Transient Current Technique (TCT) Overview

Updated 10 July 2026

Transient Current Technique (TCT) is a laser-based method used to map internal electric fields, charge transport, and depletion profiles in semiconductor detectors.
It utilizes variations such as top-, edge-, backside-, and two-photon TCT to achieve high-resolution, three-dimensional mapping of sensor responses and timing performance.
Analysis of TCT signals through prompt current and integrated charge measurements enables extraction of key parameters like depletion depth and electric field distribution.

Searching arXiv for recent and foundational papers on Transient Current Technique and related TPA-TCT/edge-TCT applications. arXiv search query: "Transient Current Technique TCT semiconductor detector edge-TCT TPA-TCT" Transient Current Technique (TCT) denotes, in detector physics, a well-established pulsed-laser method used to map the internal electric field, carrier transport properties, and depletion profile of reverse-biased semiconductor devices by recording the current transient induced when a localized cloud of electron–hole pairs drifts through the sensor (Wüthrich et al., 2022). Its major experimental variants include top-TCT, edge-TCT, backside-TCT, and Two-Photon-Absorption TCT (TPA-TCT), the last of which confines carrier generation to a micrometric focal volume and enables true three-dimensional mapping of depletion boundaries, charge sharing, Time-of-Arrival (ToA), and in-pixel structure (Pape et al., 2022, Palomo et al., 12 May 2026). The same acronym also appears in other research domains, notably transient eddy-current flow metering in conducting fluids and first-principles transient-current calculations for molecular devices, so the meaning of “TCT” is context dependent (Forbriger et al., 2015, Zhang et al., 2013).

1. Core principle and scope

In semiconductor-detector usage, TCT rests on the Shockley–Ramo theorem. When a carrier of charge $q$ moves with instantaneous velocity $v(t)$ in the weighting field $E_w$ , the induced current is written as

$i(t)=q\,v(t)\,E_w,$

or, in position-dependent form,

$i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$

For many simultaneously generated carriers, the measured transient is the superposition of the electron and hole contributions (Wüthrich et al., 2022, Debevc et al., 2023).

The total collected charge is obtained by time integration,

$Q=\int i(t)\,dt,$

and the transient shape contains direct information on drift dynamics, weighting-field non-uniformity, depletion extent, and trapping. In edge-TCT studies of strip sensors, the initial transient or integrated prompt signal is used to construct position-dependent velocity profiles and, by fitting, position-dependent electric fields $E(y)$ and space-charge densities $\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ (Klanner et al., 2019). In monolithic pixel and LGAD studies, the same framework supports extraction of depletion depth, charge-collection efficiency, gain-layer depletion voltage, and timing observables such as ToA and jitter (Palomo et al., 12 May 2026, Velkovska et al., 11 Sep 2025).

A common point of confusion is that TCT is not itself a single geometry. It is a family of transient-measurement methods that differ mainly by carrier-generation mechanism and illumination geometry. This suggests that the most precise definition of TCT is methodological rather than instrumental: it is the reconstruction of device-internal transport and field properties from laser-induced transient currents.

2. Experimental geometries and instrumentation

Reported TCT implementations span several illumination geometries and laser regimes. Linear-absorption TCT typically uses visible or near-infrared pulsed diode lasers, whereas TPA-TCT uses femtosecond infrared pulses at wavelengths for which single-photon absorption in silicon is negligible (Wüthrich et al., 2022, Pape et al., 2022).

Variant	Illumination and generation	Representative use
Top-TCT	Surface illumination with red or IR laser	Near-surface response, strip and LGAD scans
Edge-TCT	Beam injected through a polished or cleaved edge	Depletion depth, $E(y)$ , one-sided depletion
Backside-TCT	Beam enters from wafer backside	In-pixel timing and bulk response in DMAPS
TPA-TCT	Femtosecond sub-bandgap IR focused inside bulk	True 3D mapping, charge sharing, ToA, jitter

Concrete apparatus details vary by study but show a stable pattern. Edge-TCT systems for bonded Si–Si diodes employed a turnkey scanning rig with 660 nm and 1064 nm pulsed diode lasers, a telecentric objective, an $x$ – $v(t)$ 0 stage with $v(t)$ 1 steps, a Keithley 2410 SMU, a Particulars AM-02A amplifier with 53 dB gain, and a LeCroy Waverunner 8104 oscilloscope with 1 GHz bandwidth and 20 GS/s (Wüthrich et al., 2022). RD50-MPW2 measurements combined a 980 nm backside-TCT setup at Nikhef and a 1064 nm edge-TCT setup at JSI, both with sub–10 $v(t)$ 2 laser spots and GHz-class readout (Debevc et al., 2023). The H35DEMO HV-CMOS study used a 1064 nm infrared laser delivering $v(t)$ 3 ps pulses at 1 kHz, a minimum spot FWHM of $v(t)$ 4 in silicon, 1 $v(t)$ 5 positioning precision, and 53 dB current amplifiers feeding a 5 GS/s DRS4 digitiser (Anders et al., 2018).

