A Low-Power High-Sensitivity Photocurrent

Sensory Circuit with Capacitive Feedback

Transimpedance for Photoplethysmography
Sensing
Neethu Mohan, Falah Awwad, Nabil Bastaki, and Mohamed Atef, Senior Member, IEEE
Electrical and Communication Engineering Department, United Arab Emirates University, 15551 Al Ain, United Arab Emirates.
Corresponding author: Mohamed Atef (moh_atef@uaeu.ac.ae).
This research received partial funding from the United Arab Emirates University through grant G00003441, and partial support by grant G00004546.

ABSTRACT This paper proposes an integrated analog front end (AFE) for photoplethysmography (PPG) sensing.
The AFE module has a novel transimpedance amplifier (TIA) incorporating capacitive feedback techniques with
common drain feedback (CDF) TIA. The proposed TIA topology achieves a high gain and high sensitivity at low
power consumption. The designed PPG sensor module has an input noise current of 4.81 pA/sqrt Hz, transimpedance
gain of 18.43 MΩ and has a power consumption of 68 µW. The sensory system integrated an LED driver with an
automatic light control (ALC) to reduce the LED power based on the received signal strength. The AFE is implemented
using 0.35 µm CMOS technology and occupies 1.98 mm x 2.475 mm.

Index Terms: Transimpedance amplifier, PPG sensor, Optical receiver, Automatic light control
the benefit of achieving an improved trade-off relationship
between the noise, gain, power, and bandwidth.
I. INTRODUCTION The rest of the paper is structured as follows: Section II
Optical sensing and its related circuits have significant introduces the proposed PPG sensory system and the Analog
attention in the field of biomedical applications [1]. In today’s front-end performance. Section III explains the LED driver
medical and clinical practices, different health monitoring and
devices which are photoplethysmography (PPG) based are
developed. These devices provide a very convenient mode for
physicians to measure blood pressure, monitor heart rate and
even evaluate the oxygen saturation level without the automatic light control (ALC) module. Section IV presents the
requirement of invasive mode of measurement of these post layout simulation results. Section V manifests a
parameters [1], [2]. comparison of the proposed work with recently published
A PPG sensory device consists of optical transmitter and works. The last section concludes the paper.
optical receiver [3]. The main building block in the optical
receiver is the Transimpedance Amplifier (TIA). The widely
II. THE PROPOSED PPG SENSORY SYSTEM
adopted resistive feedback TIA (RF-TIA) bandwidth is
restricted by the feedback resistor and its associated parasitic Fig. 1 shows the proposed PPG sensory system. The
capacitance. To surmount this challenge, one can opt for a system has an integrated LED driver to drive an external LED
capacitive feedback TIA (CF-TIA). The CF-TIA offers from which the light is transmitted. This light transmitted
theoretically equivalent gain and bandwidth performance traverses through the human tissue, absorbed, then reflected.
compared to the RF-TIA [4] [5]. CF-TIA will reduce thermal This reflected light from the tissues gets received by the
noise and overcome the challenge of integrating high external photodiode (PD). This received signal from the tissue
resistance with CMOS technology. CF-TIA emerges as the is a very weak signal. The PD then converts this weak optical
most favorable choice for implementing a low-noise CMOS signal into a photocurrent.
TIA [6]. The CF-TIA does not result in low power when The optical receiver amplification chain then amplifies the
compared to advanced Regulated Cascode (RGC) strategy or received weak photocurrent. The signal strength is conditioned
CG TIAs [7]. An inverter with active common-drain feedback using four comparators. The outputs from these comparators
TIA (ICDF-TIA) showed a higher gain compared to RGC-TIA are fed to the LED driver to control the LED current level.
at the same power [8] [9]. The TIA uses ICDF-TIA to achieve
higher gain compared to RGC at the same power
consumption.
In this paper we present an optical receiver using
capacitive feedback with common-drain feedback TIA (CF-
CDF-TIA) to achieve high gain and low noise at low power
consumption. The proposed CF-CDF-TIA configuration offers
 C3. In addition, pseudo-resistors (Rz) in series with a constant
biasing voltage source, are connected to the gate of both M rg
and M2. This biasing circuitry helps to get the required
operating points.
Both C1 and C2 help in improving the gain of CDF-TIA in
comparison to the conventional CDF-TIA discussed in
reference [9], see equation 4. The gain varies depending on the
ratio of values chosen for capacitors C1 and C2.
From the small signal model, we can derive the transfer
function. The transimpedance gain ZTIA ( s ) of the TIA is given
by equation (1.a).

