Distinguishing Biexciton vs. Exciton Emission via Power-Dependent PL

1. The Kinetic Model: Why Scaling Laws Differ

The distinct power dependence of exciton and biexciton luminescence arises from the kinetics of their formation. We can understand this by constructing a simple rate equation model for the populations of excitons ($N_X$) and biexcitons ($N_{XX}$). Given your background in mathematical physics, you'll see this is analogous to reaction-rate problems in chemical kinetics or population dynamics.

Let's assume we are exciting the semiconductor with a continuous-wave (CW) laser, and the system is in a steady state.

• Exciton Population ($N_X$):

  • Excitons are created at a rate $G$, which is directly proportional to the excitation laser power, $P$. So, $G\propto P$.
  • They decay (radiatively or non-radiatively) with a characteristic lifetime ${\tau }_X$, at a rate of $N_X/{\tau }_X$.
  • Two excitons can bind to form a biexciton. This is a two-particle collision process, so its rate is proportional to the square of the exciton density, written as $C{N}_X^2$, where $C$ is a formation coefficient.

• Biexciton Population ($N_{XX}$):

  • Biexcitons are formed from the binding of two excitons, at the rate $C{N}_X^2$.
  • They decay with a lifetime ${\tau }_{XX}$, at a rate of $N_{XX}/{\tau }_{XX}$.

The steady-state rate equations ($dN/dt=0$) are:

1. Excitons: $\frac{d{N}_X}{dt}=G-\frac{N_X}{{\tau }_X}-2C{N}_X^2+\frac{N_{XX}}{{\tau }_{XX}}=0$
   (The factor of 2 on $C{N}_X^2$ is because two excitons are consumed; the final term is because biexciton decay produces an exciton).
2. Biexcitons: $\frac{d{N}_{XX}}{dt}=C{N}_X^2-\frac{N_{XX}}{{\tau }_{XX}}=0$

From the second equation, we immediately get a crucial relationship:
$N_{XX}=C{\tau }_{XX}{N}_X^2$
This tells us that the steady-state biexciton population is proportional to the square of the exciton population.

Now, let's consider the low-excitation regime, where biexciton formation is relatively rare. In this limit, the terms involving $N_{XX}$ and $N_X^2$ in the exciton equation are small compared to the generation and linear decay terms. The exciton equation simplifies to:
$G-\frac{N_X}{{\tau }_X}\approx 0⟹N_X\approx G{\tau }_X$
Since the generation rate $G$ is proportional to the laser power $P$, we have:
$N_X\propto P$
Combining these results, we find the scaling of the biexciton population with power:
$N_{XX}\propto N_X^2\propto P^2$

Finally, the photoluminescence intensity ($I$) for each species is proportional to its population divided by its radiative lifetime. Assuming the lifetimes don't change with power in this regime:

• Exciton PL Intensity: $I_X\propto N_X⟹I_X\propto P^1$ (Linear dependence)
• Biexciton PL Intensity: $I_{XX}\propto N_{XX}⟹I_{XX}\propto P^2$ (Quadratic dependence)

This difference in the power-law dependence is the key to experimentally distinguishing the two.

2. Experimental Analysis of Power-Dependent Spectra

The experimental procedure to verify these scaling laws is straightforward:

1. Record a series of PL spectra at different, well-controlled excitation powers ($P$).
2. For each spectrum, identify the peaks corresponding to the exciton and the candidate biexciton.
3. Calculate the integrated intensity of each peak (i.e., the area under the peak). This is more robust than just using the peak height.
4. Plot the integrated intensity ($I$) versus the excitation power ($P$) on a log-log scale.

Plotting on a log-log scale is a standard method for analyzing power-law relationships. If $I=A\cdot P^k$, then taking the logarithm gives:
$\log (I)=\log (A)+k\cdot \log (P)$
This is the equation of a straight line, $y=c+mx$, where the slope $m$ is the exponent $k$. An exciton peak should yield a slope of $\approx 1$, while a biexciton peak should yield a slope of $\approx 2$.

