Cadmium-based boules prepare to ignite the detector market

Flying offers the promise of holidaying in sunnier climes. However, getting through airports is a pain. Security checks are understandably more thorough than they were before 9/11, and the knock-on effect is far lengthier check-in times.

What s needed are better tools for exposing hidden incendiary devices. This would speed up security checks and improve safety. Millimeter-wave systems are an emerging option for detecting concealed weapons under clothing, but X-ray scanners are needed for penetrating the contents of a suitcase.

Existing versions of this technology can locate potentially dangerous packages but cannot always accurately identify the nature of the suspicious materials. Consequently there is a desperate demand for improvement in a technology that has failed to make any significant advances over the last 25 years.

Fortunately, improvements in X-ray detector performance are just around the corner, thanks to an impending switch from silicon-based detectors to those made from CdTe and cadmium zinc telluride (CZT), which can identify plastic explosives in the cluttered environment of a suitcase. These cadmium-based materials are excellent at stopping X-rays and converting the high-energy photons into an electrical signal. Both materials can also be employed in gamma-ray detection cameras for space exploration and medical procedures. The selection of one material over another comes down to personal preference and tends to divide users.

Detection with cadmium-based materials is energy selective and garners information concerning the way in which different materials, individually or in combination, interact with the X-ray tube s energy spectrum. This is a major plus point because it enables accurate identification of materials across numerous market sectors, including medical imaging, security screening and industrial inspection.

So, given CdTe and CZT s strengths, why are so few of these detectors on the commercial market today? Well, the catch is simply this: the materials are very tricky to manufacture. Supply is uncertain, material quality is generally poor and small pieces are the norm. This has meant that many equipment suppliers are wary of turning to CdTe and CZT, and they have backed alternative detector methodologies instead, such as scintillator/silicon arrangements or photomultiplier tubes. However, cadmium-based compounds are routinely employed in some high-value, cost-insensitive applications.

The problems associated with CdTe and CZT stem from the available crystal-growth techniques. Two options exist: growth from the melt and from the vapor. The latter has the edge because both compounds have significant vapor pressures below their melting points. Vapor pressure growth advantages include a far lower growth temperature (CdTe melts at 1050 °C), reduced thermal stress and less contamination. Importantly, the crystal grows in a narrow temperature window, which minimizes unwanted tellurium precipitates and inclusions. The vapor process is also self-purifying and prevents the transport of volatile impurities to the growth crystal.

Despite the advantages of vapor-based methods, today s commercially available CdTe is grown by liquid-phase Bridgman techniques. These tend to involve the transfer of an ampoule containing the molten source through a decreasing temperature gradient. At the melting point the source material freezes and ideally forms a crystal.

In its simplest configuration this method involves mechanically pulling an ampoule or crucible through the vertical temperature gradient. The downside is that the crystal is grown in contact with the crucible, which causes stresses due to thermal contraction (this was one of the findings of crystal-growth experiments carried out by NASA on the 1992 United States Microgravity Laboratory space mission).

The seed for CdTe growth via the Bridgman method is a large native polycrystalline ingot. Freezing creates a few large crystals along preferred growth directions, which can be "mined" from the resulting boule, but yield is low – typically less than 15%. In addition, CdTe suffers from extended defects, such as twinning and a subgrain boundary structure (mosaicity), which are associated with the high dislocation densities that are typically in the low 105 cm–2. Often CdTe crystals possess twin boundaries that are decorated with tellurium inclusions (1–10 µm in diameter) with dislocation densities in excess of low 104 cm–2, which cause subgrain structure. Tellurium inclusions can impact detector-grade material because they distort internal electric fields and cut the detector s sensitivity and resolution.

The first attempts at vapor growth used closed silica ampoules. Solid-source material was heated to form a gas at one end of the ampoule, before being directed to the cooler side where it crystallized.

Although the method sounds simple, it is virtually impossible to control, due to the link between sublimation, transport and growth rates. These are also incredibly sensitive to the temperature profile, which can cause unwanted fluctuations in stoichiometry and an increase in residual impurity content.

