Detection & imaging with Leak Microstructures
M. Lombardi, G. Balbinot, A. Battistella, P. Colautti, V. Conte, L. De Nardo, G. Galeazzi, G. Prete, A. Ferretti
DDetection & imaging with Leak Microstructures
M. Lombardi, ∗ G. Balbinot, A. Battistella, P. Colautti, V. Conte, L. De Nardo, G. Galeazzi, G. Prete, and A. Ferretti INFN, Laboratori Nazionali di Legnaro, viale dell’Università 2, Legnaro, Padova, Italy CERN, European Organization for Nuclear Research Università degli studi di Padova, Dipartimento di Fisica Galileo Galilei, Padova, Italy
Results obtained with a new very compact detector for imaging with a matrix of Leak Microstructures (LM)are reported. Spatial linearity and spatial resolution obtained by scanning as well as the detection of alphaparticles with 100% efficiency, when compared with a silicon detector, are stressed. Preliminary results recentlyobtained in detecting single electrons emitted by heated filament (E c < 1 eV) at 1-3 mbar of propane are reported. Keywords: Gaseous detector
I. INTRODUCTION
Some years ago we introduced a new kind of gaseous de-tector based on needles (or points) used as anodes, that is theLeak Microstructures (LM) [1–8]. It belongs to the familyof gaseous detectors based on points as anodes such the “De-tecteur Multipointes a Focalisation Cathodique” [9–11], andthe “Pin Detector” [12, 13]. The LM detector is different fromthe above detectors for the geometry and for some properties.The geometry of a LM is really simple (fig. 1): a nee-dle, 315 µ m diameter, whose point (in the order of 20 mi-crons) acts as anode, is inserted in a hole (0.35 mm) drilledon a vetronite supporting structure (G-10, commonly used forprinted circuits), which is coppered on both sides: one sideis the cathode of the detector, on the other one the needleis welded. The point of the needle emerges from the plainof the cathode 0.1-0.2 mm to work at atmospheric pressure,about 1 mm to work at low pressure [14]. The copper of thecathode surrounding the needle is removed to create an in-sulating space (b in fig. 1) for about 550-600 µ m in diame-ter to work at atmospheric pressure, 750-800 µ m to work atlow pressures (1-3 millibar). The thickness of the supportingstructure, 4.5 mm, allows a very good centering of the needlesto respect the insulating space of the cathode.At distance of 3-10 mm from the cathode, a drift electrodeis fixed, which marks the boundary of the active volume ofthe detector: this space is filled with gas. The electric field,generated between the point and the cathode, allows the pri-mary ionization electrons multiplication. For each single ion-izing radiation detected, a pair of “induced” charges (signals)are generated, one anodic and the other cathodic, with thesame amplitude and time duration, of opposite sign and intime coincidence, which are both proportional to the primaryionization. Such redundancy of information makes possible,for example, for one signal to be used to determine the en-ergy of the incident radiation or to get timing information;the other, to provide spatial information like the point addressin a matrix of LMs.Let us shortly remember some properties of these mi-crostructures. The absolute lack of isolating material betweenanode and cathode in the active volume avoids the charging- ∗ Corresponding author (e-mail: [email protected] ) Fig. 1.
Cross section (not to scale) of a LM. up phenomena that can alter the electric field and thus theresponse of the detector: the LM detector shows a very stableand repetitive behaviour.The extreme sensitivity: it is able to detect the electronsemitted by a heated filament (E c < 1 eV), clearly not ionisingparticles [7, 8]. The high gas gain: higher than 10 in detect-ing single electrons emitted by heated filament and more than6 · in detecting X photons from a 5.9 keV Fe source [6]working in proportional region.
II. IMAGING
In fig. 2 it is shown a detector with a LM matrix made by21 ×
21 LMs (441 LMs all together), pitch 3 mm, distributedon a 60 ×
60 mm surface. This last version of LMs matrixpresents some innovations with regard to that described inref. [7–10]: some experimental measures and various simu-lations have shown that, in order to avoid the imaging deadzones, the pitch of the needles has to be 3 mm or less.The cathode of each LM is divided into square pads so thateach microstructure is built with the point of the needle sur-rounded by four pads: two for the X coordinate and two forY. This structure is particularly suitable for imaging purposesbecause the cathodic charge spreads on to the four pads ac- a r X i v : . [ phy s i c s . i n s - d e t ] S e p Fig. 2.
Detector with a matrix of 441 LMs and electronic set-up. cording to the position of the striking radiation (fig. 3). Toachieve a better stability of operation with the high voltagethe thickness of the copper-pad- cathode, originally of 70 µ m,was increased to 100 µ m by galvanic deposition and has beennickel-plated in order to avoid oxidations. Fig. 3.
