ieee sensors journal, vol. 6, no. 3, june 2006 661

IEEE SENSORS JOURNAL, VOL. 6, NO. 3, JUNE 2006 661

Colorimetric Porous Photonic Bandgap Sensors WithIntegrated CMOS Color Detectors

Xiaoyue Fang, Vincent K. S. Hsiao, Vamsy P. Chodavarapu, Student Member, Albert H. Titus, Member, IEEE, andAlexander N. Cartwright, Member, IEEE

Abstract—In this paper, the development of a novel colori-metric sensor system based on the integration of complementarymetal–oxide–semiconductor (CMOS) color detectors with a mod-ified porous polymeric photonic bandgap sensor is reported. Thecolor detector integrated circuit (IC) is implemented with AMI(AMI Semiconductor) 1.5 m technology, a standard CMOSfabrication process available at MOSIS (http://www.mosis.org).The color detectors are based on the spectral responses of burieddouble junctions (BDJs) and stacked triple junctions (STJs);the ratio of the photocurrents at the junctions provides spectralinformation. Both types of color detectors are characterized witha monochromator, and the results are compared. The BDJ colordetector is used with a porous photonic bandgap reflection gratingwhose reflection spectra shifts as a function of the concentrationof vapor analyte present. The experimental results verify that thecolor change of the photonic crystal can be detected and corre-lated to the change in analyte concentration. The entire system iscompact and low power.

Index Terms—Analog circuits, color photodetectors, integratedoptoelectronics, porous plastics, sensors.

I. INTRODUCTION

COLOR is the combination of hue, brightness (lightness),and saturation of an optical source or reflection from a

surface. The detection of color provides a wealth of informationabout a scene, as illustrated by the number of animals with colorvision. Color vision arises from the existence of different typesof receptors each of which is sensitive to different peak wave-lengths of light. The combination of the responses from theseproduces color perception. Color detection in a sensor systemis also important since this can provide additional informationabout an environment. However, the extent to which the sensorsystem requires color perception as opposed to optical wave-length selectivity depends on the application. In this paper, we

Manuscript received February 7, 2005; revised April 25, 2005. We would liketo acknowledge the generous financial support of the National Science Foun-dation (NSF) IGERT: Biophotonics-Materials and Applications Award #DGE-0345408, NSF SENSORS Award #BES-0330240, and the Johnson and JohnsonFocused Giving Grant. The associate editor coordinating the review of this paperand approving it for publication was Prof. Eugenii Katz.

X. Fang was with the Department of Electrical Engineering, University atBuffalo, The State University of New York, Buffalo, NY 14260 USA. She isnow with Aerotek Automotive, Oshawa, ON L1H 8P7, Canada (e-mail: [email protected]).

V. K. S. Hsiao, V. P. Chodavarapu, A. H. Titus, and A. N. Cartwright arewith the Department of Electrical Engineering, University at Buffalo, The StateUniversity of New York, Buffalo, NY 14260 USA (e-mail: [email protected]; [email protected]; [email protected]; [email protected]).

Digital Object Identifier 10.1109/JSEN.2006.874021

use monochromatic color detectors for measuring the shift inthe wavelength of the incident light.

Previously, a novel technique for monochromatic color detec-tion using a buried double p–n junction (BDJ) structure has beenpresented [1], which is implemented with a standard comple-mentary metal–oxide–semiconductor (CMOS) process [2], [3].As example applications, the CMOS BDJ detector has been usedfor fluorescence detection in microarrays [4], and for the detec-tion and measurement of ambient light sources [5], [6]. Appli-cations in (bio)chemical/biological discrimination and seawaterpH measurement have also been reported [7]–[9].

