Radiation-induced photocurrent in semiconductor devices can be simulated using complex physics-based models, which are accurate, but computationally expensive. This presents a challenge for implementing device characteristics in high-level circuit simulations where it is computationally infeasible to evaluate detailed models for multiple individual circuit elements. In this work we demonstrate a procedure for learning compact delayed photocurrent models that are efficient enough to implement in large-scale circuit simulations, but remain faithful to the underlying physics. Our approach utilizes dynamic mode decomposition (DMD), a system identification technique for learning reduced-order discrete-time dynamical systems from time series data based on singular value decomposition. To obtain physics-aware device models, we simulate the excess carrier density induced by radiation pulses by solving numerically the ambipolar diffusion equation, then use the simulated internal state as training data for the DMD algorithm. Our results show that the significantly reduced-order delayed photocurrent models obtained via this method accurately approximate the dynamics of the internal excess carrier density-which can be used to calculate the induced current at the device boundaries-while remaining compact enough to incorporate into larger circuit simulations.

No abstract

We show how continuous-depth neural ODE models can be framed as single-layer, infinite-width nets using the Chen-Fliess series expansion for nonlinear ODEs. In this net, the output "weights" are taken from the signature of the control input -a tool used to represent infinite-dimensional paths as a sequence of tensors -which comprises iterated integrals of the control input over a simplex. The "features" are taken to be iterated Lie derivatives of the output function with respect to the vector fields in the controlled ODE model. The main result of this work applies this framework to derive compact expressions for the Rademacher complexity of ODE models that map an initial condition to a scalar output at some terminal time. The result leverages the straightforward analysis afforded by single-layer architectures. We conclude with some examples instantiating the bound for some specific systems and discuss potential follow-up work.

No abstract

The confinement of ions in a radio-frequency (RF) trap (also known as a Paul trap) has proven to be advantageous in many applications. In nearly all cases, singly- or few-times-ionized atoms are created in situ within the RF trap. Highly charged ions, on the other hand, are produced more efficiently in dedicated external sources; hence, the isolation of single highly charged species in an RF trap is more involved. In this work, highly charged ions produced by an electron beam ion trap/source are extracted in bunches via an ∼7 m long beamline, which is tuned to minimize the phase-space volume of the ion bunch. The charge-state-selected ion bunch is then captured in an RF trap constructed from cylindrically symmetric electrodes with pseudohyperbolic surfaces. The RF drive parameter space is surveyed both experimentally and computationally to investigate the dynamics and map out those regions favorable for ion capture. We find that an appreciable number of Ne10+ ions are captured using an RF frequency of 2.4 MHz and an amplitude range of 120 V–220 V, with an efficiency highly dependent on the RF field phase. An experimental capture efficiency of >20% was attained, with at least 500 ions being captured by the RF trap. This is slightly higher (∼135%) than that captured by a contiguous, compact Penning trap. However, in the absence of any cooling mechanism, the observed ion-storage lifetime in the RF trap is 69 ms, a factor of ∼30 shorter than in the Penning trap; potential improvements are discussed.