3D scheme of a LM physical process: the cathodic charge isdistributed on four pads according the position of avalanche.
Putting half a razor blade as mask on the drift electrode,which is fixed at a 3 mm distance from the cathode, and work-ing in isobutane at atmospheric pressure, lightning with anextended X-ray generator, we obtain the image of fig. 4. Itis important to underline that this image did not undergo anyfurther software elaboration: it is just the presentation of therough data. With this new detector, we can then obtain well-defined images without any imaging dead zone [10].The electronic chain to read-out the signals is very simple(fig. 5): as above said, each LM has four cathode pads, twofor the X position (left and right) and two for the Y (up anddown). We have seen that, by shortcutting all the 441 Leftpads together as well as all the 441 Right, Up and Down pads
Fig. 4.
Shadowgram of a half razor-blade obtained with a matrix ofLMs in 1 bar of isobutane. (fig. 5), so that they can all be read with only four channels(4 Preamp., 4 Main Amp. and 4 ADCs), the overall spatialresolution is still very good.
Fig. 5.
Electronic chains for imaging with a matrix of 21 ×
21 LMs.
The next step is to address the matrix points to get an ex-tended image.The charge collected by a needle (anode) is able to turn on aShottky diode and through it to reach a preamp. Two Shottkydiodes (housed in a sot23 case which is less than 3 × )are connected on the rear face of the detector to each needlein order to split the signal towards two delay lines for the ad-dressing. With this tricky solution only six analog-to-digitalconverters are required: four for the pad-channels and two toread the TAC (Time to Amplitude Converter) which providesthe address X,Y of the points.The overall mean spatial resolution across the edges offig. 4, about 460 µ m FWHM, is compatible with the oneof the image in fig. 6, obtained with a single LM and else-where described [8], taking in account that the former wasobtained with a copper anti-cathode at 20 kV while the latterwas obtained with a 5.9 keV Fe source; the mask of 7 holes,300 µ m in diameter and separated by 100 µ m, was also wellcentred on the anode. Fig. 6.
Zoom effect due to different drifting electric field; staringfrom left, the drifting electrode was at 50, 100, 200, 500 V, respec-tively.
We can then conclude that a LM matrix detector with a3 mm pitch allows a X-ray imaging on a 60 × surfacewith a quite good spatial resolution using a very simple andcompact electronic setup. III. SPATIAL LINEARITY AND SPATIAL RESOLUTION
Using an X-ray tube, with a copper anti-cathode at 20 kVas source, and using a LM detector with a 2.4 mm pitch weevaluated the spatial linearity in 1 bar of isobutane. A 0.1 mmslit has been put on to the drift electrode and shifted alongthe X-axis of constant 0.4 mm pitches. In fig. 7, we can ap-preciate the quite good spatial linearity: x-axis reports themicron-measured steps, whereas in the y-axis there is the po-sition measured by the detector. The errors reported in theplot are the FWHM of the acquired data. This measure al-lows also the evaluation of the detector spatial resolution, asa function of the slit position respect to the anodes (pointsof the needles), up to the boundary between two anodes: thebest spatial resolution is above an anode (A in fig. 7 evaluated279 µ m FWHM) and the worse on the boundary between twoanodes (B in fig. 7 evaluated 643 µ m FWHM). Fig. 7.
Detector position linearity.
IV. ALPHA PARTICLES DETECTION
To evaluate the detection efficiency of alpha particles weused a single LM with an
Am source; its efficiency is com-pared with a silicon detector which is known to have very highdetection efficiency.In 760 Torr of isobutane it is not possible to reach a count-ing plateau as a function of the increasing high voltage to thepoint because at this pressure alpha particles as well as X-raysemitted by the source are detected and an increase of the highvoltage increases also the X rays detection efficiency.At 400 Torr of isobutane, where the number of X-rays de-tected is negligible, we used the experimental set up of fig. 8.A lead collimator, 15 mm thick with a 3 mm diameter holehas been put at 3 mm from the cathode of a LM, well centredon the point.
Fig. 8.
Experimental set-up.
The alpha range, under these conditions, is enough to passthe collimator and enter the active volume of the detector. In a300 V/m drifting electric field, we obtained the curve in fig. 9:along the Y-axis is reported the relative counting efficiency torespect a silicon detector, which, when replacing the LM, isput in the output of the collimator in order to count how manyalpha particles pass through it in the same lapse of measuringtime. An efficiency to respect silicon detector of 100% isreached as well as the counting plateau.
Fig. 9.
Efficiency in alpha particles detection.
