Pixel CMOS Design Using Current-Mirror Circuit

NUT Sultan Salah uddin

Pusat Studi Multimedia dan Robotika, Universitas Gunadarmasultan@staff.gunadarma.ac.id

Abstract

This paper explains a study on several photodiodes sensors including current mirror amplifiers.These photodiodes have been fabricated using a CMOS 0.6 micrometers process from AustriaMikro Systeme (AMS). Each sensor pixel in the array occupies respectively, Imm x Immarea, 0.5mm x O.5mmarea and 0.2mm 0.2 mm area with fill factor 98 % and total chip areais 2 square millimeters. The sensor pixels show a logarithmic response in illumination andare capable of detecting very low green light emitting diode (less than 0.5 lux). These resultsallow using our sensor in new Gamma Camera solid-state concept.

Keywords: CMOS, current-mirror, sensor, pixel

INTRODUCTION

CMOS (Complementary MetalOxidizes Semiconductor) Sensors intendedfor imagery cause much interest for industrialapplications and research. Sensors based onthis technology exceed the traditional sensorsbased on CCD (Charge Coupled Device)technology. CMOS sensors offer manyadvantages. It offers weak manufacturingcost compared to a CCD sensor, integrationfacility of multiple functionalities intendedfor imagery, high space resolution, and

pixels random access. Some disadvantages ofCMOS sensors compared to CCD sensor arenoise of reading which is more important andsensitivity which is less important. CMOSimaging sensors use active or passive pixels,as depicted in Figure 1 and Figure 2. Active-pixel sensors (APS) include amplificationcircuitry in each pixel.

Active-pixel sensors make use of activetransistor to buffer signal output within thepixels and drive the readout circuitry of thepixel arrays.The array themselves canbe madevery flexible. Currently under consideration

Figure 1. CMOS Sensors with Passive Pixels

Figure 2. CMOS Sensors with Active pixels

Salahuddin, Pixel CMOS... 193

I Vreset/ dd

I~~ +idl fj=fvd~~~... 0Figure 3. Integration Mode Operation

is an array of active-pixel sensors that canbe readout one row at a time. A column of

row registers addresses each of the pixels ina row, each of which then sends output to abottom row of column registers and adjacentanalog circuitry. The pixel circuitry can usethis photo-current. The photocurrent of anordinary photodiode is very small; a chargeintegration based operation is common inmany active pixel sensors. In this context wedesigned a newCMOS image sensor array thatpresented in this article. Operation of currentmirror is introduced in the second section.The third and fourth sections describe the

design of CMOS Active Pixel Sensor usingcurrent-mirror circuit. In the fifth section,the paper concludes with fabrication and testresults about this new sensor.

RESEARCH METHODS

The first research phase of this designis the design of current-mirror to sensorphotodiode. The second phase is simulationof the current-mirror using mentor graphic.Furthermore, design of the layer structurephotodiode evaluates several sizes: I mm xI mm, 500 mm x 500 mm, 200 mm x 200mm. The objective is to identify the best lightsensitivity of these sensors measure. The nextphase is to unite the current-mirror circuitwith photodiode to form a pixel. Pixels thathave been put together and designed for thefabrication, all this is done using mentorgraphics software. The final step is testingand analysis of pixel fabrication results.

DISCUSSION

The Operation of Current Mirror

The basic idea ofthe design is illustrated

194 Jurnal Ilmiah Teknologi& Rekayasa, Volume15 No, 3, Desember 2010

in Figure 3. The capacitance C consists ofthe capacitance of any connected deviceat the node V plus the photodiode junctioncapacitance itself. When light illuminatesthe photodiode, the reverse photocurrentdischarges the output node V, therefore wecan write Equation I. For a n+ - p diode withND » NA, the junction capacitance is givenby Equation 2. In Equation 2, A is the diodearea, NA is the acceptor concentration in thesubstrate, and ND is the doping concentrationof the n+ region. Combining Equations I, 2,and integrating we find Equation 3, whereV

b' is the built-in voltage, and V t

is theI rese

reset reserve bias. Solving for Vet), we getEquation 4.

