MICROGEIGER: A MICROFABRICATED GAS-BASED BETA
RADIATION DETECTOR
Chester G. Wilson* and Yogesh B. GianchandaniDepartment of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor
* Corresponding author: [email protected].
ABSTRACT
This paper reports a micromachined Geiger counterfabricated from a glass-Si-glass stack of wafers. As a betaparticle passes through, a bias applied between twoenclosed electrodes generates electron cascades in the gasbetween them. This results in a current pulse or “count”. Asingle die of 2 cm2 had 6 independent chambers ranging insize from 8X8 mm to 1X3 mm. He, Ne, and a He/airmixture, which have different voltage bias requirements,are separately evaluated as background gases. Countingrates are lower in a Ne background gas than in He, but Ne ismore suitable for packaging. In tests the device was foundto detect incident beta particles from a Uranium-238 source.Counting rates of up to 22 counts/min were measured. Aswith conventional Geiger counters, the rates variedinversely with distance from the source. The microGeigerwas tested with pure He and Ne background gas with 90Sr,60Co, and 204Tl, all beta emitting isotopes. Rates up to 24counts/min. were measured.
I. INTRODUCTION
Radioactive materials, particularly uranium, arestockpiled in large and small quantities all over the worldwith varying degrees of security. There is a perpetual riskof their potential use in “dirty bombs,” which useconventional explosives to disperse dangerous radioactivematerials. Uranium-238 naturally decays into 234Th andthen 234mPa, emitting 0.8 MeV beta particles, which areessentially high energy electrons. Other possible dirtybomb ingredients include a number of beta sources,including 90Sr and 204Tl. The former is a particularlyhazardous material, as it is easily absorbed into the humanbody, where it displaces calcium in bone, remaining therewith a radioactive half-life of 27 years. Thus, there isconsiderable motivation to develop miniaturized sensors forradioactive materials.
Since a few radioactive materials emit X-rays, onepossibility is to exploit X-ray detectors, which havebenefited from solid-state technology in recent years [1].Unfortunately, most radioisotopes are not sources of X-rays, and the best way to detect most of the target speciesfor dirty bombs is through their emission of beta particles.Solid-state detectors for beta particle exist, but they arerelatively large, with sizes on the order of 1 cm2. Theytypically require cryogenic cooling to distinguish radiation
type and energy, and are particularly subject to radiationdamage [2]. Another type of device uses pixelated siliconstructures at room temperature to provide spatial imaging ofbeta particle flux [3].
Geiger counters, however, are the preferred sensors fordetecting beta radiation [4]. Typical Geiger counters utilizea tube held under vacuum, with a rod-like anode andconcentric cathode. (Fig. 1). The tube is biased at 500-1000volts, and a thin window − typically mica − allows passageof beta radiation. This radiation ionizes the gas at somestatistical rate, resulting in an avalanche breakdown, whichis measured by circuitry as a “count” corresponding to oneevent. These gas-based detectors are very reliable,temperature insensitive, require only simple circuitry, andmeasure over a much wider range of radiation species andenergies. Miniaturized gas-based detectors exist, but workonly for detection of photon based radiation, such as X-rays[5]. Again, as very few isotopes emit X-rays, there is acompelling need for a micromachined beta radiationdetector.
High VoltageSupply
300-1000V
High VoltageAmplifier
1-10 MΩ
1-10 pF(500V)
Speaker(or pulsecountingcircuitry)
Geigercounter
tube
Anode
Cathode
IonizingRadiation
IonsElectrons
Anode
CathodeWindow
Fig. 1: Traditional Geiger counters utilize a tube underpartial vacuum, and a high voltage supply applied through aresistor to a small capacitor.
II. DEVICE CONCEPTS AND OPERATION
The microGeiger device is fabricated by anodicbonding a glass-silicon-glass stack, where the silicon isEDP etched to form an anode-cathode configuration. WhenDC power is applied to this configuration two electrostaticregions are formed: a drift and amplification region (Fig. 2).As beta particles pass through the glass window into thedrift region, they ionize the background gas. The electronsare slowly accelerated into the amplification region. In theamplification region the electrons are quickly accelerated
through a higher field region resulting in an electroncascade. Designing the drift region to be much larger thanthe amplification region allows consistent pulses that arenot a function of the entry position of the beta particle.
