Lab on a Chip

View Article Online

PAPER View Journal | View Issue

A sheath-less combined optical and impedance

micro-cytometer†
Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

Cite this: Lab Chip, 2014, 14, 3064

Daniel Spencer,‡ Gregor Elliott‡ and Hywel Morgan*

We describe a sheath-less micro-cytometer that measures four different parameters, namely fluorescence,
large angle side scatter and dual frequency electrical impedance (electrical volume and opacity). The
cytometer was benchmarked using both size and fluorescent bead standards and demonstrates excellent
Received 20th February 2014, size accuracy (CVs ≤ 2.1%), sensitivity and dynamic range (3.5 orders of magnitude) at sample flow rates of
Accepted 5th June 2014
80 μL per minute. The cytometer was evaluated by analysing human blood, and a four part differential
leukocyte assay for accurate CD4+ T-cell enumeration was demonstrated. The integration of impedance,
DOI: 10.1039/c4lc00224e
fluorescence and side scatter into a single miniature cytometer platform provides the core information
www.rsc.org/loc content of a classical cytometer in a highly compact, simple, portable and low cost format.

1. Introduction haematology, where particle volume is measured using elec-

trical (Coulter) methods.7
Cytometry is widely used for the analysis of particles such as Microfluidic Impedance Cytometry (MIC) has been devel-
cells and beads, with application in areas as diverse as medi- oped to both count and discriminate cells. Multi-frequency
cine, diagnostics1–3 and oceanography.3–6 Particle identifica- impedance measurements are used to determine the dielec-
tion is usually performed with a combination of laser light tric properties of single particles.13–15 Cells flow between two
scatter, laser induced fluorescence and electrical volume pairs of miniature electrodes which have an AC field applied
(Coulter) analysis,1,7 or in specialist systems using fast pho- across them. As the cell passes between the electrodes, the
tography and image recognition.4,8,9 Commercial cytometers current path is disturbed and the change in current gives a
are too large or costly for use anywhere other than in dedi- single cell impedance signal. At low applied signal frequen-
cated analysis laboratories. Therefore much work has been cies the technique provides accurate cell sizing where the
undertaken to develop micro-cytometers that would be impedance signal is proportional to cell volume. Higher
smaller, cheaper and robust enough for use outside the frequency impedance measurements (1–5 MHz in saline) give
lab.3,10,11 Micro-cytometers have the potential to improve information on the cell membrane capacitance whilst much
medical diagnosis by giving fast and accurate information of higher frequencies (>10 MHz) probe the internal properties
a wide variety of conditions. Small and compact cytometers of the cell.13 Two or more frequencies can be applied simul-
are also of interest in environmental science, for example in taneously to differentiate different cell types, for example
monitoring water to detect and classify algae.12 white blood cell (wbc) subpopulations.14 However, impedance
Generally, particles are analysed using optical scatter; cytometry cannot provide the information on cell phenotype
small angle forward scattered light (FSC) provides size infor- that is obtained by labelling cells with fluorescent antibodies.
mation, whilst larger angle side scatter (SSC) gives informa- As an analogue of fluorescent labelling, small dielectric parti-
tion on internal structure and cell viability. Fluorescence cles can be used as impedance labels, but this technique can
detection provides further information on cell phenotype. For only identify a single subpopulation.16 There is therefore a
example, leukocytes are differentiated with fluorescent anti- need to combine the simplicity and robustness of single cell
bodies. Algal cells can be discriminated based on the spectral impedance measurements with fluorescent interrogation
response of fluorescence from specific photosynthetic pig- methods in a single common microfluidic platform that is
ments in a particular species. Cytometers are widely used in both easy to manufacture and has comparable performance
to conventional large-scale flow cytometers.
Faculty of Physical Sciences and Engineering, and Institute for Life Sciences, A variety of technologies have been demonstrated in the
University of Southampton, Southampton, Hampshire, SO17 1BJ, UK.
quest for a micro-cytometer that matches the performance of
E-mail: [email protected]
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
a conventional machine – for reviews see ref. 10, 11 and 13.
c4lc00224e In early work Wolff et al.17 described a miniature fluorescence
‡ Both authors contributed equally to this work. cell sorter fabricated in silicon, which used a waveguide to

3064 | Lab Chip, 2014, 14, 3064–3073 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Lab on a Chip Paper

couple laser light into a channel and a microscope to collect reported. For example Hur et al. used inertial forces to focus
fluorescence. Holmes et al.14 combined an impedance system blood cells into 256 parallel microchannels.28 High-speed
with a lab based confocal microscope to measure fluores- imaging was used to analyse blood cells at a theoretical
cence. Although the system could distinguish fluorescently throughput of up to 1 million per second, however only 8000
labelled cells, it was very sensitive to particle position in the cells were measured. The use of inertial focusing restricts
channel and consequently had a high coefficient of variation simultaneous measurement of heterogeneous populations
(CV) in fluorescence. Segerink et al. described a simple (RBCs had to be sphered prior to measurement). Curved
cytometer that detected fluorescence18 using an optical microchannels and Dean forces have also been used to focus
pickup from an HD-DVD player. They demonstrated that high particles.29 Acoustic, and dielectrophoretic focusing tech-
precision mass manufactured optical components provides a niques have also been described.30,31 However, in all these
Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

