Plane-wave Radiation Pressure Effects

This section contains derivations of analytical formulas for radiation pressure induced modulation of a plane-wave laser beam, in various setups. Test cases simulating these setups can be found in the file plane_wave/test_radiation_pressure.py, in the Finesse validation tests folder.

Mirrors

Single-Sided Free Mass

../../../_images/01_single_side_free_mass.svg

Fig. 10 Setup for the first test case, an amplitude modulated laser with two sidebands reflecting off a free mass.

Analytics

We start with an amplitude modulated input field, such as produced by connecting a SignalGenerator to a Laser’s amp port in Finesse,

(3)\[E_\mathrm{i} = E_0 \cos(\omega_0 t) \left( 1 - \frac{A}{2} \left(1 - \cos(\Omega t)\right) \right) \]

where \(A\) is the modulation index. Writing this in the complex notation used in Finesse gives

\[\begin{aligned} E_\mathrm{i} &= E_0 e^{i \omega_0 t} \left( 1 - \frac{A}{2} \left(1 - \cos(\Omega t)\right) \right) \\ &= E_0 e^{i \omega_0 t} \left( 1 - \frac{A}{2} + \frac{A}{4}\left(e^{i \Omega t} + e^{-i \Omega t}\right) \right) \end{aligned} \]

We can then separate this into three fields; the carrier field \(a_0\), and the upper and lower sidebands \(a_\pm\), such that

(4)\[E_\mathrm{i} = a_0 + a_+ + a_-, \]

where

\[\begin{aligned} a_0 &= \left(1 - \frac{A}{2}\right) E_0 e^{i \omega_0 t} \\ a_+ &= \frac{A}{4} E_0 e^{i (\omega_0 + \Omega) t} \\ a_- &= \frac{A}{4} E_0 e^{i (\omega_0 - \Omega) t}. \end{aligned} \]

Next, we consider this field when incident on a free-floating mirror. For a free mass, we have

\[F(t) = m\ddot{x}(t). \]

Taking the Fourier transform, we see that

\[\begin{aligned} \tilde{F}(\omega) &= -m \omega^2 \tilde{x}(\omega) \\ \therefore \tilde{\chi}(\omega) \equiv \frac{\tilde{x}(\omega)}{\tilde{F}(\omega)} &= -\frac{1}{m \omega^2}. \end{aligned} \]

For a force of the form \(F_0 \cos(\Omega t)\), we have

\[\tilde{F}(\omega) = \pi F_0 (\delta(\omega + \Omega) + \delta(\omega - \Omega)), \]

from which we can find \(x(t)\):

\[\begin{aligned} x(t) &= \frac{1}{2 \pi}\int_{-\infty}^{+\infty} \tilde{\chi}(\omega) \tilde{F}(\omega) e^{i \omega t} \mathrm{d}\omega \\ &= -\frac{\pi F_0}{2\pi} \int_{-\infty}^{+\infty} \frac{1}{m \omega^2} e^{i \omega t} (\delta(\omega + \Omega) + \delta(\omega - \Omega)) \mathrm{d}\omega \\ &= -\frac{F_0}{m \Omega^2} \frac{e^{i \Omega t} + e^{-i \Omega t}}{2} \\ &= -\frac{1}{m \Omega^2} F_0 \cos(\Omega t). \end{aligned} \]

A force of this form arises from the beating of the two amplitude modulation sidebands with the carrier. The force due to radiation pressure of light interacting with a mirror is given by

(5)\[\begin{aligned} F_\mathrm{rad} &= \cos{\alpha} \frac{P_{1\mathrm{i}} + P_{1\mathrm{o}} - P_{2\mathrm{i}} - P_{2\mathrm{o}}}{c}, \end{aligned} \]

where \(P_{\{1,2\}\{\mathrm{i,o}\}}\) is the power of the light on the negative and positive \(x\) sides of the mirror incoming and outgoing respectively, and \(\alpha\) is the angle of incidence. In this simple case, \(\alpha = 0\), and there is no incoming beam from the positive \(x\) direction. The force can then be calculated to leading order in \(A\) from Equation (3) as

\[\begin{aligned} F_\mathrm{rad} &= \frac{|E_\mathrm{i}| ^2 + |E_\mathrm{r}| ^2 - |E_\mathrm{t}| ^2}{c} \\ &= \frac{(1 + R - T)}{c} |E_\mathrm{i}| ^2 \\ &= \frac{(1 + R - T)}{c} \left({E_0}^2 (1 - A) + A {E_0}^2 \cos(\Omega t)\right), \end{aligned} \]

where \(R\) is the coefficient of reflectivity of the mirror. Here we are uninterested in the D.C. term as this will not contribute to the mirror’s oscillation, and so have

\[\begin{aligned} F_0 &= \frac{(1 + R - T) A {E_0}^2}{c}, \\ \therefore x(t) &= -\frac{(1 + R - T) A {E_0}^2}{m c \Omega^2} \cos(\Omega t). \end{aligned} \]

Fig. 11 shows how an incoming field changes upon reflection from the mirror.