.
Cathode
Glass 1
Glass 0
Silicon 1
Anode cathode
Betaradiation
e-
Driftregion
High-fieldAmplificationregion
Electroncascade
gas-fillhole
Optionalmountingpost
Betaradiation
Electricalconnection
Anode
Optionalmetal
Fig. 2: The microGeiger device utilizes dissolved siliconbonded to glass as the anode-cathode configuration. Betaradiation passing through the drift region creates liberatedelectrons, which travel to the amplification region, creatingan electron cascade.
Macroscale gas-based devices have been widely usedin the field of radiation detection [6]. Virtually all gas-based detectors rely on the impinging radiation ionizing thefill gas, with the resulting electrons accelerated by anelectric field, ionizing more neutral species, therebycreating an avalanche breakdown. The general form of theelectron density in a cascade of length x is given by:
n x n x( ) ( )exp( )= 0 α (1)Here, α is the first Townsend coefficient of the gas, afunction of the gasses ionizability, and electron capturecross section.
Typical detectors fall into four regimes of operation,defined by the applied electric field, electrode geometry,and the pressure and species of fill gas (Fig. 3). The fourregimes all have electric discharges with differing physicalproperties. (These regimes apply to both beta particles andphotons like X-rays and gamma particles). The regime withthe lowest voltage is the ion saturation region, where theonly charge collected is by gas directly ionized byimpinging radiation. As the voltage across the device isincreased, avalanche breakdown begins to occur, and theamount of collected current increases. This is theproportional region: the amount of current is roughlyproportional to the energy of impinging X-rays or gammaparticles, as photon radiation is completely absorbed by thebackground gas. In contrast, impinging beta particlesimpart only a portion of their kinetic energy to theionization of gas, so the resulting current created is notcorrelated to the beta energy; the proportional region ismore limited for beta particles.
As the voltage is increased, the dependence of thecurrent pulse upon the energy of the radiation is diminishedeven for photon radiation. This non-linearity is primarily
due to the difference in mobilities between ions andelectrons. In the limited proportionality regime, the muchslower ions are sufficient in quantity to create a space-charge region which distorts the local electric fields. Thislimits the total charge, such as created by avalanching andis dependent on the electric fields. (Similar space-chargeregions have been found to be the reason for lower chargedensities in previously reported microplasmas [6].) As thevoltage is increased further, the impinging radiationgenerates a self-sustaining discharge; this is the Geiger-Muller region. The total amount of current collected in thisregion for a cylindrical proportional counter is given by:
lnln
lnln
ln( / )lnM
V
b a V
V
pa b aK= ⋅ −
( / )
2
∆ (2)
Here, M is the multiplication factor, the quantity ofelectrons from a single incident. V is the applied voltage, aand b are the anode and cathode radii, respectively; p is theambient gas pressure, and ∆V and K are constants of thebackground gas, related to electron mean free path andionizability.
For the characteristic dimensions, and gas species ofthe microGeiger device, a corresponding cylindricalconfiguration would provide an M , multiplication factorranging from 10 to 105 as the operating voltages are varied.The actual multiplication of the microGeiger device is inthe range of 1-2 X1013. This is because the microGeigerdevice operates in the Geiger-Muller regime, where theinitial charge multiplication is further amplified bysecondary emission from the anode. This serves ascompensation for the small size of the microGeiger, inwhich the amplification region is smaller than in traditionaldevices.
Pul
se a
mpl
itude
(lo
g sc
ale)
Applied voltage
Ionsaturation
Proportionalregion
Limitedproportional
region
Geiger-Muellerregion
Fig. 3: Gas-based radiation detectors have four regimes ofoperation, that are defined by the applied voltage.
III. DEVICE FABRICATION
The microGeiger device is fabricated in either a two orthree mask process (Fig. 4). The first mask defines a boron-diffused etch stop. Mask two patterns an oxide whichdefines a region where silicon is etched. This silicon waferis anodically bonded to a pyrex wafer, which forms the betawindow. The silicon is etched, forming structural offsets,and boron-doped anodes and cathodes. The stack is dicedand bonded to a second glass wafer, providing the capability
of gas packaging. Photographs of the final die with sixindependent cavities is shown in Fig. 5.
