MIT Open Access Articles

Development of miniaturized, portable magnetic resonance

relaxometry system for point-of-care medical diagnosis

The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.

As Published: 10.1063/1.4754296

Publisher: AIP Publishing

Persistent URL: https://hdl.handle.net/1721.1/134501

Version: Final published version: final published article, as it appeared in a journal, conference
proceedings, or other formally published context

Terms of Use: Article is made available in accordance with the publisher's policy and may be
subject to US copyright law. Please refer to the publisher's site for terms of use.
Development of miniaturized, portable magnetic resonance relaxometry
system for point-of-care medical diagnosis
Weng Kung Peng, Lan Chen, and Jongyoon Han

Citation: Rev. Sci. Instrum. 83, 095115 (2012); doi: 10.1063/1.4754296

View online: http://dx.doi.org/10.1063/1.4754296
View Table of Contents: http://rsi.aip.org/resource/1/RSINAK/v83/i9
Published by the American Institute of Physics.

Related Articles
Toward cardiac electrophysiological mapping based on micro-Tesla NMR: a novel modality for localizing the
cardiac reentry
AIP Advances 2, 022156 (2012)
Characterization of tumors using high-Tc superconducting quantum interference device-detected nuclear
magnetic resonance and imaging
Appl. Phys. Lett. 97, 263701 (2010)
Combining magnetic resonance imaging and ultrawideband radar: A new concept for multimodal biomedical
imaging
Rev. Sci. Instrum. 80, 014302 (2009)
Improvement of temporal resolution for three-dimensional continuous-wave electron paramagnetic resonance
imaging
Rev. Sci. Instrum. 79, 123701 (2008)
Magnetic resonance–coupled fluorescence tomography scanner for molecular imaging of tissue
Rev. Sci. Instrum. 79, 064302 (2008)

Additional information on Rev. Sci. Instrum.

Journal Homepage: http://rsi.aip.org
Journal Information: http://rsi.aip.org/about/about_the_journal
Top downloads: http://rsi.aip.org/features/most_downloaded
Information for Authors: http://rsi.aip.org/authors
 REVIEW OF SCIENTIFIC INSTRUMENTS 83, 095115 (2012)

Development of miniaturized, portable magnetic resonance relaxometry

system for point-of-care medical diagnosis
Weng Kung Peng,1 Lan Chen,1 and Jongyoon Han1,2
1
BioSystems and Micromechanics IRG (BioSyM), Singapore - Massachusetts Institute of Technology Alliance
for Research and Technology (SMART) Centre, S16-05-08, 3 Science Drive 2, Singapore 117543
2
Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, 36-841,
77 Massachusetts Avenue, Cambridge, Massachusetts 02139, USA and Department of Biological Engineering,
Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, USA
(Received 9 April 2012; accepted 5 September 2012; published online 26 September 2012)
A novel, compact-sized (19 cm × 16 cm) and portable (500 g) magnetic resonance relaxometry
system is designed and developed. We overcame several key engineering barriers so that magnetic
resonance technology can be potentially used for disease diagnosis-monitoring in point-of-care set-
tings, directly on biological cells and tissues. The whole system consists of a coin-sized permanent
magnet (0.76 T), miniaturized radio-frequency microcoil probe, compact lumped-circuit duplexer,
and single board 1-W power amplifier, in which a field programmable gate array -based spectrometer
is used for pulse excitation, signal acquisition, and data processing. We show that by measuring the
proton transverse relaxation rates from a large pool of natural abundance proton-nuclei presence in
less than 1 μL of red blood cells, one can indirectly deduce the relative magnetic susceptibility of the
bulk cells within a few minutes of signal acquisition time. Such rapid and sensitive blood screening
system can be used to monitor the fluctuation of the bulk magnetic susceptibility of the biological
cells (e.g., human blood cells), where unusual state of the bulk magnetic susceptibility is related to a
number of diseases. © 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.4754296]