2
a2 s + a1 s+ a0
ZTIA ( s )= 3 2 ( 1. a )
Fig.1. PPG Sensor Block Diagram. d 3 s +d 2 s + d 1 s+d 0

Fig.2. Analog front-end circuit for the PPG optical receiver.

The automatic light control (ALC) produces four different

current levels which are used to control the LED driver by
either increasing or decreasing the optical power of the LED
transmitted signal. The ALC module is a crucial optical
transceiver circuits, where signal levels can vary widely owing
to factors, such as distance, attenuation, absorption, and
reflection. By ensuring appropriate adjustment of the LED Fig.3. The proposed CF-CDF-TIA circuit.
current, the ALC module helps maintain signal integrity and
improves the overall performance of the system and reduces d 0 = gm 2 R 0 Rz + R z+ R 0 (1.b)
the LED power at high input signals level.
d 1=
2
A. PPG Receiver Analog front end gm 2 R 0 Rz C 1 ( g mrg Rrg +1 ) + Rz R rg C1 ( gm 2 R 0 +1 ) + Ro C 1 ( Rz+ R rg ) +
The AFE depicted in Fig. 2 has three stages. The first stage
(1.c)
is the proposed capacitive feedback common drain feedback
transimpedance amplifier (CF-CDF-TIA). The second stage is d 2= Rrg Rz 2 C 1 C 2 ( g mrg R 0 +1 )+ R0 Rz2 C1 ( C 2+C pd ) +
R0 Rz C2 C pd + C 1 R0 Rz Rrg (C pd +2C 2)
a capacitive feedback post amplifier (PA) for further signal 2
amplification. The post amplifier implemented in this module
(1.d)
is an operational transconductance amplifier (OTA) with
capacitive feedback. The last stage of this PPG optical d 3= C 1 C 2 C pd R0 Rz2 Rrg
receiver is an output buffer that drives the capacitive input of (1.e)
the next stage. a 0= gmrg R0 Rz Rrg (1.f)
2
B. Proposed CF-CDF-TIA circuit. a 1= Rrg gmrg R 0 Rz ( C 1 +C2 ) (1.g)
Fig. 3 shows the proposed CF-CDF-TIA schematic. The a 2= - R0 Rz2 Rrg C 1 C 2 (1.h)
design has a common source voltage amplifier (M rg, Rrg )
which regulates the gate of the NMOS M 2. This is done by The transfer function above has two zeros and three poles.
connecting C1 between the output node, Vout of the common- This transfer function is approximated by considering the
source amplifier (Mrg, Rrg ), and the gate of M2. The capacitor following: Rz >> Ro>> Rrg , R D., and Cpd>> C2>>C1. By
C2 connects the M2 gate and the input photocurrent I ph. C1 and considering the dominant pole, the bandwidth can be
C2 create a feedback path between the input photocurrent and calculated from the higher cut off frequency ω H and can be
the output voltage. The input signal is AC coupled to M rg via
approximated as:
 g R C ( R g +1 ) The derived mathematical model for input referred noise
BW ¿ f H ≈ 1 m 2 0 1 rg mrg (2) power density of the proposed CF-CDF-TIA is derived and
2π ¿¿¿ given by:
The firs zero f Z 1 determine the lower cut-off frequency and I 2
C 2 (C 1 +C ph ) s (6)
can be approximated as:
2 2 2 2 2
n ,∈¿ =I n, MB+(I n ,Rrg + I n, Mrg ¿( ) + I n ,M 2)¿
g mrg (C 1+C 2 )