../../../_images/mirror_fields.svg

Fig. 11 Fields at a mirror’s surface.

The incident field will be multiplied by the reflection coefficient of the mirror, \(r\). The reflected light will also be phase shifted due to propagation through space by a total amount \(e^{-2i \phi}\), where

\[\begin{aligned} \phi &= \frac{2 \pi x(t)}{\lambda} \\ &= -\frac{2 \pi (1 + R - T) A {E_0}^2}{\lambda m c \Omega^2} \cos(\Omega t) \\ &= -\frac{(1 + R - T) A {E_0}^2 \omega_0}{m c^2 \Omega^2} \cos(\Omega t). \end{aligned} \]

This results in the incoming beam being phase modulated by the mirror motion,

\[E_\mathrm{r} = r E_\mathrm{i} \exp(i B \cos(\Omega t)), \]

where

\[B = \frac{2 (1 + R - T) A {E_0}^2 \omega_0}{m c^2 \Omega^2}, \]

which to first order gives

\[\begin{aligned} E_\mathrm{r} &= r E_\mathrm{i} \left( 1 + i\frac{B}{2}\left(e^{i \Omega t} + e^{-i \Omega t}\right) \right) \\ &= r(b_0 + b_+ + b_-), \end{aligned} \]

where

\[\begin{aligned} b_0 &= E_\mathrm{i} \\ b_+ &= i\frac{B}{2} E_\mathrm{i} e^{i \Omega t} \\ b_- &= i\frac{B}{2} E_\mathrm{i} e^{-i \Omega t}. \end{aligned} \]

Putting Equation (4) into this gives

\[\begin{aligned} b_0 &= a_0 + a_+ + a_- \\ b_+ &= i\frac{B}{2} E_0 e^{i (\omega_0 + \Omega) t} \\ b_- &= i\frac{B}{2} E_0 e^{i (\omega_0 - \Omega) t}, \end{aligned} \]

to leading order, where we can once again see an input field and an upper and lower sideband. The full expression for the reflected field can then be written as

\[\begin{aligned} E_\mathrm{r} &= r (a_0 + a_+ + a_- + b_+ + b_-) \\ &= r E_0 e^{i \omega_0 t} \left( 1 + \left(\frac{A + 2i B}{4}\right) \left(e^{i \Omega t} + e^{-i \Omega t}\right) \right) \\ &= r E_0 e^{i \omega_0 t} \left( 1 + \left( \frac{A}{4} + i\frac{(1 + R - T) A {E_0}^2 \omega_0}{m c^2 \Omega^2} \right) \left(e^{i \Omega t} + e^{-i \Omega t}\right) \right) \end{aligned} \]

Finesse

The analytics above are compared to the sidebands produced by the following Finesse file:

l l1 P=3
s s1 l1.p1 m1.p1 L=5
m m1 R=0.5 T=0.4
free_mass m1_sus m1.mech 0.5e-3

ad upper s1.p1.o &fsig.f
ad lower s1.p1.o -&fsig.f

sgen sig l1.amp.i 0.1
fsig(0.1)

xaxis(sig.f, log, 0.1, 1e7, 400)

The results are shown in Fig. 12.

../../../_images/01_single_side_free_mass_result.svg

Fig. 12 Comparison of the upper and lower output sidebands from Finesse with analytics.

Double-Sided Free Mass, One Modulation

../../../_images/02_two_sides_one_mod_free_mass.svg

Fig. 13 Setup for the second test case, an amplitude modulated laser with two sidebands reflecting off a free mass causing sidebands to be generated in an unmodulated laser incident on the other side.