cheap way of integrating optics. The system scans across the methods, the focusing force depends on particle size. The
channel, therefore eliminating the need for fluidic focusing, inertial migration force scales with the square of particle
but the fluidic throughput is limited to below 720 pL s−1, with radius, whilst acoustic and DEP forces scales with particle
a very low particle throughput of 1–2 beads per second. volume. In this paper we describe a micro-cytometer that
Optical fibres are commonly used to deliver and collect does not require particle focusing. It measures particle
light from micro-cytometers. For example, Wang et al.19 fabri- impedance, fluorescence and large angle side scatter, with
cated a cytometer that incorporated fibres coupled to wave- volumetric throughput, sensitivity and dynamic range compa-
guides to discriminate beads using scattered light, whilst rable to a commercial flow cytometer. Signal processing is
Tung et al.20 used micro-groves for fibre alignment to detect used to correct the impedance signal for particle position
fluorescently labelled yeast cells with PIN diodes and lock-in within the channel.32 The optical excitation volume is
amplification. designed to have uniform illumination and a thin metal
Although most micro-cytometers are manufactured using screen is fabricated in the chip that minimises stray scattered
planar lithography, devices have also been manufactured light coupling into the collection optics. Light is coupled into
from milled plastic. Ligler et al. have developed devices that the micro-channel using an integrated waveguide, and both
use chevrons for particle focusing and optical fibres to deliver fluorescence and SSC is measured using simple off-chip
and collect light at the point of interrogation.21,22 Neukammer optics. The optical, fluidic and electrical connections are all
and co-workers23,24 manufactured a cytometer using a combi- designed so that the chip can be replaced easily in a small
nation of micromachining and hot-embossing. The device used holder. Fluorescence sensitivity is evaluated using LinearFlow
sheath flow and detected fluorescence and scatter using reference intensity beads. High accuracy impedance sizing is
optical fibres. They also included electrodes to measure demonstrated using size calibration beads. The application
impedance. of the technology to haematology is demonstrated by measur-
We recently demonstrated a micro-cytometer that mea- ing different sub population of leukocytes, including CD14
sures optical and impedance parameters within a single monocytes and CD4+ T lymphocytes.
chip.25 Light was delivered and collected from the channel
using integrated optical fibres and 1D sheath flow was used 2. Experimental
to focus particles. However, the cytometer suffered from
several technological issues including misalignment of the The microfluidic chips (Fig. 1) were made as described previ-
optical fibres, incident light scatter from multiple interfaces ously.14 Platinum electrodes (30 μm wide) were patterned
and signals that strongly depended on particle position onto glass substrates; channels and integrated optics were
within the interrogation volume. All these problems degraded made from patterned SU8 with full wafer bonding in a vac-
the CV of the measured particle populations. In many uum bonder. Individual chips (25 × 10 mm) were diced from
cytometers, the optical fibres are integral to the design. the wafer. The fluidic channels had cross sectional dimen-
Inserting and aligning fibres is labour intensive, prone to sions of 30 μm high and 80 μm wide at the point of interro-
error and does not provide for a modular system where chips gation. A simple 1D hydrodynamic focusing region was
can be easily replaced if they become contaminated or included upstream from the measurement region. This was
clogged. A modular cytometer that has interchangeable chips used to examine the fluorescence measurement performance
without integrated fibres would therefore be of significant with and without hydrodynamic focusing. To avoid any
advantage and could be used as part of a simple miniature potential blockage with large particles, these channels are
system with a disposable consumable. wider than previously used. The integrated optics, patterned
Nearly all micro-cytometers use some form of particle in the SU8 layer, consisted of a waveguide running from the
focusing for high quality (low CV) data. Typically, sheath flow edge of the chip (Fig. 1a) towards the edge of the channel.
is used to focus particles, and even in a micro-system this The waveguide was terminated with a cylindrical air lens
consumes of the order of 1 mL per minute (see ref. 21–24, (Fig. 1b) patterned in SU8, which focused the light into the
26–27). This imposes significant restriction on the technology channel. The optical path was designed using free-space opti-
for use at the point-of-care or for continuous monitoring. cal calculations as described by Rosenauer et al.33 The total
Sheath-less particle focusing techniques have also been optical path length from the chip edge to the centre of the

This journal is © The Royal Society of Chemistry 2014 Lab Chip, 2014, 14, 3064–3073 | 3065
 View Article Online