1 1 Fig. 4 compares the simulated and calculated input noise

f Z 1=f L ≈
2 π ( C 1 +C2 ) Rz current for the CF-CDF-TIA. Fig. 4 shows how closely the
(3) derived theoretical noise model fits with the simulated one.
The calculated integrated noise current is equal to 6.93 x
The second zero position f Z 2 is much higher than the −10
10 A and the simulated total input referred noise current is
dominate pole which ensure the TIA stability. 7.362 x 10−10 A.
1 g mrg ( C1 +C 2)
f Z 2≈ ≫f H
2π C1C2
(4) Equivalent Input Noise
The mid-band transimpedance gain can be approximated by ( pA / √ (Hz))
the following relation:
−gmrg ( C 1+C 2 )
−C 2
ZTIA ≈ C 1+ C2 gm2 C 1 ≈
+ + gm 2 g mrg C1 g m 2 C 1
R o R rg Rrg
(5)
The CF-CDF-TIA gain is increased by ratio of two capacitors Frequency (Hz)
Fig.4. Simulated and calculated input referred noise for CF- CDF-TIA.
C2/C1, introduces no additional noise compared to ICDF-TIA
without capacitive feedback [8] [9]. The total input noise
current will be considerably reduced compared to that of the D. Post Amplifier
traditional RF-TIA. Furthermore, the system's gain is The output of the TIA is in the range of few millivolts. It is
determined by the ratio of two capacitors, making the overall required to amplify this weak signal. The post amplifier that
design less susceptible to process variations. The follows the TIA does this task by scaling the signal coming from
transconductance gm2 in equation (5) provides input signal the TIA. A post amplifier with capacitive feedback is illustrated
dependency, as the input photocurrent increases gm2 increases in Fig. 2. The transfer function for the post amplifier is given by:
providing gain compression at high input photocurrents. −sC p 2 R z
A o ( s) = (7)
C. Simulation Results for Capacitive Feedback CDF-TIA 1+ s C p 3 R z
The post layout simulation results obtained for the R z is a pseudo resistor with a resistance value of few hundreds of
proposed CF-CDF-TIA circuit is presented in this section. Giga ohms. The use of R z in the OTA design of the post
Table I describes the design parameters used for the CF-CDF- amplifier helps OTA input gate to get the required DC biasing.
TIA. The higher cut-off frequency for the post amplifier design is
Table II shows calculated and simulated values of bandwidth
1
and transimpedance gain. The analysis says that as the input and for the model implemented in this paper, F L is
photocurrent increases, the gain decreases. Furthermore, C p 3 Rz
another point that can be concluded from this parametric designed to be less than 0.5 Hz. The OTA has a mid-band gain of
analysis is that, as the photocurrent increases from 1 nA to 200 30 and consumes biasing current of 1.567 µA.
µA, the bandwidth increases. There is a good match between
the analytical model calculations and simulation results. The
calculation error is because of neglecting the parasitic III. AUTOMATIC LIGHT CONTROL (ALC) AND LED
capacitances and output resistances for the MOSFETs. DRIVER

Table I SIMULATION PARAMETERS An ALC system automatically adjusts the LED current to
Rrg R1 C1 C2 C3 Cpd WMrg, W1, W2 maintain a constant output signal level despite variations in the
input signal strength. ALC module ensures that the output signal
165KΩ 85 KΩ 1 pF 10p 50pF 20pF 1µ, 1µ, 4µ
F remains within the linear range even at wide variation in the
received optical power, thereby preventing distortion caused by
Table II. CALCULATED AND SIMULATED VALUES OF GAIN AND signal overload or excessive amplification of weak inputs. The
BANDWIDTH ALC module continuously monitors the output voltage of the
Iph Simulated Calculated Simulated Calculated TIA. For a weak input optical power, the ALC module increases
Gain (dB) Gain (dB) BW (KHz) BW (KHz) the LED current to generate higher optical power to maintain a
1nA 115.82 117.46 13.94 13.49
1µ A 114.4 114.37 16.38 15.72
100 µA 78 81.6 760.2 627.07
200 µA 74.5 79.1 886.2 698.4
stronger output voltage. Conversely, for a strong input signal,
the ALC module reduces the LED current to prevent signal
distortion owing to clipping or saturation. ALC module includes
control loop and feedback mechanisms that adjust the gain
smoothly and rapidly for tracking changes in the input signal.
Fig. 5 shows the ALC module designed in this paper. It
consists of an envelope detector and four comparators. The
output of the amplifiers chain is provided as input to the peak
detector. The output of the peak detector is fed to four different
comparators which work at four different voltage levels between
the maximum and minimum voltage levels detected at the
output of the amplifier’s module. The ALC generates four
different control signals (Ctrl1, Ctrl2, Ctrl3, Ctrl4) which
control the LED driver’s current level. If the envelope detector
output is below the threshold voltage of VDD/2 V, the
amplitude is considered to be low, and the LED current is
increased to the next level. If the envelope detector values are
above VDD/2 + 0.2 V, the signal amplitude is considered to be Fig.6. The LED Driver circuit
high, hence LED current is reduced by one step.
The LED driver supplies the LED with the required current Fig. 7 shows the transient simulation for a high input
to emits enough optical power to have a high-quality PPG photocurrent of 20 µA. Fig. 7 indicates that the switching
signal. Fig. 6 below is the LED driver designed to have four mechanism of the LED driver mentioned above reduces the
channels. The four LED channels current levels are controlled LED current at high input photocurrents levels. The LED driver
through switching transistors Mn5 , M n 6 , M n 7 ,∧M n 8 using consumes a maximum power of 1.515 mW.
Ctrl1, Ctrl2, Ctrl3, Ctrl4 coming from the ALC. The left side of
the driver is including the M p 1 ,2 and M n 1 ,2 and R1 provides IV. PPG SENSOR POST LAYOUT SIMULATION