Let's see this in action with some experimental data.

Case Study 1: Biexcitons in Perovskite Nanocrystals

Perovskites, one of your areas of interest, are an excellent platform for studying excitonic effects due to their large exciton binding energies. The following paper provides a clear demonstration of using power-dependent PL to identify biexcitons.

As you saw in the paper, the authors identified two peaks.

• The peak at 500 nm is assigned to the single exciton (X).
• A new peak emerges at 520 nm at higher power. The energy difference (95 meV) corresponds to the biexciton binding energy.
• Crucially, the log-log plot in Figure 3c shows that the intensity of the 500 nm peak scales with power with an exponent of k = 0.96 (approximately linear).
• The intensity of the 520 nm peak scales with an exponent of k = 1.83 (approximately quadratic).

This provides compelling evidence that the 520 nm peak is indeed from biexciton recombination. The slight deviation from the ideal integer values of 1 and 2 is common in real experiments due to factors like non-radiative recombination channels becoming more significant at higher powers.

The following image provides a very clear, textbook illustration of this principle in a different material system (monolayer WS₂).

Case Study 2: Disentangling Multiple Excitonic Complexes

In high-quality samples, the PL spectrum can be crowded with peaks from various excitonic species. Power-dependent spectroscopy is an essential tool for parsing this "exciton zoo."

This study in WSe₂ reinforces our understanding. The authors assign the peak labeled XX to the biexciton because its intensity scales with an exponent $\alpha =1.94$, very close to the ideal value of 2. They use this, along with the binding energy and gate-voltage dependence (a complementary technique that probes the charge of the complex), to build a complete picture.

3. Other Corroborating Evidence

While power dependence is a primary tool, a convincing identification of a biexciton relies on a collection of consistent evidence. Two other important checks are:

1. Binding Energy: As we've discussed, the biexciton peak must be red-shifted from the exciton peak by an amount ${\Delta }_{XX}$ that is consistent with theoretical expectations or other measurements of the biexciton binding energy.

2. Luminescence Lifetime: The biexciton state has two excitons, either of which can decay. This provides an additional decay channel compared to a single exciton. As a result, the biexciton decay rate is expected to be roughly twice the single exciton decay rate. This implies its lifetime should be approximately half: ${\tau }_{XX}\approx {\tau }_X/2$. This can be verified using time-resolved photoluminescence (TRPL). Indeed, in the first paper on perovskite nanocrystals (LINK), the authors measure ${\tau }_X=1170$ ps and ${\tau }_{XX}=550$ ps (see their Figure 4), providing strong corroborating evidence for their biexciton assignment.

Conclusion

Today we've established the primary experimental method for distinguishing biexciton emission from that of single excitons. By understanding the underlying kinetics of exciton-exciton interaction, we can predict and experimentally verify their distinct scaling laws with excitation power.

Key Takeaways:

• The formation of a biexciton is a two-particle process, requiring two excitons. This kinetic constraint is the origin of its unique spectroscopic signature.
• In the low-density limit, exciton PL intensity scales linearly with excitation power ($I_X\propto P^1$), while biexciton PL intensity scales quadratically ($I_{XX}\propto P^2$).
• The most effective way to analyze this is by plotting the integrated peak intensities versus power on a log-log scale and extracting the slope, which corresponds to the power-law exponent.
• A definitive identification of a biexciton is strengthened by corroborating evidence, such as a sensible binding energy (spectral red-shift) and a lifetime that is approximately half that of the single exciton.

Preview of the Next Lesson

We have now studied single excitons and two-exciton complexes (biexcitons). This naturally leads to the question: what happens when we have a dense gas of many excitons? Can they behave collectively? In the next lesson, we will begin exploring this fascinating many-body regime by addressing the learning outcome: "Establish the conditions under which excitons can be treated as composite bosons," a crucial first step toward understanding phenomena like Bose-Einstein condensation.