Jiayi (Emma) Xu

Abandoning the traditional approaches to multi-object spectrographs, DMD-MOS (Digital Micro-mirror Device-based Multi-Object Spectrograph) leverages a robust, elegant approach for mask creation. DMDs are arrays of microscopic flat mirrors where each micro-mirror can be addressed to selectively flip between two discrete angles. With this in mind, we have designed two optical channels for DMD-MOS - one for each of these two discrete positions: 1) an imaging channel that can be used for multi-object target selection and 2) a spectrograph channel. By leveraging this feature, the DMD-MOS can simultaneously image the field of view (FOV) and disperse the selected targets. In this work, we present the results of our laboratory testing of DMD-MOS including spectral resolution, line spread function, wavelength coverage, and the mapping between the DMD and imaging channel planes, the latter of which enables DMD-MOS and our control software to instantaneously configure the DMD to observe the chosen target(s) in the acquisition field.

I led the integration and calibration phase, proving system robustness through successful on-sky observations. By co-developing a homography-based tracking algorithm, I achieved excellent spectrum matching and precision target allocation on the DMD micro-mirrors, directly validating the instrument's exceptional optical reliability and high-fidelity performance.

Paper link:

• Jiayi E. Xu, Sarik Jeram, Shaojie Chen, Ting S. Li, Ziang Chen, Zain Azam, Adi Khandelwal, Mark Barnet, Yixuan Cheng, Suresh Sivanandam. "Prototyping the next-generation spectrograph for space exploration: Digital Micromirror Device based Multi-object Spectrograph (DMD-MOS) calibration and performance assessment." SPIE Astronomy Telescopes + Instruments, 2026. [PDF]

Poster link:

• Jiayi E. Xu, Sarik Jeram, Shaojie Chen, Ting S. Li, Ziang Chen, Zain Azam, Adi Khandelwal, Mark Barnet, Yixuan Cheng, Suresh Sivanandam. "Prototyping the next-generation spectrograph for space exploration: Digital Micromirror Device based Multi-object Spectrograph (DMD-MOS) calibration and performance assessment." SPIE Astronomy Telescopes + Instruments, 2026. [PDF]

GigaBIT (Giga-pixel Balloon Borne Imaging Telescope) was designed for deep, wide-field multi-band surveys. Featuring a specialized near-ultraviolet (UV) and narrowband filter system, its primary mission is to break classical stellar parameter degeneracies and reconstruct the chemo-dynamical evolutionary history of the Milky Way. In addition, it could also bring time domain astronomical studies with the observation of a few different types of transients fullfilling the observing limitation from ground-based telescopes.

I made a predictive bandpass-to-color empirical model trained with sequential machine learning methods with PHOENIX stellar spectrum libraries. To bridge ideal synthetic photometry with hardware constraints, I integrated an Exposure Time Calculator (ETC) as a dynamic noise engine, injecting realistic readout, dark current, and atmospheric/sky background limits into the simulated flux measurements across different filters. These noise-weighted color errors were then fed into a 4-dimensional Fisher Information Matrix framework. By calculating the sensitivity derivatives within the parameter space, this integrated pipeline successfully projects observational uncertainties directly back into stellar parameter constraints, systematically accounting for the multi-parameter covariances and correlations of effective temperature, surface gravity, metallicity and alpha elements abundance.

Built apong the Fisher matrix, I further implemented a time-distribution optimization engine to maximize telescope observation efficiency. Instead of employing uniform or empirical exposure times, this algorithmic optimizer interprets specified scientific precision targets—such as separating the multiple stellar populations (MSPs) within globular clusters or filtering out background field star contamination for dim targets—and reverse-engineers the mathematical criteria into custom observation strategies. The engine calculates the exact exposure time allocation required for each filter to reach the maximum scientific output per unit time, serving as a critical bridge between instrument limitations and primary science goals. This framework transforms telescope design from experiential guessing into a rigorous, automated workflow, proving that strategically allocated near-ultraviolet (UV) and narrowband observations can drastically reduce the exposure budget while preserving the multi-dimensional chemical tracking capabilities needed to reconstruct the chemo-dynamical history of the Galaxy.