A quick update on this blog article with some extra cross sections and an additional plot of Maxwell-averaged reactivities. The processes covered are:

\begin{align*}
\mathrm{D} + \mathrm{T} & \rightarrow \alpha (3.5\;\mathrm{MeV}) + n (14.1\;\mathrm{MeV}) & \\
\mathrm{D} + \mathrm{D} & \rightarrow \mathrm{T} (1.01\;\mathrm{MeV}) + p(3.02\;\mathrm{MeV})& 50\%\\
& \rightarrow \mathrm{^3He} (0.82\;\mathrm{MeV}) + n(2.45\;\mathrm{MeV}) & 50\%\\
\mathrm{D} + \mathrm{^3He} & \rightarrow \alpha (3.6\;\mathrm{MeV}) + p(14.7\;\mathrm{MeV}) & \\
\mathrm{^3He} + \mathrm{^3He} & \rightarrow 2p + \alpha + 12.86\;\mathrm{MeV} & \\
\mathrm{T} + \mathrm{^3He} & \rightarrow n + p + \alpha + 12.1\;\mathrm{MeV} & 59\% \\
& \rightarrow D (9.52\;\mathrm{MeV}) + \alpha(4.8\;\mathrm{MeV}) & 41\% \\
\mathrm{T} + \mathrm{T} & \rightarrow 2n + \alpha + 11.3\;\mathrm{MeV} & \\
p + \mathrm{^{11}B} & \rightarrow 3\alpha + 8.7\;\mathrm{MeV} &
\end{align*}

The cross section gives a probability for the reaction to occur as a function of the collision energy. In practice, the colliding particles move with a distribution of relative velocities, $f(v)$ (and hence collision energies); a useful quantity is the averaged *reactivity*:
$$
\langle \sigma v \rangle = \int_0^\infty \sigma(v)vf(v)\,\mathrm{d}v.
$$
For the Maxwell distribution of the relative speeds of the colliding particles,
$$
f_\mathrm{MB}(v) = 4\pi v^2 \left(\frac{\mu}{2\pi k_\mathrm{B}T}\right)^{3/2}\exp\left(-\frac{\mu v_i^2}{2k_\mathrm{B}T}\right),
$$
it can be shown [1, 2 and see below] that
$$
\langle \sigma v \rangle = \frac{4}{\sqrt{2\pi\mu}}\frac{1}{(k_\mathrm{B}T)^{3/2}} \int_0^\infty \sigma(E)E\exp\left(-\frac{E}{k_\mathrm{B}T}\right)\,\mathrm{d}E,
$$
where $\mu = m_1m_2/(m_1+m_2)$ is the reduced mass of the colliding particles.

The code below requires the following data files:

- D_T_-_a_n.txt
- D_D_-_T_p.txt
- D_D_-_3He_n.txt
- D_3He_-_4He_p-endf.txt
- 3He_3He_-_p_p_4He.txt
- T_3He_-_n_p_4He.txt
- T_3He_-_D_4He.txt
- T_T_-_4He_n_n.txt
- p_11B_-_3a.txt

