Category: Doing a thing

  • “Astatine – just say no!” The Thermal Fury of Astatine: Calculating the Heat Output of an element you’ve never heard of.

    This is a follow-on from the very awesome video by Randall Munroe on making a wall made of 1L samples (10x10x10cm) of every element in the periodic table.

    The question is ‘what exactly is the power output of astatine?’


    A solid $7\text{ kg}$ sample of Astatine-210 (the most stable isotope) would generate a thermal output of approximately $233\text{ MW}$ $(233,000,000\text{ watts})$ at the moment of its creation. If the sample consisted of Astatine-211—the isotope commonly created in particle accelerators—the thermal output would climb even higher, reaching roughly $495\text{ MW}$ $^{[1, 2]}$.

    To put this scale into perspective, a 1-litre block of astatine would produce as much heat as a small commercial nuclear reactor, entirely concentrated within a handful of matter.

    But don’t take my word for it; LET’S USE FIRST PRINCIPLES!!!.

    1. Count the Atoms

    To find out how much heat is released, we first determine how many atoms are present in a $7\text{ kg}$ ($7,000\text{ g}$) sample using Avogadro’s number:

    $$N = \frac{\text{Mass}}{\text{Molar Mass}} \times N_A$$

    Using Astatine-210 ($210\text{ g/mol}$) $^{[2]}$:

    $$N = \frac{7,000\text{ g}}{210\text{ g/mol}} \times 6.022 \times 10^{23}\text{ atoms/mol}$$

    $$\approx 2.0073 \times 10^{25}\text{ atoms}$$

    2. Determine the Decay Constant

    Next, we calculate the decay constant ($\lambda$) using the half-life of Astatine-210, which stands at $8.1\text{ hours}$ ($29,160\text{ seconds}$) $^{[3]}$:

    $$\lambda = \frac{\ln(2)}{t_{1/2}} = \frac{0.69315}{29,160\text{ s}}$$

    $$\approx 2.377 \times 10^{-5}\text{ s}^{-1}$$

    3. Calculate Total Radioactivity

    The total radioactivity, or activity ($A$), measured in Becquerels ($\text{Bq}$, decays per second) is the product of the number of atoms and the decay constant:

    $$A = \lambda \times N$$

    $$A = (2.377 \times 10^{-5}\text{ s}^{-1}) \times (2.0073 \times 10^{25})$$

    $$\approx 4.771 \times 10^{20}\text{ Bq}$$

    4. Convert Energy to Watts

    Every individual decay of Astatine-210 releases an average combined radiation energy—comprising alphas, electrons, and photons—of approximately $3.0515\text{ MeV}$. We convert this particle energy into Joules ($1\text{ MeV} = 1.6022 \times 10^{-13}\text{ J}$) $^{[3, 4]}$:

    $$\text{Energy per decay} = 3.0515\text{ MeV} \times 1.6022 \times 10^{-13}\text{ J/MeV}$$

    $$\approx 4.889 \times 10^{-13}\text{ J}$$

    Multiplying the total decays per second by the energy per decay yields the absolute power output:

    $$P = A \times \text{Energy per decay}$$

    $$P = (4.771 \times 10^{20}\text{ Bq}) \times (4.889 \times 10^{-13}\text{ J})$$

    $$\approx 2.333 \times 10^8\text{ W}$$

    5. Final Thermal Output Result

    The hypothetical initial thermal output of a solid $7\text{ kg}$ sample of Astatine-210 sits at $2.33 \times 10^8\text{ watts}$ ($233\text{ MW}$).

    For those who prefer programmatic verification, here is the Python script used to model these calculations based on specific activity and mean energy.

    # Calculate the thermal power of Astatine-210 based on # specific activity and mean energy
    #
    # Specific activity # = 6.817e16 Bq/g
    # Mass = 7000 g
    # Mean energy per decay = 0.0097 + 0.07962 + 2.96215 = # 3.05147 MeV
    
    spec_activity = 6.817e16  # Bq/g
    mass = 7000  # g
    total_activity = spec_activity * mass  # Bq
    
    mean_energy_mev = 3.05147
    energy_j = mean_energy_mev * 1.60218e-13
    
    power_w = total_activity * energy_j
    print(f"At-210 Power: {power_w:.4e} Watts")

