|
The basic principle of
dosimetry is that ionising radiation imparts energy to matter and a
knowledge of this energy provides information of the absorbed dose.
A knowledge of the rate of absorption of this energy gives
information on the dose rate. This is the basic principle of the
RADFET. RADFETs are normally p-type MOSFETs, and a cross-section of
a typical p-type MOSFET is shown in Figure 1. The region of interest
for the dosimeter is the gate region. The principle of the MOSFET is
such that when the gate voltage is below a certain value the device
is off and above this value (threshold) the device is on. It is this
threshold voltage which is used as the dosimetric parameter of the
RADFET.
Figure 1. Cross section of PMOS
Transistor
When the RADFET is exposed to ionising radiation, electrons and
holes are generated in the device. The important carriers generated
are those in the gate oxide as shown in Figure 2. A certain number
of the generated carriers recombine immediately after generation.
However those which do not recombine drift under the electric field
which is present in the oxide. If a positive bias is applied to the
gate, then the electrons will travel to the gate electrode very
quickly and leave the oxide.
Figure 2. Radiation induced charge mechanisms in
gate oxide region of PMOS Transistor
The holes move more slowly towards the silicon substrate. Because
of the nature of SiO2, hole traps exist as an intrinsic
part of the oxide, these hole traps are present in the oxide since
its fabrication. The density of hole traps is also greater near the
Si/SiO2 interface. As the holes move towards the silicon
a certian number of the holes get trapped and this causes the
positive charge within the oxide to increase. Positive charge at the
Si/SiO2 interface also increases as a function of dose.
As the positive charge in the bulk and interface of the oxide
increases, the device becomes harder to switch on, i.e. its absolute
threshold voltage becomes larger as shown in Figure 3. It is this
change in threshold voltage due to positive charge which allows the
absorbed dose to be measured.
Figure 3. Change in Transfer Charcteristics of a
RADFET as a result of irradiation
The typical
circuit which is used to measure the change in threshold voltage as
a function of dose is shown in Figure 4. In this circuit the device
is biased in saturation with a current of typically 10µA flowing
through the RADFET. The voltage, Vo, is the threshold voltage plus a
small component which is dependent on the biasing current. The ease
of use of this circuit is the main reason why it is used in many
dosimeter applications.
Figure 4. Reader Circuit
Configuration
Typical
applications for PMOS RADFETs are:
Application
| Dose Range (approximate)
|
| Spacecraft
| 10 rad to 0.5 Mrad
|
|
| Medicine
| 1 rad to 50 krad
|
|
| Personnel
| milli-rad to 1 krad
|
|
|
For the various applications, different dosimeter radiation
sensitivities are required. As the application goes from spacecraft
to medicine to personnel the the sensitivity of the RADFET must
increase significantly and the minimum detectable dose must also
decrease dramatically. Currently the NMRC supplies RADFET parts
commercially, the data sheets of which are available here.
However to give an indication of the sensitivities, Table 1 shows
the sensitivity and the minimum dose which may be detected for the
RADFETs with gate oxides of a thickness of either 400nm or 1µm.
Table 1. Sensitivities and Minimum Dose levels
for NMRC devices
The
radiation sensitivity of a RADFET may be increased by a number of
methods, some of which include :
- 1. Increasing the bias voltage applied to the gate during
irradiation.
- 2. Varying the processing conditions of the gate oxide during
RADFET fabrication.
- 3. Increasing the gate oxide thickness, thereby increasing the
volume of oxide which is available to trap oxide charge.
- 4. Change the gate dielectric, for a single thermal oxide to a
dual layer which consists of a thermal oxide in conjunction with a
deposited oxide.
Figure 5. Threshold Voltage shift as a function
of Dose, for a number of oxide thicknesses and irradiation
biases
Figure 5 shows a graph of the change in threshold voltage as a
function of dose. This graph also shows the affect of irradiation
bias and of oxide thickness. Three different oxide thicknesses are
examined: 100nm, 400nm and 1µm. As may be seen, both increasing the
oxide thickness and increasing the irradiation bias improves the
response to radiation. There is a practical limit to the radiation
bias which may be applied. Usually this is limited by the power
supply available in a given application. There are also limits to
how thick the oxide may be grown, due to the very long processing
times required for oxides greater than 1µm. This is one of the
reasons which prompted the evaluation of dual dielectrics. Some
results for these are shown in Figure 6 which show that there is an
advantage in using thick dual dielectrics.
Figure 6. Threshold Voltage shift as a function
of Dose, for a number of dual dielectric oxide thicknesses
Other important parameters in obtaining a RADFET suitable for
radiation dosimetry are the long term-fading and the read-time
stability of the device . The long term fading is defined as the
loss of information about dose when the RADFET is removed from the
ionising radiation. This needs to be minimised for some
applications, such as personnel dosimetry. However, in other
applications it is not as significant. Figure 7 shows the fading of
a 400nm gate oxide RADFET Fading is a function of both time and
temperature, and is in fact linear with both log(time) and
log(temperature). At room temperature the fading with time is very
small.
Figure 7. %Fading as a function of time, for a
number of temperatures and biases
The read-time stability is another important parameter. When the
dosimeter is powered-up in the measurement circuit, trapped charges
near the Si/SiO2 interface may either charge or discharge
and this causes the output voltage, Vo, to either increase or
decrease. This in turn will give a variation in the estimation of
the absorbed dose measurement, depending on the rate of change.
Normally, this is examined between 10 seconds and 20 seconds after
power up. Figure 8 shows a plot of this drift which is usually less
than 1.5%.
Figure 8. %Drift as a function of dose, for
different biases
For further information on developments in the NMRC RADFET
process, contact Alexander
Jaksic at the NMRC.
|
General Information