A 0.6V CMOS Image Sensor with In-Pixel Biphasic Current Driver for Biomedical Application Chin-Lin Lee and Chih-Cheng Hsieh Signal Sensing and Application Laboratory National Tsing Hua University Department of Electrical Engineering Hsinchu, Taiwan, R.O.C Email:
[email protected] Abstract — A 0.6V pulse frequency modulation (PFM) CMOS Image Sensor (CIS) array with in-pixel biphasic current pulse driver is presented in this paper. It achieves a photon-to- biphasic current conversion for biomedical applications like artificial 2-D vision recovery. The photon-to-biphasic-current conversion gain, the biphasic pulse width, polarity, and output rate are all tunable depends on applications and environments. A 32x32 pixel array with 30x30 um2 pixel size has been designed and fabricated in 0.18um CMOS technology providing the fill factor of 24.5%. Measurement results show a 0.63Hz/lux conversion gain of PFM sensor within 25hz~5kHz output with power consumption as 2uW~55uW depends on illumination. The maximum driving capability of biphasic neural stimulation current pulse is ±20μA with a 10kΩ electrode model. I. INTRODUCTION IOMEDICAL applications have been one of the most researched field in recent years. A lot of people have eye diseases in the world. Blindness diseases, such as retinitis pigmentosa (RP) or age-related macular degeneration (AMD), cause progressive degeneration of rods, cods and cones in the retina [1]. In previous works, MOS image sensors have been applied for helping blind people [2]. The Optacon, or optical- to-tactile converter, is probably the first use of a solid-state image sensor for blind people [3]. Nowadays, many studies have been carried out considering a number of implantation ways such as epi-retinal space (Epi.), sub-retinal space (Sub.) and suprachoroidal transretinal stimulation (STS). Low power circuit is an important issue and critical goal in portable and implantable products. In this work, we propose a low power pulse frequency modulation image sensor suitable for biomedical application. The pulse frequency modulation (PFM) photo sensor [4]-[7] has been used in high dynamic range (HDR) and retinal prosthesis. It has many advantages such as asynchronous operation, digital output, wide dynamic range and low voltage operation. A digital output of the PFM sensor eliminates the effects of column-related noise on analog performance [8] and each pixel in PFM sensor can individually produce an output pulse without other system clock, that is, asynchronously [2]. In spite of these advantages, the major limitation of this pixel architecture is its complex pixel structure which leads to a larger pixel size and a potentially higher pixel wise fixed-pattern noise (FPN) [8]. A 32×32 pixels PFM CIS array with in-pixel biphasic current output has been proposed to replace the function of damaged retinal cell. The block diagram of the pixel cell is shown in Fig. 1. The photo sensor is a pn-junction photodiode (PD) implemented by n-diffusion and p-sub. When illuminated, the photon excited electron-hole pairs are accumulated at the capacitance of pn-junction. The PFM circuit converts the slope of the voltage difference on sensor node to frequency output which is proportional to the input light intensity. A frequency-to-voltage (F-V) converter is designed to convert the output pulse to a voltage which is then used to generate biphasic current pulses [9]-[11]. The outline of this paper is as follows. In Section II, the architecture of the pixel is described. Then, the simulation and measurement results are summarized in Section III. Finally, the conclusions are drawn in Section IV. Fig. 1 Block diagram of the pixel cell II. PIXEL ARCHITECTURE AND CIRCUIT The proposed pixel contains two components: PFM photo sensor and biphasic pulse generator which includes a frequency-to-voltage converter. B 978-1-4244-9474-3/11/$26.00 ©2011 IEEE 1455 A. PFM photosensor Unlike the conventional CMOS image sensor, the main components of PFM image sensor are digital circuits that can operate at low supply voltage, and carry out analog-to-digital conversion within the pixel. The circuit structure of the proposed PFM photo sensor cell is as shown in Fig. 2. It is composed of a photodiode, a Schmitt trigger with tunable threshold, a reset transistor Mr, and a feedback loop with a delay chain (inverter chain). Before the exposure starts, the sensor node Vpd is reset to VDD through the device Mr. With light illumination, the voltage at PD as Vpd is discharged by the photon-generated carriers and decreases. When Vpd is lower than Vth, the threshold voltage of the Schmitt trigger, the output