Supplementary information:
Signal Amplification by Sensitive Control of
Bifurcation Topology
R.B. Karabalin, Ron Lifshitz2*, M.C. Cross1, M.H. Matheny1, S.C. Masmanidis1,
and M.L. Roukes1†
Kavli Nanoscience Institute and Condensed Matter Physics 114-36, California Institute of Technology,
Pasadena, CA 91125, USA
2
Raymond and Beverly Sackler School of Physics & Astronomy, Tel Aviv University, Tel Aviv 69978, Israel
*To whom correspondence on matters of theory should be addressed. E-mail: ronlif@tau.ac.il (RL).
†
To whom correspondence on matters of experiment should be addressed. E-mail: roukes@caltech.edu (MLR).
Overview
In this document we present a theoretical exposition of how bifurcation topology can be
controlled for a coupled pair of parametrically-driven nonlinear resonators. We then address the
effect of noise on the probability that the system accurately follows the topology of the
bifurcation. We subsequently provide further details about our experimental methods and on an
apparatus for realizing a coupled-NEMS BTA. Finally, we assess the effective noise in our
current implementation, and conclude by employing our noise analysis to make realistic
projections of the ultimate sensitivity limits of this implementation.
* e-mail: roukes@caltech.edu
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Theoretical Background
Response of Coupled Nonlinear Resonators to Parametric Excitation
We begin by providing the calculation of the theoretical response curves that are plotted in Fig. 1
of the main text. We consider two weakly-coupled parametrically-driven nonlinear resonators
with slightly different normal frequencies. Their dynamics are governed by a pair of coupled
equations of motion (EOM)
~
d 2~
x ± ~ d~
x±
~
~
2
3
2 dx ±
~
~
~
~
~
~
+ ω ± + h cos ω p t x ± + α x ± + η x ± ~ + D(~
x± − ~
x∓ ) = 0 ,
~ 2 + γ d~
dt
t
dt
(
)
(S1)
where all physical parameters—after having divided out the effective mass of the resonators—
are denoted with tildes to distinguish them from the scaled parameters used below. Here ~x +
denotes the displacement of the higher-frequency resonator with frequency ω + from its
equilibrium, and ~x − denotes the displacement from equilibrium of the lower-frequency resonator
with frequency ω − . The Duffing parameter α , the linear damping rate γ~ , and the coefficient of
nonlinear damping η~ are all assumed to be approximately the same for both resonators, and the
~
~
~ are the
coupling strength between the resonators is denoted by D . The parameters h and ω
p
parametric driving amplitude and driving frequency.
We rescale the units of time and space, to eliminate two additional parameters from the equation
of motion—the average resonance frequency of the resonators, and the Duffing parameter, which
are both set to 1. Because we drive the system close to twice the average resonance frequency,
we express the scaled pump frequency as ω p = 2 + Δω p . The EOM then becomes
-2-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
x± + Q −1 x± + 1 ± Δ + H cos ⎡⎣ 2 + Δω p ⎤⎦ t x± + x± 3 + η x± 2 x± + D ( x± − x∓ ) = 0 ,
(
)
(S2)
where dots denote derivatives with respect to the dimensionless time t , Q is the quality factor of
the resonators, and Δ << 1 is the scaled frequency difference between the resonators, so that
ω ± 2 are replaced by 1 ± Δ .
We calculate the response of the coupled resonators following the methods of Lifshitz & Cross
(1,2). We begin by assuming that the linear damping is weak or, equivalently, that Q is large,
and define a small expansion parameter ε , by expressing the scaled linear damping rate
as Q −1 = εγ , with γ of order unity. The parametric instability of the system then occurs for small
driving amplitudes on the order of ε near resonance. If, in addition, we consider the system near
the onset of the instability, we can assume that the effects of nonlinearity are small as well. Since
the coupling strength is the weak signal to be amplified by the BTA, it can also be considered as
a small perturbative correction. Finally, the frequency difference between the two resonators can
also be taken to be small, on the order of ε . All these perturbative corrections can be chosen to
enter the EOM in the same order of the small parameter ε by taking the leading order in x ± to be
ε , expressing the scaled parametric driving amplitude as H = εh , expressing the scaled
frequency difference as Δ = εδ , expressing the scaled coupling constant as D = ε d , and driving
the system close to twice the average resonance, taking Δω p = εΩ . The final form of the EOM is
then
x ± + εγ x ± + (1 ± εδ + εh cos[2 + εΩ ]t )x ± + x ± + ηx ± x ± + εd ( x ± − x ∓ ) = 0 .
