- Email: [email protected]
A paper-based amperometric glucose biosensor developed with Prussian Blue-modified screen-printed electrodes
Accepted Manuscript Title: A paper-based amperometric glucose biosensor developed with prussian blue-modified screen-printed electrodes Author: Nadia Chandra Sekar Seyed Ali Mousavi Shaegh Ng Sum Huan Gary Ge Liya Tan Swee Ngin PII: DOI: Reference:
S0925-4005(14)00943-5 http://dx.doi.org/doi:10.1016/j.snb.2014.07.103 SNB 17258
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
17-3-2014 3-7-2014 23-7-2014
Please cite this article as: N.C. Sekar, S.A.M. Shaegh, N.S.H. Gary, G. Liya, T.S. Ngin, A paper-based amperometric glucose biosensor developed with prussian blue-modified screen-printed electrodes, Sensors and Actuators B: Chemical (2014), http://dx.doi.org/10.1016/j.snb.2014.07.103 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A PAPER-BASED AMPEROMETRIC GLUCOSE BIOSENSOR DEVELOPED WITH PRUSSIAN BLUE-MODIFIED SCREEN-PRINTED ELECTRODES
ip t
Nadia Chandra Sekara,§, Seyed Ali Mousavi Shaeghb,#, Ng Sum Huan Garyb,‡, Ge Liyac,^, Tan Swee Ngina,*
a
cr
Natural Sciences and Science Education Academic Group, Nanyang Technological University, 1 Nanyang Walk, Singapore 637616, Singapore b
us
Singapore Institute of Manufacturing Technology, 71 Nanyang Drive, Singapore 638075, Singapore
c
M
an
Residues and Resource Reclamation Centre, Nanyang Environment and Water Research Institute, Nanyang Technological University, 1 Cleantech Loop, Cleantech One, Singapore 637141, Singapore
[email protected]
#
[email protected]
‡
[email protected]
Ac ce p
^
te
§
d
Email Addresses of Authors:
[email protected]
Corresponding Author’s Email: *[email protected]
Telephone: +65 67903810 (Swee Ngin Tan)
Page 1 of 31
1. Introduction New strategies intended for rapid detection of analytes that are of clinical and environmental importance without requiring sophisticated instrumentation is currently in high demand [1]. Lately, paper has drawn much interest as a potential material for biosensors in
ip t
analytical and clinical chemistry due to it being versatile, highly abundant and inexpensive [2]. Upon varying its pulp processing, paper can be made thin, lightweight and flexible to suit
cr
a specific application [3]. Cellulose fibers, being the key component of paper, allow liquid to infiltrate its hydrophilic fiber matrix with no active pumps or external sources required [4].
us
Hence, this makes paper a possible simple and effective matrix for enzyme immobilization in paper-based biosensors via physical adsorption. In fact, antibodies [5], DNA aptamers [6],
an
phages [7] and cells [8] have also been applied in the development of viable paper-based
M
biosensors apart from enzymes [9]. Many paper-based analytical devices are used as lowcost substitutes for medical point-of-care diagnostics due to the necessity of fast, reliable and
d
affordable diagnostic tools in poverty-stricken developing countries [10]. Lately, Whitesides’
te
group has developed techniques for creating microfluidic devices from patterned paper and demonstrated the use of paper-based microfluidic devices for colorimetric detection of
Ac ce p
glucose and protein in artificial urine [11-14]. Glucose Oxidase (GOx) is widely used in practical applications for quick and precise
glucose analysis. To date, 85% of the world biosensor market is currently dominated by electrochemical glucose biosensors [15-16] based on GOx-modified Screen-Printed Carbon Electrodes (SPCEs). As shown in equation (1), GOx catalyzes glucose oxidation in the presence of molecular oxygen (O2) producing gluconic acid and hydrogen peroxide (H2O2).
Glucose + O2
GOx
Gluconic Acid + H2O2
(1)
Monitoring of glucose levels can be based on either O2 consumption or H2O2 generation. Traditionally, most electrochemical glucose biosensors were based on O2 detection but the
Page 2 of 31
detection of H2O2 generation was found to be more sensitive [17]. At conventional electrodes, H2O2 detection occurs at a relatively high potential ca. +0.60 V versus Ag/AgCl which leads to serious interference [18]. Thus, suitable electrocatalysts are employed to resolve this issue [17].
ip t
Prussian Blue (PB) or ferric ferrocyanide (FeIII4[FeII(CN)6]3) has been used in the development of amperometric glucose biosensors because of its excellent electrocatalytic
cr
properties for H2O2 reduction [17,19-21] at low applied potential. The reduction current of H2O2 on PB-modified electrodes is two orders of magnitude higher than that of O2 reduction
us
for glucose analysis [22]. Moreover, the high catalytic activity and selectivity of PB lowers the operating potential to avoid or greatly reduce the contribution from potential interfering
an
compounds, e.g. ascorbic acid [18,20]. Previous literature has reported applied potentials
M
ranging from +0.18 to −0.40 V vs. Ag/AgCl for PB-modified glucose biosensors [23-26]. Furthermore, an oxidized state of PB, Berlin Green, also allows glucose detection but it is
d
based on electro-oxidation of H2O2 at PB-modified electrodes [17]. So far, PB-modified
te
platinum [27], carbon paste [28], graphite [29], glassy carbon [21] and SPEs [30] have been studied for glucose detection.