    References

    1. Lindegren, S., Albertsson, P., Bäck, T., Jensen, H., Palm, S., & Aneheim, E. (2020). Realizing clinical trials with Astatine-211: The chemistry infrastructure. Cancer Biotherapy and Radiopharmaceuticals, 35(6), 425–436. https://doi.org/10.1089/cbr.2019.3055 Cited by: 114
    2. ChemLin. (2024). Astatine isotopes – list and properties. ChemLin Chemical Elements. https://www.chemlin.org/chemical-elements/astatine-isotopes.php
    3. MIRDSoft. (n.d.). Astatine-210 radionuclide dosimetric data sheet. Medical Internal Radiation Dose (MIRD) Specification Sheets. https://mirdsoft.org/products/MIRDspecs/MIRDspecs_pdfs/At-210.pdf
    4. BenchChem. (2026). An in-depth technical guide to the radiotoxicity and decay products of Astatine-210. BenchChem Technical Guides. https://www.benchchem.com/pdf/An_In_Depth_Technical_Guide_to_the_Radiotoxicity_and_Decay_Products_of_Astatine_210.pdf
  • Sample quotas all up in this place

    Sample quotas all up in this place

    Catching the Risk: The Maths Behind Sample Quotas in SEGs

    In occupational hygiene, we often talk about Similar Exposure Groups (SEGs) and the need to collect enough samples to be confident in our data. But how do we actually determine if our sample size is sufficient to capture those at the highest risk?

    Standard practice is to mornally just quote Technical Appendix A in NIOSH 173/77 and go from there – but this time, that just wasn’t going to be enough.

    During a recent mentoring session, we broke down the statistics of sampling quotas. Here is how the maths of permutations and combinations helps us ensure no worker is left behind.

    The Foundation: Factorials and Choices

    Before we can calculate the probability of a successful sampling campaign, we need to understand how many ways we can “choose” workers from a group.

    Factorials

    The factorial ($n!$) is the product of all positive integers less than or equal to $n$. This value represents the total number of ways you can uniquely arrange a set of items. For example, if you have 6 items and want to know every possible sequence they could be ordered in:

    $$6! = 6 \times 5 \times 4 \times 3 \times 2 \times 1 = \mathbf{720 \text{ ways to arrange 6 items}}$$

    Permutations vs. Combinations

    Understanding the difference between these two is critical for sampling.

    Permutations

    How many ways can I choose $r$ things from $n$ things where order is important? To use our example of 6 items, let’s say I want to select 2. I can do that by:

    $$6 \times 5 = 30$$

    To get that, using only factorials, we can use some numerator/denominator cancellation. It looks like this, noting how $4 \times 3 \times 2 \times 1$ appears in both the top and bottom and so cancel out:

    $$^nP_r = \frac{6!}{(6-2)!} = \frac{6!}{4!} = \frac{6 \times 5 \times 4 \times 3 \times 2 \times 1}{4 \times 3 \times 2 \times 1} = 6 \times 5 = 30$$

    MAGIC!

    The general formula is:

    $$^nP_r = \frac{n!}{(n-r)!}$$

    Combinations

    How many ways can I choose $r$ things from $n$ things where order is NOT important? In our example of 30 permutations of 2 items, we divide by the 2 ways those items can be arranged ($2 \times 1 = 2$). So let’s divide our permutation formula by the factorial of $r$:

    $$\frac{6!}{2! \times (6-2!)} = \frac{6 \times 5 \times 4 \times 3 \times 2 \times 1}{(2 \times 1)(4 \times 3 \times 2 \times 1)} = \frac{6 \times 5}{2} = 15$$

    $$^n\mathcal{C}_r = \frac{n!}{r! \times (n-r)!}$$

    Practical Example: Choosing 3 Workers from 10

    To see why this distinction matters, imagine you have a group of 10 workers ($n=10$) and you want to select 3 ($r=3$) for monitoring.

