InGaAs Cameras See Deeper into Biological Tissue

Hank Hogan

Investigators from the Wellman Center for Photomedicine and Harvard Medical School, both at Massachusetts General Hospital in Boston, have demonstrated a full-field optical coherence microscopy system that employs an InGaAs camera. The setup offers greater imaging depths than more conventional silicon sensors provide for biological imaging applications.

Incorporating an InGaAs camera into a full-field optical coherence microscopy system enables greater imaging depths t han conventional silicon sensors for biological applications. Cross-sectional images of human thyroid tissue (far left) and cells from a frog embryo (near left) demonstrate the performance of the system. Courtesy of Wang-Yuhl Oh. ©OSA.

Full-field optical coherence microscopy is an interferometric technique in which a broad-bandwidth illuminating beam is split between the sample and reference arms. The reflected light is collected from the two paths, and it is combined to produce an image at a camera. Its advantages over other optical sectioning methods include parallel detection, high axial resolution and its use of inexpensive excitation sources.

The image penetration depth is limited by scattering in the sample. This is less of an issue in biological tissue for near-infrared wavelengths. However, silicon sensors, the imagers of choice, do not have a good responsivity for wavelengths longer than 1 µm. The researchers therefore turned to InGaAs cameras, which respond well at wavelengths of 0.9 to 1.7 µm.

They chose a 12-bit, 60-Hz, 320 × 256-pixel camera from Sensors Unlimited Inc. of Princeton, N.J., a part of Goodrich Corp.’s Optical and Space Systems Div. They used a Newport Corp. Oriel xenon arc lamp and a beamsplitter to illuminate the reference and sample arms. They used the same Optics for Research Inc. 20×, 0.45-NA microscope objectives for both arms.

Research fellow Wang-Yuhl Oh noted that the optical alignment required care. “Matching the reference arm and the sample arm for length and dispersion is one of the most critical alignment steps,” he said. Once that was done, however, the setup was stable.

The scientists compared the system’s performance with one based on a Dalsa Corp. 128 × 128-pixel, 12-bit, 490-Hz silicon camera using both a test arrangement and biological tissue samples. They found that they could image at a greater depth using the InGaAs camera, with the difference depending on the conditions. For example, the InGaAs camera achieved an imaging depth of 700 µm, versus approximately 400 µm for the silicon camera in the same situation.

Choice of sensors

Silicon imagers are the more mature technology and, thus, offer lower prices and larger formats. Because they operate in the visible, cameras that use the sensors also benefit from the wide range of available antireflection coatings for objective lenses. The same is not true for the wavelength range above 1 µm.

Employing an InGaAs camera has potential for ultrahigh-resolution imaging in highly scattering samples and in applications where large image penetration is desirable, Oh explained, but he doubted that researchers would use the sensors exclusively. The determination of whether to use a silicon or an InGaAs system for full-field optical coherence microscopy, he said, should be based on the requirements for a given application.

One challenge for any type of full-field optical coherence microscopy system is imaging speed. Its relatively slow frame rate — in the experimental setup, approximately 1 second per frame for 85-dB sensitivity — is an issue for in vivo applications.

Several possibilities to boost the speed are under investigation.

Optics Express, Jan. 23, 2006, pp. 726-735.