The mobile phone market in the Western World is considered to be mature. Most consumers possess one or more high- or mid-range camera 'phones and the sales volume is largely confined to "churn"; that is replacement before the end of the product life due to reasons of fashion, new features or new services.

Many of these handsets have two cameras, usually an outward-facing, high megapixel, camera for photography of still images and a lower resolution camera facing towards the user for video conferencing.

Early cell phones had very low resolution imagers, typically 0.1 mega pixels, and the picture quality was barely acceptable. Consequently, the imager resolution moved quickly from common intermediate format (CIF) to video graphics array (VGA). VGA provided three times the number of pixels as CIF and a very noticeable improvement in picture quality.

This cemented a firm perception for the majority of consumers that more pixels mean better image quality. This correlation was bolstered by mega pixels being a readily quantifiable metric that could be used for sales purposes. Consequently many consumers now expect the main camera on a mobile phone to be at least 2MPixel, with high end phones supporting resolutions up to 10Mpixel.

However, as can be seen from the data shown in Figure 1 below, this view ignores the much larger market comprising the rest of the World. In many societies where a wired telephone system has never been installed, a wireless infrastructure is often a better alternative for providing communications.

Figure 1. Population and mobile phone subscriber base in primary, secondary and tertiary markets. Source: Tessera

While cell phone volume growth is stagnating in Western markets it is burgeoning in the secondary and tertiary markets. China's Ministry of Information Industry announced on their website that the landmark of more than 500 million subscribers was passed in June of 2007.

But even with this vast number, which is growing by nearly 10 million per month, most Chinese still don't own a mobile phone. The current penetration rate is only 38 percent of China's 1.3 billion population, so there is plenty of scope for further growth. India is the second largest market behind China, with around 6 million new subscribers signing up each month.

Similarly, India is expected to exceed 500 million mobile phone users in 2010, its population being around 1.1 billion. A longer term market proposition is represented by Africa where only about 12% of the continents population of 1 billion are subscribers, grouped in a few countries. Expressed more starkly, the Western market represents a mere one sixth of the total addressable market.

Secondary and Tertiary Markets
Mobile phones for the secondary and tertiary markets are often manufactured and sold by indigenous companies with no international presence. Although the form factor is similar to handsets available from the Tier one manufacturers, the functionality is often limited to voice and text.

With such a limited range of functions the component count is low. Parts and assembly costs are therefore minimized. Each reduction in prices translates to affordability for more consumers and hence increase of total accessible market.

Indigenous companies also have an advantage over the larger, well-known phone manufacturers in that the handset interface and aesthetics can be tailored to domestic consumer tastes, where subtle differences can make or break a model. Consumers who cannot read or write, place much higher value on visual and verbal communication. This makes a camera an essential component of these base-level handsets.

The lower per-capita income in these geographic regions means handset demand is driven by cost, not features. In the more industrialized markets, the camera would be the highest possible resolution that the product could support.

However, in many of the newer markets consumers do not have access to personal computers. Thus, the vast majority of pictures captured by the camera are usually only viewed on the tiny cell phone screen. Even on a high-end phone, the screen is typically only 350 x 450 pixels, which falls somewhere between CIF and VGA resolution.

Most screens from low and mid-level phones have only one quarter of this resolution, meaning an image with greater detail simply cannot be displayed. If the pictures stay on the phone or are only sent to other phones, there is no benefit to having a camera with a resolution higher than VGA. Indeed, a conventional VGA camera is grossly over specified for this application.

VGA Single Lens Camera
The key factor that determines the price of a camera module is the imager diagonal; the relationship being almost a cubic power. This arises because the imager diagonal strongly influences the die size and hence the silicon cost per die, and also determines the diameter of the mating optics and hence their cost.

One way to decrease die size is to use smaller pixels. However image sensor manufacturers are already making die with the smallest possible pixels currently achievable. This is typically around 1.75 micrometers and most companies have road maps out to 0.9 micrometers.

As the die size is fixed for each generation of die, this means the diameter of the optics is also fixed. One of the few possibilities for decreasing the cost of cameras is to decrease the number of elements in the optical train. Typically a mega pixel camera will have up to four lenses, with as many as possible of these being made of glass to maximize the optical performance.