Adding a flow restrictor, such as a capillary, combats some of these weaknesses because mass transport is then governed by the pressure difference across the restrictor, and source and growth sides are thermally decoupled. This technique was first applied to the growth of mercury iodide and has subsequently been used in the NASA-funded CdTe program. The approach promised the independent optimization of the growth and source regimes, but in reality poor thermal decoupling occurred between these two, due to radiative heat transport. To make matters worse, vapor pressures could only be inferred indirectly after the growth was over.

Progress was made with vapor-based processes in the 1970s. The Russian scientists Markov and Davydov showed that it is possible to grow a crystal on a seed plate surrounded by a small ring. This avoids contact with the container walls, which leads to a high density of defects. Excess and impurity vapor constituents can be sucked down the annulus and the resulting flow can constrain the crystal geometry to the dimensions of the seed.

Researchers at Durham University, UK, extended this approach with a novel vapor growth process that has been granted a patent in the US, the EU and Japan. The Durham System, a multitube physical vapor transport tool, minimizes axial radiative coupling by physically displacing the source and growth furnace (see box, "Scaling CdTe and CZT growth"). This reactor also allows mass transport rates to be determined directly via optical absorption measurements that reveal in situ vapor pressures.

At Kromek, formerly Durham Scientific Crystals, we have acquired the intellectual property for this technique and we are working to commercialize this vapor-phase process. One of our tasks has been to scale the growth tool, and it is now capable of simultaneous growth of several crystals with various diameters at different growth rates.

Sourcing suitably sized seed crystals is a major stumbling block to the growth of cadmium-based crystals. The largest commercially available melt-grown material is just 2 inches in diameter. This has restricted our ingots to that size. But in late 2007 we started to consider a switch to GaAs seed plates. Although this invokes the penalty of heteroepitaxial growth, it is offset by the benefit of employing far larger GaAs substrates of up to 6 inches.

Turning to heteroepitaxy is not as outrageous as it may sound – other researchers have already produced CdTe-on-GaAs films by MBE, MOCVD and hot-wall epitaxy. These films were up to 10 µm thick and our aim was to scale this to several-millimeter-thick bulk single crystal.

We have already had great success and produced 2 inch diameter single-crystal boules of CdTe and CZT, which are both 12 mm thick. This is the first growth of true bulk material several millimeters in thickness by an epitaxially seeded process.

Both materials compare favorably to material grown by liquid-phase techniques, according to various characterizations. X-ray diffraction measurements show that CdTe forms excellent single-crystal material – the double-axis rocking curve full-width at half maxima is just 25 arc seconds. Resistivity profiles of this compound show good uniformity and a value of 109 Ω cm. Meanwhile, energy-dispersive X-ray analysis on a 5 mm × 5 mm dice cut from single-crystal CZT shows the uniformity of zinc composition in all directions.

Serving the markets
We can produce detectors in house because we have the facilities for slicing, polishing and adding electrodes. However, we don t just want to be a material supplier: we want to move up the chain and offer detector modules, drive electronics and subsystems.

Our current efforts include the development of prototype modules that will enable us to assess the full potential of our material. We are now developing a complete baggage-screening unit employing CdTe detectors for aviation security, thanks to a £350,000 ($0.7 million) contract from the UK s Home Office, Department for Transport, Metropolitan Police and Centre for the Protection of National Infrastructure. This is due for delivery in early 2009. Trials are likely to follow, before airport deployment.

We have also built a small unit designed to screen for liquid explosives held in containers such as bottles and cartons. This has already received a very encouraging international response, and trials are planned in the UK, the US and the Middle East.

The same CdTe technology could also be used to replace conventional X-ray detectors in industrial inspection tasks, such as checking for flaws or anomalies in assembled products and foodstuffs. Switching to cadmium-based detectors would deliver more information about these products, thanks to the material s energy selectivity. A more thorough form of quality control is also possible.

There are many exciting applications in nuclear medicine and medical imaging. In the same way that silicon advances have fueled the digital photography revolution, the CdTe family of materials should drive a similar revolution in digital radiography.

In all of these cases, cadmium-based detectors are offering one huge improvement over incumbent technologies – energy-selective detection. This is a significant step, because it frees us from the optical equivalent of monochrome images that are harder to interpret, and delivers full-color pictures.   

View pdf of article