The methods for detecting organic chemical vapors can beclassified into two broad types: electrical and optical. Opticalmeasurement techniques are considered to be easier, faster, andsafer as compared with the electrical methods [10], [11]. Oneoptical method is based on monitoring the changes in the spec-trum of light reflected from a periodic porous structure, suchas a porous silicon photonic bandgap structure [12]–[14]. Thesensor, made of the porous material, is placed in the environ-ment to be monitored. When the analyte vapor diffuses into thepores of the sensor material, the effective refractive index of thematerial changes, and there is a corresponding change in the re-flection peak (as a function of wavelength) of the sensor. Thus,the reflected light is of a different color, and the shift in detectedcolor can indicate the presence and, ultimately, concentration ofthe analyte [15], [16].

In this paper, we report the first gas-phase analyte sensorusing a CMOS-based monochromatic color detector integratedcircuit (IC) and a porous photonic bandgap structure. TheCMOS IC is fabricated through MOSIS and has a number ofcolor detectors and readout circuitry; this is described in detailin Section II. The integration of the color detectors on standardCMOS chips eliminates any additional optical analysis instru-mentation and permits signal conditioning circuitry on-chip.The photonic crystals are fabricated at the University at Buffalo,The State University of New York, using a recently developedmethod [17]. We discuss the fabrication of the photonic crystalin Section III. Results comparing these two different structuresfor different-sized active areas, as well as results of vapor de-tection using the photonic crystals, are presented in Sections IVand V, respectively.

II. CMOS COLOR DETECTOR IC

A. Detector Structure and Operation

The color detectors presented in this paper are based on de-signs that are described in [1]–[4]. In that design, two standard

1530-437X/$20.00 © 2006 IEEE

662 IEEE SENSORS JOURNAL, VOL. 6, NO. 3, JUNE 2006

Fig. 1. Buried double-junction structure (not to scale).

p–n junctions are stacked vertically in the CMOS chip. The dis-tribution of the absorbed light in the silicon layer is given by

, where is the reflection co-efficient of the air–silicon interface, is the incident light in-tensity, is the absorption coefficient of silicon, and is thedepth into the sample. The absorption coefficient of silicon isa function of the wavelength of light and is larger for shorterwavelengths of light. More important, the distribution of photo-generated electron–hole pairs will follow a corresponding expo-nential decay throughout the depth of the sample. In the designpresented above, each p–n junction is at a different depth in thesilicon substrate, so there is a different photocurrent generated ateach junction. For example, when a flux of longer wavelengthlight is incident on the detector, the deeper junction producesmore current than if it were illuminated with a shorter wave-length light. In other words, the same flux of shorter wavelengthlight will produce more current in the shallower (near surface)junction and less in the deeper one. Therefore, the ratio of thesephotocurrents indicates the wavelength of the light incident onthe chip.

In this paper, we examine two structures: the buried doublejunction (BDJ) and stacked triple junction (STJ). Our detectorsare fabricated using a standard CMOS process, so all layers usedto form the detectors are standard layers. The BDJ structure,depicted in Fig. 1, is a pair of stacked p–n junctions formedusing the p-substrate/n-well/p-base regions. The STJ structure(Fig. 2) is a triplet of p–n junctions formed using p-substrate/n-well/p-base/n-select regions.

The color detector based on the BDJ structure has two p–njunctions. Since the p-substrate is common to the entire chip, thecurrent cannot be directly measured, so the currents that aremeasured are and the sum (see Fig. 1). The STJ struc-ture stacks one additional p–n junction on top of the BDJ struc-ture, which produces one additional photocurrent. Since this topjunction is between the n-select and p-base layers, it is closer tothe surface on which the light is incident than the top junction

Fig. 2. Triple-junction structure (not to scale).

Fig. 3. Micrograph of the fabricated chip. Labels (1–9) are the detectors; seeTable I for details.

in the BDJ. As a result, the difference between the photocur-rent generated at the upper junction and lower junction is larger.Again, the currents that are measured are labeled in Fig 2.