I. INTRODUCTION
metabolic activities inside the cells, one is not only able to di-
Magnetic resonance (MR) technology, namely the mag- agnose but also to track the infection level, in non-destructive
netic resonance imaging (MRI) and nuclear magnetic reso- manner. This serves as an ideal biomarker, thereby eliminat-
nance (NMR) spectroscopy are well-established, non-invasive ing the needs for costly and tedious immuno-labeling pro-
techniques for medical imaging and biochemical structural cedures. Furthermore, since the measured band consists of
studies, respectively. The race to gain higher spectral reso- entirely RBCs pellet, one can leverage on the very much
lution and sensitivity for studies on large molecules such as faster proton relaxation rate to increase the signal-to-noise ra-
proteins, has inevitably led to harsh requirement to have ex- tio (SNR). With a commercial, bench-top (weighs a few kilo-
pensive and bulky superconducting magnet, which is perma- grams) magnetic resonance relaxometry (MRR) system, we
nently installed in dedicated laboratories. Nonetheless, low- show that by mapping out the relaxation rate from <1 μL of
field MR technology has also been widely used in geological whole blood, the infection level of Plasmodium berghei in-
studies,1–7 food sciences,8, 9 medical diagnosis,10–12 and nu- fected mouse can be classified within minutes.20
clear quadrupole based explosive detection in landmines,13–15 Unlike standard liquid state-NMR however, measuring
in which portable platform is highly desirable. relaxation rate of the “semi-solid” state of RBCs is not as
With advances in microelectronics technology, MR straightforward and poses numerous technical challenges. In
community sees the emergence of much more compact order to establish a universal baseline reading (due to varia-
MR spectrometers (pulse programmer, transmitter, receiver, tions of hematrocrit levels), therefore only the intended RBCs
and digital signal processing) on a highly integrated cir- band (which can be obtained with a standard hematrocrit cen-
cuit platform such as field programmable gate array trifuge) is measured. In order to cope with the low SNR,19–21
(FPGA),16–18 and complementary-metal-oxide semiconduc- a train of non-selective hard-pulses were applied at very short
tor (CMOS).10–12 Recently, Weissleder’s group has shown echo-time interval (usually less than 100 μs)22 to retrieve a
that small “ping-pong sized” magnet can be packed together long array spin-echoes. In the case where inhomogeneous
with a CMOS-based transceiver into a palm-sized, NMR- magnetic field (imposed by the external magnetic field as well
based biosensor.12 This transceiver is capable of producing as internal field created in the sample) is an issue, gaining
radio-frequency (rf) powers output of up to 80 mW which high SNR can further impose challenges. As a result, it in-
transmit to the sample via microcoil (Q-factor of 28), was evitably imposes harsh requirement on having strong excita-
shown to produce good signal sensitivity in most of the liq- tion pulse. Power amplifier (PA) plays a key role in generating
uid samples such as water and buffer solution. sufficiently high oscillating magnetic fields, in which the min-
We have recently shown that the presence of paramag- imally required pulse excitation field depends very much on
netic particles in parasite-infected red blood cells (RBCs) the state of the samples (e.g., solid/liquid). A much stronger
gives rise to relaxation contrast.19–21 By tracing the natural power amplifier, however, came at the cost of its physical size

0034-6748/2012/83(9)/095115/8/$30.00 83, 095115-1 © 2012 American Institute of Physics

095115-2 Peng, Chen, and Han Rev. Sci. Instrum. 83, 095115 (2012)