Ac ce p
Previously, we reported on the integration of paper disc with SPCEs for the
development of a ferrocene carboxylic acid mediated glucose biosensor [31]. As the mediator was incorporated in the paper disc, it reached saturation and leached out easily. Saturation of the mediator on the paper disc could be a limiting factor for the linearity of the biosensor response. The applied potential of the above biosensor was in the positive range, ca. 0.25 V, a relatively high operating potential which could lead to interference from other oxidizable compounds, e.g. ascorbic acid. In this study, we would like to overcome the above limitations of the paper disc mediated glucose biosensor [31] and explore the development of glucose biosensor by immobilizing GOx within porous structure of paper discs placed on top of PB-modified SPCEs, as shown in Figure (1). With this biosensor, only 0.5 µL of analyte
Page 3 of 31
was required for the detection of glucose. This results in cost saving due to a reduction in reagent consumption and prevents incessant exposure to bulk solution, as in a typical mediated glucose biosensor [31]. Therefore, leaching of PB from the SPCE, resulting in loss of catalytic activity, is significantly reduced. The various parameters for the optimization of
ip t
the biosensor development such as applied potential, pH and GOx loading were investigated. 2. Experimental
cr
2.1 Reagents and Instrumentation
All reagents used were of analytical grade. GOx from Aspergillus Niger (type X-S,
us
EC 232-601-0; 234,900 U g-1 solid) were purchased from Sigma-Aldrich (St Louis, MO, USA). β-D-glucose was purchased from Nacalai Tesque (Kyoto, Japan). For validation
an
studies, glucolin glucose powder (420g) and hydralyte were purchased from a local
M
supermarket and pharmacy store respectively. All solutions were prepared with 18 mΩ ultrapure water obtained from Millipore Alpha-Q water system (Bedford, MA, USA). All
d
electrochemical characterizations and measurements were performed using a four-channel
te
system (eDAQ QuadStat, e-Corder 8 and Echem software, eDAQ Europe, Poland). PB-SPCEs (DRP-710) and the boxed connector for SPEs (DRP-DSC) were purchased
Ac ce p
from DropSens (Asturias, Spain). The working electrode (4 mm diameter) consisted of Carbon/Prussian Blue, while Ag/AgCl and a carbon ring were the reference and counter electrodes, respectively. Paper discs were cut from Grade 1 filter papers [Whatman Asia Pacific Pte Ltd (Singapore)]. Data points were plotted using Microsoft Excel (USA) and ORIGIN (Northampton, MA, USA). 2.2 Preparation of GOx Discs and Electrochemical Characterization Grade 1 filter paper was cut into round discs with ca. 9 mm diameter using a paper punch. Then, 5 µL of GOx solution (150 U mL-1 in 0.1 M PBS, pH 7.0) was carefully added to each paper disc and allowed to dry at room temperature (25 °C). These paper discs laden
Page 4 of 31
with GOx were used for glucose analyses and stored at 4 °C for stability study of the biosensor. Before conducting any electrochemical measurements, each new PB-SPCE was preconditioned by successive cyclic scanning at 25 mV s-1 from +0.30 V to −0.50 V for 5 cycles.
ip t
The PB-SPCE was slotted into the boxed connector that was connected to the potentiostat. Typically, the paper disc with immobilized GOx was placed on top of the PB-SPCE to
cr
completely cover the working, counter and reference electrodes before each measurement.
As shown in Figure (2), 12 µL of 0.1 M pH 7.0 PBS buffer solution was added onto
us
the paper disc and good contact was formed with the PB-SPCE with this volume of buffer.
an
Amperometric experiments were carried out at a potential of −0.30 V while cyclic voltammetry scans were performed from +0.30 V to −0.50 V at a scan rate of 10 mV s-1.
M
Unless otherwise indicated, all measurements were performed in triplicates on paper discs prepared using 5 μL from a 150 U ml-1 GOx solution (0.75 U disc-1), unless stated otherwise,
d
and glucose detection was performed using 0.5 μL from a 1 mM glucose solution.
te
Furthermore, all different concentrations of GOx and glucose solutions were prepared using 0.1 M pH 7.0 PBS buffer solution, unless stated otherwise.
Ac ce p
FIGURE 1 IS PLACED HERE.
3. Results and Discussion
3.1 Cyclic Voltammetric Investigation and Electrocatalytic Properties of the PB-SPCE Typically, there are two major groups of peaks observed in cyclic voltammograms for
PB-SPCEs. The cathodic group corresponds to the Prussian Blue/Prussian White redox reaction whereas at very high anodic potentials, i.e. Prussian Blue is converted to Berlin Green [17]. Preliminary investigations on the catalytic redox behavior of PB on H2O2 were first conducted with PB-SPCEs using blank paper discs whereby wide cyclic voltammetry scans were performed from +1.00 V to −1.00 V at a scan rate of 50 mV s-1. It was observed
Page 5 of 31
that the cathodic group showed a significant increase in current when H2O2 was added whereas the anodic group showed a small increase in current. This observation has been reported in the literature [32]. Since Prussian White has been reported to catalyze the reduction of H2O2 at low applied potential, the Prussian Blue/Prussian White redox reaction
ip t
was more favorable as the effect of potential electrochemical interfering compounds can be avoided or greatly reduced. Additionally, it was also observed that the increase in reduction
cr
current was more significant as compared to the slight increase in oxidation current in the cyclic voltammograms of the Prussian Blue/Prussian White redox reaction with H2O2 (data
us
not shown). Therefore, glucose detection based on the electrocatlytic reduction of H2O2 was preferred and the cathodic peaks representing the Prussian Blue/Prussian White redox
an
reaction were focused on for this study.