3
2
(S3)
-3-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Expecting the motion of the resonators away from equilibrium to be on the order of
ε we try a
solution of the form
x ± (t ) =
ε
2
(A (T )e
it
±
)
+ c.c. + ε 3 2 x ±
(1)
(t ) + … .
(S4)
The lowest order contribution to this solution is based on the solution to the linear equations of
motion of the two simple harmonic oscillators x ± + x ± = 0 , where T = εt is a slow time
variable, allowing the complex amplitudes A± (T ) to vary slowly in time, due to the effect of all
the perturbative terms in Eq. (S3). Following the methods of Lifshitz & Cross1,2 we obtain a pair
of coupled equations for determining the amplitudes A± (T ) ,
2i
dA±
h
3 + iη
2
+ d ( A± − A∓ ) ± δA± + A±∗ e iΩT + iγA± +
A± A± = 0 .
dT
2
4
(S5)
The explicit time dependence can be removed by taking a solution of the form
A± (T ) = a ± (T )e
Ω
i T
2
,
(S6)
yielding a corresponding equation for a ± (T ) ,
h ∗
⎛ d
⎞
⎛ 3 + iη ⎞ 2
+ d ± δ − Ω + iγ ⎟a ± − da ∓ + ⎜
⎜ 2i
⎟ a± a± = − a± .
2
⎝ dT
⎠
⎝ 4 ⎠
(S7)
With the expression (S6) for the slowly varying amplitudes A± (T ) , the steady-state solution to
the scaled equations of motion (S2), for which the complex amplitudes a ± are constant in time,
becomes an oscillation at half the drive frequency 1 + ε Ω 2 . Note that we are not interested in
-4-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
the corrections x±(1) (t ) of order ε 3 2 to these oscillations, but rather in finding the fixed complex
amplitudes a ± of the lowest order terms. These are obtained by solving the coupled algebraic
equations, obtained from (S7) by requiring that da ± / dT = 0 . Setting d = 0 decouples the
equations, giving two equations that can be solved in closed from (1,2). The solution to a single
equation gives the response curve, shown in Fig. 1a in the main text. The sum of both solutions,
taking into account the relative π phase freedom, is plotted in Fig. 1b. For finite coupling d ≠ 0
we can find the roots of the coupled Eqs. (S7) numerically once values are chosen for the
different parameters. These are shown in Figs. 1c and 1d for positive and negative coupling
respectively. In all plots we use the values γ = 1 , η = 0.01 , δ = 1.1 , and h = 1.05 . In 1b we
take d = 0 , and in 1c and 1d we take d = ±0.01 .
Amplitude Equation for an Imperfect Pitchfork Bifurcation
We wish to consider in more detail what happens at the bifurcation upon an upward frequency
sweep, as the second mode starts oscillating. In the presence of weak coupling the normal modes
are slightly modified from pure motion of the individual resonators. Diagonalization of the linear
terms in (S7), keeping only corrections of order d / δ , yields the modified modes
v (1)
⎛ d
⎜
= ⎜ 2δ
⎜ 1
⎜
⎝
⎞
⎟
⎟,
⎟
⎟
⎠
v ( 2)
⎛
⎜ 1
=⎜
⎜− d
⎜
⎝ 2δ
⎞
⎟
⎟,
⎟
⎟
⎠
with
⎛ a+ ⎞ 2
⎜⎜ ⎟⎟ = ∑ a n (T )v ( n ) ,
⎝ a − ⎠ n =1
(S8)
where the mode frequencies are unchanged to first order in d / δ . Substitution into Eq. (S7)
yields a set of nonlinearly-coupled equations of motion for the modified mode amplitudes a n (T ) ,
-5-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
d ⎛1
1 2 ∗ ⎞⎤
h ∗
2
2
⎛ d
⎞
⎛ 3 + iη ⎞ ⎡ 2
− 2δ − Ω 2 + iγ ⎟a1 + ⎜
⎜ 2i
⎟ ⎢ a1 a1 + ⎜ a 2 a 2 − a1 a 2 − a1 a 2 ⎟⎥ = − a1 , (S9)
2
2
δ ⎝2
⎝ dT
⎠
⎝ 4 ⎠⎣
⎠⎦
d⎛ 1 2
1 2 ∗ ⎞⎤
h ∗
2
⎛ d
⎞
⎛ 3 + iη ⎞ ⎡ 2
− Ω 2 + iγ ⎟ a 2 + ⎜
⎜ 2i
⎟ ⎢ a 2 a 2 + ⎜ − a1 a1 + a 2 a1 + a 2 a1 ⎟⎥ = − a 2 ,
2
2
δ⎝ 2
⎝ dT
⎠
⎝ 4 ⎠⎣
⎠⎦
(S10)
where we now measure the drive frequency with respect to the second mode frequency by
defining Ω 2 = Ω − (d + δ ) .