Vout changes to HIGH and is feedback through a delay chain to reset the sensor node. Then, Vout is pulled to LOW again. It results in a pulse generation behavior and the frequency is proportional to the light intensity. Since the full- well of sensor node is self-reset and re-used through the feedback loop, the dynamic range of PFM sensor is much higher than the conventional image sensor. The delay chain is used to implement the necessary reset pulse width and maintain the oscillation condition of the loop. The output frequency f shows a linear function of the input light intensity and can be expressed as Equation (1) )( thddpd pd VVC I f − ≈ (1) Vpd Delay Chain VDD VoutPD Mr IpdCpd Fig. 2 Basic circuit of a PFM photosensor Fig. 3 The transient response of the sensor node Vpd and the pulse output of Schmitt trigger The transient response of Vpd and pulse output of Schmitt trigger are shown in Fig 3. To reduce power consumption and pixel size, a six-transistor (6T) Schmitt trigger with tunable threshold is implemented instead of using the conventional analog comparator. As shown in Fig. 4, the VthH and VthL are the upper and lower thresholds of the Schmitt trigger which can be varied by changing external bias voltage VH and VL respectively. The photon-generated pulse frequency is tunable by adjusting the thresholds, which results in a front-stage tunable gain. With this front-stage tunable gain, the dynamic range and sensor SNR can be optimized based on applications. Fig. 4 Schematic of the Schmitt trigger Fig.5 Schematic of the F-V converter Fig.6 Schematic of the biphasic current generator 1456 B. Biphasic current generator For the artificial retina application, the image sensing chip needs to generate an in-pixel biphasic current for neural stimulation. An in-pixel frequency-to-voltage (F-V) converter followed by a voltage-to-biphasic-current generator is proposed to provide the electric stimulus. As shown in Fig. 5, the F-V converter is implemented by a switching current Isw and an integration capacitor Cint for low operating voltage availability and minimized circuit complexity. Within a certain exposure period controlled by “reset”, the voltage on Cint is proportional to the frequency output of PFM sensor. Through the level shifter (Mpsf), the converted voltage is then fed to the voltage controlled current source as shown in Fig. 6. The generated current is then mirrored to as a current sink and a source of electrode, respectively. In this design, the biphasic pulse width, polarity, and output rate are adjustable by controlling the timing of the φ1 and φ2. The impedance of electrode and a tissue is modeled by the equivalent circuit as shown in Fig. 6. (10kΩ in this work) The biphasic current pulse is used for generating the electric stimulus to a nerve cell [12]. An optimized matching between the negative and positive current level is important to deliver a good charge balance [13]. The switch SWref, controlled by φ3, is used to remove the charge residue on the electrode after every stimulation cycle. It can avoid the neuron memory effect and possible physiological damage. III. PROTOTYPE CHIP AND MEASUREMENT RESULT The microphotograph of the fabricated chip (1.8 mm × 1.8 mm) in TSMC 0.18 μm CMOS technology is shown in Fig. 7. The pixel layout is shown in Fig. 8. Each 30x30 μm2 pixel size contains a 220.5μm2 n-diffusion/p-substrate photodiode results in a 24.5% fill factor. An in-pixel electrode pad window is embedded for further electrode formation. Fig.7 Microphotograph of the fabricated chip Figure 9 shows the measured sensitivity of the PFM photo sensor. It shows a linear response between input light intensity (lux) and output frequency (Hz) as expectation. The measured conversion gain (frequency/illumination) is 0.63 (Hz/lux). Figure 10 shows the output waveform of this chip with an electrode model (10kΩ) under illumination. The measured maximum PFM output frequency is 5.3 kHz at 0.6V supply. The maximum driving capability of biphasic neural stimulation current pulse is ±20μA limited by 0.6V supply and 10kΩ loading. Fig. 8 Pixel layout PFM output response Illumination (lux) 0 2000 4000 6000 8000 10000 Pu ls e Fr eq ue nc y (H z) 0 1000 2000 3000 4000 5000 6000 Fig. 9 Sensitivity of the PFM photosensor Fig. 10 The measured PFM sensor output, biphasic current output and F-to-V output waveforms. 