M
The performance of SPCE and PB-SPCE for the detection of glucose using GOx disc was investigated via cyclic voltammetry. Figure (2) shows two sets of cyclic voltammograms
d
for GOx disc on the SPCE (I) and PB-SPCE (II). For set I, no redox peaks were observed
te
and addition of 5 mM glucose did not produce any change to the cyclic voltammogram. However, for set II, redox of PB was observed in (A) and addition of 5 mM glucose (B)
Ac ce p
caused an increase in the reduction current at ca. -0.24 V. Thus, it was demonstrated that PB was a good electrocatalyst for the reduction of H2O2. FIGURE 2 IS PLACED HERE.
Figure (3) shows the cyclic voltammograms of the GOx disc on PB-SPCE with buffer
alone (A), addition of 1 mM (B) and 5 mM of (C) glucose solutions, respectively. It can be noted that upon adding glucose, significantly well-defined reduction peaks were observed at a low applied potential of ca. −0.24 V. The reduction current increased with increasing concentration of glucose.
Thus, it could be deduced that with increasing glucose
concentration, more H2O2 was formed and thus resulting in increasing reduction current. On
Page 6 of 31
the other hand, the oxidation peaks did not significantly increase as much as in comparison to the reduction peaks. It can also be inferred that for a paper disc on the PB-SPCE, a small sample volume of 0.5 µL was sufficient in producing excellent signals. As with previous literature, these scans are characteristic of the enzymatic reaction of
ip t
GOx on glucose and H2O2 reduction was catalyzed by PB [32]. Moreover, the enhancement of reduction current in the scans clearly demonstrates that the close proximity of PB and the
cr
GOx disc allowed for its efficient electrocatalytic role for the H2O2 reduction reaction.
us
FIGURE 3 IS PLACED HERE.
3.2 Selection of the Applied Potential
an
The selection of the applied potential at the working electrode plays a significant role
M
in the response of the paper-based glucose biosensor. The effect of applied potential on the amperometric response of the paper-based biosensor to glucose is shown in Figure (4). The
te
−0.36 V.
d
signals produced in response to glucose were measured in the potential range of −0.24 V to
It was found that the sensitivity of the paper-based glucose biosensor increased
Ac ce p
progressively with as potential was changed from −0.24 V to −0.30 V, with a maximum at −0.30 V and declined sharply when the potential was brought more negative from −0.33 V to −0.36 V. This increased sensitivity with applied potential from −0.24 V to −0.30 V was attributed to the upsurge in the driving force for the reduction of H2O2. On the basis of the above results, a potential of −0.30 V was selected thereafter for the subsequent experiments in order to attain paramount sensitivity. FIGURE 4 IS PLACED HERE. 3.3 Selection of pH An optimum pH range is vital to the sensitivity of glucose biosensors as it influences both the bioactivity of the GOx and electrochemical behavior of PB. It has been reported that
Page 7 of 31
extreme pH may possibly modify the kinetics of glucose measurements as a result of disturbing the redox state of the GOx reaction [33]. The effect of pH on the biosensor response was investigated in the range of 5.0 to 9.0 and the trend is shown in Figure (5). From Figure (5), it can be deduced that the paper-based glucose biosensor exhibited
ip t
an optimum response at pH 7.0 over the experimented pH range of 5.0 to 9.0 in 0.1M PBS buffer solutions. It can also be noted that the electrochemical response of the paper-based
cr
glucose biosensor was very poor when exposed to strong acidic or alkaline environments. This is because high acidity causes a decrease in GOx activity and strong alkalinity causes a
us
drop in PB stability and decrease of GOx activity too [34]. On the basis of the above results, a pH of 7.0 was selected thereafter for the subsequent experiments in order to attain paramount
an
sensitivity.
M
FIGURE 5 IS PLACED HERE.
d
3.4 GOx Loading
te
Typically enzyme loading affects the amperometric response of the paper-based glucose biosensor [31]. Furthermore, to improve the biosensor performance, the amount of
Ac ce p
GOx should be sufficient so as to achieve a broad linear response range. The trend is such that with increasing GOx loading, more H2O2 is formed and hence leading to an increase in current. Different concentrations of GOx (0.25 to 1 U disc-1) immobilized on paper discs were used to determine the optimum response for the paper-based glucose biosensor. As shown in Figure (6), it can be observed that the sensitivity of the biosensor
increased steadily with increasing concentration of glucose oxidase. GOx concentrations of 0.75 and 1.0 U disc-1 on the PB-SPCE displayed maximum sensitivity with a wide linear range in the concentration range of 0.5 to 2 mM glucose. Figure (6), showed that for various GOx concentrations, the reduction current increased rapidly and linearly initially at low substrate concentrations but gradually levelled off towards high glucose concentrations. This
Page 8 of 31
trend is in accordance with enzyme kinetics whereby the rate becomes zero order because all the enzyme molecules are bound to the substrate molecules. On the basis of the above results, an optimized enzyme concentration of 0.75 U disc-1 was selected in order to attain an optimum sensitivity. However, a drop in current was
ip t
observed in this graph when compared to the other graphs obtained previously. For instance in Figures (4) and (5), a 0.75 U disc-1 with 1 mM glucose concentration produced an average
cr
current of 2.35 μA at a potential of −0.3 V. Whereas in Figure (6), a 0.75 U disc-1 with 1 mM glucose concentration produced a current response of 1.16 μA at the same potential. This
us
lower response can be accounted for the continuous leaching of PB from the PB-SPCE due to
an
the increased length of time needed when investigating this particular parameter [24,29].