Our aim is to obtain an equation for the growth of the amplitude of the second mode at the
bifurcation, as the frequency is swept upward. At that point the first mode will have already
reached a certain non-zero amplitude, which can be determined analytically (2) by solving (S9)
after setting a 2 = 0 . To find the initial growth of the second mode we linearize (S10) in a 2 .
Taking the coupling d / δ to be weak, and assuming an initial growth of the form
a 2 = ae
⎛π
⎞
i ⎜ +φ ⎟
⎝ 4 ⎠ σT
e
,
with
a, φ , σ ∈ ℜ ,
(S11)
we find that the phase φ and the growth rate σ satisfy the relations
2σ = −γ +
h
cos 2φ ,
2
(S12)
h
0 = Ω 2 + sin 2φ .
2
Thus, the bifurcation occurs as the drive frequency Ω 2 is increased and reaches a critical value
of Ω C = − (h / 2) − γ 2 .
2
-6-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Next, we wish to include nonlinearity to saturate the growth of the second mode, and to include
the coupling to the first mode, which is already oscillating, to affect the topology of the
bifurcation. Performing a calculation similar to that found in section 1.3.3 of Lifshitz & Cross
(2), we find that
a 2 = B(T )e
⎛π
⎞
i ⎜ +φ ⎟
⎝4 ⎠
,
(S13)
where the real-valued saturated amplitude B (T ) satisfies the equation
2
⎞
ΩC
1 ⎛ 3 ΩC
π⎞
dB
d h 9 +η
3
⎛
= (Ω 2 − Ω C )
+ η ⎟⎟ B 3 +
a1 sin⎜ φ + φˆ1 − ⎟ ,
B − ⎜⎜
2γ
8⎝ γ
4⎠
δ 32γ
dT
⎝
⎠
(S14)
where a1 is the amplitude of the first mode determined earlier, and φˆ1 is the phase φ1 of the
first mode plus arctan(η / 3) . It can be shown that for weak nonlinear damping (η << 1 ),
φˆ1 ≅ φ + π / 4 .
Scaling back to the parameters of Eq. (S2), and dropping terms of order η 2 we find that at the
bifurcation the modes [with eigenvectors given by (S8)] are oscillating as
⎞
⎛ ⎡ Δω p ⎤
x1 (t ) = X 1 cos⎜⎜ ⎢1 +
⎥ t + φ1 ⎟⎟,
2 ⎦
⎠
⎝⎣
⎞
⎛ ⎡ Δω p ⎤ π
x 2 (t ) = X 2 (t ) cos⎜⎜ ⎢1 +
⎥ t + + φ ⎟⎟.
2 ⎦
4
⎠
⎝⎣
(S15)
The amplitude of the second mode X 2 (t ) satisfies the equation
-7-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Qω C
dX 2
1
3 D
3
3
= (Δω p − [D + Δ − ω C ])
X 2 − (3Qω C + η )X 2 +
Qω C X 1 ,
dt
2
8
16 Δ
(S16)
which has the typical form of an imperfect pitchfork bifurcation, with Δω p acting as the control
⎛ QH ⎞
parameter, and where Qω C = ⎜
⎟ −1 .
⎝ 2 ⎠
2
Analysis of the Effects of Noise
Noise in the system will have its largest effect on the measurement process as the pump
frequency passes through the pitchfork bifurcation of the second mode. The dynamics in this
vicinity can be analyzed using Eq. (S16) for the amplitude of the second mode, supplemented
with a noise term. For small signal and noise the important time range is when X2 is small, so
that the nonlinear term in Eq. (S16) is not involved. We write the linearized Eq. (S16) in the
form
dX 2
= rtX 2 + s + f (t )
dt
(S17)
where rt is the linear ramp of control parameter (the parameters in front of X2 in the first term on
the right hand side of Eq. (S16)) choosing to measure time t from the bifurcation point, s is the
term leading to the imperfect bifurcation proportional to the coupling D (the last term in Eq.