Paper Lab on a Chip

channel is 5 mm. The waveguide is 300 μm wide at the edge 1.05%, 0.28%, 0.055%. Non-fluorescent 6 μm diameter beads
of the chip and tapers to 60 μm wide before the lens. The were also used.
lens design is a double concave air lens (Fig. 1b). In our Beads were suspended in phosphate buffered saline (PBS)
previous design,25 light scatter from the waveguide and lens at an approximate number density of 200 beads per microliter.
was a significant problem. In this design these elements were The sample was loaded into a syringe and pushed through the
screened using the same metal layer that formed the chip at a constant flow rate with a syringe pump (Chemyx
electrodes. A small window was created for detection (Fig. 1b). fusion 200). The device was design to operate without sheath
Impedance detection requires separate electrical contacts flow, but for a full evaluation of the device, samples were also
for the top and bottom electrode pairs. The design shown in measured using a 1-D sheath flow of PBS. The ratio between the
Fig. 1 solves this problem and only requires connection to sample and sheath flow was varied, but the total flow rate was
kept constant at 80 μL min−1 for ease of comparison. In each
Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

one side of the chip, with the top to bottom interconnection

made individually using silver epoxy (see Fig. S1 in the ESI†). case, the same sample was also analysed using a BD FACSAria,
The chip was mounted in a plastic holder fabricated with with the flow rate set to give similar particle throughput.
a 3D printer (UP! Mini). The holder includes the fluidic and The sizing accuracy of the impedance micro-cytometer
electrical connections with access ports for the fibre and was evaluated using size calibration beads (3 μm, 4.5 μm,
detection optics (see Fig. S2†). Laser light is directed onto the 6 μm and 10 μm in diameter from Polysciences), pumped
side of the chip from a fibre, and an objective lens (×20), through the chip at 80 μL min−1 without sheath flow. For
filter set and PMT collect the fluorescence and side scattered leukocyte measurements, blood was collected from a finger
light. An image of the laser excitation light in the micro- prick and pipetted directly into EDTA tubes (MiniCollect
fluidic channel, together with a cross sectional profile is also 0.5 mL K3EDTA) which were placed on a roller to prevent
shown in Fig. S3.† The image shows that the light is uniform aggregation. Leukocytes were labelled with fluorescent anti-
across the channel and has a Gaussian profile along the bodies as follows. 50 μL of blood was incubated with 10 μL
direction of flow. of CD14-APC or CD4-APC (Miltenyi) antibody for 10 minutes
Single cell impedance was measured using an impedance at room temperature (on a roller). Erythrocytes were subse-
spectroscope (Zurich instruments HF2IS). The electrical quently removed by addition of 600 μL of lysis solution
signal from the electrodes in the channel was measured with (0.05% saponin 0.12% formic acid) and agitating for 6 seconds.
a transimpedance amplifier (Zurich instruments HF2TA). The reaction was stopped with the addition of 265 μL of
The fluorescence signal from the PMT was plugged into the quench solution (3% NaCL, 0.6% Na2CO3), as described previ-
auxiliary port of the impedance spectroscope and sampled ously.14 Unbound antibody label was removed from the sample
simultaneously with the electrical signal. Signals were sam- by centrifuging at 300 g for 3 minutes, removing the superna-
pled at 230 ksps and post-processing was carried out using tant and re-suspending the cells in PBS. This sample was
custom software written in Matlab. Signal processing was measured with the microcytometer and a BD FACS Aria with a
used to correct for variation in impedance with particle posi- 633 nm laser.
tion in the channel.32
Fluorescence performance was evaluated using AlignFlow 3. Results and discussions
beads (L14819), which are used to calibrate and align conven-
tional flow cytometers. These beads are 6 μm diameter, fluo- 3.1 Fluorescence performance
resce at 660 nm (when excited at 635 nm), and have relative The micro-cytometer was designed to operate without sheath
fluorescence intensities corresponding to 100%, 25%, 5%, flow, but the BD FACSAria uses a constant pressure to provide

Fig. 1 (a) Schematic diagram of the micro-cytometer chip showing the electrodes, the integrated waveguide that terminates on the edge of the
chip, butt-coupled to an optical fibre for light delivery. Fluid inputs are shown for sample delivery and optional sheath flow. (b) Close up of the
measurement region, with the lens, waveguide and impedance detection electrodes, together with the metal screen.

3066 | Lab Chip, 2014, 14, 3064–3073 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Lab on a Chip Paper

sheath flow. The sample pressure is varied on a relative scale Table 1 shows that for both systems, the CVs increase as the
of 1–11, which effectively varies the diameter of the sample flow rate increases for all three flow rates. Importantly the CVs
stream. These figures correspond to sample volumetric flow of the micro-cytometer are comparable to the FACSAria across
rates from approximately 10 μL to 120 μL min−1. For the all flow rates, as is the percentage count for each bead popula-
experiments, the flow rates on the FACSAria were set to tion. The data also shows that the optimum results for the
approximately match the throughput on the micro-cytometer. micro-cytometer is obtained when using sheath flow to focus
Fig. 2 shows a section of a data stream showing simulta- the particles. However, removing the sheath has very little
neous impedance and fluorescence data for AlignFlow fluo- influence on the CV data for fluorescence, as seen by compar-
rescent beads. Fig. 2a–b shows raw un-processed signals for ing the data in Fig. 3(e) with (f), which shows a minor degra-
different beads (100%, 25%, 5%, 1.05%, 0.28%, 0.055%, 0% dation in performance. This data should be compared with
literature (e.g. Ligler,21,22 Neukammer23,24 and Huang26,27)
Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

relative intensities) demonstrating excellent discrimination.