Analytics

This is similar to the previous test, however there are no amplitude modulation sidebands present in the input field. As a result, we only expect to see the output containing the phase modulation sidebands. The second laser will not contribute to the oscillation of the mirror, as it is a purely D.C. force. The sidebands will also be \(180^\circ\) out of phase, as the mirror’s motion is inverted when looking from the other side. In addition, we assume that the mirror is opaque i.e. \(T = 0\). For a given input field

\[E_\mathrm{i} = E_1 \cos(\omega_0 t), \]

we therefore have

\[E_\mathrm{r} = r E_1 \left( 1 - i\frac{(1 + R) A {E_0}^2 \omega_0}{m c^2 \Omega^2} \left(e^{i \Omega t} + e^{-i \Omega t}\right) \right). \]

Finesse

The analytics above are compared to the sidebands produced by the following Finesse file:

l l1 P=3
s s1 l1.p1 m1.p1 L=5
m m1 R=1 T=0
s s2 m1.p2 l2.p1 L=5
l l2 P=2

free_mass m1_sus m1.mech 0.5e-3

ad upper s2.p2.o &fsig.f
ad lower s2.p2.o -&fsig.f

sgen sig l1.amp.i 0.1
fsig(0.1)

xaxis(sig.f, log, 0.1, 1e7, 400)

The results are shown in Fig. 14.

../../../_images/02_two_sides_one_mod_free_mass_result.svg

Fig. 14 Comparison of the upper and lower output sidebands from Finesse with analytics.

Double-Sided Free Mass, Two Modulations

../../../_images/03_two_sides_two_mod_free_mass.svg

Fig. 15 Setup for the third test case, two amplitude modulated lasers with two sidebands each incident on each side of a mirror, with some arbitrary phase difference.

Analytics

As the second beam is now also amplitude modulated, it will contribute to the mirror’s motion, so we must recalculate the radiation pressure force from Equation (5). For the left and right input fields,

\[\begin{aligned} E_\mathrm{Left} &= E_0 \cos(\omega_0 t) \left( 1 - \frac{A_0}{2} \left(1 - \cos(\Omega t)\right) \right) \\ E_\mathrm{Right} &= E_1 \cos(\omega_0 t) \left( 1 - \frac{A_1}{2} \left(1 - \cos(\Omega t - \psi)\right) \right), \end{aligned} \]

we then have

\[F_\mathrm{rad} = \frac{1 + R}{c} \left( {E_0}^2 (1 - A_0) + A_0 {E_0}^2 \cos(\Omega t) - {E_1}^2 (1 - A_1) - A_1 {E_1}^2 \cos(\Omega t - \psi) \right), \]

or ignoring the D.C. terms,

\[F_\mathrm{rad} = \frac{1 + R}{c} \left( A_0 {E_0}^2 \cos(\Omega t) - A_1 {E_1}^2 \cos(\Omega t - \psi) \right). \]

To write this as a single cosine, we use some trigonometry:

\[\begin{aligned} F_\mathrm{rad} &= \frac{1 + R}{c} \left[ A_0 {E_0}^2 \cos(\Omega t) - A_1 {E_1}^2 (\cos(\Omega t)\cos(\psi) + \sin(\Omega t)\sin(\psi)) \right] \\ &= \frac{1 + R}{c} \left[ (A_0 {E_0}^2 - A_1 {E_1}^2 \cos(\psi)) \cos(\Omega t) - A_1 {E_1}^2 \sin(\psi) \sin(\Omega t) \right]. \end{aligned} \]

Writing the term in brackets in the form \(D \cos(x + \theta)\), we get

\[\begin{aligned} D^2 &= {(A_0 {E_0}^2 - A_1 {E_1}^2\cos(\psi))}^2 + {(A_1 {E_1}^2\sin(\psi))}^2 \\ D &= \sqrt{ {A_0}^2 {E_0}^4 - 2 A_0 A_1 {E_0}^2 {E_1}^2 \cos(\psi) + {A_1}^2 {E_1}^4 } \end{aligned} \]

and

\[\theta = \tan^{-1} \left( \frac{ A_1 {E_1}^2 \sin(\psi) }{ A_0 {E_0}^2 - A_1 {E_1}^2 \cos(\psi) } \right). \]

We can now proceed in the same way as in Single-Sided Free Mass, to obtain an expression for the reflected field on the left hand side of the mirror,

\[E_\mathrm{r} = r E_{0} e^{i \omega_0 t} \left[ 1 + \left( \frac{A_0}{4} \left(e^{i\Omega t} + e^{-i\Omega t}\right) + i\frac{(1 + R) D \omega_0}{m c^2 \Omega^2} \left( e^{i (\Omega t + \theta)} + e^{-i (\Omega t + \theta)} \right) \right) \right]. \]

To get the field on the right hand side of the mirror, we can apply the following (remembering that \(\theta\) will also be affected):

\[\begin{aligned} E_0 &\Longleftrightarrow E_1 \\ A_0 &\Longleftrightarrow A_1 \\ \psi &\Longrightarrow -\psi. \end{aligned} \]