(S16)), and f(t) is the noise force term, assumed to be Gaussian white noise of strength F defined
by
f (t ) f (t ' ) = 2 Fδ (t − t ' ) .
(S18)
-8-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
The conventional force spectral density S F is related to this noise strength by S F = 4 F . The
Fokker-Planck equation for the probability distribution P(X2) of X2 at time t corresponding to Eq.
(S17) can be solved to give the Gaussian distribution
P( X 2 ) =
1
2π σ ( t )
e −[ X 2 − X D ( t )]
2
/ 2σ 2 ( t )
,
(S19)
where X D (t ) is the deterministic solution, given by equation (S17) without the noise term, and
σ (t ) is the time dependent width. X D (t ) grows away from zero due to the signal s. The explicit
expressions, assuming the control parameter ramp starts at a value far below the bifurcation
point, are
X D (t ) = s ∫ e r ( t
t
σ (t ) = 2 F ∫ e
2
−t '2 ) / 2
−∞
2
t
r (t 2 −t '2 )
−∞
dt ' ,
(S20)
dt '.
For long times t >> r −1 / 2 these expressions give
⎛ 2π ⎞
X D (t ) = ⎜
⎟
⎝ r ⎠
⎛ 4π ⎞
σ (t ) = ⎜ ⎟
⎝ r ⎠
1/ 2
s e rt
2
1/ 2
rt 2 / 2
/2
,
(S21)
1/ 4
F
e
.
Note the super-exponential growth of both the center and width of the distribution due to the
increasing amplification rate, proportional to the bifurcation parameter rt. We now calculate the
probability P↑ that at long times X2 falls in the basin of attraction of the ↑ branch (and
P↓ = 1 − P↑ ). The shift − s / rt in the basin boundary away from X2=0 is small compared with XD
for t >> r −1 / 2 , and so P↑ ≈ ∫ P( X 2 )dX 2 giving
∞
0
⎡⎛ π ⎞1 / 4
⎤
s
⎛ XD ⎞
P↑ − P↓ ≈ erf ⎜
.
⎟ = erf ⎢⎜ ⎟
1/ 4 1/ 2 ⎥
⎝ 2σ ⎠
⎣⎢⎝ 4 ⎠ r F ⎦⎥
(S22)
-9-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Note that the expression for P↑ − P↓ is independent of time, and we may choose any time for its
evaluation that is long enough compared with r −1/ 2 so that the approximations we have made are
good, but short enough so that the nonlinear term in the evolution equation is not yet important.
Such a time always exists for the limit of small signal and noise of interest. Also note that the
effective bandwidth for the noise appearing in the “signal to noise” ratio in the argument to the
error function in Eq. (S22) is r1 / 2 and is determined by the frequency ramp-rate.
As we show below in Eq. (S27), the BTA output signal is proportional to P↑ − P↓ . For the limit of
small signals Eq. (S22) reduces to
⎛4⎞
P↑ − P↓ ≈ ⎜ ⎟
⎝π ⎠
1/ 4
r
1/ 4
s
.
F 1/ 2
(S23)
For M sweeps through the bifurcation, the distribution of the fraction of up traces follows
Poisson statistics, which for large M reduces to a Gaussian with mean P↑ , and standard deviation
1 / 2M for P↑ close to 1/2. For a sweep rate ν sw = ω sw / 2π , this leads to the error estimate for
the measurement of the signal s of
( π rS
F
/ 4ν sw
)
1/ 2
/ Hz .
The noise term f(t) in Eq. (S17) ultimately derives from physical noise forces on the beams
~ ~
f ± ( t ) with
~ ~
~ ~ ~ ~
~
f ± ( t ) f ± ( t ′) = 2 Fδ ± δ ( ~
t −~
t ′) leading to terms f ± ( t ) / m on the right hand side of
Eq. (S1). Proceeding through the transformations as in Eqs. (S2)-(S16) leading to the evolution
equation for X2, but now including this noise term, relates the force strength in Eq. (S17) to these
fundamental forces. For small beam coupling, the dominant noise source is just from the higher
frequency beam, and then we find
F=
α
1
2 (mω 0 2 ) 3
⎛ QH ⎞ ~
⎟ F.