IV. EXPERIMENTAL RESULTS
Unsealed devices were tested in a flow chamber topermit the comparison of various fill gasses (Fig. 7). Thebackground gas flows into the microGeiger device throughinput ports machined into the glass window. The gasseswhich were evaluated included He, Ne, and a He/Air gasmixture. Helium was chosen for its lower ionizationenergy; this allows a larger current pulse at microscaledimensions. Neon also has a lower ionization energy, butdoes not leak through metal. In the test set-up DC powerwas provided to a capacitor which powers the anode; thecathode was grounded. Figure 8 shows a uranium oresample that was used to test the devices. For all uraniumtests a He/Air gas mixture was used as the background fillgas. Figure 9 illustrates the counting rate of one chamber inthe microGeiger as a function of the distance from thedetector. The normalized decline in counts is similar to thatmeasured with a conventional detector, an Electro-Neutronics CDV-700.
Si1
G1
Si1
G1
Si1
G1
Si1
G2
Silicon wafer boron implanted, oxidized
Silicon wafer boron implanted, oxidized
Silicon wafer EDP etched
Wafer diced and bonded to glass window
Fig. 4: Process flow of the microGeiger device. Mask 1 defines aboron etch stop, mask 2 a patterned oxide. EDP etching, anodicbonding, and additional glass processing define the microGeigercavities.
Fig. 5 (a-upper) ThemicroGeiger diecontains multipledetector cavities and isorders of magnitudesmaller than traditionaltubes. (b-lower)Backside of themicroGeiger device.
Figure 10 shows the counts per minute as a function ofthe thickness of the glass windows. As the thickness of thewindow increases, more beta radiation is absorbed before itreaches the encapsulated gas as the glass thickness isincreased. This illustrates that standard glass wafers from500 µm – 750 µm can be utilized for fabrication.
As the discharge capacitor that powers the microGeigerdevice is charged to increasing voltages the chargedelivered in each pulse increases, however, the final voltageon the discharge capacitor after ignition remains fairlyconstant (Fig. 11). This provides evidence that a self-sustaining discharge is created in the device that terminateswhen the capacitor reaches a certain voltage.
V
Flow chamberHeliumtank
Gas regulatorChamber lid
Outgas regulator
DischargecapacitorBeta
sourceMicro-geiger
Va
Rl
Fig. 7: Preliminary test setup for the microGeiger device; tankallows for helium filling, and pulse discharge measurement. FinalmicroGeiger devices have the gas sealed within them, and do notrequire the chamber.
Fig. 8: Uranium ore sample used to test device.
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40
Cou
nts
per
min
ute
(nor
mal
ized
)
Distance from source (cm)
Source 1
Source 2
Referencegeigercounter
(Source 2)
Fig. 9: Normalizedcounts per minute forthe microGeiger and areference Geigercounter as a functionof source distance.
0
5
10
15
20
25
400 500 600 700 800Glass Thickness (µm)
Cou
nts
per
min
ute
Source 1
Source 2
Fig. 10: Over 20counts per minute arerealized on themicroGeiger. 200µm thickness of glassreduces the countingrate by about a factorof 2.
0
100
200
300
400
500
600
700
800 900 1000 1100 1200 1300
Vol
tage
afte
r di
scha
rge
(V)
Charged capacitor voltage (V)
Cha
rge
deliv
ered
from
cap
acito
rdu
ring
disc
harg
e (µ
C)
Voltage
Charge
6.565.554.543.532.52.01.5
Fig. 11: Increasing the capacitor bias increases the amount ofcharge delivered to the discharge. The residual voltage on thecapacitor remains fairly constant
The microGeiger device was tested with a variety ofradioactive isotopes, and fill gases. The correspondingresults are shown in Figs. 12-14. Figure 12 illustrates datafor a 500 µm thick Pyrex window, and He fill gas. Countrates were measured for three different cavity dimensions(8X8 mm, 4X4.5 mm, and 2X2 mm), and two isotopes, 90Srand 60Co. 90Sr emits 0.546 MeV beta particles, and 60C oprovides a range of beta particles, with a maximum energy of0.314 MeV. The count rate can be seen to increase as thecavity size increases, providing more collection area, for bothisotopes. Figure 13 provides data for Ne fill gas, using 90Srand 204Tl isotopes. 204Tl produces a spectrum of beta particleswith a maximum energy of 0.776 MeV. Again, the countingrate is seen to increase as the collection area increases.