The fluorescence signals from individual beads correlates where high volumetric sheath flows are required (typically
with impedance signals. An example is shown in Fig. 2c–d, of the order of 1 mL min−1) to give CVs comparable with
for a 1.05% intensity bead. The small 200 μs offset in time reference cytometers.
arises from the displacement of the waveguide from the
centre of the electrode pair by 245 μm (see Fig. 1b). This
offset was introduced to eliminate any potential effect of 3.2 Size accuracy
laser light on the impedance signals (through fluid heating). High accuracy electrical volume measurement was demon-
The data in Fig. 2 was obtained at a volumetric flow rate of strated using a mixture of 3, 4.5, 6 and 10 μm diameter beads
80 μL min−1, corresponding to a bead velocity of 0.8 m s−1. suspended in PBS at a density of 1000 beads per μL. The
Fig. 3 shows histograms for a single sample containing all bead suspension was flowed through the chip at 80 μL min−1
different bead intensities measured on the micro-cytometer with no sheath flow. In all cases the CV of the diameter (cube
and the BD FACSAria. Impedance signals were used to trigger root of volume) is better than that quoted by the manufac-
event detection on the micro-cytometer, and scatter on the turers (Table 2). This demonstrates the high accuracy particle
FACSAria. The ratio of sheath to sample in the micro- sizing capability of the micro-cytometer even without sheath
cytometer was varied as shown in the figure, with the total flow. Although forward scattered light is frequently used to
volumetric flow kept constant at 80 μL min−1. The flow rate measure particle size in flow cytometry, the dependence of
of the BD FACSAria was varied to give a similar sample Mie scatter with particle volume is non-linear, unlike imped-
throughput. The coefficient of variance (CV) and total particle ance which scales linearly with particle volume. Additionally,
counts for this data set are compared in Table 1. it is technologically challenging to incorporate collection
These histograms show that the lowest intensity fluores- optics into micro-devices that can collect low-angle scattered
cence (0.055%) beads could be distinguished from the zero light. Although this has been achieved, hydrodynamic
level on both systems, however, the dynamic range was insuf- particle focusing is required (Fig. 4).34,35
ficient to allow all 6 different intensities to be observed at the As demonstrated by this data, the device operates excep-
same time. Therefore the gain was set to allow the lowest tionally well without sheath flow. This brings significant
intensity particles to be detected, meaning that the 100% advantages since the fluidics are much easier and robust
intensity particles saturated the detectors, (data not included (one inlet and outlet), and the system does not require addi-
in Table 1). tional large volumes of clean and sterile sheath liquid.

Fig. 2 Raw data for 6 μm AlignFlow beads. (a) Impedance signals and (b) the corresponding fluorescence signals for six different bead intensities
(b). Fig. (c) and (d) show examples of impedance and fluorescence signals from a single particle, showing the very small offset in time between the
two signals.

This journal is © The Royal Society of Chemistry 2014 Lab Chip, 2014, 14, 3064–3073 | 3067
 View Article Online

Paper Lab on a Chip

Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

Fig. 3 Fluorescence histograms of a suspension of all 7 bead intensities, measured on the micro-cytometer and the FACSAria. The
micro-cytometer total sample rate was kept at 80 μL min−1 with variable sheath to sample ratio. Flow rates were (a) sample = 20 μL min−1, sheath =
60 μL min−1, (c) sample = 40 μL min−1, sheath = 40 μL min−1, (e) sample = 80 μL min−1, sheath = 0 μL min−1. The FACSAria sample flow rate was:
(b) 2/11 (approx. 20 μL min−1), (d) 4/11 (approx. 40 μL min−1), (f) 8/11 (approx. 80 μL min−1); see inset images.

3.3 Leukocytes from granulocytes. Discrimination of monocytes from neutro-

The utility of the device for cell analysis was demonstrated by phils requires a second higher frequency (2 MHz) that mea-
analysing CD14 labelled monocytes and CD4 T-lymphocytes sures cell membrane capacitance. Fig. 5a shows a scatter plot
with a combination of fluorescence, impedance and large for wbc differential measured on the micro-cytometer with
angle scatter. electrical volume plotted on the x-axis and electrical opacity
3-Part White Blood Cell (wbc) differential: simultaneous (ratio of high to low frequency) on the y-axis. Opacity mea-
impedance & fluorescence. We have previously shown that sures the cell membrane capacitance scaled with cell volume.
impedance cytometry can perform a simple 3-part wbc differ- The figure shows that the three major leukocyte populations
ential based on measurement at two separate frequencies.14 can be distinguished solely on the basis of their electrical
A low frequency signal (0.5 MHz) provides information on properties. A conventional FSC vs. SSC plot measured on a
cell size (volume) and discriminates smaller lymphocytes FACSAria is shown in Fig. 5b.