the reference current which is mirrored using the M n 3 ,4 The complete system layout is implemented using AMS
¿ M p 3 ,4 ,5to the four channels M p 6 ,7 , 8 ,9 . Different mirroring 0.35 µm CMOS technology. Fig. 8 depicts the layout of the
sensor module with chip area of 1.98 mm x 2.475 mm. The
ratios 2:4:8:16 are used to supply various LED current ranges
post layout frequency response simulation for different
from 400 μA to 11 mA for each channel.
photocurrents from 1 nA to 200 µA is depicted in Fig. 9. As
input photocurrent increases, the gain decreases. This gain
reduction will preserve the signal linearity at high input photo
current. The overall bandwidth for the proposed system is 1.6
kHz for a low input photocurrent. The overall bandwidth is
limited by the post amplifier and output buffer bandwidths not
by the TIA bandwidth.
Fig. 10 illustrates the post layout transient simulation for the
PPG sensor module. The transient analysis used an input
photocurrent of 1 µA and 200 µA. The output is amplified by
145.3 dB at low input photocurrent while maintaining the
linearity at high input photocurrent of 200 µA by reducing the
gain to 98.73 dB.
The effect of process variation on the proposed PPG sensor
bandwidth and gain has been analyzed using Monte-Carlo
simulation. The bandwidth and gain changes are calculated for
1000 Monte-Carlo runs and plotted in Fig. 11 (a) and (b). The
Monte-Carlo simulation histogram for bandwidth has a mean
bandwidth of 1.61 KHz and a standard deviation of 25.29 Hz.
The histogram for gain variation depicts a mean gain of 18.44
Fig.5. The AGC module MΩ and a standard deviation of 597.9 KΩ.
 Fig.10. Output PPG voltage for input photocurrent 1 µA and 200 µA
respectively.

Fig.7. The LED Driver circuit current for 20 µA input photocurrent.

(a)
Fig.8. The PPG sensor layout.

Fig.9. Frequency response parametric analysis for the proposed system for
different photocurrent from 1 nA to 200 µA. (b)
Fig. 11. Monte-Carlo simulation for (a) Bandwidth and (b) Gain.
V out (V )

V. COMPARISON OF PROPOSED WORK RESULTS

This section introduces a comparison of different performance
time(s) parameters discussed in this work with state-of-the-art PPG
sensors. Table III shows the comparison between state-of-the-
V out (V ) art PPG sensors and the proposed one.
In [10], H. Aminah et.al. predicts the sugar level using single
wavelength photoplethysmography. The system is accoutered
using 180 nm process. The system has an input referred current
noise of 7.3 pA/ √Hz.

time( s)
In [3], L. Binghui et.al. presents a PPG sensory system for REFERENCES
continuous health monitoring. The PPG chip is fabricated using
350 nm standard CMOS technology. The average power
consumption of the receiver analog front-end is 50.75 µW. The
PPG sensor has a input noise current of 41.3 pA/ √Hz and a
gain of 11.9 MΩ.
S. Wala et.al. in [11] proposed a 180 nm process for
photoplethysmography-based non-invasive glucose sensing.
The work in [11] has an input photocurrent range up to 65 µA.
The system proposed in [12] by L. Qiuyang et.al. implemented
in 180 nm process. It is a PPG based non-invasive light-to-
digital converter with a maximum gain of 4 MΩ.
The performance of the proposed sensor in this study is better
than the state-of-the-art PPG sensors performance. The
proposed CF-CDF-TIA succeeded in reaching high-gain and
low-noise level at the low power consumption.