A VGA camera module may have three or two lenses, both of which are usually plastic. Condensing the optical train to a single plastic lens provides significant saving in the cost of the camera module but impacts the image quality.

Normally the result would be considered to be visually unacceptable. However for video streaming, or where the picture is viewed on a very low resolution screen, the difference cannot be discerned. This makes VGA single lens cameras ideal for handsets intended for secondary and tertiary markets.

Traditionally, camera modules are manufactured using chip-on-board assembly. In this approach the image sensor die is attached to a substrate and interconnected to it by wire bonds. An enclosure is then glued in place over the imager.

Meanwhile the lenses are made as discrete components and precisely aligned and positioned in a lens barrel. The imager is then powered up and the lens barrel is screwed in to the enclosure until a well focused image is obtained, then the barrel is locked in position.

Clearly, this is a slow and labor intensive process with plenty of opportunity for yield loss. The opportunity to use only a single lens opens the way to a radically different methodology to manufacture camera modules.

Wafer Scale Integrated Optics
By restricting the camera optics to just one element it is possible to switch the lens manufacture to wafer scale processing and realize a fully integrated optical component.

Wafer scale assembly is inherently much cheaper than discrete assembly because the materials and process cost are divided among the good parts on the wafer, which can number many thousands on a 200mm diameter wafer.

By using semiconductor-based processes and equipment the optics can be made with sufficient precision that they can be mated with the imager without the need for manual adjustment of focus, resulting in significant savings in assembly cost.

The optical element is typically fabricated in two parts by replication processes, where an optical surface is formed on each side a flat substrate. This does not mean that the substrate has to be optically passive. For example, it can provide a filter action like infra-red cut. Another possibility is to apply to the substrate a thin layer of metal, containing a hole, which then functions as an aperture

Alternatively, both apertures and infra-red filter coatings can be applied to one or both of the optical surfaces. The net result is a completely integrated optical component that has all the functionality necessary for a solid state camera, an example of which is shown in Figure 2 below.

Figure 2. Exploded and (inset) perspective view of an integrated optical component. It comprises two optical surfaces, aperture disc and a substrate incorporating an infra-red filter. Note the two lenses can have different diameter and may be made of different materials. This type of optical is expensive to manufacture using discrete components but is easily realised in high volume at low cost using wafer scale manufacturing techniques. Source: Tessera.

As a result of wafer level manufacturing, the lens is compact, rugged, low cost, with highly predictable optical performance and negligible part-to-part variation.

Through judicious choice of the materials chosen for the integrated optical component, it is possible for the VGA single lens camera to survive the lead-free solder reflow thermal cycle. Conventional camera modules are unable to withstand this thermal excursion and must be interfaced to the phone by a flexible lead and connector.

Figure 3. Example of an integrated optical component for a VGA single lens camera, manufactured at the wafer scale. Source: Tessera.

Notably, this arrangement is mechanically not robust and interconnect failures are one of the leading causes of camera phone returns. A reflowable camera (Figure 3, above) can be attached and interconnected to the main printed circuit board of the handset at the same time as the components using a standard surface mounting process.

The benefits are improved reliability and decreased piece part and assembly costs. This interconnect technology is best suited to small die with low numbers of interconnects, so favoring VGA over higher resolution cameras.

Conclusions
Future growth in mobile phone handset volume will come from the secondary and tertiary markets. In these geographic areas, the way the phones are used and the expectations of consumers are very different. In particularly, the majority of still and video images are only ever viewed on the screen of a handset.

The low resolution of these screens makes it possible to engineer compatible VGA format cameras using a single lens. Because handset cost is a key factor in this subscriber base, conventional camera module manufacturing techniques are inappropriate.

Instead, by using wafer scale manufacturing it is possible to produce a fully integrated optical component containing two optical surfaces, aperture and infra-red filter. This can be married to an image sensor to realize a reflowable camera module that is extremely compact, robust and above all, very low cost.

Humpston is Director, Applications and Yehudit Dagan is VP Marketing, Wafer-Level Camera, at Tessera.