TABLE I. Comparison between our MRR system developed in this work (BMS) of the cells is strongly correlated to oxidation state
and the current state-of-the-art miniaturized MRR developed by Sun et al.12 of the cells. With increased portability and much lower cost
Weight does not include the power supplies. Although their system is much
lighter and smaller in size, our system incorporated much stronger power
(cost of the equipment as well as cost per test), we hope that
amplifier which allows more diverse detection applications (liquid and semi- such rapid, sensitive blood screening system can one day be
solids), much better Q-factor of the probe, and stronger static magnetic field. relevant in clinical diagnostic settings, whereby an unusual
state of the BMS of the RBCs is correlated to a number
This work Sun et al. of diseases such as malaria,19, 20 methemoglobinemia,24, 25
Processor FPGA CMOS oxidation/oxygenation level of the blood,26–28 and possibly
Weight (g) 500 100 many other diseases which are known to be related to oxida-
Size 19 cm × 16 cm Palm-sized tive stress such as cardiovascular,29, 30 and neurodegenerative
Static magnetic field (T) 0.76 0.5 diseases.31–34 Besides, the amount of blood needed is less
Q factor of the probe 35 28 than 1 μL and can be obtained with minimally invasive
Power amplifier (mW) 1000 80 procedures. Unlike current technology35 which measure
oxidative stress by means of electron spin resonance (ESR),
spin trapping, or electrochemistry in the laboratory setting
and power consumptions, which inevitably present a bottle- due to hardware constriction, we hope that a portable MRR
neck in realization of portable system. Typically, commercial- system can be employed in field settings in near future.
ized liquid-state NMR spectrometer is equipped with power Similarly to ESR, MRR can measure the cells in vivo in a
amplifier which is capable of few watts rf-outputs, while non-destructive manner, which opens up a new opportunity
solid-state NMR requires much higher rf-power (>100 W).23 for drug screening and metabolism study.
Another key strategy is to leverage on the reduced micron
sized rf-coil to generate much higher rf-nutation field per unit
current. This, however, comes at the expense of reduced sam- II. DESIGN AND CONSTRUCTION
ple volume. We balanced up the needs and based on our ex-
A. Overall architecture
perience, in order to attain high signal sensitivity, rf-powers
ranging to several hundreds of mW allow a good leeway to The overall architecture of the system is shown in
work on semi-solids biological sample (e.g., RBCs) on top of Figure 1. A compact, miniaturized, and yet sufficiently strong
normal liquid state MRR. (1 W) power amplifier is integrated into a single (4 cm
Hence, in this work we designed, miniaturized, and × 4 cm) printed circuit board (PCB). A high Q factor,
built all the necessary parts (trans-receiver, power ampli- hand-wound type solenoid microcoil (ID = 950 μm) is
fier, rf-probe, permanent magnet, and duplexer) to develop a developed to generate strong oscillating magnetic field, B1 .
compact-sized (19 cm × 16 cm), and portable (500 g) home- This important strategy increases the mass-sensitivity per unit
built MRR system, which is capable of producing good sen- volume and improves power efficiency. To overcome the chal-
sitivity for both liquid and semi-solids biological cells, in lenge of routing the higher rf transmission power and much
particular RBCs. The whole system consists of a coin-sized smaller MR signal, a new, compact form of lumped-circuit
permanent magnet (0.76 T), miniaturized radio-frequency mi- duplexer was designed and developed. Both the rf-probe and
crocoil probe, compact lumped-circuit duplexer, and single the duplexer were integrated into a single PCB board (4 cm
board 1-W power amplifier, in which a FPGA-based spec- × 3 cm). To generate polarizing magnetic field, Bo , coin-sized
trometer is used for pulse excitation, signal acquisition, and and light-weight permanent magnets were constructed in our
data processing. The novelty and the details of the construc- laboratory. The whole system is controlled by a single chip
tion are described in Sec. II, and the distinct features of our FPGA-based spectrometer,17, 18 in which the digital circuit
work and the current state-of-art palm-NMR12 are highlighted is built by writing hardware description languages known
in Table I. as very high speed integrated circuit (VHSIC) hardware
In Sec. III, we calibrated the dynamic working range of description language (VHDL). We refer the readers for the
the system, measured the transverse relaxation rate of proton details on FPGA and VHDL developments to previously
nuclei in RBCs, in which the bulk magnetic susceptibility published work.17, 18 The FPGA chip used in this work is

fpga

DDS TRANS PA

to pc USB ppg DDS Dup rf-probe

RCVR p-amp

permanent
magnet

FIG. 1. A block diagram of overall architecture of MRR system developed in this work. The glossary is as following: Pulse programmer (PPG), Direct Digital
Synthesizer (DDS), Receiver (RCVR), Transmitter (TRANS), power amplifier (PA), field array programmable array (FPGA), pre-amplifier (p-amp), universal
serial bus (USB), and duplexer (Dup).
095115-3 Peng, Chen, and Han Rev. Sci. Instrum. 83, 095115 (2012)

microcapillary tube
(a)

acrylic transparent
frame

Neodynium Disc
Magnet

Integrated duplexer
and rf-detection probe

10 mm

(b) Magnetic field, Bo variations

Bo (Tesla)

Bo (Tesla)

Permanent magnet y

FIG. 2. (a) A snapshot of the constructed permanent magnets with a gap size of 2700 μm. An integrated duplexer and rf-probe is shown slotted inside the gap
between the magnets. (b) Magnetic field, Bo variations along the cross section of the magnet as measured by gaussmeter (GM08, Hirst magnetic Instruments
Ltd., UK). The “sweet spot” has the magnetic field of about 0.76 T.

