Ma N
Abstract
We have
poly(vinylpy
by a microdi
weight cut o
rates in the d
pretreatmen
and reprodu
our device, t
© 2005 Elsevi
Keywords: M
1. Introdu
The bed
tant for the
tions in int
of chronic
undertaken
using relat
autoanalyz
immunosor
time to obt
The real-tim
ical treatm
sure biomo
(Ramsey an
∗ Correspon
E-mail ad
0956-5663/$
doi:10.1016/jBiosensors and Bioelectronics 21 (2006) 1649–1653
Short communication
iniaturized one-chip electrochemical sensing device integrated
with a dialysis membrane and double thin-layer flow
channels for measuring blood samples
Ryoji Kurita a,∗, Norikuni Yabumoto b, Osamu Niwa a
ational Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki, Japan
b NTT Advanced Technology, 3-1 Morinosato-Wakamiya, Atsugi, Kanagawa, Japan
Received 22 April 2005; received in revised form 14 July 2005; accepted 15 July 2005
Available online 29 August 2005
developed a microfluidic device consisting of a gold film working electrode modified with lactate oxidase and Os-
ridine) mediator containing horseradish peroxide, and reference and counter electrodes in a microflow detection channel separated
alysis membrane from another microflow channel used for sample injection. The dialysis membrane is cellulose with a molecular
ff of 10 kDa. We achieved control over a wide recovery rate range of 3–94% because the device is capable of controlling both flow
ual thin-layer channels. We were able to measure the lactate concentration in blood samples within a few minutes without any
t because biomolecules are simultaneously separated by molecular weight and detected in the device. We achieved quantitative
cible measurements of the lactate concentration in blood samples, and obta
he lactate concentration in dog whole blood was measured with high stab
er B.V. All rights reserved.
icrodevice; Microfluidics; Electrochemical; Biosensor; Dialysis; Lactate
ction
side monitoring of in vivo biomolecules is impor-
medical treatment of patients with acute condi-
ensive care units or for the periodic monitoring
cases. However, clinical analysis has mainly been
in laboratories located nowhere near the patient
ively large and expensive instruments such an
er or costly techniques such as enzyme-linked
bent assay. As a result it takes a relatively long
ain the results and apply them to the treatment.
e monitoring of a patient’s state and timely med-
ent might be realized if it were possible to mea-
lecules frequently or continuously at the bedside
d Michael, 1999; Tu¨do˝s et al., 2001).
ding author. Tel.: +81 29 861 6166; fax: +81 29 861 6177.
dress: r.kurita@aist.go.jp (R. Kurita).
Recentl
designed to
use or beds
2000; Coll
1996). How
quantitative
1997; Kros
the sensitiv
a result of
surface. Se
sors that ar
et al., 1987
ferrocene (
Although t
sor stability
at the bedsi
above poly
the polyme
sensing ele
– see front matter © 2005 Elsevier B.V. All rights reserved.
.bios.2005.07.016ined a relative standard deviation of 1.5% (n = 8). With
ility without any pretreatment.
y, research has focused on miniaturized biosensors
allow simple, frequent and inexpensive personal
ide monitoring (Reyes et al., 2002; Suzuki et al.,
ison and Meyerhoff, 1990; Ma˘da˘ras¸ and Buck,
ever, it is difficult to measure in vivo biomolecules
ly using these biosensors (Reddy and Vadgama,
et al., 2001; Kurita et al., 2004). This is because
ity of the sensor is unstable for blood samples as
the protein adsorption or thrombus on the sensor
veral groups have reported relatively stable sen-
e coated with polymers such as cellulose (Fischer
), polyethylene glycol (Quinn et al., 1997), vinyl-
Kurita et al., 2004) or Nafion (Turner et al., 1990).
his approach has led to an improvement in biosen-
, it is still unsatisfactory for continual monitoring
de. When the sensing electrode is coated with the
mers, sophisticated skills are required to control
r thickness and the diffusion of analytes to the
ctrode.

1650 ics 21 (2
Recentl
ysis system
systems th
sample han
al., 1990).
take clinica
Some rese
brane in a m
molecular
device with
probes (Pe
al., 1998).
nylon mem
micro flow
fabrication
detected am
mal lens m
membrane
laser beam
was poor b
pared with
is because
cal to the o
of the mole
condensati
sensor with
on the dial
nels separa
area.