Finesse

The analytics above are compared to the sidebands produced by the following Finesse file:

l l1 P=3
s s1 l1.p1 m1.p1 L=5
m m1 R=1 T=0
s s2 m1.p2 l2.p1 L=5
l l2 P=4

free_mass m1_sus m1.mech 0.5e-3

ad upper_left s1.p1.o &fsig.f
ad lower_left s1.p1.o -&fsig.f

ad upper_right s2.p2.o &fsig.f
ad lower_right s2.p2.o -&fsig.f

fsig(0.1)
sgen sig1 l1.amp.i 0.1
sgen sig2 l2.amp.i 0.1 phase=29

xaxis(fsig.f, log, 0.1, 1e7, 400)

For comparison with the analytical results, we must also include an extra factor of \(e^{\pm i\psi}\) in the upper / lower sidebands respectively when looking at the right hand side field. This is because Finesse measures phase relative to the first laser in the file, which in this case is the left hand side laser. The results are shown in Fig. 16 & Fig. 17.

../../../_images/03_two_sides_two_mod_free_mass_result_left.svg

Fig. 16 Comparison of the upper and lower output sidebands from Finesse and analytics for the reflected field on the left of the mirror.

../../../_images/03_two_sides_two_mod_free_mass_result_right.svg

Fig. 17 Comparison of the upper and lower output sidebands from Finesse and analytics for the reflected field on the right of the mirror.

Beamsplitters

The beamsplitter behaves similarly to the mirror, with two key differences. Firstly, the angle of incidence \(\alpha\) in Equation (5) is no longer necessarily 0. This also affects the phase change on reflection due to the beamsplitter tuning, \(\phi\). Secondly, there are now 4 ports at which light can enter and exit, rather than two. Here we repeat similar experiments to the previous section, taking these factors into account.

Single-Sided Free Mass

../../../_images/04_beamsplitter_single_side_free_mass.svg

Fig. 18 Setup for the first beamsplitter test case, an amplitude modulated laser with two sidebands reflecting off a free mass at some non-zero angle.

Analytics

The steps here are much the same as in the equivalent mirror case, with the only changes being the inclusion of \(\alpha\) in the radiation pressure force calculation, and a change in the phase shift due to mirror tuning \(\phi\) — the total phase shift of reflected light will now be \(e^{-2i\phi\cos{\alpha}}\), as the tuning \(\phi\) is defined relative to the surface normal.

We therefore have

\[\begin{aligned} x(t) &= -\frac{(1 + R - T) A {E_0}^2}{m c \Omega^2} \cos(\Omega t) \cos{\alpha}.\\ \therefore \phi &= -\frac{(1 + R - T) A {E_0}^2 \omega_0}{m c^2 \Omega^2} \cos(\Omega t) \cos{\alpha}.\end{aligned}\]

This results in the incoming beam being phase modulated by the mirror motion,

\[E_\mathrm{r} = r E_\mathrm{i} \exp(i B \cos(\Omega t)),\]

where

\[B = \frac{2 (1 + R - T) A {E_0}^2 \omega_0}{m c^2 \Omega^2} \cos^2{\alpha},\]

Following the same steps as in the mirror case, the full expression for the reflected field can then be written as

\[\begin{aligned} E_\mathrm{r} &= r E_0 e^{i \omega_0 t} \left( 1 + \left(\frac{A + 2i B}{4}\right) \left(e^{i \Omega t} + e^{-i \Omega t}\right) \right) \\ &= r E_0 e^{i \omega_0 t} \left( 1 + \left( \frac{A}{4} + i\frac{(1 + R - T) A {E_0}^2 \omega_0}{m c^2 \Omega^2} \cos^2{\alpha} \right) \left(e^{i \Omega t} + e^{-i \Omega t}\right) \right) \end{aligned}\]

Finesse

The analytics above are compared to the sidebands produced by the following simple Finesse file:

l l1 P=3
s s1 l1.p1 bs1.p1 L=5
bs bs1 R=0.5 T=0.4 alpha=37
s s2 bs1.p2 out.p1 L=5
nothing out

free_mass bs1_sus bs1.mech 0.5e-3

ad upper s2.p2.o &fsig.f
ad lower s2.p2.o -&fsig.f

fsig(0.1)
sgen sig l1.amp.i 0.1

xaxis(fsig.f, log, 0.1, 1e7, 400)

The results are shown in Fig. 19.