EP3C80F780C8N, Cyclone III (Altera Inc., US) embedded (Fig. 2(a)). Three dipole-dipole disks shaped neodymium
on a bread-board (ACM-202-80C8, HumanData, Japan). magnets (Master Magnetic Inc., USA) were stacked on each
The processor communicates with a personal computer via side separated by a gap of 2700 μm. An acrylic transparent
USB FT2232C (FTDI). All the developed parts were finally type of frame measuring 3 cm × 3 cm × 2 cm was constructed
mounted on a single acrylic board (19 cm × 16 cm) and the to hold these magnets in place. Since the gap is much smaller,
whole system weighs less than 500 g. a relatively homogenous high field concentrated in the middle
of the gap (Fig. 2(b)) producing a “sweet-spot” which peaked
at about 0.76 T. While smaller gap is desirable in gaining
B. Micro permanent magnet
higher magnetic field, it also inevitably reduces the sample
A relatively high static field (0.76 T), and yet portable and volume that can be measured, thereby reducing the volume
light-weight (about 60 g) permanent magnet is constructed. It sensitivity. This shortcoming is however compensated by
operates at proton NMR frequency of about 31.9 MHz. The employing a rf-microcoil, which increases the mass sensitiv-
main magnet is a construction of a conventional dipole design ity, as described in Sec. II C. The magnetic field realized in
095115-4 Peng, Chen, and Han Rev. Sci. Instrum. 83, 095115 (2012)

FIG. 3. (a) Schematic diagram of the duplexer and rf-probe which were both tuned to resonate at 31.5 MHz for proton NMR. The parameters for LC components
are L1 = 0.47 μH, C1 = C2 = 28 pF, Cm = 870 pF, Ct = 50 pF, and the measured Q-factor of the probe is about 35. (b) Insertion loss between the rf-probe to
the RCVR, as measured by spectrum analyzer (FSL, Rodhe Schwarz, Germany). Applied powers of −45 dBm were used.

this work is much stronger than other miniaturized portable NMR experiment (below 35 MHz), quarter wavelength trans-
magnets reported in other literatures.10–12, 36 At this stage of mission cable can be inconveniently long and not suitable
work, shim coil and temperature controller were deliberately for the development of bench-top system. In this work,
excluded in the system, which allows us to further increase lengthy and bulky quarter wavelength transmission cable is
the portability and compactness of the system. The effects of replaced with a much simpler and compact resonant circuit
field-inhomogeneity across the sample were studied and the (Fig. 3(a)).37 The lumped-circuit was constructed by wire-
dynamic working range of the relaxation rate is reported in wound chip inductor (Bourns CM453232, USA) and fixed
Sec. III. We found no substantial temperature drift occurred capacitor (Voltronic, USA) and ceramic trimmer capacitor
during the time scale of measurement as indicated in the (Murata TZ03, Japan).
calibration test (Fig. 6(a)). However, depending on the need Typically, signal from free induction decay ranges
of the users, adding an additional commercial temperature from −45 dBm or lower, while much higher transmission
controller (e.g., by mean of air flow) is straightforward. rf-powers (in the order of 30 dBm) were used in this work.
The duplexer is tuned to operate at a narrowband frequency
of 31.5 MHz (Fig. 3(b)). Insertion losses of less than 1 dB
C. Compact and integrated passive duplexer
were recorded at the receiver. In the transmission mode (at
and rf-detection probe
30 dBm), by employing a combination of RF PIN switch
Conventional passive duplexer employs pairs of cross diodes (HSMP-389C, Avago Technologies) and Schottky-
diode and quarter wavelength cable in order to isolate the type diodes (HSMS-286F, Avago Technologies), power
high rf power as it exceeds the threshold value of the cross leakage to the receiver is suppressed to about −2 dBm (an
diodes. In practice, however, especially for low frequency isolation of −32 dB). This is sufficiently small as compared
095115-5 Peng, Chen, and Han Rev. Sci. Instrum. 83, 095115 (2012)