⎜
⎝ 2 ⎠
2
(S24)
-10-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Fabrication and Methods
Device Fabrication: Coupled-NEMS BTA
Fabrication is based on a GaAs-based multilayer grown by molecular beam epitaxy (MBE) upon
a GaAs substrate. The device structural layer itself is a 200nm thick multilayer structure,
comprising a stack of three layers forming a vertically-oriented (i.e. out of the wafer plane) p-i-n
diode: a topmost 100 nm n-GaAs (1019 cm-3) layer, a 50 nm i-GaAs layer, and a 50 nm p-GaAs
(1018 cm-3). The i-GaAs layer has a p-type background concentration of ~5x1015cm-3 arising
from natural impurities, which is negligible compared to the intentionally doped regions. These
structural p-i-n diode layers are grown on a sacrificial p-Al0.8Ga0.2As (1018 cm-3) layer, which in
turn is grown on a p-doped (1018 cm-3) GaAs (001) substrate by MBE. The lead frame structure
and wire bond pads are patterned by photolithography. This is followed by deposition of a thin
~5nm Ti adhesion layer and a 50nm Au layer, subsequently standard liftoff is employed. The
backside of the p+-doped wafer is coated with Ti/Au in order to provide a bottom electrical
contact. The NEMS devices themselves are defined by electron beam lithography, which is
followed by deposition of a 60nm Ti layer and liftoff. This Ti mask layer protects the desired
structural regions during a dry etch using argon ion-beam milling to a depth of 250 nm.
Subsequently, the patterned devices are suspended by removing the sacrificial Al0.8Ga0.2As layer
using a selective, wet chemical etch in dilute hydrofluoric acid. This step also removes the Ti
masks, exposing the Au electrodes and pads.
Coupling
Two different physical phenomena contribute to the coupling between beams: elastic mechanical
coupling mediated through the substrate, and an electrostatic dipole-dipole interaction between
-11-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
adjacent p-i-n diode
structures in the
voltage-biased beams.
We discuss each
mechanism in turn;
they are pictorially
Figure S1. We model the interactions between beams as arising from
two mechanisms: elastic coupling through the substrate (right) and
electrostatic dipole-dipole interaction (left). Device geometry is
designed so that the coupling forces are of the same order.
depicted in Figure S1.
Electrostatic Coupling. When a DC bias voltage is applied to both beams, charges of opposite
sign accumulate on their top and bottom surfaces forming dipole moments. Two such identical
dipoles interact electrostatically with a force (3)
Fdipoles
(ε 0 εAbeam ) 2
=−
4πε 0
d5
9
⎛
t ⎞
⎜⎜Vbias + ψ 0 m ⎟⎟ ( x1 − x 2 ) ,
ti ⎠
⎝
2
(S25)
where d is the distance between the beams, Abeam is a beam’s surface area L × w, x1 and x2 are
the out-of-plane displacements at the center of the beams, ψ0 ≈ 1.2V is a built-in potential,
ti = 50nm is the thickness of the insulating layer, and tm ≈ 78nm is the depletion width of the p-i-n
diode (4).
This coupling has a simple linear form with a negative coefficient, meaning a repulsive
interaction ensues. This arises from the fact that, with the same voltage applied to both beams,
dipole moments of the same orientation develop in the beams. For the geometry of the devices
used in these experiments ( d=400nm, Abeam = L×w = 6µm×0.5µm, and Voffset ≈ 1.1V ), a force
of approximately −1pN is generated from a 1nm difference in displacements.
-12-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Mechanical Coupling. Without any applied voltage there is residual attraction between the
beams due to elastic coupling through their shared elastic support, even though no ledge is
shared by the beams. Finite element numerical simulations are used to estimate the magnitude of
elastic coupling mediated through the substrate, the displacement colormap of the mechanically
interacting beams is shown in Figure S1. We find the interaction to be attractive and linearly
dependent on displacement difference. We design the geometry of the system so that at the
difference in displacements of ~1nm, the effective mechanical coupling force is approximately
~1pN, as a result the dipole–dipole interaction compensates the elastic coupling within
experimentally accessible voltage range.
Summary of Measurement Apparatus and Methods
Samples are mounted in a room temperature vacuum chamber, which is pumped down to a
typical pressure of 5 mTorr for experiments. This chamber is fitted with a transparent sapphire
optical window to enable optical interferometry, however to minimize the impact of spurious
light on device performance (for example, due to heating and inadvertent generation of
photocarriers) we place a neutral density filter with a 10-fold extinction factor in front of the
optical port. Illumination is provided by an infrared laser diode emitting 2 mW at 904 nm. The
laser is focused to a spot of ~10 μm in diameter upon the device. The reflected signal is detected
by a low-noise, high-bandwidth photoreceiver (New Focus 1801, bandwidth=125 MHz, optical
noise power spectral density = 30 pW/√Hz, referred to input).