../../../_images/04_beamsplitter_single_side_free_mass_result.svg

Fig. 19 Comparison of the upper and lower output sidebands from Finesse with analytics.

Double-Sided Free Mass, One Modulation

../../../_images/05_beamsplitter_two_sides_one_mod_free_mass.svg

Fig. 20 Setup for the second beamsplitter test case, an amplitude modulated laser with two sidebands reflecting off a free mass causing sidebands to be generated in an unmodulated laser incident on the other side.

Analytics

This is similar to the equivalent mirror case, with the inclusion of the \(\cos^2{\alpha}\) term from the previous section. For a given input field

\[E_\mathrm{i} = E_1 \cos(\omega_0 t),\]

we therefore have

\[E_\mathrm{r} = r E_1 \left( 1 - i\frac{(1 + R) A {E_0}^2 \omega_0}{m c^2 \Omega^2} \cos^2{\alpha} \left(e^{i \Omega t} + e^{-i \Omega t}\right) \right).\]

Finesse

The analytics above are compared to the sidebands produced by the following Finesse file:

l l1 P=3
s s1 l1.p1 bs1.p1 L=5

l l2 P=2
s s2 l2.p1 bs1.p4 L=5

bs bs1 R=1 T=0 alpha=37
s sout bs1.p3 out.p1 L=5
nothing out

free_mass bs1_sus bs1.mech 0.5e-3

ad upper sout.p2.o &fsig.f
ad lower sout.p2.o -&fsig.f

fsig(0.1)
sgen sig l1.amp.i 0.1

xaxis(fsig.f, log, 0.1, 1e7, 400)

The results are shown in Fig. 21

../../../_images/05_beamsplitter_two_sides_one_mod_free_mass_result.svg

Fig. 21 Comparison of the upper and lower output sidebands from Finesse with analytics.

Double-Sided Free Mass, Two Modulations

../../../_images/06_beamsplitter_two_sides_two_mod_free_mass.svg

Fig. 22 Setup for the third beamplitter test case, two amplitude modulated lasers with two sidebands each incident on each side of a beamsplitter, with some arbitrary phase difference.

Analytics

Once again, this is similar to the procedure in the equivalent mirror case, with the addition of a \(\cos^2{\alpha}\) term on the amplitude of the generated sidebands. The output at the top detector will therefore be

\[\begin{aligned} E_\mathrm{r} &= r E_{0} e^{i \omega_0 t} \left[ 1 + \left( \frac{A_0}{4} \left(e^{i\Omega t} + e^{-i\Omega t}\right) + i\frac{(1 + R) D \omega_0}{m c^2 \Omega^2}\cos^2{\alpha} \left( e^{i (\Omega t + \theta)} + e^{-i (\Omega t + \theta)} \right) \right) \right], \end{aligned}\]

with \(D\) defined as in the mirror case. The results for the right detector can be obtained with the same transformations as before.

Finesse

The analytics above were compared to the sidebands produced by the following Finesse file:

l l1 P=3
s s1 l1.p1 bs1.p1 L=5

l l2 P=4
s s2 l2.p1 bs1.p4 L=5

bs bs1 R=1 T=0 alpha=37

s sout_left bs1.p2 out_left.p1 L=5
nothing out_left

s sout_right bs1.p3 out_right.p1 L=5
nothing out_right

free_mass bs1_sus bs1.mech 0.5e-3

ad upper_left sout_left.p2.o &fsig.f
ad lower_left sout_left.p2.o -&fsig.f

ad upper_right sout_right.p2.o &fsig.f
ad lower_right sout_right.p2.o -&fsig.f

fsig(0.1)
sgen sig1 l1.amp.i 0.1
sgen sig2 l2.amp.i 0.1 phase=29

xaxis(fsig.f, log, 0.1, 1e7, 400)

As in the mirror case, for comparison with the analytical results, we must also include an extra factor of \(e^{\pm i\psi}\) in the upper / lower sidebands respectively when looking at the right hand side field. This is because Finesse measures phase relative to the first laser in the file, which in this case is the left hand side laser. The results are shown in Fig. 23 & Fig. 24.

../../../_images/06_beamsplitter_two_sides_two_mod_free_mass_result_left.svg

Fig. 23 Comparison of the upper and lower output sidebands from Finesse and analytics for the reflected field on the left of the beamsplitter.

../../../_images/06_beamsplitter_two_sides_two_mod_free_mass_result_right.svg

Fig. 24 Comparison of the upper and lower output sidebands from Finesse and analytics for the reflected field on the right of the beamsplitter.