In this p
formance o
device inte
trodes. Thi
recovery ra
the use of d
ysis memb
of the devic
blood samp
2. Materia
2.1. Fabric
Fig. 1 s
membrane
of two acr
nium was
sputtering
metal mask
out breakin
thickness o
100 nm. W
Os-poly(vi
al., 1995; K
Systems, W
. Schematic representation of a microfluidic device. Sample solution
erfusate are introduced from (A and B), respectively. The potential
rking electrode is held at 450 mV or −50 mV vs. the inner reference
de.
at room temperature for 1 h, we modified the electrode
1
L of lactate oxidase (Kanto-kagaku, Tokyo, Japan).
actate oxidase solution contained 1 unit/
L lactate oxi-
2% bovine serum albumin (Sigma, St. Louis, MO),
.2% glutaraldehyde (Kanto-kagaku), which we used toR. Kurita et al. / Biosensors and Bioelectron
y, many researchers have reported integrated anal-
s called lab-on-a-chip or micro total analysis
at are small, light, and capable of performing all
dling steps using micro flow channels (Manz et
These techniques may make it possible to under-
l measurements simply and rapidly at the bedside.
archers have reported the fabrication of a mem-
icro channel and the separation of molecules by
weight using the membrane in a micro analytical
out connection to the conventional microdialysis
trou et al., 2003; Lunte et al., 1991; Kaptein et
Zhao et al. (2002) reported the fabrication of a
brane by polycondensation from monomers in a
channel. Hisamoto et al. (2003) also reported the
of a nylon membrane in a micro channel and they
monia species through the membrane with a ther-
icroscope. Song et al. (2004) fabricated a dialysis
in a micro channel by polymerization with a UV
. However, their recovery rate on the membrane
ecause the inner volume was relatively large com-
the surface area of the dialysis membrane. This
the membrane was fabricated so that it was verti-
blong flow channel. Furthermore, precise control
cular weight cut off is difficult with nylon poly-
on. A suitable way to develop a highly sensitive
Fig. 1
and p
of wo
electro
film
with
The l
dase,
and 0a fast response is to obtain a high recovery rate
ysis membrane using dual thin-layer flow chan-
ted by a membrane with a relatively large surface
aper, we report the fabrication procedure and per-
f a miniaturized electrochemical one-chip sensing
grated with a dialysis membrane and sensing elec-
s integrated device is capable of controlling the
te, and its response is both rapid and stable through
ual thin-layer flow channels separated by the dial-
rane. We also report results on the reproducibility
e response when we used it to measure lactate in
les.
ls and methods
ation of dialysis membrane integrated device
hows schematic representations of our dialysis
integrated device. The device consists mainly
ylic plates and a cellulose membrane. Thin tita-
deposited on the acrylic plates by using an RF
equipment (Seed Lab, Tokyo, Japan) through a
, and then we formed three gold electrodes with-
g the vacuum (Kurita et al., 2000). The total
f the gold and titanium film electrode was about
e modified the working electrode with 0.5
L of
nylpyridine) wired horseradish peroxide (Yang et
urita et al., 2002a) (Os-gel-HRP) (Bioanalytical
est Lafayette, IN). After drying the Os-gel-HRP
crosslink th
paste as a r
To fab
10 mm × 2
a cutting p
each flow c
fabricated a
the cutting
Tokyo, Jap
tapes. Fina
B) (120
m
2.2. Measu
We inst
pump (CM
a sample s
ple solutio
becco’s pho
Rockville,
from (B). T
4C ampero
in a compu
dopamine
we investig
unmodified
active mole
the device
electrode p
and at −50
nal referen006) 1649–1653e film. We coated one of the electrodes with silver
eference electrode.
ricate the adhesive spacer, we obtained a
7 mm piece of 50
m thick adhesive tape using
lotter (Mimaki, Nagano, Japan). The thickness of
hannel was same as that of the spacer. We then
rectangular flow channel (1 mm × 19 mm) using
plotter. We inserted a cellulose membrane (BAS,
an) and fixed two acrylic plates with two adhesive
lly, we connected two inlet and outlet tubes (A and
i.d., 700
m o.d.) to the acrylic plates.
rements
alled two syringes in a CMA 102 dual syringe
A, Stockholm, Sweden) to allow us to introduce
olution and purfusate. We introduced the sam-
n into the sampling channel from (A) and Dul-
sphate buffered saline (PBS) (Life Technologies,
MD) as the perfusate into the detection channel
he three electrode pads were connected to an LC-
metric detector (BAS) and the data was stored
ter. In this paper, we measured two chemicals,
(Tokyo Kasei, Japan) and lactate (Sigma). First,
ated the electrochemical characteristics of the
device by using dopamine, which is an electro-
cule. We then measured lactate after modifying
with enzyme and mediator. We held the working
otential at 450 mV for the dopamine measurement
mV for the lactate measurement versus the inter-
ce electrode.