3.5 Biosensor Response Characteristics
M
FIGURE 6 IS PLACED HERE.
d
By employing the optimum conditions from the above parameters investigated, the
te
calibration graph of the sensor was obtained as shown in Figure (7). The PB-modified screenprinted glucose biosensors exhibited a linear range between 0.25 to 2.00 mM. The linear
Ac ce p
regression equation was y=2.1387x + 0.489, where y signifies the current in μA and x is the glucose concentration in mM with R2 value was found to be 0.996. In addition, the relative standard deviation of the biosensor at 1 mM glucose was 4.3% (n=6) thereby showing good reproducibility and the limit of detection (LOD), calculated based on IUPAC definition was 0.01 mM glucose, with n = 5 at 90% confidence level [35]. FIGURE 7 IS PLACED HERE.
The analytical performance of the present biosensor was compared against other types of PB-based glucose biosensors as shown in Table (1). It can be noted that the LOD for the present biosensor is lower than our earlier paper-based mediated glucose biosensor [31] and other PB-modified glucose biosensors [26,28,36,]. Despite the simplicity of the present
Page 9 of 31
biosensor, it produced a lower LOD compared to the highly complex paper-based electrochemical device fabricated with an integral battery and PB electrochromic read-out, i.e. colour change of PB due to different redox state in the presence of glucose [37]. Moreover, the linear range of this simple biosensor is comparable with other more complex
ip t
PB-modified glucose biosensors [26,28,36] and is wider compared to our previous paperbased mediated glucose biosensor [31].
cr
TABLE 1 IS PLACED HERE
The storage stability of the biosensors in solution was investigated at 4°C. The
us
response of the paper-based glucose biosensor as a function of the storage time was shown in Figure (8). The response to 1 mM glucose of the paper-based glucose biosensors, stored at
an
4 °C when not in use and tested for 45 days retained ca. 72% of the initial current value. The
M
lifetime of the present paper-based biosensor was compared with other types of PB-based glucose biosensors as shown in Table (1). Our shelf life study showed that the present low
d
cost, simple biosensor was comparable to other more complex PB-modified biosensors based
te
on carbon paste electrodes with solid paraffin, i.e. 0% drop in the signal after 30 days of storage [28]. Also, the sensitivity of a glucose biosensor based on PB-modified screen-
Ac ce p
printed electrodes and GOx immobilized on nafion dropped by 50% within half a day [26]. In this work, the immobilization technique has been extremely low cost and straightforward as no additional chemical reagents, apart from pH 7.0 PBS buffer solution, and thus the fiber matrix in paper proved to be an effective stable microenvironment for GOx. FIGURE 8 IS PLACED HERE.
3.6 Selectivity against Interferences Electroactive interferences have been a problem when biological or industrial samples are assayed by amperometric biosensors. The effect of electroactive interferents that might affect the response of the paper-based glucose biosensor was investigated. Nine possible
Page 10 of 31
interferents were used to evaluate the selectivity of the glucose biosensor. The current obtained for each interfering substance at a concentration of 1 mM (unless otherwise stated) in the presence of 1 mM glucose (Iθ) is compared to the current obtained in the presence of 1mM glucose alone (I) as a criterion for the selectivity of the biosensor, shown in Table (2).
ip t
Acetic acid, ascorbic acid, citric acid, fructose, oxalic acid, sucrose, urea and uric acid do not cause any observable interference. Only L-cysteine interfered significantly.
cr
TABLE 2 IS PLACED HERE.
us
3.7 Validation Study TABLE 3 IS PLACED HERE.
an
To ascertain the applicability of the paper-based glucose biosensor, glucose concentrations in two commercial beverages were measured by the paper-based biosensor
M
and the results were compared quantitatively with the results obtained from Benedict’s Test along with UV/Vis spectrophotometry, shown in Table (3). Beverage A was a pure glucose
d
energy supplement (Glucolin) and beverage B was an oral rehydration solution comprising of
te
electrolytes and glucose only (Hydralyte). The original concentrations were obtained from
Ac ce p
the nutritional information stated on container of beverages. The glucose concentration determined by the two methods corroborates rather well and it can be inferred that the results obtained by the biosensor had lower RSD compared to the Benedict’s test.
4. Conclusions
In the present study, an amperometric glucose biosensor based on PB-SPCEs was
developed by using paper as a matrix for GOx immobilization via a simple physical adsorption method. This immobilization method on a green matrix is extremely beneficial as complex immobilization methods involving the use of multiple chemicals can be avoided and this also ensures that the original enzyme configuration is not compromised thus allowing optimal activity. It was demonstrated that the integration of paper disc as a matrix for enzyme
Page 11 of 31
immobilization with SPEs was highly advantageous due to the simplicity and costeffectiveness of its preparation procedures, easy control of enzyme loading with good shelflife. This novel paper-based glucose biosensor provides a robust and convenient platform with promising prospects for applications which require miniturization, small volumes of
ip t
sample, high sensitivity and selectivity such as point-of-care medical diagnostics in resourcelimited settings as well as field-deployable environmental analysis and monitoring.
cr
Sensitivity could be enhanced by incorporating nano-electrocatalysts into paper discs.
us
5. Acknowledgements
The team acknowledges the generous funding provided by National Institute of Education-
Ac ce p
te
d
M
an
Nanyang Technological University (NIE-NTU) (RI7/12 TSN).
