RadFET Irradiation at TIT ========================= Setup and source activity ------------------------- On Friday 7-Sep-2001 we irradiated 3 RadFET chips (5 working K range radfets) at TIT Co-60 irradiation facility. Co-60 emits gamma rays with E=1.17 MeV (50%) and 1.33 MeV (50%), and expected dose in air (calc. by TIT Co-60 centre) 10cm from the source is 0.861 kGy/h (86.1 kRad/h). Source activity: March 2001: 27.0 TBq March 2002: 23.7 TBq => Sept. 2001: 25.3 TBq Dose rate in air, 10 cm from source in Sept. 2001 (according to Isotope center): 861 Gy/h = 86.1 kRad/h =========== Mass attenuation coefficients (mu/rho) for 1.25 MeV photons (NIST X-ray attenuation database, http://physics.nist.gov/PhysRefData/XrayMassCoef/cover.html): Si: 5.688e-2 cm2/g air: 5.687e-2 cm2/g That means that the dose rate for 1.25 MeV photons is the same for air and Si. In a rough approximation, assuming a point source, dose rate at 10cm distance would be estimated to be 86.5 kRad/h (i.e. same as provided by TIT Co-60 center) RadFETs were irradiated 10cm away from the source for 5 hours and 12 min, thus estimated dose in Si is 450 kRad. They were placed as close togeather as possible, the distance between them was about 5mm. Annealing Wed Sep 11 2002 ------------------------- Sample irrad. to 485MRad on 2002/03/14 -------------------------------------- ch# Voltage (V) ------------------- 1 6.6 2 6.7 3 12.9 4 6.545 5 6.284 6 6.324 7 6.642 8 6.654 ------------------- Sample irrad. to 472MRad on 2001/09/07 -------------------------------------- ch# Voltage (V) ------------------- 1 5.761 2 6.083 3 13 4 7.702 5 6.117 6 6.135 7 13 8 13 ------------------- Absolute dose measurement ------------------------- Date: Mon, 17 Sep 2001 09:17:36 +0900 From: Toru Tsuboyama will be done with an AMINOGRAY dosimeter, cylindrical shaped, with D=2mm and L=20 mm. AMINOGRAY is expected to give +/- 5 value since this was calibrated with 60Co source. AMINOGRAY measures the amount of un-paired electrons which are produced by irradiation. Therefore, AMONOGRAY should be kept clean (not touched by hand). AMINOGRAY dosimeter was been calibrated within the container, so they should be irradiated in the container for correct results. Exact dosimetry was performed on 18-Sep-2001 when two Aminogray dosimeters were irradiated in the same location as previously the RadFETs for 2.11 hours. Total dose received by the dosimerers was #239 195 206 185 kRad #240 193 kRad --------------- ave. 194.17 kRad --> 92.02 kRad/h Averaging the measured dose rate and rate given by TIT we get the average dose rate at 10cm rate_20010907_10cm = 89.1 kRad/h Other experiments ----------------- 1. VA+DSSD irradiation at TIT (Kaneko, Ishino) Date: Mon, 17 Sep 2001 18:45:00 +0900 (JST) From: Hirokazu Ishino When Kaneko-san irradiated VA+DSSD, there was an aluminum boad (0.5mm thickness) in front of the target which was located at a distance of 10cm from the radiation source. He calculated the radiation dose using the EGS simulation and found that the dose difference between with and without the boad is a few percent and the obtained value is consistent with the quoted value withing 10%. So we think the 86.5kRad/h is a reliable value with an accuracy of 10%. 2. RadFET calibration at UH (Varner, Rosen) Date: Mon, 17 Sep 2001 09:29:14 -1000 (HST) From: Gary S. Varner > By comparing our voltage drop vs. dose curves (see plot) to the one made > in 1999 at Univ. of Hawaii (up to 200 kRad at dose rate about 23 kRad/h), > we can see that in our case we have some 25% smaller voltage drop at > (what we believe to be) the same dose. It will be interesting to compare your integrated dose with that of the Aminogray. Based on the deviation of the 3 measurement devices we used previously, I assigned a rather generous 20% systematic error to our measurement. To protect our RadFETs from ESD, as I recall, they were placed inside plastic housings with a thin plexiglass window. As far as the rest of the details, my memory isn't very good, so I took detailed notes in my notebook. Unfortunately, that notebook is in Hawaii and I now currently reside in Silicon Valley. Your suspicion may have some merit: > We suspect it might be due to the fact that we did not put any > material in front of our monitors in which gammas could develop full > showers, and at UH they might have done that. To understand the > difference, we would like to get more information about how > exactly was the irradiation