B. Integrated Circuit

A micrograph of the integrated circuit (chip) is shown inFig. 3. The advantage of fabrication through a standard CMOSprocess means that the ultimate cost of such a device will be low,and additional signal processing can be put on-chip. However,we have no control of the depth of each junction. The proto-type chip has a number of different devices on it to allow us tocharacterize the structures, which are detailed in Table I. The

FANG et al.: COLORIMETRIC POROUS PHOTONIC BANDGAP SENSORS WITH INTEGRATED CMOS COLOR DETECTORS 663

TABLE IRESPONSE OF DETECTORS AT 632 nm (HeNe LASER)

Fig. 4. Acetone sensor response for difference concentrations.

peripheral area of each detector is covered by metal to preventthe generation of unwanted photocurrents.

The chip is fabricated in 1.5 m AMI process availablethrough MOSIS. This process has two metal layers and twopolysilicon layers. Our prototype is a 2 mm 2 mm silicon diethat is packaged in a 40-pin DIP ceramic package. Since thechip is fabricated through a commercial process, the junctiondepths and doping concentration of each layer are not publiclyavailable.

III. MODIFIED PHOTONIC CRYSTAL

The photonic crystal used for our sensor is a made froma polymer material; the periodic structure is formed simplythrough a holographic process [18], [19]. This has manybenefits over the more complex process of etching the peri-odic structure in materials such as porous silicon [20]. Theprepolymer syrup composed of a monomer, photoinitiator,co-initiator, and liquid crystal (LC) are sandwiched betweentwo glasses, and the periodic structure is formed by the opticalinterference pattern all in one step. The porous grating structureis created by adding a nonreactive solvent, such as acetone ortoluene, into the prepolymer syrup. The grating forms when thenonreactive solvent evaporates after opening the cover glassslide [20]. More important, the Bragg reflection notch shiftswhen the voids inside the grating structure fill with solvent

Fig. 5. Experimental setup.

vapor. This changes the average refractive index of the material,and, subsequently, the grating spacing according to the formula

where is the reflected wavelength, is the averagerefractive index of grating film, and is the grating spacing. Thesimplicity of the fabrication, combined with the CMOS colordetectors, creates a straightforward, low-cost sensor system.

Fig. 4 is the reflection spectra of grating cell filled withdifferent concentrations of acetone vapor. The Bragg wave-length shifts from 540 to 580 nm with only 9% of acetonevapor. Increasing the concentration of the acetone vapor notonly changed the average index of the grating but also theeffective grating spacing. The change in the reflected light canbe observed even with the naked eye.

IV. EXPERIMENTAL RESULTS AND MEASUREMENTS

A. Monochromator Characterization

The color detectors are first characterized using a monochro-mator to provide a narrowband optical input [full-width at half-maximum (FWHM) of approximately 5 nm) and using a tung-sten lamp light source. The incident light intensity on the detec-tors is wavelength dependent, and the average intensity is ap-proximately 1 W/mm . There are four BDJ structures on thechip and two STJ structures (see Fig. 3 and Table I). For theBDJ structures, we plot the responses for the two larger detec-tors. The two junctions are biased as shown in Fig. 5. is thecurrent through the top junction, and is the current throughthe bottom junction. To reiterate, since the p-substrate is alwaysconnected to ground, current readout can only be taken from


Fig. 6. (a) Experimental results for the BDJ structure with the 500 �m �

500 �m- and the 300 �m � 300 �m-sized detectors; (b) comparison ofnormalized values; currents are normalized to the peak value for each size.

P-base and N-well, which are currents and , respec-tively. Fig. 6(a) shows the current versus wavelength for both thesizes. We can see that the bottom junction produces a larger re-sponse for longer wavelengths, while the top junction producesa larger current at shorter wavelengths, as expected. In addition,the larger detector produces a larger current.