Page 12 of 31
6. References [1]
S.M. Hossain, R.E. Luckham, A.M. Smith, J.M. Lebert, L.M. Davies, R.H. Pelton, C.D. Filipe, J.D. Brennan, Development of a bioactive paper sensor for detection of neurotoxins using piezoelectric inkjet printing of sol-gel-derived bioinks, Anal. Chem.
ip t
81 (2009) 5474-5483. W.A. Zhao, A. Van Den Berg, Lab on paper, Lab Chip. 8 (2008) 1988–1991.
[3]
D.D. Liana, B. Raguse, J.J. Gooding, E. Chow, Recent advances in paper-based
cr
[2]
sensors, Sensors (Basel). 12 (2012) 11505-11526.
A.W. Martinez, S.T. Phillips, G.M. Whitesides, E. Carrilho, Diagnostics for the
us
[4]
developing world: Microfluidic paper-based analytical devices, Anal. Chem. 82 (2010)
W. Lewis, E. Keshavarz-Moore, J. Windust, D. Bushell, N. Parry, Construction and
M
[5]
an
3–10.
evaluation of novel fusion proteins for targeted delivery of micro particles to cellulose
S. Su, R. Nutiu, C.D. Filipe, Y. Li, R. Pelton, Adsorption and covalent coupling of
te
[6]
d
surfaces, Biotechnol. Bioeng. 94 (2006) 625-632.
ATP-binding DNA aptamers onto cellulose, Langmuir. 23 (2007) 1300-1302. M. Tolba, L.Y. Brovko, M.W. Griffiths, Engineering of bacteriophages displaying
Ac ce p
[7]
affinity tags on its head for biosensor applications, Nanotechnology: Life Sciences, Medicine & Bio Materials. 2 (2008) 449-452.
[8]
S.J. Craig, A. Shu, Y. Xu, F.C. Foong, R. Nordon, Chimeric protein for selective cell
attachment onto cellulosic substrates, Protein Engineering, Design & Selection. 20 (2007) 235-241.
[9]
R. Pelton, Bioactive paper provides a low-cost platform for diagnostics, Trac-Trends in Anal. Chem. 28 (2009) 925-942.
Page 13 of 31
[10] P. Yager, T. Edwards, E. Fu, K. Helton, K. Nelson, M.R. Tam, B.H. Weigl, Microfluidic diagnostic technologies for global public health, Nature. 442 (2006) 412418. [11] A.W. Martinez, S.T. Phillips, M.J. Butte, G.M. Whitesides, Patterned paper as a
ip t
platform for inexpensive, low-volume, portable bioassays, Angew. Chem. Int. Ed. Engl. 46 (2007) 1318–1320.
cr
[12] A.W. Martinez, S.T. Phillips, E. Carrilho, S.W. Thomas, H. Sindi, G.M. Whitesides, Simple telemedicine for developing regions: camera phones and paper-based
us
microfluidic devices for real-time, off-site diagnosis, Anal. Chem. 80 (2008) 3699– 3707.
an
[13] A.W. Martinez, S.T. Phillips, G.M. Whitesides, Three-dimensional microfluidic
M
devices fabricated in layered paper and tape, Proc. Natl. Acad. Sci. USA. 105 (2008) 19606–19611.
d
[14] A.W. Martinez, S.T. Phillips, B.J. Wiley, M. Gupta, G.M. Whitesides, FLASH: A
(2008) 2146-2150.
te
Rapid Method for Prototyping Paper-Based Microfluidic Devices, Lab on a Chip. 8
Ac ce p
[15] J. Wang, Electrochemical glucose biosensors, Chem. Rev. 108 (2008) 814-825. [16] J.D. Newman, S.J. Setford, Enzymatic biosensors, Mol. Biotechnol. 32 (2006) 249268.
[17] A.A. Karyakin, E.E. Karyakina, L. Gorton, On the mechanism of H2O2 reduction at Prussian Blue modified electrodes, Electrochem. Commun. 1 (1999) 78–82.
[18]
A.A. Karyakin, O.V. Gitelmacher, E.E. Karyakina, Prussian Blue-based first-
generation biosensor. A sensitive amperometric electrode for glucose, Anal. Chem. 67 (1995) 2419–2423.
Page 14 of 31
[19] A.V. Mokrushina, M. Heim, E.E. Karyakina, A. Kuhn, A.A. Karyakin, Enhanced hydrogen peroxide sensing based on Prussian Blue modified macroporous microelectrodes. Electrochem. Commun. 29 (2013) 78-80. [20] A.A. Karyakin, E.E. Karyakina, L. Gorton, Prussian-Blue based amperometric
ip t
biosensors in flow-injection analysis, Talanta. 43 (1996) 1597–1606. [21] A.A. Karyakin, E.E. Karyakina, L. Gorton, The electrocatalytic activity of Prussian
cr
Blue in hydrogen peroxide reduction studied using a wall-jet cell with continuous flow, J. Electroanal. Chem. 456 (1998) 97-104.
applications, Electroanalysis. 13 (2001) 813-819.
us
[22] A.A. Karyakin, Prussian Blue and its analogues: electrochemistry and analytical
an
[23] A. Boyer, K. Kalcher, R. Pietsch, Voltammetric behavior of perborate on Prussian
M
Blue modified carbon paste electrodes, Electroanalysis. 2 (1990) 155–161. [24] A.A. Karyakin, M. Chaplin, Polypyrrole Prussian Blue films with controlled level of
te
(1994) 85–92.
d
doping: codeposition of polypyrrole and Prussian Blue, J. Electroanal. Chem. 56
[25] F. Ricci, A. Amine, G. Palleschi, D. Moscone, Prussian Blue based screen printed
Ac ce p
biosensors with improved characteristics of long-term lifetime and pH stability, Biosens. Bioelectron. 18 (2003) 165– 174.