Mechanical and electrical packaging are often as consequential as the optical chain. For bonded pad diodes, a bespoke “sandwich” holder replaced wire-bond fixtures by clamping the diode between two custom 4-layer PCBs with conductive-foam gaskets forming a Faraday cage, while openings allowed laser access from top, bottom, or edge (Wüthrich et al., 2022). For segmented sensors, grounding of neighboring cells, impedance-controlled lines, and calibration of metallization-induced reflections are recurrent requirements (Lasaosa et al., 19 May 2025, Palomo et al., 12 May 2026).

3. Two-photon absorption and three-dimensional localization

TPA-TCT exploits the fact that for wavelengths above approximately $v(t)$ 6 in silicon, linear absorption is negligible, while two-photon absorption remains possible. The local carrier-generation rate is

$v(t)$ 7

so charge is created only where the instantaneous optical intensity is maximal, namely in the focal volume of a tightly focused ultrafast beam (Palomo et al., 12 May 2026). In the RD50-MPW4 study, for $v(t)$ 8 and $v(t)$ 9 in silicon with $E_w$ 0, the measured beam waist was $E_w$ 1 and the Rayleigh length $E_w$ 2, corresponding to lateral resolution of about $E_w$ 3, depth resolution of about $E_w$ 4, and an excitation voxel approximated by

$E_w$ 5

(Palomo et al., 12 May 2026).

This localization changes the epistemic status of the measurement. In linear-absorption top- or edge-TCT, the laser deposits charge along an extended path, so deconvolution with absorption and weighting effects is central. In TPA-TCT, carrier generation is confined to the focal point, which enables direct volumetric scans. In the RD50-MPW4 DMAPS, backside illumination was possible because silicon is transparent for near-infrared light at $E_w$ 6; the chip was mounted face down, no polishing was needed, and any pixel element in the $E_w$ 7 matrix could be characterized by moving the device relative to the focal spot (Palomo et al., 12 May 2026).

Segmented detectors introduce TPA-specific artifacts. In top-illuminated strip sensors, beam clipping by front-side metal and reflection from a metallized backside modulate the generated charge. To compensate, the weighted prompt current,

$E_w$ 8

was introduced to divide out intensity variations and recover a quantity proportional to $E_w$ 9 rather than to the raw local light intensity (Pape et al., 2022). The same study introduced the mirror technique, which uses backside reflection to probe beneath front-side metal with an effective mirrored depth

$i(t)=q\,v(t)\,E_w,$ 0

A plausible implication is that TPA-TCT is most powerful when optical-system characterization and signal normalization are treated as part of the field-reconstruction problem rather than as secondary corrections.

The method is not restricted to silicon. In synthetic single-crystalline diamond, two-photon absorption with 400 nm pulses was used to generate carriers in a true 3D focal volume, map prompt-current variations, and correlate the resulting electric-field inhomogeneities with X-ray diffraction topography (Dorfer et al., 2019).

4. Reconstruction methods and derived observables

TCT data analysis proceeds from a transient waveform to physically interpretable observables. The most basic are prompt current, integrated charge, and drift time, but the literature shows several more specialized reconstruction strategies (Klanner et al., 2019, Lasaosa et al., 19 May 2025).

Depletion depth is often extracted from spatial scans. In an XZ scan of RD50-MPW4 at $i(t)=q\,v(t)\,E_w,$ 1, the transient amplitude as a function of depth showed a plateau over the depleted region and fell to zero outside it, yielding a full-signal interval $i(t)=q\,v(t)\,E_w,$ 2 (Palomo et al., 12 May 2026). In H35DEMO edge-TCT, the depth profile $i(t)=q\,v(t)\,E_w,$ 3 was fitted with an $i(t)=q\,v(t)\,E_w,$ 4 term plus a Gaussian term accounting for charge sharing, and the depletion depth was defined as the average FWHM across a fiducial region. For non-irradiated samples, the depth was then fitted as

$i(t)=q\,v(t)\,E_w,$ 5

with $i(t)=q\,v(t)\,E_w,$ 6 (Anders et al., 2018).