In order to better compare the two sizes, we normalize thecurrents from each detector, i.e., set the largest current valuefor each detector to unity and calculate the other values propor-tionally. This is shown in Fig. 6(b). Both diodes have nearlythe same response at the top junction while the smaller diodehas a larger normalized deeper junction response. This resultsin the larger size detector having a larger ratio range, which isshown in Fig. 6(b). The difference due to size is because thecontacts for the detectors are located around the edge of each;therefore, fewer of the carriers generated in the larger detectorat the lower junction may be collected at the contacts. Based onthese experimental results, the 300- m detector is acceptable,but the 500- m detector is better since the ratio is higher. Thus,the current ratio is a good measure of spectral information in

the range of 425 to 700 nm. We used a low-pass filter for mea-surements below 525 nm to prevent frequency doubling in themonochromator, and this was removed at 525 nm. Thus, whatappear to be prominent sidelobes at 570 nm are a result of thechange in the light intensity hitting the detectors; this producesthe change in current. The peak at 470 nm for the top junctioncurrents ( ) occur because the responsivity for silicon drops forshorter wavelengths [21].

We also analyze the STJ structures with the monochro-matic light input. There are two sizes of STJ-based detectors:200 m 200 m and 500 m 500 m. In this device, is thecurrent through the top p–n junction (n-select/p-base), is thecurrent through the middle p–n junction (p-base/n-well), andis the current through the bottom p–n junction (n-well/p-sub-strate). In this triple-junction structure, the P-substrate andP-base are internally connected and grounded (connected toground on-chip). The output currents , , and areread directly from the chip, and the current ratios are calculatedfrom these. Fig. 5(a) shows the current versus wavelength foreach size. For both sizes of the triple-junction detector, the topdiode has maximum response at approximately 450 nm whilethe maximum for the bottom diodes occurs between 500 and700 nm.

We use the ratio as the indicator of wavelength;this ratio is good in a spectral range of 425 to 600 nm [seeFig. 5(b)]. The double-junction detector working range beginsat wavelengths around 425 nm. At wavelengths longer than 600nm, the currents are still wavelength dependant, but since littleis absorbed at the top junction, the ratio becomes nearly con-stant.

B. Comparison of Two Structures

In order to function properly, the detectors must produce acurrent ratio that is spectrally dependent. For the BDJ detector,we obtain and from the chip directly, calculate

from this, and use as the color/wavelength measure.For the STJ detector, since it is not possible to get ordirectly, the ratio is used as the wavelength in-dicator. Comparing Fig. 6 with Fig. 7, it is clear that the STJstructure produces larger currents than the BDJ structure for thesame size device. However, the current ratio, not absolute value,is used to indicate color. The graph in Fig. 8 depicts the cur-rent ratios for both structures. From this, we see that the BDJdetectors have a larger change in ratio across the wavelengths425–670 nm; this indicates that the BDJ detectors have muchbetter resolution than the STJ detectors. This is true for thesmaller 300 m 300 m BDJ detector as well. In addition, atlonger wavelengths, the STJ detectors change in current ratio isextremely small. However, the additional p–n junction enablesthe STJ structure to respond better to lower intensity light.

C. Photonic Crystal Acetone Sensor

The photonic crystal exhibits a shift in the reflection peakas a function of acetone concentration in air. Fig. 3 showsthe response of the photonic-crystal sensor when illuminatedwith white light. Clearly, the shift occurs from 540 to 635 nmwhile the acetone concentration increases from 0% to 22%.A drop of acetone (0.01 g, with a relative evaporation rate of


Fig. 7 (a) Experimental results for the STJ structure with the 200 �m �

200 �m- and the 500 �m � 500 �m)-sized devices; (b) comparison of thenormalized currents.

7.7 ) is dropped onto 25-mm square crystalsurface [22].