to safety input level of the pre-amplifier (e.g., AMP-75+,

Mini Circuits, USA and AU-1114, MITEQ, USA) which
are typically able to tolerate input power of 10 dBm. For
further protection, a surface mount-type high-power, high
isolation rf switch (e.g., HSWA2-30DR+, Mini-Circuits) can
be employed at the preamplifier. This high-speed switch is
controlled by a transistor-transistor logic (TTL) signal which
synchronizes to the acquisition mode.
Together with the duplexer, a tank circuit for single res-
onance probe is constructed on a single PCB (Fig. 2(a)). A
four-turn microcoil (inner diameter 950 μm) made of sin-
gle strand enameled copper wire (FA-NR 927943, Block,
Germany) of diameter 400 μm was closely woven. Micro-
capillary tube (outer diameter 730 μm) (Drummond Scientific
Co., Broomall, PA) is used to contain liquid sample under de-
tection. As little as approximately 350 nL of effective volume
is measured. Microcoil increases the volume sensitivity, hence
reduces the need to have high-field homogeneity over a large
area, which is beneficial especially for home-built magnetic
system where obtaining high homogeneity poses additional
engineering difficulties. Microcoil also increases the power
efficiency, hence reducing the need to achieve bulky power-
amplifier. The measured Q-factor of the probe which is about
35, is much higher than the Q-factor reported by other liter-
atures; 28.5 by Demas et al.,38 10 by Sillerud et al.,39 24 by
Massin et al.,40, 41 and 28 by Sun et al.12

FIG. 4. (a) Schematic diagram of the power amplifier. All the electronic
D. Single board 1-watt power amplifier
components were mounted on a single printed circuit board. (b) A snapshot,
Figure 4(a) is a schematic diagram describing a 1-W showing front-view (left) and back-view (right) of the power amplifier con-
struction. (c) Performance of the power amplifier as measured by the spec-
power amplifier. A cascaded three stages of surface mount trum analyzer (HM 1010, HAMEG, Germany). The y axis shown in the graph
type of power amplifier (HELA-10D+, 8-300 MHZ, was scaled back 20 dB due to the attenuator. An attenuator of −20 dB was
50 , Mini-Circuits) were packed onto a two-sided (40 mm used to protect the receiver of the spectrum analyzer, resulting in 1 dB of
insertion loss.
× 40 mm) printed circuit board (Fig. 4(b)). Each core com-
ponent of the power amplifier (HELA-10D+) is capable of
producing a typical gain of 11 dB and typical maximum out- III. ASSEMBLY OF THE SYSTEM
put power of 30 dBm, which is powered by a dc power supply AND MRR DEMONSTRATION
of +12V. Our rf-transmitter which is controlled by the FPGA A. Methods and MRR parameters
pulse programmer produces rf-power of 0 dBm, therefore a
1
three cascaded system was needed to produce a maximum H MRR measurements were carried out at the resonance
amplification to 30 dBm (1 watt) of rf-output (Fig. 4(c)). frequency of 31.45 MHz inside a micro permanent magnet, Bo
A low pass filter (LFCN-105+, Mini-Circuits) with cut- = 0.76 T. All samples were measured at room-temperature.
off frequency at 105 MHz is inserted before the input of each The transverse relaxation rates were measured by standard
amplifier stage and output of the PA. A surface mount type Carr-Purcell-Meiboom-Gill (CPMG) train pulses consisting
high-power switch (HSWA2-30DR+, Mini-Circuits), em- of 2000 echoes. Interecho times used for blood experiments
ployed before the output port is an absorptive rf switch. This and DI-water doped with copper sulfate were 80 μs and
switch is controlled by a signal (TTL) to synchronize with 100 μs, respectively. The transmitter power output is main-
the rf input pulse sequence, which can further reduce noise tained at 0.18 W for a single 90◦ -pulse with pulse length
leakage to the receiver, especially in the receiving mode. 4 μs, which corresponds to a nutation frequency of 62.5 kHz.
Figure 4(b) shows a power amplifier with a frequency range of A recycle delay of 2 s which was set between each pulse is
8–100 MHz. The actual working frequency of HELA-10D+ sufficiently long enough to allow all the spins to return to ther-
ranges from 8 MHz to 300 MHz. In current work however mal equilibrium. A total of 80 scans were acquired for signal
the upper frequency limit in the current setup is limited by averaging.
the low pass filter, which has cut-off frequency at 105 MHz.
For heat dissipation purposes, a heat sink is attached directly
B. System stability and consistency
on HELA 10D+. Additional mini-fan (40 mm × 40 mm) is
mounted on top of the power amplifier to improve the heat In attempt to calibrate the stability and consistency of
dissipation. the entire integrated MRR system (Fig. 5(a)), the proton
095115-6 Peng, Chen, and Han Rev. Sci. Instrum. 83, 095115 (2012)

