Paper Lab on a Chip
microscopy was employed as these particles were too small to
be quantified in the Multisizer 3 Coulter Counter used in
this study. Visually, it could be observed that 0.5 μm diameter
fluorescent polystyrene particles could be focused in
both the square (Fig. 6, Table 2) at 2 μL min−1 and the rectangular
cross-section channels at 0.5 μL min−1 when using
two-dimensional focusing (data not shown). When using
one-dimensional focusing in the rectangular cross-section
channel, the particles could not be completely focused, which
is consistent with our previous results. The focusability of
0.24 μm fluorescent particles was also investigated, but these
particles could only be seen to stream and they could not be
focused in any channel (data not shown), placing the new
critical particle diameter somewhere between 0.24 μm and
0.5 μm.
Bacteria focusing
A suspension of E. coli was also investigated to evaluate the
biological relevance of the systems. The bacteria showed a relative
focusability R = 0.95 ± 0.35 in the square channel with
two-dimensional focusing, whereas it was only R = 0.40 ± 0.13
in the rectangular channel using one-dimensional focusing
(Table 3).
In these experiments, we deliberately kept the concentration
of particles and bacteria below 109 mL−1 to avoid the complication
of particle-particle interaction due to hydrodynamic
Table 3 Highest relative focusability achievable for E. coli and 1 μm
and 0.6 μm diameter polystyrene particles
Particle 1D rectangular 2D rectangular 2D square
E. coli 0.40 ± 0.13 0.95 ± 0.35
1 μm 0.52 ± 0.17 0.87 ± 0.1 0.95 ± 0.08
0.6 μm 0.48 ± 0.07 0.92 ± 0.34 1.04 ± 0.1
coupling of the particles.41 Also, in future microbiological
applications, the need for bacterial enrichment is most evident
in samples with very low concentrations of bacteria. In
contrast, previously reported focusing of E. coli based on a
one-dimensional standing wave used a high sample concentration
of 1010 mL−1, which caused the bacteria to agglomerate
and effectively act as larger particles.42
Comparison of the experimental and numerical streaming flow
The experimental data demonstrate that sub-micrometer
particles as small as 0.5 μm can be focused using twodimensional
acoustic focusing, which indicates that these
systems are dominated by a streaming velocity field similar
to that in Fig. 2(d) rather than Fig. 2(c). Importantly, the
centred streaming roll derived in Fig. 2(d) has also been
observed visually in some parts of the channel (video S1 of
ESI†), while other parts appear to be “quieter” (i.e. not
showing much streaming activity). This is consistent with the
assumption that the vibration of the walls most likely
changes along the channel. At different positions along the
channel, the streaming rolls observed did not all move in the
same direction: both clockwise and counter-clockwise streaming
rolls were seen. Analogous streaming patterns have also
been observed in acoustic resonance cavities with almost
square geometry.43
Based on the experimental data and the numerical simulation,
we hypothesise that the centred streaming roll in combination
with two-dimensional focusing is the predominant
effect along the full length of the channel, which enables
focusing of sub-micrometer particles in the experimental
square channel system presented in this paper.
Conclusions
This paper reports the successful use of acoustophoresis
to focus sub-micrometer cells and particles. The use of twodimensional
actuation of a square channel was found to
enable two-dimensional focusing of E. coli and polystyrene
particles as small as 0.5 μm in diameter with recovery above
90%, something that could not be achieved using onedimensional
focusing. This sets the experimental limiting
particle diameter for continuous-flow half-wavelength resonators
operated at about 3 MHz to somewhere between 0.25 μm
and 0.5 μm for particles and bacteria with acoustic properties
similar to those of polystyrene suspended in water. The
focusing of sub-micrometer particles is enabled by a streaming
velocity field consisting of a large centered flow roll
that does not counteract the weak two-dimensional focusing,
Fig. 6 Fluorescent image of 0.5 μm particles (red) focusing in the
square-cross-section channel at a flow rate of 2 μL min−1 and at the
same voltage as used for the 0.6 μm particles in the square-crosssection
channel focusing experiments. The broken gray lines show the
edges of the channel.
Table 2 Highest relative focusability achieved for 0.5 μm and 0.6 μm
diameter polystyrene particles
Focusing method Particle Relative focusability Flow rate Q
1D rectangle 0.6 μm 0.48 ± 0.07 3 μL min−1
2D rectangle 0.6 μm 0.92 ± 0.34 3 μL min−1
2D square 0.6 μm 1.04 ± 0.1 5 μL min−1
2D square 0.5 μm 1a 2 μL min−1
SAW14,c 0.5 μm 0.79b 1.8 μL min−1
SAW15 0.5 μm 1a 0.2 μL min−1
a No recovery data obtained. Visual focus of particles shown.
b Recovery, no focusability data available. c Device uses a combination
of dielectrophoretic and acoustic forces.
2798 | Lab Chip, 2014, 14, 2791–2799 This journal is © The Royal Society of Chemistry 2014