dV(t) .C(V)- = -~phot". (1)dt

C(V) = A[

2qE8iNA

]

1/2

2 V(t) . (2)

A 1/~[

' m]

V(t)+\1,-(2qESiNA) 2v V = -iphotot. (3)2 Vr...t+v.,

, t

]

2

1/2 IphotD 2 'Vet) = [vr...t - A(2q€8iNA)1/(4)

When Vet) is plotted as a function oftime, linearity in the output is observed for a

time period. Worthy noting is that since ipholooc loA, the dependence on photodiode arecancels out. In general, a large photodiodearea with high fill factor is desired. These areseveral advantages in using integration-mode:charge integration, low parasitic on sensenode, voltage multiplexing, no kT/c noiseand low dark current. On the other hand, thedisadvantages are smaller fill factor and morecomplicated addressing.

VI

hef

V2

lout

M2

~nm05 n~057b-' Slave mr

Figure 3. Current Mirror Principles in nMOS

~2

1Current-Readout Pixels Operation

Figure 4. Current Mirror Principles in pMOS

Current readout active pixel sensors areinherently advantageous in terms of readoutspeed because the fixed output line voltageat input of trans-resistance amplifier preventscharge-discharge phenomena. Another benefitof current readout is current-mode processingwhich is relatively compact in size andsimple in its operations. In current-mode, thephoto-current of detector can be mirrored andreadout directly.

The Current Mirror

The current mirror is one of the most

useful basis block in analog design. In itsmost simple configuration, it consists in twoMOS devices as in Figure 3 and Figure 4.

A current Iref flowing through thenMOS device MI is copied to the nMOSdevice M2. If the size of MI and M2 are

identical, in most operating conditions, thecurrent are the same. The remarkable pointis that the current is almost independent ofthe drain voltage of the M2 V2. if the rationWIL of the M2 is 10 times the ratio of the M I,the current on the right branch is 10 timesthe current on the left branch. The formula

Salahuddin, Pixel CMOS...

when MI operates in saturation is shown inEq 5 below. Equation 6 and 7 shows if M2operates in saturation.

Iref=II=~'(~)(vGSI-VTH j(I+AVDsl) (5)

VDSI =VGSI =VGS2

k '=current gain.

Ioul=h=~('r)~GS2-VTH )(l+AVDS2) (6)

loo'~IJ If) Q+1.VDS2}

(!)(I+AVDSI)

(7)

The Design of Current-mirror CMOSPixels

Figure 5 shows a block diagram andFigure 6 shows layout of small test chipimplemented in a standard 0.6 ~m CMOSprocess. This is the first step to realize agamma-camera on-a-chip. The chip consistsmainly of current-mirror active pixels size:(a) Ix I mm2 (the two pixels), (b) 500x500~m2 (the four pixels) and (c) 200x200 ~m2(the eight pixels), each constituted of a

195

c

~output

"'Put

b

/'

\IDI)!!. It!!!. coms COmJ ,!!comt Sa VB

"a

.A...

output

Figure 5. Block Diagram of Current-mirrorCMOS Pixels

Figure 6. Layout of Current-mirror CMOS Pi;

photodiode with control switch and current cathode of the photodiode consist of a squareoutput amplifying circuitry. of diffusion connected on all its circum-

ference by a whole of contacts connectedin crown. This provision makes it possibleto reduce the resistance of the contacts. The

anode was carried out by surrounding thephotodiode of a ring of polarization of thesubstrate, this to limit the resistive effectsof the substrate. The contacts of these twocrowns are carried out with the minimum

step authorized by technology.

Pixel Circuit

The pixel circuit is shown in Figure7. Here, the current mirror (MI and M2) isdesigned such that W2 =2 xWI. The photo-current is thus amplified 10 times.

The output current gain of this simplecurrent mirror is finite, as a result. Forincreasing of current gain, in this designwe added on nMOS-pMOS combination ofcurrent-mirror circuits. M9 is used as switch,when COM (M9) is on, pixel current is outputto the column, as in Figure 8.