Electric-field extraction is more subtle. In heavily irradiated strip sensors, velocity profiles derived from edge-TCT were fitted on a grid of depths with the bias constraint

$i(t)=q\,v(t)\,E_w,$ 7

and the recovered field was differentiated to obtain the local space-charge density (Klanner et al., 2019). In segmented sensors, weighted prompt current was introduced because raw prompt current inherits optical clipping and reflection artifacts (Pape et al., 2022). In LGADs, the usual velocity-profile approach can fail because the rising edge is dominated by avalanche build-up; the diffusion profile method therefore uses the maximum of the time derivative of the edge-TCT waveform to infer field-dependent carrier-cluster dispersion and reconstruct $i(t)=q\,v(t)\,E_w,$ 8 (Fu et al., 2023).

Timing analysis has also become a major TCT application. In RD50-MPW2, time resolution was extracted from the spread of

$i(t)=q\,v(t)\,E_w,$ 9

with the total timing budget written as

$i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 0

where electronic jitter was parameterized as $i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 1 (Debevc et al., 2023). For 3D pixel sensors, ToA was defined by a constant-fraction threshold,

$i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 2

with $i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 3, and a TPA-TCT-based jitter method used two consecutive pulses in the same acquisition window, fitted the distribution of $i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 4, and obtained

$i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 5

under the assumptions of Gaussian, uncorrelated jitter and negligible laser-cavity jitter between consecutive pulses (Lasaosa et al., 19 May 2025).

A recurrent misconception is that a single scalar observable, such as peak current, always maps directly to the electric field. The published methods indicate otherwise: in segmented sensors intensity artifacts bias prompt-current maps; in LGADs avalanche gain distorts the velocity-profile method; and in irradiated strip detectors finite electronics response limits the depth range over which fitted fields remain reliable (Pape et al., 2022, Fu et al., 2023, Klanner et al., 2019).

5. Semiconductor-detector applications

The most developed use of TCT is the characterization of silicon detectors for particle physics. In the RD50-MPW4 Depleted Monolithic Active Pixel Sensor, TPA-TCT resolved the boundaries of the sensitive volume for $i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 6 pixels, measured a depletion depth of $i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 7 at $i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 8, found $i(t)=q\cdot E_w(r(t))\cdot v(r(t)).$ 9 charge-collection efficiency via hit-detection efficiency, and directly observed charge sharing at the pixel periphery; the 3D sensitivity map also resolved in-pixel electronics and the limits of the depletion region (Palomo et al., 12 May 2026). The device itself is a $Q=\int i(t)\,dt,$ 0 DMAPS matrix developed by the HV-CMOS working group within the CERN RD50 collaboration (Palomo et al., 12 May 2026).

TCT has also been used to study timing in monolithic pixels. In RD50-MPW2, backside- and edge-TCT showed a uniform central timing region with approximately $Q=\int i(t)\,dt,$ 1 at $Q=\int i(t)\,dt,$ 2 of deposited charge, while performance degraded near pixel edges and at lower deposited charge. Edge-TCT maps showed more than $Q=\int i(t)\,dt,$ 3 degradation at pixel edges due to charge sharing, and comparison with direct charge injection indicated that front-end electronics jitter dominated the TCT timing at $Q=\int i(t)\,dt,$ 4 (Debevc et al., 2023).

For 3D pixel sensors, TPA-TCT has become a tool for geometry optimization. Measurements on IMB-CNM devices showed that, at $Q=\int i(t)\,dt,$ 5 and $Q=\int i(t)\,dt,$ 6, a square cell had a median-corrected ToA spread of $Q=\int i(t)\,dt,$ 7, while a hexagonal cell had $Q=\int i(t)\,dt,$ 8 under the same conditions; the hexagonal sample jitter was measured as $Q=\int i(t)\,dt,$ 9, best achieved with a CFD fraction of $E(y)$ 0 at $E(y)$ 1 (Lasaosa et al., 19 May 2025). In CNM double-sided 3D sensors, top-TCT with red and infrared lasers showed mostly homogeneous charge collection in the bulk, while source-timing measurements reached down to about $E(y)$ 2 after irradiation to $E(y)$ 3 $E(y)$ 4 neutron equivalent fluence (Lex et al., 11 Jun 2026).

LGAD production quality control provides a different application mode. In the ATLAS-HGTD production program, a TCT-based irradiation test on a $E(y)$ 5 LGAD structure extracted gain-layer depletion voltage $E(y)$ 6, internal gain $E(y)$ 7, leakage current through baseline subtraction, and effective interpad distance. These quantities were cross-calibrated with C–V and $E(y)$ 8 measurements and used to define radiation-hardness acceptance criteria for wafer production (Velkovska et al., 11 Sep 2025).