V. PHOTONIC CRYSTAL WITH INTEGRATED CMOSCOLOR DETECTORS

We use the 500 m m BDJ detector to detect thereflected signal from the photonic-crystal sensor. In our experi-ment, the CMOS color detector is positioned facing the photoniccrystal. A tiny drop of acetone is dropped onto the crystal sur-face. At this time, the acetone concentration in air surroundingthe crystal reaches a maximum, and it reflects in the red/orangewavelengths; the detector IC produces a response that indicatesa longer wavelength spectrum. As the acetone evaporates, i.e.,the acetone concentration decreases, the reflected color of thelight changes to yellow, and then to green (to shorter and shorterwavelengths). The test setup is shown in Fig. 5. Thus, we mea-sure a change over time of the acetone concentration.

Fig. 9 is a plot of the measured and calculated currents as afunction of time. The three curves are and that are

Fig. 8. Comparison of the current ratios for both types of detectors.

Fig. 9. Measurements with the photonic-crystal sensor. The plot shows typicalchanges in the currents and current ratio with time. The “zero” time on the x axisis actually 55–65 s after the acetone drop is placed into the chamber. Thus, thetotal time of measurement until the response stops changing can be up to 100 s.

directly read from the chip output and that is calculated fromthe measured currents. The axis is time, with the time startingwhen the reflected light changes to red. During the time pe-riod shown, the reflected light changes from red (wavelength ofapproximately 635 nm) to green (wavelength of approximately540 nm). We see that the top-junction current increases as thecolor changes from longer wavelengths to shorter wavelengths,while the bottom-junction photocurrent decreases. Therefore, asindicated in Fig. 9 with respect to the right-side scale, the cur-rent ratio increases. Since the spectrum of the reflectedlight is not as narrow as the monochromator output, the ratioscannot be directly compared. However, the change in the cur-rent and current ratio is correct and correctly indicates the colorshift, and in turn, the acetone concentration in the air.

VI. CONCLUSION

In this paper, a color detector has been designed and fabri-cated with AMI 1.5 technology. This detector utilized burieddouble-junction (BDJ) or triple-junction structure. Since junc-tions at different depths have different spectral responses, the


ratio of the photocurrents at shallower and deeper junctions pro-vides spectral information. Both of the structures are tested withmonochromator. The experimental results from the chip showthat the BDJ detectors function well over a spectral range of425 to 700 nm. The BDJ detectors excel the triple-junction de-tectors with better resolution and wider usable range, but thetriple-junction detectors have better responsivity, which is betterfor detecting in lower light intensity. However, it is possible fora different BDJ detector structure to approach the responsivityof the STJ, but for commercial process-based CMOS detectors,the STJs will always be larger. The size of the detector is a func-tion of the diffusion and well regions.

The 500 m 500 m BDJ color detector was chosen to usein an acetone sensor. The acetone sensor is a photonic crystalthat reflects different wavelengths of light at different acetoneconcentrations in the air when illuminated with white light. Thecolor detector chip produces an output signal (current ratio) thatindicates a change in reflected wavelength from the crystal.

REFERENCES

[1] G. N. Lu, M. B. Chouikha, G. Sou, and M. Sedjil, “Color detectionusing a buried double p–n junction structure implemented in the CMOSprocess,” Electron. Lett., vol. 32, pp. 594–596, 1996.

[2] G. N. Lu, G. Sou, F. Devigny, and G. Guilland, “Design and testing ofa CMOS BDJ detector for integrated microanalysis systems,” Micro-electron., vol. 32, pp. 227–234, 2001.

[3] G. N. Lu, J. M. Galvan, C. Jeloyan, G. Goumy, and V. Marcoux, “Sen-sitivity estimation of CMOS optical BDJ detector,” Mater. Sci. Eng. C,Biomimetic Mater., Sensors Syst., vol. C21, pp. 203–210, 2002.

[4] G. N. Lu, G. Guillaud, G. Sou, F. Devigny, M. Pitaval, and P. Morin,“Investigation of CMOS BDJ detector for fluorescence detection in mi-croarray analysis,” in Proc. 1st Annu. Int. IEEE-EMBS Special TopicConf. Microtechnologies in Medicine Biology, 2000, pp. 381–386.