We create of photodiode type N as infigure 9. The cathode of photodiode N, the

196 Jurnal Ilmiah Teknologi & Rekayasa, Volume15 No.3, Desember20/0

Simulation and Layout Current-MirrorCMOS circuit

For the simulation of our circuit

current-mirror CMOS, we called upon asoftware of computer-aided design called

11>I"

'"

I!k!!Is

,as"OJ { output

I9tI!k

id.....

Vdd

Vout

R

Figure 7. Pixel circuit.

Figure 9. Layout of Photo diode Type N.

Mentor Graphics. By the intennediary ofthe software Design Architect version AMS(Austria MilcroSystem), we can carry out theseizure of our diagrams electronics directlyin technology CMOS. Accusim enables usto simulate our diagrams carried out. Afterthat we use the software IC-Station versionLY or PR for the design of the layout ofelectronic diagrams. Given the complexityof the software IC-Station, we have createdmanual, allowing the user to familiarize itselfwith the manual routing suggested by AMS.The technology used is a technology CMOSof 0.6 micro. To be able to process the dataprovided by the photodiode, we tested circuitsallowing the amplification and the selectionof our signal.

In Figure 10, we added two moreCMOS (nMOS (MlO) and pMOS (MIl». Itsaim is to see the output of this circuits, beforeor after the amplifying. The circuit has twoswitch (sw and com) and two output (directand sortie). Figure 11 is diagram of layout thecurrent-mirror with combination nMOS and

pMOS.

Salahuddin, Pixel CMOS...

....

Vdd

Figure 8. Current-Mirror with AmplifyingNmos-Pmos Combination.

Iphoto

Direct

xSW

Sortie

I' ~Com

Figure 10. Current-Mirror Circuits withCombination Nmos and Pmos.

The circuit operation is very simple.First, the current Iphoto passes through MIand is mirror to M2 using current mirror. IftheHIGH signal is given to switch SW and COM,the results obtined at the output DIRECT, thatis before the current.Iphoto is amplified by thecurrent-mirror with combination nMOS and

pMOS. If the LOW signal is given to switchSW and COM, the results at output SORTIE,that is after the current Iphoto is amplified bythe current-mirror with combination nMOSand pMOS.

For simulation purpose, we used acalibrated current of OA with 200nA. Fromthe results of simulation, it is visible thatthe current-mirror circuit with combination

nMOS and pMOS to achieve a high currentgam.

Fabrication and Test Results

Figure 12 shows a block diagram andFigure 13 shows photograph of small testpixels implemented in a standard 0.6 J.lmCMOS process from AMS. The chip consists

197

lfUI]

~

Figure 11. Diagram of Layout The Current-Mirorwith Combination Nmos and Pmos

.~L~J

~ MI~. ~~

~

~

Figure 11. Diagram of Layout The Current-Mirorwith Combination Nmos and Pmos

Figure 13. Photograph of photodiode Sensors

198Jurnaillmiah Teknologi & Rekayasa, Volume15No.3. Desember 2010

Table J .The measurement characteristies

Parameter I ResultTechnology 0.6 ~m CMOS. 2-layer Metalet I-layer PolyDiffusion N+/ P Substrat2mmx2mm

( mm x (mm, 0.5 mm x 0.5mm

and 0.2 mm x 0.2 mm.98%700 nm (RED) and 585 nm(GREEN) .5 VoltLoearitmie

Photodetector

Sensor AreaPixel Pitch

Fill FactorSpectral Respon

Power SupplayResDonse in Illumination

1

Figure 14. Output Signal of Pixels with Green LED lux

Figure 15. Output Signals of Pixels with Red LED Lux

I'" ~1--- "" + L-""-- -==-t-=; + ~:

. ; . r tr t t......... -:; I .,Ii

~

-+ ... -.:---r ,' -

1t

Ii,' I

.1,; I

S I " r

Figure 16.Photocurrentin DynamicModewith 1mm xl mmArea Using Green LED

mainly of current-mirror active pixels with Figure 14 and 15 shows variationsdifferent sizes: (a) lxl mm2 (the two pixels: output voltage with light intensity of greenu and v), (b) 500x500 J!m2 (the four pixels: and red lights emitting diode. It shows aq, r, sand t) and (c) 200x200 J!m2 (the logarithmic relation of output signal insixteen pixels: a - p), each constituted of a function of emitting light (Lux).photodiode with control switch and current The test result also shows variations ofoutput amplifier circuits. photo current in dynamic mode. In this mode,

