Previous: Reticon - Bob Maddoux.
Up: A Chronological Report of the USA Visit, 17-28 May 1993.
Next: Other U.S.A. Companies.
Previous Page: Reticon - Bob Maddoux.
Next Page: Other U.S.A. Companies.
Tektronix were the last company visited during our USA visit. The impending devolution of the CCD operation was discussed. The new company would be forming from the expertise in the CCD technology section of Tek and would be called Scientific Imaging Technology - SITe.
The individuals concerned at that stage indicated that there would be a possibility of producing and , CCDs in a foundry type effort. It was hoped that the devices would be 3-edge buttable but nothing, at that time, had been committed to hardware in terms of production. They maintained that CCDs produced in their production facilities had been conservatively designed and that it should be possible to obtain devices with pixel sizes of the order of (cf. current pixel sizes of ). The devices would have a full well of and a read-noise of . The outputs would be of slightly different design with a slightly higher gain.
The Tek process results in a high impedance to the parallel clock phases which makes it necessary to drive the thinned devices at slower (vertical) clock speeds. IN most slow-scan operations this does not present a major problem.
They quoted an expected QE for the new devices of 70% at 400nm and a few % at 320nm; the AR coating to be used would be peaked at 650-700nm with a coating centered at 500nm. Some experimentation with the coatings they were ultimately going to use was to be undertaken. They have been active recently in developing a custom coating for the STIS, CCD - specified to achieve an optimum response of 50% at 300nm. They did not however reveal details of their process for peaking the response of their thinned CCDs; there did however, appear to be additional processes involved in obtaining the superior broad-band response of their devices. The materials mentioned which were used for their AR coatings were:- and a newer material of their own:- Herb12, this appeared to produce a response at 330nm of . The device tested had been subject to a 5 minute UV flood from a pen-ray lamp () which resulted in a response increase of 30% (15% pre-flood, 45% post-flood). The broad-band response was also very good, being quite flat from around 400nm to 700nm.
The dark current in these devices seemed to be high, in non MPP mode (at room-temperature) this was quoted as being () with a UV AR coating and () without a coating. They agreed this was rather high and that the goal was .
Tek have had problems with the flatness of their and devices. An approximately bow has been observed in their devices, this bow originating in the wafer itself during processing. The wafers take on the shape of part of a spherical surface and this results in a centre-to-edge change in flatness of, at worse, . This curvature should be easily seen in all our instruments, especially in optical systems with fast beams.
We were assured that they were attempting to remedy the situation but it appeared that all we may expect, in the immediate future, is for the bow to be reduced by about 50%. The TK2048 chips, including 36 on order for the Sloane digital sky survey, will remain as they are (almost bow edge-centre). The TK1024 devices have a smaller bow, in proportion to their dimension; process improvements on this device should reduce the bow to about .