[26] M.P. O’Halloran, M. Pravda, G.G. Guilbault, Prussian Blue bulk modified screenprinted electrodes for H2O2 detection and for biosensors, Talanta. 55 (2001) 605–611.
[27] A. M. Vinu Mohan, K. K. Aswini, A. Maria Starvin, M. Biju, Amperometric detection of glucose using Prussian blue-graphene oxide modified platinum electrode, Anal. Methods. 5 (2013) 1764-1770. [28] D. Moscone, D. D’Ottavi, D. Compagnone, G. Palleschi, Construction and Analytical Characterization of Prussian Blue-Based Carbon Paste Electrodes and Their Assembly as Oxidase Enzyme Sensors, Anal. Chem. 73 (2001) 2529-2535.
Page 15 of 31
[29] Q. Deng, B. Li, S. Dong, Self-gelatinizable copolymer immobilized glucose biosensor based on Prussian Blue modified graphite electrode, Analyst. 123 (1998) 1995-1999. [30] J. Wang, X. Zhang, Screen-printed cupric-hexacyanoferrate modified carbon enzyme electrode for single-use glucose measurements, Anal. Lett. 32 (1999) 1739-1749.
glucose biosensor, Sensor Actuat. B-Chem. 193 (2014) 536-541.
ip t
[31] C.S.L. Kuek, S.N. Tan, C.Z. Floresca, A “green” cellulose paper based amperometric
cr
[32] Q. Chi, S. Dong, Amperometric biosensors based on the immobilization of oxidases in a Prussian blue film by electrochemical codeposition, Anal. Chim. Acta. 310 (1995)
us
429-436.
[33] Z. Tang, X. Du, R.F. Louie, G.J. Kost, Effects of pH on Glucose Measurements With
an
Handheld Glucose Meters and a Portable Glucose Analyzer for Point-of-Care Testing,
M
Arch. Pathol. Lab. Med. 124 (2000) 577-582.
[34] S. Banerjee, P. Sarkar, A.P.F. Turner, Amperometric biosensor based on Prussian
d
Blue nanoparticle-modified screen-printed electrode for estimation of glucose-6-
te
phosphate, Analytical Biochemistry. 439 (2013) 194-200. [35] J. Mocak, A.M. Bond, S. Mitchell, G. Scollary, A statistical overview of standard
Ac ce p
(IUPAC AND ACS) and new procedures for determining the limits of detection and quantification: application to voltammetric and stripping techniques, Pure Appl. Chem. 69 (1997) 297–328.
[36] F. Ricci, C. Gonçalves, A. Amine, L. Gorton, G. Palleshi, D. Moscone, Electroanalytical study of prussian blue modified glassy carbon paste electrodes, Electroanal. 15 (2003) 1204-1211.
[37] H. Liu, R.M. Crooks, Paper-based electrochemical sensing platform with integral battery and electrochromic read-out, Anal. Chem. 84 (2012) 2528-2532.
Page 16 of 31
Authors’ Biography Nadia Chandra Sekar received her BSc (Hons) in Biomedical Sciences from the University of Bradford and is currently a Master of Science (Research) student from the Nanyang Technological University (NTU), Singapore. Dr Tan Swee Ngin is an Associate Professor at NTU, Singapore. Her current research areas include electroanalytical chemistry and separation science.
ip t
the Singapore Institute of Singapore with focus on activities are focused on fuel cells; and design and
cr
Dr Seyed Ali Mousavi Shaegh is a Research Scientist at Technology (SIMTech). He completed his PhD at NTU, Microfluidic Fuel Cells at 2012. Currently, his research electrochemical microdevices including biosensors and micro fabrication of microfluidic devices for biomedical applications.
an
us
Dr Sum Huan Ng received his BEng and MEng in mechanical engineering in 1997 and 2000 respectively from the National University of Singapore. He obtained his PhD from the Georgia Institute of Technology. He is currently a research scientist with SIMTech working on the following areas: microfluidics, microfabrication techniques and abrasive removal processes.
Ac ce p
te
d
M
Dr Liya Ge obtained her PhD from NTU, Singapore. She has been a research fellow working at NTU since 2005.
