Charge-Carrier Mobility in Diamond: Review, Data Compilation, and Modelling for Detector Simulations

This paper resolves the wide dispersion in reported charge-carrier mobility data for diamond by attributing it to experimental and modeling variables, then benchmarks and recommends specific parameterizations for electrons and holes to improve the accuracy of detector simulations.

The Gist

The Big Picture: Why Diamond is the "Super-Runner" of Sensors

Imagine you are building a race track for tiny particles (electrons and holes) to run on. You want this track to be incredibly tough, able to handle extreme heat, radiation, and heavy traffic without breaking. Diamond is the ultimate material for this job. It's not just a gemstone for jewelry; in the world of physics, it's a "super-athlete" used to detect radiation in nuclear plants, space telescopes, and particle colliders.

However, there's a problem. For decades, scientists have been arguing about how fast these particles can actually run on a diamond track. Some say they run at 20 mph, others say 40 mph. Some say they speed up quickly, others say they hit a wall and stop. This confusion makes it hard to build accurate sensors.

This paper is like a detective story where the authors gather all the conflicting reports, figure out why everyone is disagreeing, and finally agree on the "true" speed limits.

---

The Mystery: Why the Confusion?

The authors found that the "speed" of the particles depends on three main things, which they call the "Three Culprits":

1. The Measuring Stick (Electric Field): Imagine trying to measure a runner's speed. If you only watch them for the first 10 meters, they look fast because they are accelerating. If you watch them for a mile, they might slow down or hit a top speed. Different scientists measured the particles over different distances (electric fields), leading to different results.
2. The Math Formula (The Model): Scientists use different math equations to calculate speed from their data. It's like using a ruler vs. a tape measure. Some formulas are better for short distances, others for long ones. The paper tested three different "rulers" (models) to see which one fits best.
3. The Starting Gun (Excitation Source): How do you start the race?
   • Alpha particles are like a heavy, dense crowd of runners starting all at once.
   • Lasers are like a single, precise spark starting a few runners.
   • The paper found that the "crowd" (Alpha) often makes the runners look faster than the "spark" (Laser) because of how they interact with each other.

---

The Solution: The "Piecewise" and "Caughey-Thomas" Models

The authors tested three main ways to describe the runners' speed:

1. The Old School (Trofimenkoff/TK): A simple formula that worked okay for short races but got messy over long distances.
2. The Silicon Standard (Caughey-Thomas/CT): A formula famous in silicon chips. It worked great for the "holes" (positive runners) in diamond.
3. The New Hybrid (Piecewise/PW): This is the authors' new invention. It's like a smart GPS.
   • For short distances (low energy): It acts like a simple, straight-line runner.
   • For long distances (high energy): It switches to a more complex curve that accounts for the runners getting tired or changing lanes.

The Verdict:

• For Electrons (Negative runners): The new Piecewise (PW) model is the winner. It explains the weird "speed bumps" and lane changes that happen when electrons get hot.
• For Holes (Positive runners): The Caughey-Thomas (CT) model is the winner. Holes are more predictable; they don't do the weird lane changes, so the standard formula works perfectly.

---

The "Valley" Analogy: Why Electrons are Weird

To explain why electrons are so tricky, the paper uses a concept called Intervalley Repopulation.

Imagine the diamond crystal is a landscape with six valleys.

• The "Cool" Valleys: These are deep, wide, and heavy. Running here is slow but stable.
• The "Hot" Valleys: These are shallow, narrow, and light. Running here is fast but unstable.

What happens?
When you apply a gentle push (low electric field), the electrons hang out in the "Hot" valleys because they are easy to get to. They run fast!
But as you push harder (high electric field), the electrons get scared and jump back into the "Cool" valleys to hide. Suddenly, they slow down.

This "jumping" creates a weird dip in the speed graph that old math models couldn't explain. The new Piecewise model accounts for this "jumping" behavior, making it the most accurate description for electrons.

---

The Takeaway: What Does This Mean for the Future?