3068 | Lab Chip, 2014, 14, 3064–3073 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Lab on a Chip Paper

Table 1 Comparison of CVs for fluorescent populations as measured by the micro-cytometer and FACSAria. Micro-cytometer measurements were
taken at 80 μL min−1 total flow with ratio of sheath to sample varied. The sample throughput on the FACSAria was similar to the micro cytometer. Refer
to Fig. 3

Micro Micro Micro

cytometer FACS cytometer FACS cytometer FACS
Fig. 3 (a) (b) (c) (d) (e) (f)
Sample to sheath flow ratio 1:3 NA 1:1 NA No sheath NA
Measured throughput (beads s−1) 58 83 163 143 338 248

Fluorescence intensity (relative) CV (%) CV (%) CV (%) CV (%) CV (%) CV (%) Count (% total) Count (% total)
100 — — — — — — 2.30 2.21
25 6.04 6.49 6.70 7.80 10.7 8.3 14.4 14.0
Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

5 6.00 7.32 6.87 8.41 10.6 9.35 15.8 15.6

1.05 6.20 7.27 7.04 8.62 11.0 9.17 19.1 18.7
0.28 7.61 7.62 8.15 9.47 11.5 9.78 15.1 14.4
0.055 24.5 20.2 24.3 20.0 27.9 20.2 21.0 22.3
0 46.0 36.0 44.3 36.2 46.3 36.8 12.3 12.6

Table 2 Comparison of data for size calibration beads as given by the 200 cells μL−1, and increased frequency of clinical monitoring
manufacturer and measured with the micro-cytometer
is advised in patients with CD4 counts between 200 and
Manufacturer Measured 350 cells μL−1. Either the absolute CD4 count or the CD4%
Nominal
size (μm) CV (%) CV (%) can be used for diagnosis, but the absolute CD4 count is
much preferred.36–38 Disease is diagnosed by counting the
3 2.83 2.1
4.5 3.89 1.53 T-cells that express the CD4 antigen on their surface. Mono-
6 3.18 1.68 cytes also express the CD4 antigen, but at lower surface den-
10 2.99 1.7 sities than the CD4+ T-cells.39 It is important to distinguish
the CD4+ T-lymphocytes from these other CD4 expressing
Additionally the monocytes were labelled with a fluores- cells to obtain accurate CD4 T-cell counts. In optical cytome-
cent antibody (CD14-APC) and the fluorescence measured try, a fluorescent label is used to identify the CD4+ cells and
simultaneously with impedance. These fluorescent events are FSC and SSC discriminates monocytes from lymphocytes.
shown in red (Fig. 5), with the histograms of fluorescence Fig. 6(a) shows impedance-fluorescence scatter data for
intensities shown in Fig. 5c–d. The data for the micro- leukocytes labelled with CD4-APC. There are 5 classes of lym-
cytometer was collected without sheath flow and compares phocytes, of which approximately 50% express CD4 (Helper
favourably with the FACS data. The fine structures in the T-cells). This population is clearly identifiable in the figure.
histogram of Fig. 5d is an artefact which arises due to The monocytes also fluoresce but at a lower level, since these
quantisation of the data from the instrument and the subse- cells express five times fewer CD4 antibodies.39 Impedance
quent histogram bins. enables the lymphocytes to be distinguished from the larger
The ability to further distinguish sub-populations was neutrophils and monocytes. In optical cytometry FSC is used
demonstrated by counting CD4+ T lymphocytes. An absolute to size cells, but as shown in the FACSAria data (Fig. 6b), this
CD4+ T-lymphocyte count is widely used to monitor the does not differentiate lymphocytes from neutrophils, which
progression of HIV-AIDS. Initiation of antiretroviral treatment is normally done with SSC. However, the FSC signal is first
is recommended before the CD4 count falls below used to gate the smaller particles and debris (Fig. 6b) which
overlaps with the lymphocyte population in SSC, as shown in
the Fig. 6c. The final gates that are used to enumerate the
leukocyte sub-populations is shown in Fig. 6d. Finally, Table 3
compares the counts and ratios (%) of the micro-cytometer
with the FACSAria, and shows excellent agreement.
The micro-cytometer can also measure side scatter, providing
an additional parameter for cell identification (see Fig. S2 ESI†).
Fig. 7 shows a 3-D scatter plot for a different CD4 labelled
leukocyte sample. This plot shows side scatter, fluorescence
(APC) and low frequency impedance (particle volume), with
the impedance opacity (cf. Fig. 5) defining colour. The figure
shows the same four wbc sub-populations, each identifiable
according to one or more of the measured criteria.
Fig. 4 Histogram of the size (cube root of impedance) for a mixture
As shown by the bead and cell data, low-frequency imped-
of calibration beads. The sample was measured at a volumetric flow ance provides a very simple, yet accurate method for measur-
rate of 80 μL min−1 with no sheath flow. ing particle volume. Unlike optical analysis, the method is

This journal is © The Royal Society of Chemistry 2014 Lab Chip, 2014, 14, 3064–3073 | 3069
 View Article Online

Paper Lab on a Chip

Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

Fig. 5 Scatter plots of a wbc differential measured on (a) the micro-cytometer and (b) the FACSAria (total number of events = 15 000). The three
different cell sub-populations can be discriminated in both cases. The monocytes were labelled with fluorescent antibody (CD14-APC) and the
fluorescence histograms are shown in (c) and (d). The threshold level for fluorescent events (red) was set to −2.3.