    The Permutation Approach (Order Matters)

    If we cared about the order (e.g., who gets the pump first, second, and third), the maths looks like this:

    $$^{10}P_3 = \frac{10!}{(10-3)!} = \frac{10!}{7!} = 10 \times 9 \times 8 = \mathbf{720 \text{ ways}}$$

    The Combination Approach (Order Does Not Matter)

    In a typical hygiene survey, we don’t care about the sequence—only the final set of people monitored. We divide the permutations by the ways those 3 people can be rearranged ($3!$):

    $$^{10}\mathcal{C}_3 = \frac{720}{3 \times 2 \times 1} = \frac{720}{6} = \mathbf{120 \text{ ways}}$$

    Why this matters: There are far fewer unique “groups” than “orders.” We use combinations because we are looking for the probability of a specific subset of the population being captured but what order we select them is not important.

    Defining the Sampling Problem

    When sampling a population ($N$), we are often looking for a specific “Target” group ($N_0$), such as a Highly Exposed Risk Group (e.g., the top 10% of exposed workers).

    The critical question is: What is the probability that I get ZERO samples from my high-risk group?

    To find this, we look at the three components of the probability calculation:

    Combinations of $n$ samples from the non-target group $(N – N_0)$.

    Combinations of 0 samples from the target group $(N_0)$.

    Total possible combinations of $n$ samples from the entire population ($N$).

    The maths behind this derivation follows a logical flow:

    $$P(r=0) = \frac{^{(N-N_0)}\mathcal{C}_n \times ^{N_0}\mathcal{C}_0}{^N\mathcal{C}_n}$$

    Since any combination of 0 items ($^{N_0}\mathcal{C}_0$) is equal to 1 (there is only 1 way to select nothing; it makes maths-sense, go with it), we expand the factorials as follows:

    $$P(r=0) = \frac{(N-N_0)!}{n!(N-N_0-n)!} \times \frac{N_0!}{0!(N_0-0)!} \times \left[ \frac{N!}{n!(N-n)!} \right]^{-1}$$

    By simplifying the expressions (flipping the $^N\mathcal{C}_n$ with a $-1$ power) and cancelling out terms (specifically $n!$), we arrive at the final simplified formula:

    $$P(r=0) = \frac{(N-N_0)!}{(N-N_0-n)!} \times \frac{(N-n)!}{N!}$$

    Case Study: Sampling a Population of 18

    Let’s apply this to a group of 18 workers ($N=18$). We want to ensure we haven’t missed the top 10% of exposed workers ($18 \times 0.1 = 1.8$, which we call 2 workers, so $N_0 = 2$).

    If we collect different sample sizes ($n$), how does your risk of “missing” ($r=0$) that high-risk group change?

    Scenario A: Collect 9 Samples ($n=9$)

    Using our formula $\frac{(N-N_0)!}{(N-N_0-n)!} \times \frac{(N-n)!}{N!}$:

    $$P(r=0) = \frac{(18-2)!}{(18-2-9)!} \times \frac{(18-9)!}{18!} = \frac{16!}{7!} \times \frac{9!}{18!} = \mathbf{0.235}$$

    With 9 samples, there is a 23.5% chance you will not capture a single person from that top 10% high-risk group.

    Scenario B: Collect 12 Samples ($n=12$)

    $$P(r=0) = \frac{(18-2)!}{(18-2-12)!} \times \frac{(18-12)!}{18!} = \frac{16!}{4!} \times \frac{6!}{18!} = \mathbf{0.098}$$

    By increasing to 12 samples, your risk of missing the high-risk group drops to roughly 9.8%.

    Scenario C: Collect 15 Samples ($n=15$)

    $$P(r=0) = \frac{(18-2)!}{(18-2-15)!} \times \frac{(18-15)!}{18!} = \frac{16!}{1!} \times \frac{3!}{18!} = \mathbf{0.0196}$$

    At 15 samples, the probability of missing the target group is less than 2%, providing a very high level of certainty.

    The “Flip It” Rule

    In hygiene, we usually want to know the confidence level—the probability of getting at least one sample from the target group ($r \ge 1$). To find this, simply “flip” the result:

    $$\text{Confidence} = 1 – P(r=0)$$

    With 12 samples, you have a 90.2% confidence level

    i.e. $(1-0.098) \times 100$.

    With 15 samples, you have a 98.04% confidence level 

    i.e. $(1-0.0196) \times 100$.

    By using these calculations, you can move away from “guessing” your sample quotas and start providing a statistical justification for your monitoring programs.

    BONUS: And I made a calculator for you!!

    Sample Quota calculator

  • ARPAB @ ARPS2025

    I’m doing a thing again.