1. Stop Arguing: The authors have created a "Universal Speed Limit" for diamond sensors. They standardized all the old data so that a measurement from 1980 can be compared fairly with one from 2024.
2. Better Simulations: Engineers who design radiation detectors (for hospitals or space) can now use these new, accurate formulas. This means their computer simulations will be much closer to reality, leading to better, safer, and more efficient devices.
3. The "Alpha" Standard: They decided to use the "Alpha particle" measurements as the gold standard reference. If a new sensor is tested with a laser, they now know exactly how to mathematically convert that result to match the Alpha standard.

In a Nutshell

This paper is the rulebook update for diamond sensors. It admits that previous measurements were messy because everyone was using different rules and measuring different things. By introducing a smarter math model (Piecewise for electrons, Caughey-Thomas for holes) and a standard way to compare data, they have cleared the fog. Now, engineers can build diamond detectors with confidence, knowing exactly how fast the particles will run.

Technical Summary

1. Problem Statement

Despite diamond's superior properties for high-radiation and high-temperature applications (e.g., in High-Energy Physics detectors and power electronics), there is a significant and unexplained dispersion in reported values for charge carrier mobility ($\mu$) and saturation velocity ($v_{sat}$) in the literature.

• Reported Ranges: Electron low-field mobility ($\mu_{0,e}$) varies from 1,300 to 4,500 cm² V⁻¹ s⁻¹, and hole mobility ($\mu_{0,h}$) from 2,000 to 3,800 cm² V⁻¹ s⁻¹. Saturation velocities show similar wide spreads.
• Root Causes: The authors attribute this inconsistency to three primary factors:
  1. Experimental Windows: The specific electric-field ranges probed in Transient Current Technique (TCT) measurements vary widely.
  2. Model Selection: Different semi-empirical models (e.g., Trofimenkoff vs. Caughey-Thomas) are used to fit the same data, leading to different extracted parameters.
  3. Excitation Sources: Systematic differences exist between measurements using $\alpha$-particles, lasers, or electron beams, which affect carrier generation profiles and space-charge effects.

2. Methodology

The authors employed a comprehensive data-driven approach to resolve these inconsistencies:

• Data Aggregation: They digitized drift velocity ($v_d$) vs. electric field ($E$) data from 17 published papers (covering 14+ experiments) using PlotDigitizer, creating a unified dataset (available in the TUDOdata repository).
• Data Grouping:
  • Group 1 (Experiment-wise): Analyzed individual experiments to assess model performance per dataset.
  • Group 2 (Pooled Data): Aggregated data into Electron and Hole subgroups, further categorized by excitation source ($\alpha$, laser, electron) and electric field range (Full range: 0.009–12 V/$\mu$m; Detector Operational (DO) range: 0.1–4.0 V/$\mu$m).
• Source Normalization: To compare datasets directly, the authors calculated multiplicative scaling factors to normalize laser-induced data to the $\alpha$-particle baseline (e.g., $x_{LA} \approx 1.09$ for electrons, $1.22$ for holes), accounting for differences in ionization density and space-charge effects.
• Model Comparison: Three models were benchmarked using statistical metrics ($\chi^2$/ndf, AIC, BIC, $R^2$, and standardized residuals):
  1. Trofimenkoff (TK): Widely used in diamond literature.
  2. Caughey-Thomas (CT): Standard for silicon, includes a shaping parameter $\beta$.
  3. Piecewise (PW): A new model proposed by the authors, combining a linear Drude regime at low fields with a CT-like non-linear regime at higher fields.
• Temperature Analysis: The study analyzed temperature dependence, specifically investigating the "Repopulation Effect" in electrons (intervalley scattering) and proposing a superposition model for low-temperature transport.

3. Key Contributions

A. Identification of Systematic Source Effects

The study quantified a systematic offset between excitation sources. $\alpha$-TCT measurements consistently yield higher drift velocities than laser-TCT.

• Mechanism: $\alpha$-particles create a dense, localized electron-hole cloud that undergoes ambipolar diffusion, preferentially depleting slower carriers and leaving faster ones to dominate the transient signal. Laser excitation creates a surface-near distribution prone to surface scattering and exciton formation.
• Quantification: A global scaling factor was derived to normalize laser data to the $\alpha$-scale, enabling direct comparison of material quality across different experimental setups.