done at UH, as well as how exactly were > the VA1 irradiated at TIT - (was there anything in front of them, > how far from the source they were, etc). however, we measured dose rates with both cover on and one time with cover off and found little appreciable difference. It would take a bit of explaining to describe the water tank and aluminum/steel chamber set up for testing at UH. Perhaps you can ask Marc Rosen about the details. I believe all of the details of the measurement that we did at Tokyo was presented in a long-ago SVD meeting (distances, dose rates, etc.), and so it might be worth trying to excavate. One important thing to keep in mind is that there were a couple of different _batches_ of RadFET, each of which have different gate Tox and therefore different response curves. Important to make sure we are comparing the same sensors. Voltage drop data ----------------- During irradiation, one of the old RadFETs (hyb. 2, k2) died - we checked it is not due to logger/read-out problems by changing the channel on RadFET logger. After irradiation the data is being taken again and annealing effect can be clearly seen. After about 1 day one more RadFET become behaving strangely (hyb. 3, k1) - the voltage drop started increasing smoothly for a while. as can be seen on the annealing plot (also checked, it is not a logger problem). During irradiation, we found out that the reader starts to behave strangely when environment temperature exceeds 28 deg (R range voltages start to raise from -13.4V even if nothing is connected). First we switched the reader off and back on a couple of times after a short pause, but later we found out the problem dissapears after aplying a fan. The reader problem persists also during annealing measurement, so now the reader is cooled also. We did not notice it before since the room temperature in B4 clean room is 24.5 deg. RadFET chips ------------ Date: Tue, 18 Sep 2001 11:43:27 +0100 From: Andrew Holmes-Siedle RADFETs for BELLE xxxxxxxxxxxxxxxxx 1. I am very glad to hear that you may want an upgrade of your RADFET detector heads during your machine upgrade next summer. I will be very glad to quote on bare chips for this. I have a good stock of such chips. Can you give me an idea of the MONTH of 2002 in which you need them to arrive in Japan? 2. AGENT. I do not have an agent in Japan but direct communication and transport from England to Japan is very good. Professor Koya Abe of Tohoku procured some encapsulated RADFETs from REM. His college office found a method of paying REM in GB pounds sterling. Alternatively, I can quote in dollars. Delivery by air freight worked OK. 3. PREVIOUS CONTACTS. Yes, I work closely with Dr. Rosenfeld on a clinical system development. Gary Varner and I worked together very well on the sensor head development but I understand he is on leave from the University, working in California. 4. My "CATALOG" is the WORD 95 document attached to this email, file rfi2001-W95.doc. Please let me know if you need more precise feedback. 5. PRICES. I will give you a quotation for 50 bare chips when we have established specifications (see below). The price will lie somewhere between 1500 and 2000 GB pounds exclusive of import charges or other tax. Exact price depends on technical specification and the amount of labour involved in preparation and import permit administration. 6. DETECTOR DATA SO FAR?. Feedback on the record of the RADFETs so far is very important to REM in deciding what to recommend to you for the upgrade. Can you cite any reports? Also, have the wire bonds made by Hawaii survived well in the severe environment of the BELLE vertex detector? As part of the technical specification, becuse of your need to make very reliable metal-to-metal contact, we need to establish clearly the chip/substrate bonding process planned (alternatives, Al ultrasonic, gold ball-bond, bump bond, beam lead etc.). The final criterion is "pull strength" of the metal-metal bond. 7. DETECTOR RESPONSES. The "catalog" attached explains the non-linear response of the RADFET under zero bias. Was the Hawaii version under ZERO or POSITIVE BIAS? If I know the conditions, I can advise on the best bias method - chip type combination for good performance up to 2 megarad. Possibly use Type K under zero bias but let us not be dogmatic at this stage. I hope to hear from you after you have looked at the attachment. regards Dr Andrew Holmes-Siedle Technical Director, REM Oxford Ltd.