Eventually, as the voltage is increased in themicroGeiger device, a continuous arc breakdown occursacross the anode-cathode gap. Pulsed breakdown due to theimpinging beta particle occurs at voltages that are lower, butstill sufficiently high to develop the avalanche breakdown.This defines the allowed voltage operating range for thedevice, which is shown for three different background gassesin Fig. 14. Helium provides the lowest operating voltagesbut can be difficult to package long term, as it leaks throughmetal. Neon eliminates this problem with only a slightincrease in operating voltage. He/Air mixtures provide amuch wider range of allowed voltages, however the countingrate becomes a function of the operating voltage.
0
5
10
15
20
25
30
0 20 40 60 80
He Background
Sr90
Co60
Cavity area (mm2)
Cou
nts
per
min
.
Fig. 12: Countingrates for a Hepackaged device,for varying cavitysizes.
0
4
8
12
16
20
0 20 40 60 80
Ne Background
Sr90
Tl204
Cavity area (mm2)
Cou
nts
per
min
.
Fig. 13: Countingrates seen in a Nepackaged device,with varying cavitysizes.
He/Air mix
400
600
800
1000
1200
1400
Ne
He
Allo
wed
Ope
ratin
g V
olta
ge
Fig. 14: Operationalrange of voltage onthe microGeigerdevice for various fillgasses.
V. CONCLUSIONSThe microGeiger device was found to measure beta
radiation produced from Uranium-238 90Sr, 60Co, and 204Tl,all beta emitting isotopes. Counting rates of up to 24 countsper minute were measured, and were found to fall as thesource moved away in distance with similarly to alaboratory grade Geiger counter. He and Ne were evaluatedas background gasses, as they provide a reduced operatingvoltage. The voltage required for a He/Air mixture ishigher, but it provides a larger range of biasing.
ACKNOWLEDGEMENTSThis work was supported primarily by the Engineering Research
Centers Program of the National Science Foundation under AwardNumber EEC-9986866. The facilities used for this research include theSolid State Electronics Laboratory (SSEL) at the University of Michigan.The authors are grateful to Mr. Ramon Torres-Isea, for use of theAdvanced Physics Laboratory. Travel support has been generouslyprovided by the Transducers Research Foundation and by the DARPAMEMS and DARPA BioFlips programs.
REFERENCES[1] S. A. Audet, E. M. Wouters, and M. H. Kim, “High-Purity Silicon SoftX-Ray Imaging Sensor”, Sensors and Actuators, (A22) nos. 1-3, Mar. 1990pp. 482-486[2] R. Wunstorf, “Radiation hardness of silicon detectors: current status”,IEEE Transactions on Nuclear Science, (44) nos. 3, June 1997 pp. 806-14.[3] E. Bertolucci, et. al., “BETAview: A Digital β-Imaging System forDynamic Studies of Biological Phenomina,” Nucl. Inst. and Methods,A381, 1996 pp. 527-530[4] W. J. Price, Nuclear Radiation Detection, McGraw-Hill, 1964, pp. 123.[5] G. Charpak, et. al., “Micromegas, a multipurpose gaseous detector,”Nucl. Inst. and Methods, A478,2002 pp. 26-36[6] G.F. Knoll, “Radiation Detection and Measurement,” John Wiley andSons, 2000, pp 161[7] C.G. Wilson, Y.B. Gianchandani R.R. Arslanbekov, V. Kolobov, andA.E. Wendt, “Profiling and Modelling of DC Nitrogen Microplasmas,”Jour. of Applied Physics, 94(5) Sept 2003, pp. 2845-51