TCT is equally important when the device under study behaves unexpectedly. Edge-TCT on low-temperature covalently bonded Si–Si P–N diodes showed that only the P-side depletes under reverse bias, that no charge is collected when the IR spot is placed in the N-side bulk, and that the bonded interface behaves electrically like an $E(y)$ 9 region. The same waveforms displayed a characteristic tail with $\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 0, absent in a reference device, indicating trapping or slow de-trapping at the covalent interface (Wüthrich et al., 2022).

6. Limitations, artifacts, and interpretive boundaries

The methodological strengths of TCT are explicit in the literature: it provides direct, time-resolved access to carrier drift dynamics and collected charge; edge-TCT gives high-resolution depletion profiling; and femtosecond TPA-TCT yields true three-dimensional, micrometric mapping of timing response and charge collection (Wüthrich et al., 2022, Lasaosa et al., 19 May 2025). These strengths are balanced by equally explicit limitations.

Optical artifacts are central. Metallization can introduce reflections and amplitude modulations, requiring calibration by preliminary scans over metallized and non-metallized regions (Palomo et al., 12 May 2026). In segmented sensors, beam clipping by front-side metal and back-reflection from a metallized backside distort raw prompt-current and collected-charge maps unless compensated (Pape et al., 2022). In 3D sensors, beam clipping by metal pads and columnar electrodes can locally reduce charge generation and bias ToA extraction (Lasaosa et al., 19 May 2025). Backside reflections at the Si/air interface can enhance measured peaks relative to simulations, especially near the back side of the device (Pape et al., 2024).

Geometrical corrections are also intrinsic to the method. In backside TPA-TCT, the focal depth inside silicon must be corrected for refraction through a conversion factor, reported for RD50-MPW4 as

$\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 1

with beam parameters linked to

$\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 2

(Palomo et al., 12 May 2026). This suggests that depth calibration and optical modeling are part of the measurement, not merely preprocessing.

Interpretive limits depend on sensor class. In irradiated strip sensors, finite electronics rise time implies that the first and last tens of micrometers in depth are less reliable for field extraction (Klanner et al., 2019). In LGADs, the authors of the diffusion-profile method explicitly argue that the velocity-profile method produces a spurious peak and cannot map $\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 3 because avalanche build-up dominates the rising edge (Fu et al., 2023). In TPA timing studies, absolute ToA values carry an arbitrary offset set by cable and trigger delays, and two-photon excitation does not perfectly replicate minimum-ionizing-particle charge deposition (Lasaosa et al., 19 May 2025).

Equipment complexity is a practical boundary. TPA-TCT requires a femtosecond infrared laser system and high-NA optics, and signal interpretation demands careful volume and refraction corrections (Palomo et al., 12 May 2026). Reported studies therefore combine TCT with C–V, $\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 4, TCAD, or source timing when validation or absolute calibration is needed (Velkovska et al., 11 Sep 2025, Pape et al., 2024).

7. Terminological ambiguity and uses beyond detector physics

Although TCT overwhelmingly denotes the laser-based detector-characterization method in semiconductor-device research, the acronym is not exclusive to that domain. In electrically conducting fluids, “Transient Current Technique” refers to a calibration-free, contactless flow-measurement method in which a pulsed magnetic field induces a transient ring-shaped eddy-current system that is advected by the flow. Pickup coils track the motion of the corresponding magnetic-field pattern, and the local velocity is inferred from the time dependence of the pole position $\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 5 (Forbriger et al., 2015). The governing equation is a diffusion–advection equation,

$\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 6

rather than a carrier-drift problem in a semiconductor (Forbriger et al., 2015).

In molecular electronics, the phrase “Transient Current Technique” appears in a different theoretical sense. The NEGF-DFT-CAP formalism for molecular devices calculates the transient current under a step-like pulse from first principles by combining non-equilibrium Green’s functions, density functional theory, and a complex absorbing potential. The reported scaling is $\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 7 in the number of time steps, compared with at least $\rho(y)=\epsilon_{\mathrm{Si}}\,dE/dy$ 8 for earlier first-principles time-dependent calculations (Zhang et al., 2013). Here the core object is the time-dependent current in a quantum transport calculation, not a laser-induced drift transient in a bulk detector.

The coexistence of these usages matters for literature search and citation practice. A precise reading of “TCT” therefore requires the surrounding disciplinary markers: detector physics implies laser-induced transient-current spectroscopy; liquid-metal metrology implies eddy-current flow tracking; molecular electronics implies time-dependent quantum transport. In current arXiv usage relevant to silicon and wide-bandgap detectors, however, TCT most commonly denotes the family of top-, edge-, backside-, and two-photon transient-current methods used to reconstruct internal electric fields, depletion structure, carrier transport, gain, timing, and charge-sharing behavior in semiconductor sensors (Wüthrich et al., 2022, Palomo et al., 12 May 2026).