1457 Four test patterns have been applied to do the image reconstruction verification. The reconstructed sample images from biphasic current output level is shown in Fig. 11. It shows that an English letter “A”, an arrow, five circles and a number 7 have been successfully rebuilt and recognized. The measured results of the fabricated prototype chip are summarized in Table I. Fig. 11 The captured image from prototype chip TABLE I Specifications Technology TSMC CMOS 0.18-um Supply voltage 0.6V Number of pixels 32x32 Pixel size 30 μm × 30 μm Fill factor 24.5% Frame rate 20 Hz PFM output Frequency 25Hz~5kHz Power consumption 2uW~55uW(without current driver) PFM conversion gain 0.63Hz/lux Biphasic current ±20μA (with 10kΩ loading) IV. CONCLUSION The design and test of a 1024 pixels PFM CMOS image sensor array with in-pixel biphasic current pulse driver has been presented. The main work is a photon-to-biphasic current conversion system on chip operated at 0.6V supply voltage. A tunable front-gain by adjusting the threshold voltages of Schmitt trigger is implemented. The resulting biphasic pulse width, polarity, and output rate are all tunable depends on applications and environments. It is verified that the patterns can be reconstructed and recognized from the photon-to- biphasic current conversion of the proposed sensor chip. It shows a low-voltage and low-power solution for retinal implantation. ACKNOWLEDGMENT The authors would like to thank Prof. L.S. Fan, Prof. K.T. Tang, S.F. Yeh, and M.T. Chung for supporting this work. REFERENCES [1] A. Rothermel,L. Liu, N.P. Aryan, M. Fischer, J. Wuenschmann, S. Kibbel, and A. Harscher, “A CMOS chip with active pixel array and specific test features for subretinal implantation,” IEEE J. Solid-State Circuits, vol. 44, pp. 290–300, 2009. [2] J. Ohta, Smart CMOS Image Sensor and Applications, CRC Press, Boca Raton, FL, 2007. [3] J.G. Linvill, and J.C. Bliss, “A direct translation reading aid for the blind,” Proc. IEEE, 54(1):40-51, January 1966 [4] K.P. Frohmader, “A novel MOS compatible light intensity-to- frequency converter suited for monolithic integration,” IEEE J. Solid- State Circuits, vol. 17, pp. 588-591, 1982. [5] J.G. Nicholls, A.R. Martin, B.G. Wallace, and P.A. Fuchs, From Neuro to Brain, Sinauer Associates, Inc., Sunderland, MA, 4th edition, 2001. [6] K. Tanaka, F. Ando, K. Taketoshi, I. Ohishi, and G. Asari, “Novel digital photosensor cell in GaAs IC using conversion of light intensity to pulse frequency,” Jpn. J. Appl. Phys., vol.32, pp. 5002-5007, Nov. 1993. [7] J. Ohta, N. Yoshida, K. Kagawa, and M. Nunoshita, “Proposal of application of pulsed vision chip for retinal prosthesis,” Jpn. J. Appl. Phys. 41(4B), pp. 2322-2325, 2002. [8] X. Wang, W. Wong, and R. Hornsey, “A high dynamic range CMOS image sensor with inpixel light-to-frequency conversion,” IEEE Transactions on electron devices, vol.53, no.12, Dec. 2006, pp. 2988- 2992. [9] M. Sivaprakasam, W. Liu, M.S. Humayun, and J.D. Weiland, “A variable range Bi-phasic current stimulus driver circuitry for an implantable retinal prosthetic device,” IEEE J. Solid-State Circuits, vol. 40, pp. 763–771, 2005. [10] S. Kavusi, K. Ghosh, and A.E. Gamal, “A per-pixel pulse-FM background subtraction circuit with 175ppm accuracy for imaging applications,” IEEE Int. Solid-State Circuit Conf. Dig. Tech. papers, Feb. 2007, pp. 504-618. [11] D.C. Ng, T. Furumiya, K. Yasuoka, A. Uehara, K. Kagawa, T. Tokuda, M. Nunoshita, and J. Ohta, “Pulse frequency modulation based CMOS image sensor for subretinal stimulation,” IEEE transactions on circuits and system-II: express briefs, vol. 53, Jun. 2006, pp. 487-491. [12] T. Furumiya, D.C. Ng, K. Yasuoka, F. Shiraishi, K. Kagawa, T. Tokuda, J. Ohta, and M. Nunoshita “A 16x16-pixel pulse-frequency- modulation based image sensor for retinal prosthesis,” Sensors,2004. Proceedings of IEEE,vol.1, Oct. 2004, pp. 276-279. [13] A. Rothermel, L. Liu, N.P. Aryan, M. Fischer, J. Wuenschmann, S. Kibbel, and A. Harscher, “A CMOS chip with active pixel array and specific test features for subretinal implantation,” IEEE J. Solid-State Circuits, vol. 44, pp. 290–300, 2009. 1458 /ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 200 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 2.00333 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 400 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 600 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.00167 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /CreateJDFFile false /Description > /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ > /FormElements false /GenerateStructure false /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles true /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /NA /PreserveEditing false /UntaggedCMYKHandling /UseDocumentProfile /UntaggedRGBHandling /UseDocumentProfile /UseDocumentBleed false >> ] >> setdistillerparams > setpagedevice