B. Development and Validation of the Piecewise (PW) Model

For electrons, the authors introduced a Piecewise (PW) model that outperforms traditional models over the full electric field range.

• Structure: It uses a linear Drude model ($v_d = \mu_0 E$) below a threshold field ($E_t \approx 0.08$ V/$\mu$m) and transitions to a Caughey-Thomas form above $E_t$.
• Physical Basis: The PW model is shown to be the room-temperature limit of a more general Superposition Model. This superposition model explicitly accounts for the Repopulation Effect, where electrons redistribute between "cool" (longitudinal mass) and "hot" (transverse mass) valleys in the conduction band.
• Performance: The PW model accurately captures the "hill-type" non-linearity and saturation of electron drift, whereas TK and CT models systematically overestimate drift velocity at high fields.

C. Resolution of Hole Mobility Discrepancies

For holes, the study found that the Caughey-Thomas (CT) model is statistically preferred.

• Unlike electrons, holes do not exhibit a significant repopulation effect in the accessible data.
• The CT model (which reduces to TK when $\beta \approx 1$) provides a robust description of hole transport across all field ranges, with performance comparable to the PW model in the detector operational range.

D. Temperature Scaling and Regimes

The paper details three distinct transport regimes for electrons based on temperature:

1. $T \lesssim 120$ K: Two-step falling edges in TCT signals indicating distinct "cool" and "hot" electron populations; negative differential mobility is observed.
2. $120 \text{ K} \lesssim T \lesssim 220$ K: A pseudo-saturation plateau appears due to intervalley scattering competition.
3. $T \gtrsim 220$ K: Monotonic behavior where the repopulation effect is suppressed by phonon scattering, allowing standard models (PW/CT) to apply.

4. Key Results

• Model Selection Recommendations:
  • Electrons: Use the Piecewise (PW) model for the broadest accuracy, especially at high fields. In the standard detector operating range (0.1–4.0 V/$\mu$m), the PW model effectively collapses to the CT model.
  • Holes: Use the Caughey-Thomas (CT) model. It offers the best balance of simplicity and accuracy.
• Parameter Sets: The authors provide optimized parameter sets (saturation velocity $v_s$, critical field $E_c$, low-field mobility $\mu_0$, and shaping parameter $\beta$) for implementation in device simulation frameworks (e.g., Allpix²).
  • Example (Electrons, PW, AllLA scaled): $v_s \approx 26.6 \times 10^6$ cm/s, $\mu_0 \approx 1880$ cm²/Vs.
  • Example (Holes, CT, AllLA scaled): $v_s \approx 15.1 \times 10^6$ cm/s, $\mu_0 \approx 2744$ cm²/Vs.
• Source Normalization: The study establishes that laser data must be scaled by $\approx 1.09$ (electrons) and $\approx 1.22$ (holes) to match $\alpha$-particle baselines to avoid underestimating mobility in simulations.
• Material Quality Indicator: The "upper envelope" of $\alpha$-TCT data represents near-ideal material. Samples falling below this envelope (or below the laser-scaled curve) indicate the presence of defects, impurities, or polarization effects.

5. Significance and Impact

• Reconciliation of Literature: The paper resolves decades of conflicting mobility values by demonstrating that the spread is largely methodological (source and model choice) rather than intrinsic to the material.
• Standardization for Simulation: By providing standardized, statistically validated parameter sets and scaling factors, the work enables more accurate TCAD simulations of diamond detectors and power devices.
• Experimental Guidance: The authors recommend that future TCT campaigns cover a wider electric field range (up to 10 V/$\mu$m) and temperature range (up to 1000 K) with consistent excitation sources to further refine temperature scaling laws.
• Physical Insight: The explicit link between the Piecewise model and the underlying intervalley repopulation physics offers a deeper theoretical understanding of electron transport in diamond, particularly for low-temperature applications in magnetic fields.

In conclusion, this work provides a definitive framework for modeling charge transport in diamond, moving the field from fragmented, experiment-specific parameters to a unified, physics-based standard suitable for next-generation detector and power electronics design.