The code:

```
import numpy as np
from matplotlib import rc
import matplotlib.pyplot as plt
from scipy.constants import e, k as kB
from scipy.integrate import quad
rc('font', **{'family': 'serif', 'serif': ['Computer Modern'], 'size': 18})
rc('text', usetex=True)
# To plot using centre-of-mass energies instead of lab-fixed energies, set True
COFM = True
# Reactant masses in atomic mass units (u).
u = 1.66053906660e-27
masses = {'D': 2.014, 'T': 3.016, '3He': 3.016, '11B': 11.009305167,
'p': 1.007276466620409}
# Energy grid, 1 – 100,000 keV, evenly spaced in log-space.
Egrid = np.logspace(0, 5, 1000)
class Xsec:
def __init__(self, m1, m2, xs):
self.m1, self.m2, self.xs = m1, m2, xs
self.mr = self.m1 * self.m2 / (self.m1 + self.m2)
@classmethod
def read_xsec(cls, filename, CM=True):
"""
Read in cross section from filename and interpolate to energy grid.
"""
# Read in the cross section (in barn) as a function of energy (MeV).
E, xs = np.genfromtxt(filename, comments='#', skip_footer=2,
unpack=True)
if CM:
collider, target = filename.split('_')[:2]
m1, m2 = masses[target], masses[collider]
E *= m1 / (m1 + m2)
# Interpolate the cross section onto the Egrid (keV) and
# convert from barn to cm2.
xs = np.interp(Egrid, E*1.e3, xs*1.e-28)
return cls(m1, m2, xs)
def __add__(self, other):
return Xsec(self.m1, self.m2, self.xs + other.xs)
def __mul__(self, n):
return Xsec(self.m1, self.m2, n * self.xs)
__rmul__ = __mul__
def __getitem__(self, i):
return self.xs[i]
def __len__(self):
return len(self.xs)
xs_names = {'D-T': 'D_T_-_a_n.txt', # D + T -> α + n
'D-D_a': 'D_D_-_T_p.txt', # D + D -> T + p
'D-D_b': 'D_D_-_3He_n.txt', # D + D -> 3He + n
'D-3He': 'D_3He_-_4He_p-endf.txt', # D + 3He -> α + p
'p-B': 'p_11B_-_3a.txt', # p + 11B -> 3α
'T-T': 'T_T_-_4He_n_n.txt', # T + T -> 4He + 2n
'T-3He_a': 'T_3He_-_n_p_4He.txt', # T + 3He -> 4He + n + p
'T-3He_b': 'T_3He_-_D_4He.txt', # T + 3He -> 4He + D
'3He-3He': '3He_3He_-_p_p_4He.txt', # 3He + 3He -> 4He + 2p
}
xs = {}
for xs_id, xs_name in xs_names.items():
xs[xs_id] = Xsec.read_xsec(xs_name)
# Total D + D fusion cross section is due to equal contributions from the
# above two processes.
xs['D-D'] = xs['D-D_a'] + xs['D-D_b']
xs['T-3He'] = xs['T-3He_a'] + xs['T-3He_b']
def get_reactivity(xs, T):
"""Return reactivity, <sigma.v> in cm3.s-1 for temperature T in keV."""
T = T[:, None]
fac = 4 / np.sqrt(2 * np.pi * xs.mr * u)
fac /= (1000 * T * e)**1.5
fac *= (1000 * e)**2
func = fac * xs.xs * Egrid * np.exp(-Egrid / T)
I = np.trapz(func, Egrid, axis=1)
# Convert from m3.s-1 to cm3.s-1
return I * 1.e6
T = np.logspace(0, 3, 100)
xs_labels = {'D-T': '$\mathrm{D-T}$',
'D-D': '$\mathrm{D-D}$',
'D-3He': '$\mathrm{D-^3He}$',
'p-B': '$\mathrm{p-^{11}B}$',
'T-T': '$\mathrm{T-T}$',
'T-3He': '$\mathrm{T-^3He}$',
'3He-3He': '$\mathrm{^3He-^3He}$',
}
DPI, FIG_WIDTH, FIG_HEIGHT = 72, 800, 900
fig = plt.figure(figsize=(FIG_WIDTH/DPI, FIG_HEIGHT/DPI), dpi=DPI)
ax = fig.add_subplot(111)
for xs_id, xs_label in xs_labels.items():
ax.loglog(T, get_reactivity(xs[xs_id], T), label=xs_label, lw=2)
ax.legend(loc='upper left')
ax.grid(True, which='both', ls='-')
ax.set_xlim(1, 1000)
ax.set_ylim(1.e-20, 1.e-14)
xticks= np.array([1, 10, 100, 1000])
ax.set_xticks(xticks)
ax.set_xticklabels([str(x) for x in xticks])
ax.set_xlabel('$T$ /keV')
ax.set_ylabel(r'$\langle \sigma v\rangle \; /\mathrm{cm^3\,s^{-1}}$')
# A second x-axis for energies as temperatures in millions of K
ax2 = ax.twiny()
ax2.set_xscale('log')
ax2.set_xlim(1,1000)
xticks2 = np.array([15, 150, 1500, 5000])
ax2.set_xticks(xticks2 * kB/e * 1.e3)
ax2.set_xticklabels(xticks2)
ax2.set_xlabel('$T$ /million K')
plt.savefig('fusion-reactivities.png', dpi=DPI)
plt.show()
# Plot resolution (dpi) and figure size (pixels)
DPI, FIG_WIDTH, FIG_HEIGHT = 72, 800, 600
fig = plt.figure(figsize=(FIG_WIDTH/DPI, FIG_HEIGHT/DPI), dpi=DPI)
ax = fig.add_subplot(111)
for xs_id, xs_label in xs_labels.items():
ax.loglog(Egrid, xs[xs_id], lw=2, label=xs_label)
ax.grid(True, which='both', ls='-')
ax.set_xlim(1, 1000)
xticks= np.array([1, 10, 100, 1000])
ax.set_xticks(xticks)
ax.set_xticklabels([str(x) for x in xticks])
if COFM:
xlabel ='E(CM) /keV'
else:
xlabel ='E /keV'
ax.set_xlabel(xlabel)
ax.set_ylabel('$\sigma\;/\mathrm{m^2}$')
ax.set_ylim(1.e-32, 1.e-27)
ax.legend(loc='upper left')
plt.savefig('fusion-xsecs2.png', dpi=DPI)
plt.show()
```