[5] G. de Graaf and R. F. Wolffenbuttel, “Smart optical sensor systemsin CMOS for measuring light intensity and color,” Sens. Actuators A,Phys., vol. A67, pp. 115–119, 1998.

[6] ——, “Optical CMOS sensor system for detection of light sources,”Sens. Actuators A, Phys., vol. A110, pp. 77–81, 2004.

[7] M. Sedjil, G. N. Lu, G. Michard, and F. Prevot, “A colorimetric methodwith the use of BDJ detector for seawater pH measurement,” AnalyticaChimica Acta, vol. 377, pp. 179–184, 1998.

[8] D. P. Poenar, T. M. Siu, and T. O. Kiang, “Color sensor for(bio)chemical/biological discrimination and detection,” Mater. Sci.Semiconductor Process., vol. 5, pp. 17–22, 2002.

[9] Y.-J. Kook, J.-H. Lee, Y.-J. Park, and H.-S. Min, “A new CMOS pixelfor a biochip,” in Proc. 9th Korean Conf. Semiconductors, Chunan,Korea, pp. 209–210.

[10] T. Gao, J. Gao, and M. J. Sailor, “Tuning the response and stability ofthin film mesoporous silicon vapor sensors by surface modification,”Langmuir, vol. 18, pp. 9953–9957, 2002.

[11] Y. Y. Li, F. Cunin, J. R. Link, T. Gao, R. E. Betts, S. H. Reiver, V. Chin,S. N. Bhatia, and M. J. Sailor, “Polymer replicas of photonic poroussilicon for sensing and drug delivery applications,” Science, vol. 299,pp. 2045–2047, 2003.

[12] F. Cunin, T. A. Schmedakes, J. R. Link, Y. Y. Li, J. Koh, S. N. Bhatia,and M. J. Sailor, “Biomolecular screening with encoded porous-siliconphotonic crystals,” Nature Mater., vol. 1, pp. 39–41, 2002.

[13] V. S.-Y. Lin, K. Motesharei, K.-P. S. Dancil, M. J. Sailor, and M.R. Ghadiri, “A porous silicon-based optical interferometric biosensor,”Science, vol. 278, pp. 840–843, 1997.

[14] J. Diener, N. Künzner, D. Kovalev, E. Gross, V. Y. Timoshenko, G.Polisski, and F. Koch, “Dichroic Bragg reflectors based on birefringentporous silicon,” Appl. Phys. Lett., vol. 78, pp. 3887–3889, 2001.

[15] L. Pancheri, C. J. Oton, Z. Gaburro, G. Soncini, and L. Pavesi, “Im-proved reversibility in aged porous silicon NO2 sensors,” Sens. Actua-tors B, Chem., vol. 97, pp. 45–48, 2004.

[16] Z. Gaburro, L. Pavesi, C. Baratto, G. Faglia, and G. Sberveglieri, “Aporous silicon microcavity as an optical and electrical multiparam-teric chemical sensor,” NATO Science Series, II: Mathematics, Physicsand Chemistry, Frontiers of Multifunctional Nanosystems vol. 57, pp.399–412, Springer, 2002, E. V. Buzaneza and P. Scharff, Eds..

[17] V. K. S. Hsiao, T. C. Lin, G. S. He, A. N. Cartwright, P. N. Prasad, L.V. Natarajan, V. P. Tondiglia, and T. J. Bunning, “Optical microfab-rication of highly reflective volume Bragg gratings,” Appl. Phys. Lett.,vol. 86, pp. 13113–1–13113-3, 2005.

[18] L. V. Natarajan, C. K. Shepherd, D. M. Brandelik, R. L. Sutherland,S. Chandra, V. P. Tondiglia, D. Tomlin, and T. J. Bunning, “Switch-able holographic polymer-dispersed liquid crystal reflection gratingsbased on thiol-ene photopolymerization,” Chem. Mater., vol. 15, pp.2477–2484, 2003.