-13-
R.B. Karabalin et al.
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
We calibrate the displacement response of
the
interferometer
using
the
20
known
amplitude at the onset of nonlinear response
for doubly-clamped beams, which arises
from the Duffing instability.
Direct
frequency response measurements for the
|a1+a2| (nm)
15
10
5
0
13.08
coupled resonators are made using an RF
vector network analyzer (Hewlett Packard
3577A) as shown in Figure S2.
The
amplitudes of the two peaks are different
13.12
13.16
13.20
13.24
Frequency (MHz)
Figure S2. Duffing nonlinearities dominate at
high excitation levels. The onset of nonlinear
behavior depends only on quality factor and
geometry of the beam, hence it provides a effective
way of calibrating the optical interferometer
responsivity.
due to slight variations in their actuation efficiencies. The family of curves displayed represents
drive amplitudes from 50mVrms to 1.2Vrms. The onset of nonlinearity occurs for a ~600mVrms
drive level, which yields an optical signal of approximately 70µVrms. The estimated accuracy of
this calibration is of order 10%.
We have configured the sample geometry and experimental apparatus so that the laser spot
illuminates both NEMS resonators simultaneously.
In this case the output from the
photodetector represents the summed contribution from the coupled beams. As described in the
main text, the response to a parametric pump signal in the 26MHz range is for both beams to
become excited when the pump is roughly twice their natural resonance frequency. Depending
on the sign of Vin (t ) the pump induces coupled vibrations that are either in-phase (yielding
strong optical reflection) or out-of-phase (giving a weak optical response). Parametric frequency
sweeps are measured with a spectrum analyzer (Agilent 4395A).
-14-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
BTA Measurement Protocol
To evaluate the BTA’s performance we first null the beam-beam coupling with a DC input signal
set to Vin( DC ) = 0V. Then, to simulate a small-signal input to be amplified, we add to this static
input a simple single-tone, square-wave periodic waveform of the form Vin( AC ) (t ) = vin sq(ω in t ) ,
where sq(ω in t) is a square-wave function that changes from -1 to +1 with a period of
Tin = 1 / ω in .
We use typical input frequencies ω in / (2π ) ~ 170Hz.
This can be increased
without significant change to the output signal if ω in remains a factor of 2 lower than the sweep
rate discussed below.
A function generator (Agilent 33250A) is used to source both the
(summed) DC and AC input voltages simultaneously; this summed signal is applied to both
NEMS actuation electrodes via a DC/RF bias tee.
The function generator (Agilent 33250A) used to provide the aforementioned ~26MHz
parametric pump signal, provides is own internal frequency modulation in the form of a
triangular sweep signal (ramp waveform) to provide an output with instantaneous frequency
ω (t) = ω bif + ⎡⎣ −ω offset + Δω sw rmp(ω sw t)⎤⎦ .
(S26)
Here, rmp(ω sw t) is a triangle function that changes from 0 to 1 and back again with period
Tsw = 1 / ω sw .
This frequency modulation serves to sweep the pump signal through the
bifurcation point, ω bif , starting from a frequency ω bif − ω offset below it, to a frequency
ω bif + ⎡⎣ Δω offset − Δω sw ⎤⎦ above it. In most of our measurements an FM rate ω sw / (2π ) = 557Hz is
used. However, for our study of BTA amplification bandwidth, as described in the main text,
this was varied between 70Hz and 3kHz. As mentioned, a 3dB decrease in gain was observed for
ω sw / (2π ) ~ 2kHz.
-15-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
A reference signal at half the
pump frequency (~13MHz),
in the range of the natural
resonance frequencies of the
NEMS
is
required
to
demodulate the photodetector
output.
This signal, of the
form VLO cos [ω (t)t + ϕ LO ], is
generated by phase locking a
separate
voltage-controlled
oscillator to the pump, as
shown in Figure S3. To lock
Figure S3.