Table III. COMPARISON BETWEEN STATE-OF-THE-ART AND THE

PROPOSED WORK
CMOS I(noise) Gain I ph Power
Process (pA/√Hz) (MΩ) µW
µA
[10] 180 nm 7.3 1 20 1620

[3] 350 nm 11.9 12.5 24 50.75

[11] 180 nm 9.4 1 65 186

[12] 180 nm 5.7 4 50 89

This work 350 nm 4.81 18.43 200 68

VI. CONCLUSION
This paper presents a PPG sensory chip which has a high
sensitivity and a low power. To attain the low power sensory
system along with a high gain and low noise, a novel CF-
CDF-TIA was designed using a common drain feedback TIA
with capacitive feedback. An automatic light control loop was
also integrated for reducing the LED current at high input
photocurrent. The validity of the new configuration has been
confirmed through the remarkable consistency observed
among the analysis, and simulation results. The achieved high
sensitivity and low power consumption enable the proposed
integrated PPG sensor to be used for wearable health
monitoring systems.

[1] Z. Bielecki, T. Stacewicz, J. Wojtas, J. Mikolajczyk, D. Szabra and A. Prokopiuk, "Selected Optoelectronic Sensors in Medical Applications," Opto-
Electron, pp. 122-133, 2018.
[2] Z. Chen, S. Obaid and L. Lu, "Recent advances in organic optoelectronic devices for biomedical applications," Opt. Mater. Express, vol. 9, p. 3843–
3856, 2019.
[3] L. Binghui, M. Zhouchen and M. Atef, "Low-Power High-Sensitivity Photoplethysmography Sensor for Wearable Health Monitoring System," IEEE
SENSORS JOURNAL, vol. 21, no. 14, pp. 16141-16150, July 15, 2021.
[4] J. D. Noh, "Frequency-Response analysis and Design Rules for Capacitive Feedback Transimpedance Amplifier," IEEE Transactions on
Instrumentation and Measurement, vol. 69, no. 12, pp. 9408-9416, Dec. 2020.
[5] A. Romanova and V. Barzdenas, "Programmable-Gain Capacitive Feedback TIA for Low-Noise Applications in CMOS," IEEE Open Conference of
Electrical, Electronic and Information Sciences (eStream), pp. 1-6, 2020.
[6] Z. Ebrahimi and B. Gosselin, "Ultralow-Power Photoplethysmography (PPG) Sensors: A Methodological Review," IEEE Sensors Journal, vol. 23,
no. 15, pp. 16467-16480, 2023.
[7] A. Romanova and V. Barzdenas, "A review of Modern CMOS Transimpedance Amplifiers for OTDR Applications," Electronics 2019, 2019.
[8] M. Atef and H. Zimmermann, "10Gbit/s 2mW inductorless transimpedance amplifier," IEEE International Symposium on Circuits and Systems
(ISCAS), pp. 1728-1731, 2012.
[9] A. Atef, M. Atef, E. Esam M Khaled and M. Abbas, "CMOS Transimpedance Amplifiers for Biomedical Applications: A Comparative Study," IEEE
 CIRCUITS AND SYSTEMS MAGAZINE, pp. 12-31, 2020.
[10] H. Aminah and S. Wala, "A noninvasive glucose monitoring SoC based on single wavelength photoplethysmography," IEEE Trans. Biomed. Circuits
Syst., vol. 14, no. 3, pp. 504-515, 2020.
[11] H. Aminah and S. Wala, "A 186 μW Photoplethysmography-Based Noninvasive Glucose Sensing SoC," IEEE SENSORS JOURNAL, vol. 22, no.
14, pp. 14185-14195, 2022.
[12] L. Qiuyang, X. Jiawei, S. Shuang, B. Arjan, K. Mario, V. H. Chris and T. Filip, "A 119dB Dynamic Range Charge Counting Light-to-Digital
Converter for Wearable PPG/NIRS Monitoring Applications," IEEE Trans. Biomed. Circuits Syst., vol. 14, no. 4, pp. 800-810, 2020.