FIG. 5. (a) A snapshot of the whole MRR system measuring 19 cm × 16 cm. As a comparison, a screw driver is placed by the side of the system. The name of
each part is indicated in (b).

relaxation rates of de-ionized (DI)-water doped with cop- used in our system. The main focus here is to be able to mea-
per(II)sulfate hydrate (Cu2 SO4 .xH2 O, Sigma Aldrich) over sure and compare relative relaxation measurements instead
a wide range of doping concentrations, were measured. In- of comparing the absolute value of the relaxation rate with
crease in the proton relaxation rate (R2 > 0.99) with in- other system. Nonetheless, it is clearly shown that despite op-
creasing of doping concentration (Fig. 6(a)) conform to the erating in “less-than-perfect” field homogeneity, such home-
theory,42 and thus confirming the functionality of the system. built permanent magnet allows us to operate at a large dy-
To evaluate the reproducibility of the system, five measure- namic working range with high reproducibility. Importantly,
ments were recorded for the same concentration, and its vari- this range works fine for the blood measurements related to
ations were reported as standard error measurement (s.e.m). work as shown in Sec. III C.
The reproducibility are much higher at lower concentration
(<500 mM), while for concentration of 500 mM and higher,
it gradually reduced. There is a slight discrepancy between C. Effect of sodium nitrite on redox
chemistry of red blood cells
our results and the one reported in literature,43 in which for
a linear fit, the slope reported in their work (0.69 mM−1 s−1 ) We demonstrate that MRR system can be a portable
is much larger than our work (0.51 mM−1 s−1 ). We attribute platform for medical diagnosis by measuring the transverse
this discrepancy to the effect of field inhomogeneity across relaxation rate, R2 of actual biological cells. A sample of
the sample (and possibly the effect of temperature drift and 150 μL of whole blood taken from healthy donor was exposed
pulses imperfection43 ), due to the fact that shimming is not to a drop (about 10 μL) of 10 mM of sodium nitrite, NaNO2
095115-7 Peng, Chen, and Han Rev. Sci. Instrum. 83, 095115 (2012)

(a)
300
(500, 276.3)
250
Relaxation rate (s-1)
200

150

100

50 (100, 56.5 )
(50, 31.2 )
(10, 11.1 )
0
(1, 6.5 )

-50
-100 0 100 200 300 400 500 600
Concentration of CU2SO4 (mM)
(b) 25
Relaxation rate (s-1)

(60, 22.01)
20
(30, 20.37 )

(0, 17.9 )

15
0 10 20 30 40 50 60
exposure time (s)

(c) Normal blood Blood with 1min of NaNO2 exposure

1.6
Signal intensity (arbitary)

1.2

0.8 R2 = 21.5 s-1

R2 = 17.4 s-1
0.4

0.0

-0.4
0 40 80 120 160 0 40 80 120 160
Time (ms) Time (ms)
FIG. 6. (a) Plot depicting the relaxation rate of DI-water doped with Cu2 SO4 (in mM concentration). The parentheses indicate concentration (mM) and relaxation
rate(s−1 ). The relaxation rates in this work (black dot) is compared to the one reported in the literature43 (red dot). The error bar shown is the standard error
measurement (s.e.m). Note that for lower concentration (100 mM and below) the error bars were too small to be visible in the plot. (b) Proton relaxation rate
of RBCs as a function to NaNO2 exposure time (in seconds). The parentheses indicate exposure time (s) and relaxation rate (s−1 ). The error bar shown is the
standard error measurement. (c) The actual echo trains measured for a normal healthy blood (left) and blood exposed to 1 min of NaNO2 (right).
 095115-8 Peng, Chen, and Han Rev. Sci. Instrum. 83, 095115 (2012)

