Sensor-Aware Modulator in Integrated Systems

• Sensor-aware modulators are integrated devices that co-optimize the sensing interface and modulation block, offering enhanced precision and adaptability in signal acquisition.
• They employ delta-sigma techniques and noise mitigation methods like chopping and transconductance mismatch to suppress thermal and flicker noise.
• State-of-the-art designs achieve high SNR, wide dynamic range, and efficient electromagnetic sensing in both analog front-ends and photonic circuits.

A sensor-aware modulator is a class of signal acquisition and conversion device that seamlessly integrates sensing interface functionality with signal modulation, typically in an analog front-end (AFE) or integrated photonic context. Sensor-aware modulators are distinguished by their explicit co-optimization of the sensing interface (e.g., impedance, noise, gain) and the modulation or digitization block, resulting in enhanced precision, adaptability, and physical compactness. Notable architectures include delta-sigma $\Delta\Sigma$ modulator-based sensor AFEs with dual-role differential-difference amplifiers (DDA), and silicon-organic hybrid photonic crystal waveguide (PCW) modulators with antenna-mediated field enhancement (Basu et al., 2019, Zhang et al., 2015).

1. Sensor-Aware Modulators in Analog Front-End Architectures

Delta-sigma modulator-based AFEs leverage a DDA designed for dual purposes—serving simultaneously as a high-impedance instrumentation amplifier (IA) for sensor inputs, and as an integrator within the $\Delta\Sigma$ modulation loop. The sensor interface is thus merged with the main signal processing chain, eliminating external IA circuits and enabling both filtering and digitization within a unified block. The DDA features two input transistor pairs: one couples to the sensor input, providing high input impedance ($>100\,\mathrm{M}\Omega$), while the other receives loop-feedback signals from an on-chip programmable feedback DAC.

The input path supports direct dc-to-1 kHz acquisition with zero loss of DC coupling, allowing faithful transduction of signals from high-impedance sources such as thermocouples or biosensors. The programmable feedback network controls the closed-loop gain, supporting on-chip analog preamplification without external PGAs or discrete-gain amplifiers. These features enable adaptation to a wide dynamic range of sensor outputs—from sub-microvolt to hundreds of millivolts—while maintaining low noise and high linearity (Basu et al., 2019).

2. Noise Mitigation Techniques and Dual-Role Amplifier Design

Minimization of input-referred noise is central in sensor-aware modulators. The DDA incorporates a transconductance mismatch strategy: by deliberately choosing the feedback-side input pair’s transconductance $g_{m2}$ less than the sensor-side $g_{m1}$, the thermal noise contributed by the continuous-time (CT) integrator’s front-end resistor is suppressed by the factor $A_2/A_1 = g_{m2}/g_{m1} < 1$, where $A_1$ and $A_2$ are the respective midband gains.

Flicker noise in the DDA is addressed by chopping, which inverts the input/output polarity of the DDA at the modulation oversampling clock frequency ($f_{\mathrm{ch}} = f_s$). This up-converts the $1/f$ noise to out-of-band frequencies, which are subsequently rejected by the integrator’s low-pass roll-off. The input-referred spectral density after chopping and integration is expressed as

$$S_{n,\rm in}(f)=\frac{8 k T R}{1 + [2\pi f (A+1)RC]^2} + \frac{S_{\rm DDA,th}}{A^2}\frac{1 + (f/f_c)}{1 + [2\pi f (A+1)RC]^2}$$

where $R$ and $C$ implement the R–C integrator, and $A$ is the DDA gain (Basu et al., 2019).

3. Integrated Device and Programmability

The unified modulator combines the instrumentation amplifier, CT integrator, discrete-time switched-capacitor integrator, 1-bit quantizer, and programmable feedback network on-chip. The programmable gain is realized via selectable reference voltages in the 1-bit feedback DAC, controlling the feedback factor $\alpha = V_{\mathrm{fb}}/V_{\mathrm{in}}$ and hence the loop gain. The transfer function for signal gain becomes

$\text{input gain} = \frac{1}{\alpha}$

Amplification within the modulator loop improves small-signal SNR without sacrificing dynamic range. Detailed implementation in 0.18-$\mu$m CMOS achieves simulated $80.1$ dB SNR, $109$ dBFS dynamic range, and $98.4\,\mu$W power for $1$ kHz bandwidth (OSR $=500$), with measured SNR only $9$ dB lower as limited by device and process noise (Basu et al., 2019).