Page 17 of 31
Figure Captions: Figure 1: Schematic illustration of GOx paper disc preparation and integration with the PBSPCE. 5 μL GOx solution was dropped on the surface of the paper disc with subsequent drying in air at room temperature; the paper disc with immobilized GOx was placed on top of the PB-SPCE surface for complete coverage of working, counter and reference electrodes; 12 µL PBS buffer solution with 0.5 μL glucose solution was dropped on the paper disc for electrochemical detection.
cr
ip t
Figure 2: Two sets of cyclic voltammograms for GOx disc on the SPCE (I) and PB-SPCE (II) in which each set contains one voltammogram with 12 μL of 0.1 M PBS (pH 7.0) buffer solution only (A) and another with the addition of 5 mM glucose solution (B). The cyclic voltammetry scans were performed from +0.30 V to -0.50 V at a scan rate of 10 mV s-1, [GOx]=0.25 U disc-1.
us
Figure 3: Cyclic Voltammograms for GOx disc on the PB-SPCE with 12 µL of 0.1 M PBS (pH 7.0) buffer solution (A), 1 mM (B) and 5 mM (C) glucose solutions. The cyclic voltammetry scans were performed from +0.30 V to -0.50 V at a scan rate of 10 mV s-1, [GOx]=0.75 U disc-1.
M
an
Figure 4: Effect of applied potential on the amperometric response of the GOx paper-based glucose biosensor on PB-SPCE with 1 mM glucose solution (pH 7), [GOx]=0.75 U disc-1. The error bars show standard deviations for n=3~4. Figure 5: Effect of pH on the amperometric response of the GOx paper-based glucose biosensor on PB-SPCE with 1 mM glucose solution at an applied potential of -0.30 V (vs Ag/AgCl), [GOx]=0.75 U disc-1. The error bars show standard deviations for n=3~4.
Ac ce p
te
d
Figure 6: Effect of GOx concentration (U disc-1) on the amperometric response of the paperbased glucose biosensor on PB-SPCE with various glucose concentrations at an applied potential of -0.30 V (vs Ag/AgCl). Paper discs with GOx concentrations of 0.25 U disc-1 (A), 0.50 U disc-1 (B), 0.75 U disc-1 (C) and 1 U disc-1 (D), n=3~4. Figure 7: Calibration curve of the amperometric response of the paper-based glucose biosensor on PB-SPCE with various glucose concentrations at an applied potential of -0.30 V (vs Ag/AgCl), pH=7.0, [GOx]=0.75 U disc-1. The error bars show standard deviations for n=3~4. Figure 8: Stability of the amperometric response of the paper-based glucose biosensor on PB-SPCE during 30 days determined with 1 mM glucose in PBS buffer (pH 7) at an applied potential of -0.30 V (vs Ag/AgCl), [GOx]=0.75 U disc-1. The error bars show standard deviations for n=6.
Page 18 of 31
Linear Range (mM)
Detection Limit (mM)
Lifetime
Paper-based/PB-SPCE
0.25 to 2.0
0.01
72% for 45 days
−
PB-based carbon paste electrodes with solid paraffin
0.1 to 20
0.10
cr
Table 1: Comparison of the analytical performance of the present biosensor against other related biosensors.
[28]
an
0.22
ip t
Lost 50% sensitivity after 4 hours
[26]
0.05 to 0.8
0.05
−
[36]
M
PB-glassy carbon paste electrodes Paper-based/SPCE with ferrocene carboxylic acid as mediator Paper-based electrochemical sensing platform with integral battery and PB electrochromic readout
Until 3
100% for 30 days
0.18
98% for 120 days
[31]
0.10
−
[37]
1.0 to 5.0
−
Ac ce p
Paperbased glucose biosensors
PB-modified screenprinted electrodes with GOx immobilized on nafion
References
d
PBmodified electrodes for glucose analysis
te
Current Work
us
Types of PB-based glucose biosensors
Page 19 of 31
Table 2: Possible Interferences Tested With The Paper-Based Glucose Biosensor
Acetic Acid
0.9
Citric Acid
1.0
Fructose
1.0
L-Cysteine
0.6
Oxalic Acid
1.1
Sucrose
0.9
cr
Ascorbic Acid
ip t
Current Ratioa (Iθ/I) 1.0
Possible Interfering Substance
Urea
1.0
1.2
us
Uric Acid
Current ratios (Iθ/I) for mixtures of 1mM interfering substance and 1mM glucose (C*) in comparison to that of 1mM glucose alone (C), [GOx]=0.75 U disc-1, n=3~4.
Ac ce p
te
d
M
an
a
Page 20 of 31
Table 3: Glucose Beverage Analysis Based On Amperometry and Benedict’s Tests
Benedict's Testb (g/L)
Paper-Based Glucose Biosensorc (g/L)
A
89.5
105.0±0.0003
90.4±0.1
B
14.4
11.9±0.001
14.1±0.1
cr
ip t
Beverage
Original Concentrationa (g/L)
us
a
Ac ce p
te
d
M
an
The original glucose concentrations were calculated based on the nutritional information stated on the beverage packaging. b.c Values are the mean (±standard error) of n=6~8, c [GOx]=0.75 U disc-1.
Page 21 of 31
Graphical Abstract (for review)
Ac ce
pt
ed
M
an
us
cr
ip t
Graphical Abstract
Page 22 of 31
Research Highlights
Highlights •
Ac ce
pt
ed
M
an
us
cr
ip t
• •
Glucose oxidase (GOx) was physically adsorbed on paper without use of chemicals. This biosensor configuration was cost-effective and showed good stability. Simple platform for applications that require small sample volumes.