(Fluka Analytical, Switzerland) for less than 1 min. 10 μL 3 J. D. Seymour, J. P. Gage, S. L. Codd, and R. Gerlach, Adv. Water Resour.
of blood is transferred into microcapillary tube (Drummond 30, 1408–1420 (2007).
4 G. H. Sorland, H. W. Anthonsen, J. G. Seland, F. Antonsen, H. C. Wideroe,
Scientific Co.) via capillary action, and then centrifuged at and J. Krane, Appl. Magn. Reson. 26, 417–425 (2004).
3000 g for 3 min to separate the plasma from the RBCs. The 5 Y. Q. Song, I. Saha, J. Franck, T. Hopper, and B. Q. Sun, Petrophys. 50,

capillary is then loaded into the rf-probe as such that only the 345–351 (2009).
6 Y. Q. Song, L. Venkataramanan, and L. Burcaw, J. Chem. Phys. 122,
RBCs portion is measured. About 350 nL effective volume
104104 (2005).
of RBCs under detection gives sufficiently strong spin echo 7 Y. Q. Song, L. Venkataramanan, M. D. Hurlimann, M. Flaum, P. Frulla, and
trains signal. C. Straley, J. Magn. Reson. 154, 261–268 (2002).
8 I. V. Mastikhin, O. V. Petrov, J. Hay, and B. J. Balcom, Food Res. Int. 41,
Sodium nitrite, a form of strong oxidant, if present in
unusually high amount or concentration, can be harmful to 758–764 (2008).
9 H. T. Pedersen, S. Ablett, D. R. Martin, M. J. Mallett, and S. B. Engelsen,
human blood cells.31, 32 We showed (Figs. 6(b) and 6(c)) J. Magn. Reson. 165, 49–58 (2003).
that exposure of whole blood to higher than normal levels 10 H. Lee, T. Yoon, J. Figueiredo, F. Swirski, and R. Weissleder, Proc. Natl.

of sodium nitrite will lead to oxidation of the iron heme, Acad. Sci. U.S.A. 106, 12459 (2009).
11 H. Lee, E. Sun, D. Ham, and R. Weissleder, Nat. Med. 14, 869–874 (2008).
from ferrous (Fe2+ ) to ferric (Fe3+ ). This simple oxidation 12 N. Sun, T. J. Yoon, H. Lee, W. Andress, R. Weissleder, and D. Ham, IEEE
reaction, transformed predominantly the oxy-hemoglobin (in J. Solid-State Circuits 46, 342–352 (2011).
healthy whole blood, which is in diamagnetic state, (Fe2+ )) 13 J. B. Miller and G. A. Barrall, Am. Sci. 93, 50–57 (2005).
14 A. N. Garroway, M. L. Buess, and J. B. Miller, Abstr. Pap. – Am. Chem.
into paramagnetic state (Fe3+ ). Such a redox transformation
Soc. 220, U98–U99 (2000).
of RBC hemoglobin, however, increased the bulk magnetic 15 M. Ostafin and B. Nogaj, Measurement 40, 43–54 (2007).
susceptibility of the RBCs significantly, which is measurable 16 W. Tang and W. Wang, Meas. Sci. Technol. 22, 015902 (2011).

by MRR system.42, 44 Measurements as depicted in Figs. 17 K. Takeda, Rev. Sci. Instrum. 78, 033103 (2007).
18 K. Takeda, J. Magn. Reson. 192, 218–229 (2008).
6(b) and 6(c) show that on reacting with nitrite, irons in 19 W. K. Peng, Biomedical Engineering Society 2011 Annual Meeting,
hemoglobin are progressively turning into paramagnetic state Hartford, Boston, USA (2011).
which in return increases the transverse relaxation rate of 20 W. K. Peng, in Proceedings of Experimental Nuclear Magnetic Resonance

nearby proton nuclei. Conference 2011, Asilomar Conference Center, Pacific Grove, CA, 2011.
21 T. F. Kong, W. K. Peng, T. D. Luong, N. T. Nguyen, and J. Han, Lab Chip

12, 287–294 (2011).

IV. CONCLUSION 22 C. Eccles, in Encyclopedia of Spectroscopy and Spectrometry (Second Edi-

tion), edited by Â. Â. John (Academic, Oxford, 2010), p. 1357–1371.

In summary, we have shown that by bringing down the 23 T. Mizuno and K. Takegoshi, Rev. Sci. Instrum. 80, 124702 (2009).
barrier in engineering aspects, magnetic resonance technol- 24 S. Aime, M. Fasano, S. Paoletti, A. Arnelli, and P. Ascenzi, Magn. Reson.

ogy can be scaled down (size, volume, weigh and cost) to a Med. 33, 827–31 (1995).
25 T. H. Huang, J. Biol. Chem. 254, 11467–74 (1979).
level which is suitable for on-site applications such as point- 26 J. M. Gomori, R. I. Grossman, C. Yu-Ip, and T. Asakura, J. Comput. Assist.
of-care medical diagnosis/monitoring purposes. Such home- Tomogr. 11, 684–90 (1987).
built system allows one to tailor made additional hardware 27 K. R. Thulborn, J. C. Waterton, P. M. Matthews, and G. K. Radda, Biochim.