[19] V. P. Tondiglia, L. V. Natarajan, R. L. Sutherland, D. Tomlin, and T.J. Bunning, “Holographic formation of electro-optical polymer-liquidcrystal photonic crystals,” Adv. Mater., vol. 14, pp. 187–191, 2002.

[20] V. Mulloni and L. Pavesi, “Porous silicon microcavities as opticalchemical sensors,” Appl. Phys. Lett., vol. 76, pp. 2523–2525, 2000.

[21] S. M. Sze, Physics of Semiconductor Devices, 2nd ed. New York:Wiley, 1981.

[22] VWR Material Safety Data Sheet [Online]. Available: http://www.vwrsp.com/msds/10/JT9/JT9002-2.htm

Xiaoyue Fang received the B.Eng. degree fromTongji University, China, in 1998 and the M.S.degree from the University of Buffalo (UB), TheState University of New York, Buffalo, in 2004, bothin electrical engineering.

She was a Research Assistant in the Analog VLSILaboratory at UB. She is currently with Aerotek Au-tomotive, Oshawa, ON, Canada. Her interests includeanalog VLSI and integrated sensor systems.

Vincent K. S. Hsiao is currently working toward thePh.D. degree in the Department of Electrical Engi-neering of the University of Buffalo (UB), The StateUniversity of New York, Buffalo.

Currently, he is a Research Assistant in the Insti-tute for Lasers, Photonics and Biophotonics, UB. Hisresearch areas of interest are one-dimensional peri-odic porous polymer structures and nanoparticle pat-terning in polymer matrix using holographic interfer-ence pattern.

Vamsy P. Chodavarapu (S’03) received the B.E.degree in electronics and instrumentation engi-neering from Osmania University, India, in 2001 andthe M.S. degree in electrical engineering from theUniversity at Buffalo (UB), The State University ofNew York, Buffalo, in 2003. Currently, he is workingtoward the Ph.D. degree in electrical engineering atthe same university.

His research interests are CMOS integrated chem-ical and biological sensors, optical sensor integration,and data acquisition and processing.

Albert H. Titus (S’86–M’97) received the B.S. andM.S. degrees from the University at Buffalo (UB),The State University of New York, Buffalo, in 1989and 1991, respectively, and the Ph.D. degree from theGeorgia Institute of Technology, Atlanta, in 1997.

Prior to joining the faculty at Buffalo, he wasan Assistant Professor at the Rochester Instituteof Technology, Rochester, NY. Currently, he is anAssistant Professor in the Department of ElectricalEngineering at the University at Buffalo, The StateUniversity of New York, Buffalo. His research

interests include analog VLSI implementations of artificial vision, hardwareand software artificial neural networks, hardware implementations of deci-sion-making aids, optoelectronics, and integrated sensor systems.

Dr. Titus is a member of the INNS, ASEE, and SPIE. He has been a reviewerfor many journals and conferences and has numerous research grants from fed-eral and private sources, including an NSF CAREER Award.


Alexander N. Cartwright (S’94–M’95) received theB.S. and Ph.D. degrees in electrical and computer en-gineering from the University of Iowa, Iowa City.

He is an Associate Professor of Electrical En-gineering, Director of the Institute for Lasers,Photonics and Biophotonics and Co-Director of theElectronics Packaging Laboratory at the Universityat Buffalo (UB), The State University of New York,Buffalo. His research is focused on III–Nitride mate-rials, quantum dot materials, optical nondestructivetesting of stress and strain for device reliability,

optical sensors, biophotonics, nanophotonics, and nanoelectronics.

Dr. Cartwright received an NSF CAREER Award in 1998 and a Depart-ment of Defense Young Investigator Award for research on optical propertiesof III–Nitride materials in 2000. More recently, he was the recipient of the 2002State University of New York’s Chancellor’s Award for Excellence in Teaching.

ieee sensors journal, vol. 6, no. 3, june 2006 661

Documents