easy to implement in a micro-device. Size or volume measure- 20 μL min−1,21–24 but similar to conventional cytometers
ment is an essential parameter in flow-cytometric analysis of (typical sample flow rates of 100 μL min−1). Our particle
cells. However, measuring particle size using small angle for- throughput is up to 1000 per second. Although this is
ward scattered light is very difficult to implement in micro- lower than the quoted maximum throughput of many
systems due to the small angles involved and the fact that bench-top cytometers, there is always a trade-off between
the incident light can couple into the detection optics, satu- speed and accuracy. At high throughputs (e.g. 10 000 per
rating the FSC detector. Godin et al.34 used blackened baffles second) coincidence becomes a problem. Simmonet and
in a chip, and Watts et al.35 fabricated a notch in the delivery Groisman were the first to demonstrate a high speed
optics to prevent incident light entering the FSC detection fluorescence-based micro cytometer with a maximum
waveguide. However in both cases the SNR was very low, and throughput of 17 000 per second.40 CV's were comparable
the CV in the FSC signal was much poorer than measure- to a commercial cytometer, however the device utilised a
ments using conventional cytometers. Additionally, both complex 3D hydrodynamic focusing architecture which
devices required sheath flow to centre particles in the required precisely balanced pressure driven flows for each of
channel. By contrast, impedance is very simple and can size the four inlets. To achieve maximum throughput the bead
particles with very high accuracy, for example with CVs of concentration was very high (~2.8 × 108 per mL) and the
1.5–2.1%, better than manufacturer's quoted data (2.8–3.9%) cytometer was not sensitive enough to measure fluorescently
and in the absence of sheath flow. Our micro-cytometer marked live cells.
has similar performance to the FACSAria at volumetric flow Throughput can be increased in many ways. The sampling
rates of up to 80 μL min−1 with cell counts of hundreds of rate of the electronics hardware places an upper bound on
cells per second despite our comparatively low sampling rate. the flow-rate. Conventional cytometers sample at around 10
The volumetric throughput is higher than comparable minia- Msps, whereas our electronics is limited to 0.23 Msps.
ture cytometers which mostly operate around 10 μL to Utilising high speed electronics would increase the maximum

3070 | Lab Chip, 2014, 14, 3064–3073 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Lab on a Chip Paper

Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

Fig. 6 Four part differential of white blood cells measured on (a) micro-cytometer and (b) FACSAria. (b) to (d) demonstrates the process flow used to
gate using FSC and SSC to determine the CD4 lymphocytes from the monocytes. Sample measured at a flow rate of 80 μL min−1 (no sheath flow).

Table 3 WBC differential measured on the micro-cytometer and

FACSAria

Micro-cytometer FACSAria
Count Relative % Count Relative %
Neutrophils 5580 55.8 5749 57.5
Monocytes 1116 11.2 1087 10.9
CD4+ lymphocytes 1600 16.0 1476 14.8
CD4− lymphocytes 1704 17.0 1688 16.9
Lymphocyte (CD4+/CD4−) 0.48 0.47

flow-rate by at least one order of magnitude (up to the

Nyquist limit). Fundamentally, throughput is limited by parti-
cle coincidence, which for a random distribution of particles
is defined by Poisson statistics. Particle concentration could Fig. 7 3-D scatter plot for CD4-APC labelled wbcs for side scatter,
be increased but to avoid coincidence, particles need to be fluorescence (natural log) and low-frequency impedance (electrical
laterally ordered using techniques such as inertial focus- volume). Each data point is coloured according to electrical opacity.
sing.41 Alternatively, throughput could be increased by using This multi-parameter plot demonstrates discrimination of the different
cell sub-populations.
multiple parallel streams.42

4. Conclusions
The device was benchmarked against a BD FACSAria using
A high accuracy modular micro-cytometer that incorporates fluorescent and size calibration beads. Application to
impedance, side scatter and fluorescence has been described. haematology analysis was demonstrated by enumerating

This journal is © The Royal Society of Chemistry 2014 Lab Chip, 2014, 14, 3064–3073 | 3071
 View Article Online

Paper Lab on a Chip

antibody labelled white blood cells. A combination of multi- High-throughput single-microparticle imaging flow analyser,
frequency impedance, SSC and fluorescence provides all the Proc. Natl. Acad. Sci. U. S. A., 2012, 109(29), 11630–11635.
primary metrics required for high speed single cell analysis. 10 D. A. Ateya, J. S. Erickson, P. B. Howell, L. R. Hilliard,
Since the cytometer is sheath-less, all that is required is a J. P. Golden and F. S. Ligler, The good, the bad, and the
single syringe pump for the sample. The system is modular, tiny: a review of microflow cytometry, Anal. Bioanal. Chem.,
with easy optical, fluidic and electric interconnects to the 2008, 391, 1485–1498.
chip. Impedance-based particle sizing outperforms particle 11 S. H. Cho, J. M. Godin, C. H. Chen, W. Qiao, H. Lee and
sizing from on-chip FSC, with excellent CVs ≤ 2.1%. The Y. H. Lo, Review Article: Recent advancements in optofluidic
device also has a large dynamic range in fluorescence with flow cytometer, Biomicrofluidics, 2010, 4, 043001.
CVs comparable to a BD FACSAria at the same high sample 12 N. Hashemi, J. S. Erickson, J. P. Golden and F. S. Ligler,
flowrate of 80 μL per minute. The integration of impedance,
Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