Schematic diagram of the bifurcation-topology
amplifier setup. A half-frequency reference is obtained by
synchronizing an additional voltage-controllaboe RF source
through a phase locked loop. The amplitude demodulation of the
output signal is performed by mixing it with the reference. An
oscilloscope is used in place of a lock-in amplifier for time domain
response measurements.
the pump and reference generators, an error signal is developed by homodyne converting the
doubled reference output against the pump signal. The reference signal generated in this manner
serves as the local oscillator for homodyne conversion of the photodetector’s radio frequency
output signal. The reference phase ϕ LO is set by a phase shifter to maximize the mixer’s phasesensitive baseband output.
Consequently, the measured output of the BTA gives the r.m.s. summed displacements of the
two resonators, which we express in nanometers. The summed displacement is averaged over the
frequency sweep range, and over many sweeps, to give
Vout = P↑ a↑ + P↓ a↓ =
a↑ + a↓
2
+ (P↑ − P↓ )
a↑ − a↓
2
,
(S27)
-16-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
where P↑ ( P↓ ) is the probability of following the upper (lower) branch, and a↑ ( a↓ ) is the
averaged summed displacement of the two resonators along the upper (lower) branch. Thus, to
within a constant shift, the BTA output is proportional to the difference in probabilities P↑ − P↓ ,
and to the area that is confined between the two branches.
For time-domain measurements of the demodulated signal, a digital oscilloscope (Agilent
54622D) is employed, as shown in the inset of Fig. 3b in the main text. For frequency and
amplitude sweeps to characterize BTA performance, the baseband signal is fed to a lock-in
amplifier (Stanford Research System SR830), which is externally synched to Vin(t).
-17-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
Noise and Sensitivity Analysis
Noise and Sensitivity Analysis for the Current Implementation
In our experimental implementation of the BTA, the output signal is proportional to the
difference in probabilities P↑ − P↓ (S27). We fit the measured output signal to the predicted error
function given by Eq. (S22). From the fit, shown in Fig. 4(c) of the main text, using the known
parameters of the beam, we extract the effective force noise strength S F1 / 2 acting on the beam.
We find an effective force noise of 1.8 × 10−13 N / Hz in our experimental setup, which
corresponds to effective charge sensitivity of 39 e/√Hz.
Alternatively we can determine the sensitivity of our measurement setup by direct noise
measurements. The inset of Fig. 4a displays the smallest input signal characterized in this study,
Vin
( AC )
= 1 mV. This corresponds to about 70e (e=1.6×10-19 C) in the conducting layers of the p-
i-n stack in each resonator. A Gaussian fit shows that the magnitude of the induced displacement
response is ~1.22 nm. From this we obtain a displacement-to-charge responsivity for the BTA of
∂x / ∂q ~ 17 pm/e. For the measurement bandwidth Δ f employed (see below), intrinsic noise
measurements with no input signal yield a displacement noise in the output signal of
(S x Δf )1 / 2 ~ 126pm ,
which is consistent with the Poisson statistics of equal probabilities of
following the up and down branches. This corresponds to an electronic charge noise of
(S Δf )
1/ 2
q
= (S x Δf )
1/ 2
(∂q / ∂x ) ~ 7.3e .
For these measurements a four-pole Butterworth filter
(roll off = 24dB/octave) with a time constant of τ =0.3s was employed. The resulting sweep was
averaged N=22 times, thus yielding an equivalent noise bandwidth Δ f = 5 / (64τ N ) ~ 0.056
Hz. Accordingly, the demonstrated charge sensitivity is Sq1/2 ~ 31 e/√Hz. This result is consistent
with our previous estimate obtained by fitting the amplifier transfer function to an error function.
-18-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
In our present implementation, the noise in the system is primarily limited by non-idealities in
the measurement system; we believe significant further optimization is possible.
Principal Fundamental Noise Mechanisms of a Coupled-NEMS BTA
Assuming the displacement sensitivity of the measurement setup is improved to the extent that
the thermomechanical noise of both resonator can be measured we then analyze the limits of the
sensitivity of coupled-NEMS BTA. Among the fundamental noise sources are:
I. Thermomechanical motion of the NEMS resonators. The thermomechanical noise force is
caused by the Brownian motion of the NEMS resonator. Near the bifurcation point one
resonator is excited while the other is still dormant, and so the contribution of the
thermomechanical motion of the first resonator is negligible. The noise floor of the second
resonator is set by the thermomechanical force noise power spectral density,
S F1 /,th2 = 4k B Tmω + / Q , where kB is a Boltzman constant, T is room temperature, and m is the
effective
mass
of
the
resonator.
