers, such as through-silicon vias (TSVs) and bumps, must be built using FEOL tools and processes.

Figure 2. TSV-MEOL/BEOL process in overall TSV process flow. (Source: ICEP-IAAC 2012 Proceedings; STATS ChipPAC-SW Yoon et al[1])

Companies in the OSAT industry, however, have traditionally provided relatively low-margin, commoditized test and packaging services in support of BEOL requirements. To remain competitive in the wafer-level packaging (WLP) era, they must now provide their customers with higher-level engineering resources and fablike manufacturing capabilities.

In addition, advanced packaging applications are growing faster than the semiconductor industry and carry relatively high margins. So new, formidable competitors are now encroaching on OSAT turf.[2] Several leading wafer foundries and integrated device manufacturers (IDMs)—seeing MEOL as a profitable extension of their existing capabilities— either already have, or are building, wafer fab-like packaging facilities to address it.

One way OSAT companies can overcome these technological and competitive challenges is through greater use of automation to reduce errors and waste, provide increased flexibility and responsiveness, and drive higher levels of output.

AUTOMATION STRATEGIES FOR WLP

WLP factories at all wafer sizes face the challenges of delivering multiple products at acceptable yield levels while managing varying levels of complexity, multiple test methodologies, and a range of time commitments. Thus, they must more closely resemble wafer fabs in structure, operations, and performance than traditional test and assembly factories.

WLP factories require tools for photolithography, etch, CMP, dielectric deposition, sputter, plating, cleaning, inspection, measurement and test, and each of these equipment-types may be used for single-wafer, lot-by-lot or batch operations as production requirements dictate. In addition to production tools, automated material handling systems are required, especially for high-volume 300mm operations.

Monitoring and controlling such complex production operations is orders of magnitude beyond what OSAT companies have had to do in the past. Before, they could operate their factories with spreadsheet-based applications. Now, WLP requires much higher levels of statistical data analysis and precise, automated control of equipment, processes and factory operations, similar to wafer fabs.

Realistically, meeting these much more complex requirements can only be achieved by implementing modern automation and control strategies (see figure 3).

Figure 3. Automation strategies implemented using Applied Materials solutions can help OSAT factories manage complexity and meet production goals and timetables.

At the equipment level, the essential capabilities needed for WLP competitiveness are automated recipe management (RM) to reduce human errors and achieve better yields, and integrated fault detection and classification (FDC) to improve equipment availability and reduce scrap.

At the process level, statistical process control (SPC) is needed for faster data analysis and better quality. Advanced process control (APC) solutions can lead to higher yields and reduced scrap.

At the factory level, automated real-time product dispatching is a key enabler for high-volume production (see figure 4). The need for real-time dispatching occurs when a tool is available and lots are in a queue waiting to be processed.

Figure 4. Uses and benefits of real-time dispatching

Software such as Applied’s APF Real-Time Dispatcher (RTD) can determine which lot to process first to achieve the overall highest throughput. It takes into account factors such as equipment capability, lot due dates and priorities, desired cycle time, equipment setup and maintenance requirements, and ancillary resources such as masks or reticles.

Short-interval scheduling automation strategies can help OSAT companies further increase overall productivity by eliminating process and equipment inefficiencies known as “white space.” White space is the term for small gaps in processing that, in aggregate, sap factory productivity.

To eliminate white space, a comprehensive and realistic look at anticipated factory production is required. Short-interval scheduling can accomplish this. The scheduling methodology is based on the processing of large amounts of good data, highly realistic mathematical modeling of factory operations, and the assumption that tasks will be executed according to prescribed schedules.

A FOUNDATION FOR THE FUTURE

Automated solutions provide OSAT companies not only with a framework for flexible, high-quality, high-output, and profitable production today, they also provide a foundation that can accommodate future changes in production strategies.

An example is the introduction of mobile technology into WLP factories to enhance manufacturing productivity and quality. One Applied Materials customer, a global leader in flash memory storage solutions for a wide range of applications and devices, implemented mobility-based solutions to increase productivity, quality, and reliability in one of its Asian assembly and test facilities.

A large proportion of that company’s products were manufactured using 2.5D and 3D packaging technologies, applying factory automation techniques such as manufacturing execution systems (MES), statistical and advanced process controls, equipment automation, and advanced scheduling solutions. However, the effective use of these techniques was hindered by reliance on manual methods and procedures. In such a complex environment, this can lead to procedural errors in the execution of the manufacturing process, resulting in product scrap and quality issues.

In this particular factory, the customer’s manufacturing operation involved processing hundreds of part numbers at multiple machines under different conditions. The processing parameters for each operation and each part-type were encapsulated into a process recipe, and hundreds of different recipes could be qualified on any particular machine. Using an incorrect recipe invariably caused product scrap, leading to higher manufacturing costs and potential customer satisfaction issues.

A key opportunity for scrap reduction and quality improvement came from the deployment of an automated recipe management system (RMS) that extended to mobile devices. This eliminated human processing errors caused by using incorrect process recipes. The application was installed on both tablets and smartphones.

A user interface was provided to capture the sequence of automated and manual events. Initially this was meant only for configuration and diagnostics, but it provided enough value that it eventually became the universal operator interface, enabling rapid user training.

Alarm management was a key component of the application. Until the deployment of the RMS, operators were performing many unnecessary steps. The RMS alarm management function captured all these events along with their frequencies, and the information was used to educate operators about the need to eliminate unnecessary steps.

Some of the key results seen from the first phase of the project were:

  • Product cycle time reduction and fab throughput improvement
  • Streamlined equipment monitoring and alert management
  • Dramatic reduction in product quality incidents related to equipment unit issues
  • Increased operator e˜ffciency thanks to mobile access to key transaction information right next to their machines

In addition, throughput improvement and MES transaction automation alone provided a payback for the first phase of the project in just a few months per machine.

CONCLUSION

The need for WLP is growing strongly, and is introducing great complexity into the operations of OSATs. Successfully addressing this complexity requires the use of innovative automation strategies to increase manufacturing flexibility, efficiency, and quality to meet customer demands and remain competitive.

For additional information, contact shekar_krishnaswamy@amat.com.

[1]http://www.statschippac.com/~/media/Files/DocLibrary/whitepapers/2012/STATSChipPAC_ICEP2012_TSV_MEOL_and_Pkg.ashx

[2] https://www.gartner.com/doc/3130518/semiconductorpackaging-assembly-test-wafer