The Maxwell distribution for the relative speeds of two particles in equlibrium at temperature $T$ (in three dimensions) is: $$ f_\mathrm{MB}(v) = 4\pi v^2 \left(\frac{\mu}{2\pi k_\mathrm{B}T}\right)^{3/2}\exp\left(-\frac{\mu v_i^2}{2k_\mathrm{B}T}\right), $$ where $\mu = m_1m_2/(m_1+m_2)$ is the reduced mass of the colliding particles. Hence the reaction rate is: $$ \langle \sigma v \rangle = 4\pi \left(\frac{\mu}{2\pi k_\mathrm{B}T}\right)^{3/2} \int_0^\infty v^3\sigma(v) \exp\left(-\frac{\mu v^2}{2k_\mathrm{B}T}\right) \,\mathrm{d}v $$ To convert this integral to one over the relative kinetic energy of the particles, $E = \frac{1}{2}\mu v^2$, note that $\mathrm{d}E = \mu v \,\mathrm{d}v \Rightarrow v\,\mathrm{d}v = \mathrm{d}E / \mu$ and $v^2 = 2E/\mu$. The integral can therefore be written

\begin{align*} \langle \sigma v \rangle &= 4\pi \left(\frac{\mu}{2\pi k_\mathrm{B}T}\right)^{3/2} \int_0^\infty \frac{2E}{\mu} \sigma(E) \exp\left(-\frac{E}{k_\mathrm{B}T}\right) \,\frac{\mathrm{d}E}{\mu}\\ &= \frac{4}{\sqrt{2\pi \mu}} \frac{1}{\left(k_\mathrm{B}T\right)^{3/2}} \int_0^\infty \sigma(E) E \exp\left(-\frac{E}{k_\mathrm{B}T}\right) \,\mathrm{d}E. \end{align*}

- D. D. Clayton, "Principles of Stellar Evolution and Nucleosynthesis", University of Chicago Press (1983).
- S. Atzeni and J. Meyer-ter-Vehn, "The Physics of Inertial Fusion", OUP (2004).

## Comments

Comments are pre-moderated. Please be patient and your comment will appear soon.

## Stephen 2 years, 7 months ago

Where do you calculate or plot the integrated reactivity?

Link | Reply## christian 2 years, 6 months ago

Oops – I think I posted the wrong code. Fixed now!