for more specific experiments, and the cost of development is Biophys. Acta 714, 265–270 (1982).
28 S. Ogawa, T. M. Lee, A. R. Kay, and D. W. Tank, Proc. Natl. Acad. Sci.
very much lower than a commercialized system. To our best
U.S.A. 87, 9868–72 (1990).
knowledge, this is the first work that describes the develop- 29 B. Halliwell, Saudi Med. J. 12, 84–90 (1991).
ment of a portable MRR system that provides sufficiently high 30 B. Halliwell, Br. J. Exp. Pathol. 70, 737–757 (1989).
31 H. Kosaka, K. Imaizumi, K. Imai, and I. Tyuma, Biochim. Biophys. Acta
radio-frequency power and yet portable enough which enable
581, 184–188 (1979).
measurement transverse relaxation rates of RBCs from a vol- 32 H. Kosaka, I. Tyuma, and K. Imaizumi, Biomed. Biochim. Acta 42, S144–
ume of less than 1 μL blood, in a rapid manner. We hope that S148 (1983).
with increased portability and accessibility to MRR system, it 33 A. Samouilov, Y. Y. Woldman, J. L. Zweier, and V. V. Khramtsov, Nitric

will accelerate research into discovery of unique characteris- Oxide 16, 362–70 (2007).
34 B. Halliwell, Drugs Aging 18, 685–716 (2001).
tic (e.g., T1/T2 correlation,7 membrane diffusivity45 ) of each 35 B. Halliwell and M. Whiteman, Br. J. Pharmacol. 142, 231–255 (2004).
disease, which share the same class of “MRR-active nuclei” 36 N. Sun, Y. Liu, H. Lee, R. Weissleder, and D. Ham, IEEE J. Solid-State

biomarkers. Circuits 44, 1629–1643 (2009).

37 L. A. Mclachlan, J. Magn. Reson. 39, 11–15 (1980).
38 V. Demas, J. L. Herberg, V. Malba, A. Bernhardt, L. Evans, C. Harvey, S.
ACKNOWLEDGMENTS C. Chinn, R. S. Maxwell, and J. Reimer, J. Magn. Reson. 189, 121–129
(2007).
This work is supported by Singapore-MIT Alliance for 39 L. Sillerud, A. McDowell, N. Adolphi, R. Serda, D. Adams, M. Vasile, and
Research and Technology (SMART) Centre, BioSystems and T. Alam, J. Magn. Res. 181, 181–190 (2006).
40 C. Massin, G. Boero, F. Vincent, J. Abenhaim, P. A. Besse, and R. S.
Micromechanics (BioSyM) IRG. W. K. Peng acknowledges
the support from the SMART Postdoctoral Research Fellows Popovic, Sens. Actuators, A 97-8, 280–288 (2002).
41 C. Massin, F. Vincent, A. Homsy, K. Ehrmann, G. Boero, P. A. Besse, A.
Programme, and Kenry, Huang Yong Xue and Kong Tian Daridon, E. Verpoorte, N. F. de Rooij, and R. S. Popovic, J. Magn. Reson.
Fook for assistance. We are greatly in debt to Dr. Kazuyuki 164, 242–255 (2003).
42 A. Roch, R. N. Muller, and P. Gillis, J. Chem. Phys. 110, 5403–5411
Takeda (Kyoto University) for his kind, and selfless assistance
in sharing his Opencore NMR project with us. (1999).
43 J. M. Pope and N. Repin, Magn. Reson. Imaging 6, 641–646 (1988).
44 P. Gillis, S. Peto, F. Moiny, J. Mispelter, and C. A. Cuenod, Magn. Reson.
1 P.
T. Callaghan, Magn. Reson. Imaging 14, 701–709 (1996). Med. 33, 93–100 (1995).
2 P. 45 G. Benga, V. I. Pop, O. Popescu, and V. Borza, Biosci. Rep. 10, 31–36
T. Callaghan, C. H. Arns, P. Galvosas, M. W. Hunter, Y. Qiao, and K. E.
Washburn, Magn. Reson. Imaging 25, 441–444 (2007). (1990).

View publication stats