4. Sensor-Aware Modulation in Integrated Photonics

In photonic implementations, sensor-aware modulation employs co-designed slot PCW modulators integrated with broadband RF antennas. The silicon-organic hybrid device utilizes a polymer (SEO125) with high EO coefficient ($r_{33} \approx 100$ pm/V), embedded within a $320$ nm-wide slot PCW. Engineered lattice shifts in the PCW yield a nearly flat group index $n_g \approx 20.4$ across an $8$ nm bandwidth, entering a slow-light regime which increases the effective interaction between RF-induced refractive index changes and the optical field.

The device’s EO modulation efficiency is quantified by the in-device $r_{33,\mathrm{eff}}=1230$ pm/V, determined from the measured $V_\pi\cdot L = 0.282$ V·mm, with an optical interaction length $L=300\,\mu$m and overlap factor $\sigma\approx0.33$. The slow-light effect scales the phase shift per unit $\Delta n$ as $n_g$, enabling efficient modulation even in ultra-short devices (Zhang et al., 2015).

5. Antenna Coupling and Electromagnetic-Field Sensing

An integrated bowtie antenna, fabricated from $5\,\mu$m thick Au and designed for resonance at $f_0\approx10$ GHz (arm length $3$ mm, flare angle $60^\circ$, gap $8.4\,\mu$m), is used to locally enhance incident EM fields within the PCW slot. Simulation predicts a peak RF field enhancement $G(f_0) \approx 10^4$ and $1$-dB bandwidth exceeding $9$ GHz about the central frequency. The modulated optical output encodes the incident electromagnetic signal, making the system a high-sensitivity, broadband electromagnetic field sensor.

Experimental results indicate a minimum detectable power density $S_{\mathrm{avg}} = 8.4$ mW/m$^2$ at $8.4$ GHz, corresponding to a minimum detectable field $E_{\min}=2.5$ V/m and sensitivity $S=2.7\times10^{-5}$ V·m$^{-1}$·Hz$^{-1/2}$, surpassing prior integrated photonic sensors (Zhang et al., 2015).

6. Performance Metrics and Comparison

Key metrics for representative sensor-aware modulators are summarized below:

Metric	$\Delta\Sigma$ Modulator AFE (Basu et al., 2019)	Silicon-Organic Hybrid PCW (Zhang et al., 2015)
Input Impedance	$>100$ M$\Omega$	N/A (photonic)
Bandwidth	DC–1 kHz	up to 11 GHz (3 dB)
SNR	80.1 dB simulated, 70 dB measured	N/A
Dynamic Range	109 dBFS	N/A
Power	$\sim 100\,\mu$W	N/A (photonic)
$r_{33,\mathrm{eff}}$	N/A	1230 pm/V
$V_\pi \cdot L$	N/A	0.282 V·mm
Sensitivity $S$	N/A	$2.7 \times 10^{-5}$ V·m$^{-1}$·Hz$^{-1/2}$

The $\Delta\Sigma$ AFE achieves state-of-the-art noise and precision for DC-to-medium frequency sensors at sub-mW power, while the photonic modulator sets records for EO efficiency and EM field sensitivity in a monolithically integrated device.

7. Outlook and Adaptability

Sensor-aware modulator architectures are adaptable to a diverse set of low-frequency (electrical) and electromagnetic (photonic) sensor modalities. In CMOS circuits, programmable gain, high-impedance input, on-chip preamplification, and comprehensive noise suppression democratize precision sensor interfacing for biomedical, environmental, and industrial signal sources (Basu et al., 2019). In photonics, integration of slow-light PCWs, high-$r_{33}$ polymers, and nanoantenna field concentrators enables compact and spectrally agile electromagnetic wave sensors, with prospects for expansion to terahertz-frequency plasmonic slot configurations for even higher bandwidth and sensitivity (Zhang et al., 2015).

A plausible implication is that continued development of sensor-aware modulators will converge toward architectures in which sensing, amplification, noise conditioning, and signal modulation coexist within one reconfigurable, highly integrated device, further minimizing form factor and power consumption while maximizing precision and operational flexibility.