Optofluidic characterization of marine algae using a

fluorescence and side scatter into a single miniature microflow cytometer, Biomicrofluidics, 2011, 5, 032009.
cytometer platform provides the core information content of 13 T. Sun and H. Morgan, Single-cell microfluidic impedance
a classical cytometer in a highly compact, simple, portable cytometry: a review, Microfluid. Nanofluid., 2010, 8, 423–443.
and low cost format. A CD4 lymphocyte count from human 14 D. Holmes, D. Pettigrew, C. H. Reccius, J. D. Gwyer,
blood demonstrates that this device could be used for C. van Berkel, J. Holloway, D. E. Davies and H. Morgan,
haematology analysis away from a centralised lab. Leukocyte analysis and differentiation using high speed
microfluidic single cell impedance cytometry, Lab Chip,
2009, 9, 2881–2889.
Acknowledgements 15 S. Gawad, L. Schild and P. H. Renaud, Micromachined
impedance spectroscopy flow cytometer for cell analysis and
The authors acknowledge funding from the European Union
particle sizing, Lab Chip, 2001, 1, 76–82.
(DIMID project) and the Royal Society.
16 D. Holmes and H. Morgan, Single Cell Impedance Cytometry
for Identification and Counting of CD4 T-Cells in Human
References Blood Using Impedance Labels, Anal. Chem., 2010, 82, 1455–1461.
17 A. Wolff, I. R. Perch-Nielsen, U. D. Larsen, P. Friis,
1 H. M. Shapiro, Practical Flow Cytometry, John Wiley & Sons, G. Goranovic, C. R. Poulsen, J. P. Kutter and P. Telleman,
Hoboken, New Jersey, 2003. Integrating advanced functionality in a microfabricated
2 R. C. Sobti, Advanced flow cytometry: applications in high-throughput fluorescent-activated cell sorter, Lab Chip,
biological research, Kluwer Academic Publishers, Dordrecht, 2003, 3, 22–27.
The Netherlands, 2003. 18 L. I. Segerink, M. J. Koster, A. J. Sprenkels and A. van den Berg,
3 J. S. Kim and F. S. Ligler, The Microflow Cytometer, Pan A low-cost 2D fluorescence detection system for μm sized
Stanford Publishing, Singapore, 2010. beads on-chip, Lab Chip, 2012, 12, 1780–1783.
4 L. Campbell, R. J. Olson, H. M. Sosik, A. Abraham, 19 Z. Wang, J. El-Ali, M. Engelund, T. Gotsaed, I. R. Perch-Nielsen,
D. W. Henrichs, C. J. Hyatt and E. J. Buskey, First Harmful K. B. Mogensen, D. Snakenborg, J. P. Kuter and A. Wolff,
Dinophysis (Dinophyceae, Dinophysiales) Bloom in the U.S. Measurements of scattered light on a microchip flow
Is Revealed by Automated Imaging Flow Cytometry, cytometer with integrated polymer based optical elements,
J. Phycol., 2010, 46, 66–75. Lab Chip, 2004, 4, 372–377.
5 M. Thyssen, G. A. Tarran, M. V. Zubkov, R. J. Holland, 20 Y. C. Tung, M. Zhang, C. T. Lin, K. Kurabayashi and
G. Gregori, P. H. Burkill and M. Denis, The emergence of S. J. Skerlos, PDMS-based opto-fluidic micro flow cytometer
automated high-frequency flow cytometry: revealing tempo- with two-color, multi-angle fluorescence detection capability
ral and spatial phytoplankton variability, J. Plankton Res., using PIN photodiodes, Sens. Actuators, B, 2004, 98, 356–367.
2008, 30, 333–343. 21 A. L. Thangawng, J. S. Kim, J. P. Golden, G. P. Anderson,
6 R. J. Olson, A. Shalapyonok and H. M. Sosik, An automated K. L. Robertson, V. Low and F. S. Ligler, A hard microflow
submersible flow cytometer for analyzing pico- and nano- cytometer using groove-generated sheath flow for
phytoplankton: FlowCytobot, Deep Sea Res., Part I, 2003, 50, multiplexed bead and cell assays, Anal. Bioanal. Chem.,
301–315. 2010, 398, 1871–1881.
7 W. H. Coulter, High Speed Automatic Blood Cell Counter 22 L. C. Shriver-Lake, J. Golden, L. Bracaglia and F. S. Ligler,
and Cell Size Analyser, Proc. Natl. Electron. Conf., 1956, 12, Simultaneous assay for ten bacteria and toxins in spiked
1034–1040. clinical samples using a microflow cytometer, Anal. Bioanal.
8 R. J. Olson and H. M. Sosik, A submersible imaging-in-flow Chem., 2013, 405, 5611–5614.
instrument to analyze nano- and microplankton: Imaging 23 A. Kummrow, J. Theisen, M. Frankowski, A. Tuchscheerer,
FlowCytobot, Limnol. Oceanogr.: Methods, 2007, 5, 195–203. H. Yildirim, K. Brattke, M. Schmidt and J. Neukammer,
9 K. Goda, A. Ayazi, D. R. Gossett, J. Sadasivam, Microfluidic structures for flow cytometric analysis of
C. K. Lonappan, E. Sollier, A. M. Fard, S. C. Hur, J. Adam, hydrodynamically focussed blood cells fabricated by
C. Murray, C. Wang, N. Brackbill, D. Di Carlo and B. Jalali, ultraprecision micromachining, Lab Chip, 2009, 9, 972–981.