V. PRELIMINARY RESULTS ON SINGLE ELECTRONSDETECTION AT LOW PRESSURE (1-3 MBAR)
With a small matrix of 4 × lm ) for three different voltages applied tothe grid (V d ) at 3 mbar of propane. As calculated in ref. [14]at 3 mbar of propane, with the height of the point of 1 mm, theTownsend ionisation coefficient a is bigger than 1 everywherein the active volume of the detector starting from the grid:the avalanche spreading depends on the voltage applied to thegrid (V d ), which therefore will influence not only the drift butalso the process of multiplication.The current preamp. used has 15 mV/ µ A dynamic charac-teristic. Signals were well recorded also at 1 mbar of propane.The behaviour in proportional region is evident.
Fig. 10.
Average pulse (mV), Gas gain in 3 mbar of propane.
In fig. 11 an example of the average of pulses recorded by aLeCroy 8300A digital oscilloscope at V lm
480 V;V d
410 V in 3 mbar of propane. In order to test if the leak microstructure(LM) counter can be used to detect single-electron in STAR-TRACK nanodosimeter, single-electron pulse-height spectrawere measured in propane gas at low pressure (3 mbar) [18].Experimental data show good prospects for this single stepdetector: LM detects single-electrons operating in propor-tional mode also at low pressure, the pulse-height spectraare well fitted by theoretical Polya distribution, allowing tocalculate the single-electron multiplication efficiency, whichcan reach a value of 96%, quite similar to that of the MSACdetector proposed by A. Breskin.
VI. CONCLUSIONS
Up to now we developed the part of multiplication of pri-mary electrons and the rough handling of the signals obtainedin a very simple compact and reliable way to obtain imageswith quite good spatial resolution but with low X-ray conver-sion efficiency. In fact the conversion efficiency of soft X raysin 1 bar of isobutane is less than 1% [15]. Next part of this job,to enhance the detection efficiency, is to use a suitable pho-tocathode which could be of Secondary Electrons Emission(SEE) type [16, 17] or some other type. The detection capa-bility of single electrons at very low gas pressure (1-3 mbar)working in proportional region is achieved.
Fig. 11.
Average of pulses of single electron detected in 3 mbar ofpropane. [1] M. Lombardi et al.,
Proceedings of the InternationalConference-Brolo (Messina) 15-19 Oct. 1996, World Scientific459-465[2] M. Lombardi et al., Nucl. Instr. and Meth., A 392 (1997) 23-27[3] M. Lombardi et al., Nucl. Instr. and Meth., A 388 (1997) 186-192[4] M. Lombardi et al., Nucl. Instr. and Meth., A 409 (1998) 65-69[5] M. Lombardi et al., Nucl. Instr. and Meth., A 461 (2001) 91-95[6] M. Lombardi et al., Pramana (Indian Academy of Sciences)Vol. 57, No. 1 (July 2001) 115-123 [7] M. Lombardi et al., Nucl. Instr. and Meth., A 477 (2002) 64-71[8] M. Lombardi et al., Nucl. Instr. and Meth., A 518 (2004) 448-451[9] G. Comby et al., Nucl. Instr. and Meth., A 174 (1980) 77-92[10] G. Comby et al., Nucl. Instr. and Meth., A 174 (1980) 93-107[11] G. Comby et al., Nucl. Instr. and Meth., A 176 (1980) 313-321[12] J. E. Bateman, Nucl. Instr. and Meth., A 238 (1985) 524-532[13] J. E. Bateman, Nucl. Instr. and Meth., A 240 (1985) 177-187[14] L. De Nardo et al., Nucl. Instr. and Meth., A 562 (2006) 127-135(Messina) 15-19 Oct. 1996, World Scientific459-465[2] M. Lombardi et al., Nucl. Instr. and Meth., A 392 (1997) 23-27[3] M. Lombardi et al., Nucl. Instr. and Meth., A 388 (1997) 186-192[4] M. Lombardi et al., Nucl. Instr. and Meth., A 409 (1998) 65-69[5] M. Lombardi et al., Nucl. Instr. and Meth., A 461 (2001) 91-95[6] M. Lombardi et al., Pramana (Indian Academy of Sciences)Vol. 57, No. 1 (July 2001) 115-123 [7] M. Lombardi et al., Nucl. Instr. and Meth., A 477 (2002) 64-71[8] M. Lombardi et al., Nucl. Instr. and Meth., A 518 (2004) 448-451[9] G. Comby et al., Nucl. Instr. and Meth., A 174 (1980) 77-92[10] G. Comby et al., Nucl. Instr. and Meth., A 174 (1980) 93-107[11] G. Comby et al., Nucl. Instr. and Meth., A 176 (1980) 313-321[12] J. E. Bateman, Nucl. Instr. and Meth., A 238 (1985) 524-532[13] J. E. Bateman, Nucl. Instr. and Meth., A 240 (1985) 177-187[14] L. De Nardo et al., Nucl. Instr. and Meth., A 562 (2006) 127-135