Link | ReplyThanks, Christian

## Kemal 2 years, 6 months ago

Thank you Christian. The post will be useful for me to understand and visualize Big Bang Nucleosynthesis, I was solving approximated version.

Link | Reply## christian 2 years, 6 months ago

Hi Kemal,

Link | ReplyI'm glad you found it useful!

Cheers, Christian

## Joseph 2 years, 6 months ago

Thank you so much for your steps and code here! May I know where did you obtain your p-B11 cross-section data? It seems that the ENDF library does not include the said data.

Link | Reply## christian 2 years, 6 months ago

Dear Joseph,

Link | ReplyThank you – the p + B-11 cross section is on JANIS (https://www.oecd-nea.org/janisweb/).

Cheers, Christian

## Henri 1 year, 7 months ago

Thanks for the beautiful plots and code although I use matlab. I suggest you remove the temperature scale at the top of the cross section plot. It perpetuates a typical student mistake, which is to mix up temperatures and particle energies! The meaning of energy in the upper plot is that of just one reacting particle pair, not a thermal population!

Link | Reply## christian 1 year, 7 months ago

That's true, of course, but I'm enough of a chemist to have trouble remembering how much energy an electronvolt is, so I find the unphysical temperature scale helpful!

Link | Reply## Samuele 1 year, 7 months ago

Well done Christian, very usefull!

Link | ReplyLine:

fac = 4 * np.pi / np.sqrt(2 * np.pi * xs.mr * u)

shouldn't be:

fac = 4 / np.sqrt(2 * np.pi * xs.mr * u)

i.e., no pi multiplication? Could you doublecheck?

Cheers!

## christian 1 year, 7 months ago

Thanks, Samuele – you're right, there should be no pi multiplication: I guess I didn't fix the online code when I corrected this in the plots last year. It stems from a typo in the textbook by Atzeni and Meyer-ter-Vehn.

Link | ReplyIt's corrected now!

Cheers, Christian

## Zhiqin 1 year, 1 month ago

Hi! It seems to me that particles with higher velocities would have a higher rate of collision relative to particles with lower velocities. Is this true? If so, how should it be taken into account?

Link | Reply## Steve Xiao 1 year ago

Liked your discussion on nuclear cross section. How do we reference or credit it in publications?

Link | Reply## christian 1 year ago

I'm glad you found it useful. You can reference the blog article title, URL, name and date if you would like to cite it.

Link | ReplyCheers, Christian

## Axel 6 months, 1 week ago

I just found this great blog post, but have problems to understand how sigma*v is calculated in the code. Whereever the Boltzmann constant is used in the equation, it seems that the code (inside get_reactivity) does something different using the elementary charge 'e' !?

Link | ReplyCan somebody explain what's going on ??

Many thanks,

Axel

## Martin Nieto 5 months, 3 weeks ago

This article has analytic expressions for the fusion cross sections that you can use to make the rate calculation even more accurate.

Link | Replyhttps://iopscience.iop.org/article/10.1088/0029-5515/32/4/I07/meta

If you cannot access it, I can pass it to you.

## christian 5 months, 3 weeks ago

Dear Martin,

Link | ReplyThank you – I can access this paper. I'll try to get round to comparing the predicted reactivities for D-3He with those obtained from the cross section fit to 4900 keV.

Cheers, Christian

## Justin 3 months, 3 weeks ago

There is an option to plot the cross sections in the center-of-mass frame or not. But it seems that this flag only changes the plot axis label. Shouldn't you also convert the energy grid ?

Link | Reply## christian 3 months, 3 weeks ago

Dear Justin,

Link | ReplyI think the CM flag defaults to True in the read_xsec classmethod definition, so the energy grid transformation is performed without having to pass COFM explicitly. Perhaps the line in which the cross sections are read should be

xs[xs_id] = Xsec.read_xsec(xs_name, CM=COFM)

to make this more explicit?

Cheers, Christian

## New Comment