For
our
beam
dimensions
this
value
is
S F1 /,th2 = 1.6 × 10 −15 N / Hz .
II. Thermal charge fluctuations on the beam electrodes. These serve to induce a fluctuating
force on the BTA which can drive transitions at the bifurcation point. Their magnitude can
be estimated using the equipartition theorem for the charge noise in the beam electrode,
< q 2 > /C = k BT . This energy is distributed over the RC bandwidth, with R the load
resistance and C the interlayer electrode capacitance. By separating the measurement
electrodes from the drive electrodes (not done in the present implementation), and by tuning
the external load resistance, the charge fluctuation noise can be confined to a narrow region
-19-
R.B. Karabalin et al.
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
outside the measurement bandwidth, so that this noise source becomes negligible in the
measurements.
III. Noise in the beam coupling. To estimate this contribution we calculate the electrostatic
interaction force to next order in the relative displacement, O( x1 − x 2 ) 3 . This is given by
Fdipole ~ Δ1 (x1 − x2 ) + Δ 3 (x1 − x2 )3 , where we find from an analysis of the electrostatics
Δ 3 = 75Δ 1 / (2d 2 ) .
Therefore
the
equivalent
force
noise
spectral
density
is
x2 , where x1 and x2 are the r.m.s. displacements of the first and
SF1/2
= Δ 3 x1 S1/2
x1
dipole
second resonators respectively, and Sx1/2
is the spectral density of the thermomechanical
1
displacement thermal noise of the first resonator. For the system of coupled beams employed
2
= 8.2 × 10 −18 N / Hz , which
in our BTA implementation the coupling force noise is S F1 /,dipole
is 200 times smaller than the thermomechanical noise calculated in I, above.
Projections of the Ultimate Sensitivity of a Coupled-NEMS BTA
The dominant thermomechanical noise source, calculated in I of the previous section, is used to
estimate the fundamental limits to the charge sensitivity of our first implementation of the
coupled-NEMS BTA that are quoted in the main text. For our present device we estimate that a
thermodynamically limited sensitivity of 0.46e/√Hz should be achievable at room temperature.
For a 1GHz device (6), with reduced capacitance due to its smaller geometry, operating at
300mK as a base temperature (7), we estimate that a sensitivity of order 1×10-5 e/√Hz is feasible.
In order for such device to be implemented at cryogenic temperatures the phase sensitive
transduction scheme needs to be realized. At room temperature we used optical interferometry
-20-
Supplementary information: “Signal amplification by sensitive control of bifurcation topology”
R.B. Karabalin et al.
which is capable of distinguishing in and out of phase motion. On the one hand optical detection
techniques proved to be compatible with cryogenic measurements(8). Alternatively there are
wide variety of all-electric transduction techniques ranging from electrostatic, to piezoelectric, to
magnetomotive that are phase sensitive, so that BTA architecture can be implemented.
References
1. R. Lifshitz, M. C. Cross, Response of parametrically driven nonlinear coupled oscillators
with application to micromechanical and nanomechanical resonator arrays. Physical
Review B 67, 134302 (2003).
2. R. Lifshitz, M. C. Cross, Nonlinear dynamics of nanomechanical and micromechanical
resonators. in Annual Review of Nonlinear Dynamics and Complexity. Edited by H.G.
Schuster (Wiley, New York, 2008) pp. 1–52.
3. H. X. Tang, X. M. H. Huang, M. L. Roukes, M. Bichler, W. Wegscheider, Two-dimensional
electron-gas actuation and transduction for GaAs nanoelectromechanical systems.
Applied Physics Letters 81, 3879-3881 (2002).
4. S. C. Masmanidis et al., Multifunctional nanomechanical systems via tunably-coupled
piezoelectric actuation. Science 317, 780 (2007).
5. A. N. Cleland, M. L. Roukes, Noise processes in nanomechanical resonators. Journal of
Applied Physics 92, 2758–2769 (2002).
6. X. M. H. Huang, C. A. Zorman, M. Mehregany, M. L. Roukes, Nanodevice motion at
microwave frequencies. Nature 421, 496 (2003).
7. A. Naik et al. Cooling a nanomechanical resonator with quantum back-action. Nature
443, 193–196 (2006).
8. D. Rugar, R. Budakian, H. J. Mamin, B. W. Chui, Single spin detection by magnetic
resonance force microscopy. Nature 430, 329(2004)
-21-