worked on the improvement of quantum dot optical properties. He received

his PhD in 2010 from Vienna University of Technology, Institute of
Electrodynamics, Microwave and Circuit Engineering, and then worked as a
post-doctoral researcher until the end of 2012. Dr. Atef visited the School of
Microelectronics, Shanghai Jiao Tong University, China, from 2015 to 2017.
NEETHU MOHAN is a Ph.D. candidate in the Mohamed Atef has been an Associate Professor at Assiut University, Egypt
Electrical Engineering Department of United Arab since 2016. In 2020 he joined Electrical Engineering Department, United Arab
Emirates University of UAE. She has a bachelor’s Emirates University, United Arab Emirates, and he is still serving there as an
degree in Electronics and Communications Associate Professor. His research interests are in the areas of optoelectronic
Engineering from Vidya Academy of Science and integrated circuits, and biomedical circuits and system. He is an author of two
Technology in the year 2015 and master’s degree in Springer books, "Optical Communication over Plastic Optical Fibers:
applied Electronics and Communications Systems Integrated Optical Receiver Technology" and "Optoelectronic Circuits in
from A P J Abdul Kalam Technical University, India Nanometer CMOS Technology” and an author and co-author of more than
in the year 2019. Her research interests include the 100 scientific publications. Mohamed Atef served as a TPC member in many
intersection of Neural Networks, Machine Learning IEEE conference, as a guest editor for Sensors Journal special issue (2019-
and Embedded systems. 2020), and for several TBioCAS special issues, and as an associate editor for
TBioCAS (2020-2023). Mohamed Atef is a member of the Biomedical and
Life Science Circuits and Systems Technical Committee (2018-present). He
Falah Awwad received MSc. and Ph.D. degrees in was awarded the State Encouragement Award in Advanced Technological
Electrical and Computer Engineering from Sciences Serving the Engineering Sciences for 2018 from the Egyptian
Concordia University (Montreal, QC, Canada) in Academy of Scientific Research and Technology (ASRT).
2002 and 2006, respectively. He was a Post-
Doctoral Fellow at Ecole Polytechnique de
Montréal and Concordia University, Montreal, QC,
Canada. Between Aug. 2007 and Feb. 2013, he was
an assistant professor with the College of
Information Technology at United Arab Emirates
University. Currently, he is a professor with the Department of Electrical and
Communication Engineering - College of Engineering (UAE University). He
has published over 100 research articles in peer-reviewed journals and
international conferences. He is a member of the editorial boards of several
journals. He is the principal investigator and co-PI of more than 20 research
projects and has supervised several postgraduate students. His primary
scientific research interests include sensors, circuits, and devices, in addition
to hardware security and biomedical applications. He is a member of the
steering committee of ICM and session chair and a technical program
committee member of several international conferences, including the
International Conference on Design and Technology of Integrated Systems in
the Nanoscale Era, ISCAS, IT Innovations Conference, MWSCAS, EMBC,
and IEEE Sensors conference. He is also a regular reviewer for several
journals such as Microelectronics Journal, Journal of Nanomaterials &
Molecular Nanotechnology, TVLSI, and Biosensors and Bioelectronics.

NABIL BASTAKI is an assistant professor in the Department of Electrical

Engineering at the United Arab Emirates University (UAEU). He is also the
assistant dean for research and graduate studies in the College of Engineering
(COE). He received his BSc degree in Computer Engineering from the
University of Arizona in May 1989. He joined UAEU as a Teaching Assistant
in 1990. He received his MEng. degree in electrical engineering from Cornell
University in May 1993 and his Ph.D. degree in Computer Engineering in
May 2001 from the University of Southern California. His research interests
are in Embedded Systems, Robotics, and Digital Systems Design. He has held
several positions while at UAEU, the Director of the Continuing Education
Center, the Assistant Dean for Student Affairs at the College of Engineering,
and the Head of Industrial Training and Graduation Projects Unit.

MOHAMED ATEF (Senior Member, IEEE) received

the B.Sc. and M.Sc. degrees in electrical engineering,
electronics and communications from Assiut University,
Egypt, in 2000 and 2005, respectively. From 2006 to 2007,
he got a research scholarship from the Czech Technical
University in Prague, Department of Microelectronics, he