3072 | Lab Chip, 2014, 14, 3064–3073 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Lab on a Chip Paper

24 M. Frankowski, J. Theisen, A. Kummrow, P. Simon, hydrodynamic particle focusing and integrated optical
H. Ragusch, N. Bock, M. Schmidt and J. Neukammer, elements applying silicon photodiodes, Microfluid.
Microflow cytometers with integrated hydrodynamic Nanofluid., 2011, 10, 761–771.
focussing, Sensors, 2013, 13, 4674–4693. 34 J. Godin and Y. H. Lo, Two-parameter angular light scatter
25 D. Barat, D. Spencer, G. Benazzi, M. C. Mowlem and collection for microfluidic flow cytometry by unique
H. Morgan, Simultaneous high speed optical and impedance waveguide structures, 2010, Biomed. Opt. Express, 2010, 1(5),
analysis of single particles with a microfluidic cytometer, 1472–1479.
Lab Chip, 2012, 12, 118–126. 35 B. R. Watts, Z. Zhang, C. Q. Xu, X. Cao and M. Lin, A
26 X. Mao, A. A. Nawaz, S. C. S. Lin, M. I. Lapsley, Y. Zhao, method for detecting forward scattering signals on-chip with
J. P. McCoy, W. S. El-Deiry and T. J. Huang, An integrated, a photonic-microfluidic integrated device, Biomed. Opt.
multiparametric flow cytometry chip using “microfluidic
Published on 05 June 2014. Downloaded by FAC DE QUIMICA on 6/22/2021 1:47:07 PM.

Express, 2013, 4(7), 1051–1060.

drifting” based three-dimensional hydrodynamic focusing, 36 Laboratory Guidelines for enumerating CD4 T-Lymphocytes in
Biomicrofluidics, 2012, 6, 024113. the context of HIV/AIDS, WHO, June, 2007, ISBN 9789290222989.
27 A. A. Nawaz, X. Zhang, X. Mao, J. Rufo, S. C. S. Lin, F. Guo, 37 D. T. Dunn, HIV Paediatric Prognostic Markers Collaborative
Y. Zhao, M. Lapsley, P. Li, J. P. McCoy, S. J. Levine and Study: Use of total lymphocyte count for informing when to
T. J. Huang, Sub-micrometer-precision, three-dimensional start antiretroviral therapy in HIV-infected children: a meta-
(3D) hydrodynamic focussing via “microfluidic drifting”, Lab analysis of longitudinal data, Lancet, 2005, 366, 1868–1874.
Chip, 2014, 14, 415. 38 Revised classification system for HIV infection in children
28 S. C. Hur, H. T. Kwong Tse and D. Di Carlo, Sheathless less than 13 years of age; Official authorized addenda:
inertial cell ordering for extreme throughput flow cytometry, human immunodeficiency virus infection codes and official
Lab Chip, 2010, 10, 274–280. guidelines for coding and reporting ICD-9-CM. MMWR 1994,
29 A. A. S. Bhagat, S. S. Kuntaegowdanahalli, N. Kaval, 43, RR-12.
C. J. Seliskar and I. Papautsky, Inertial microfluidics for 39 B. Lee, M. Sharron, L. J. Montaner, D. Weissman and
sheath-less high-throughput flow cytometry, Biomed. Micro- R. W. Doms, Quantification of CD4, CCR5, and CXCR4 levels
devices, 2010, 12, 187–195. on lymphocyte subsets, dendritic cells, and differentially
30 Y. Chen, A. A. Nawaz, Y. Zhao, P. H. Huang, J. P. McCoy, conditioned monocyte-derived macrophages, Proc. Natl.
S. J. Levine, L. Wang and T. J. Huang, Standing surface Acad. Sci. U. S. A., 1999, 96, 5215–5220.
acoustic wave (SSAW)-based microfluidic cytometer, Lab 40 C. Simonnet and A. Groisman, High-Throughput and High-
Chip, 2014, 14, 916–923. Resolution Flow Cytometry in Molded Microfluidic Devices,
31 D. Holmes, H. Morgan and N. G. Green, High throughput Anal. Chem., 2006, 76, 5653–5663.
particle analysis: Combining dielectrophoretic particle 41 D. Di Carlo, D. Irimia, R. G. Tompkins and M. Toner,
focussing with confocal optical detection, Biosens. Continuous inertial focusing, ordering, and separation of
Bioelectron., 2006, 21, 1621–1630. particles in microchannels, Proc. Natl. Acad. Sci. U. S. A.,
32 D. Spencer and H. Morgan, Positional dependence of 2007, 104(48), 18892–18897.
particles in microfludic impedance cytometry, Lab Chip, 42 M. E. Piyasena, P. P. A. Suthanthiraraj, R. W. Applegate,
2011, 11, 1234–1239. A. M. Goumas, T. A. Woods, G. P. Lopez and S. W. Graves,
33 M. Rosenauer, W. Buchegger, I. Finoulst, P. Verhaert and Multinode acoustic focussing for parallel flow cytometry,
M. Vellekoop, Miniaturized flow cytometer with 3D Anal. Chem., 2012, 84, 1831–1839.

This journal is © The Royal Society of Chemistry 2014 Lab Chip, 2014, 14, 3064–3073 | 3073