The photo response of test photodiodes light emitting diode is driven with a pulseon the chip is obtained by measuring the generator. Results presented in Figure 16 andphoto current under illumination (Lux) from 17 are obtained with a green LED calibratedgreen and red light emitting diode. Table for a 0.4 lux illumination. Results shown in1 summarizes the overall measurement Figure 18and 19are obtained with a red LEDcharacteristics. calibrated for a 0.3 lux illumination. In these

Salahuddin, Pixel CMOS... 199

5,05

'r

4.9 .1. 4.85 ...i5 4.8 .. . .J 4.15

... .4,1 ... .4.854.6 ... ... ...

4.550 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5

Lux of G,..n LED

5,255

arc171mmx1m1

4,15 . .2mm x 2mm

1. 4,5.. .

1154,25 . .13.1:

. .3.5 . .3,25

0 0.5 1 1,5 2 2,5 3 3.5 4 4,5 5 5.5Lux of RedL£D

. .4 ~1- - +--1- t-o +-

r=~-4- _

Figure 17. Photo current in Dynamic Modewith 2 mm x 2 mm Area Using Green LED

I

+-~ ~

+ --+---<

Figure 18. Photocurrent in Dynamic Modewith 1 mm x 1mm Area Using Red LED

Figure 19. Photocurrent in Dynamic Modewith 2 mm x 2 mm Area Using Red LED

figures, the upper curves represent the LED DeRenzo, S.E. Holland, S.E., Pedrati-voltage input and the lower curves represent Noy, M., Kriege, B., Mandelli, E.,the sensor output. Meddeler, G., Wang~N.W.,and Witt, E.K.

"A Compact 16- Module Camera Using64-Pixel Csi(TI)/Si PIN PhotodiodeImaging Modules," IEEE Trans. Nuc/.SeL, vol.49, pp.2228-2235, 2002.

Erard, Olivier and Valaud,Fabien. InstructionManual of IC_Station in Layout Modeand Reation of a Photodiode of The TypeP. University of Burgundy.

Hamamatsu Corporation. PhotomultiplierTubes:PMT Handbook.

Majewski, S., Kieper, D., Curran, E., Keppel,c., Kross, B., Palumbo, A. "OptimizationOf Dedicated Scintimam-MographyProcedure Using Detector PrototypesAnd Compressible Phantims," IEEETrans. Nuc/. Sci., vol.48, pp.822-829.2001.

SCIONIX USA. SCIONIX Scintillation

I

.1Hi'

" 1I

d

f

Conclusions and Perspectives

A 2 square millimeters area of CMOSactive photodiode sensor with current mirroramplifier has been fabricated using a 0.6 IlmCMOSprocess.Theexperimental results showthat this sensor has logarithmic response inillumination and is capable of detecting verylow green lights emitting diode. These resultsallow us to consider using of this technologyin new solid state gamma cameras.

REFERENCES

Baschirotto, A. Bias Circuit. University ofLeece.

Choong, W.S., Gruber, G.J., Moses, W.W.,

1--"- I

-1

200 Jurnal Ilmiah Teknologi & Rekayasa, Volume 15 No.3, Desember 2010

i" 1 '"

.. 1II1 .

!I r 1 v

.

,'I" I .1 /1j1

J

Detector. http://www.scionixusa.com.Stateikina, Irina. Optoelectronic Semi-

conductor Devices-Principals andCharacteristics. Concordia University.

Wilamowski, B.M., Ferre-Pikal, E.S.,Kaynak, O. "Low Power, Current ModeCMOS Circuits for Synthesis of Arbitrary

Salahuddin, Pixel CMOS...

Nonlinear Fuctions," Proceedings of9th NASA Symposium on VLSI Design.2000.

William, M., Goode, A., Galbis-Reig, V. etal. "Performence of a PSPMT Based

Detector for Scintimammography," Phys.Med. Bioi., vol.45, pp.781-800. 2000.

201