Page 23 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 1
Page 24 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 2
Page 25 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 3
Page 26 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 4
Page 27 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 5
Page 28 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 6
Page 29 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 7
Page 30 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 8
Page 31 of 31
S0925-4005(14)00943-5 http://dx.doi.org/doi:10.1016/j.snb.2014.07.103 SNB 17258
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
17-3-2014 3-7-2014 23-7-2014
Please cite this article as: N.C. Sekar, S.A.M. Shaegh, N.S.H. Gary, G. Liya, T.S. Ngin, A paper-based amperometric glucose biosensor developed with prussian blue-modified screen-printed electrodes, Sensors and Actuators B: Chemical (2014), http://dx.doi.org/10.1016/j.snb.2014.07.103 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A PAPER-BASED AMPEROMETRIC GLUCOSE BIOSENSOR DEVELOPED WITH PRUSSIAN BLUE-MODIFIED SCREEN-PRINTED ELECTRODES
ip t
Nadia Chandra Sekara,§, Seyed Ali Mousavi Shaeghb,#, Ng Sum Huan Garyb,‡, Ge Liyac,^, Tan Swee Ngina,*
a
cr
Natural Sciences and Science Education Academic Group, Nanyang Technological University, 1 Nanyang Walk, Singapore 637616, Singapore b
us
Singapore Institute of Manufacturing Technology, 71 Nanyang Drive, Singapore 638075, Singapore
c
M
an
Residues and Resource Reclamation Centre, Nanyang Environment and Water Research Institute, Nanyang Technological University, 1 Cleantech Loop, Cleantech One, Singapore 637141, Singapore
[email protected]
#
[email protected]
‡
[email protected]
Ac ce p
^
te
§
d
Email Addresses of Authors:
[email protected]
Corresponding Author’s Email: *[email protected]
Telephone: +65 67903810 (Swee Ngin Tan)
Page 1 of 31
1. Introduction New strategies intended for rapid detection of analytes that are of clinical and environmental importance without requiring sophisticated instrumentation is currently in high demand [1]. Lately, paper has drawn much interest as a potential material for biosensors in
ip t
analytical and clinical chemistry due to it being versatile, highly abundant and inexpensive [2]. Upon varying its pulp processing, paper can be made thin, lightweight and flexible to suit
cr
a specific application [3]. Cellulose fibers, being the key component of paper, allow liquid to infiltrate its hydrophilic fiber matrix with no active pumps or external sources required [4].
us
Hence, this makes paper a possible simple and effective matrix for enzyme immobilization in paper-based biosensors via physical adsorption. In fact, antibodies [5], DNA aptamers [6],
an
phages [7] and cells [8] have also been applied in the development of viable paper-based
M
biosensors apart from enzymes [9]. Many paper-based analytical devices are used as lowcost substitutes for medical point-of-care diagnostics due to the necessity of fast, reliable and
d
affordable diagnostic tools in poverty-stricken developing countries [10]. Lately, Whitesides’
te
group has developed techniques for creating microfluidic devices from patterned paper and demonstrated the use of paper-based microfluidic devices for colorimetric detection of
Ac ce p
glucose and protein in artificial urine [11-14]. Glucose Oxidase (GOx) is widely used in practical applications for quick and precise
glucose analysis. To date, 85% of the world biosensor market is currently dominated by electrochemical glucose biosensors [15-16] based on GOx-modified Screen-Printed Carbon Electrodes (SPCEs). As shown in equation (1), GOx catalyzes glucose oxidation in the presence of molecular oxygen (O2) producing gluconic acid and hydrogen peroxide (H2O2).
Glucose + O2
GOx
Gluconic Acid + H2O2
(1)
Monitoring of glucose levels can be based on either O2 consumption or H2O2 generation. Traditionally, most electrochemical glucose biosensors were based on O2 detection but the
Page 2 of 31
detection of H2O2 generation was found to be more sensitive [17]. At conventional electrodes, H2O2 detection occurs at a relatively high potential ca. +0.60 V versus Ag/AgCl which leads to serious interference [18]. Thus, suitable electrocatalysts are employed to resolve this issue [17].
ip t
Prussian Blue (PB) or ferric ferrocyanide (FeIII4[FeII(CN)6]3) has been used in the development of amperometric glucose biosensors because of its excellent electrocatalytic
cr
properties for H2O2 reduction [17,19-21] at low applied potential. The reduction current of H2O2 on PB-modified electrodes is two orders of magnitude higher than that of O2 reduction
us
for glucose analysis [22]. Moreover, the high catalytic activity and selectivity of PB lowers the operating potential to avoid or greatly reduce the contribution from potential interfering
an
compounds, e.g. ascorbic acid [18,20]. Previous literature has reported applied potentials
M
ranging from +0.18 to −0.40 V vs. Ag/AgCl for PB-modified glucose biosensors [23-26]. Furthermore, an oxidized state of PB, Berlin Green, also allows glucose detection but it is
d
based on electro-oxidation of H2O2 at PB-modified electrodes [17]. So far, PB-modified
te
platinum [27], carbon paste [28], graphite [29], glassy carbon [21] and SPEs [30] have been studied for glucose detection.
Ac ce p
Previously, we reported on the integration of paper disc with SPCEs for the
development of a ferrocene carboxylic acid mediated glucose biosensor [31]. As the mediator was incorporated in the paper disc, it reached saturation and leached out easily. Saturation of the mediator on the paper disc could be a limiting factor for the linearity of the biosensor response. The applied potential of the above biosensor was in the positive range, ca. 0.25 V, a relatively high operating potential which could lead to interference from other oxidizable compounds, e.g. ascorbic acid. In this study, we would like to overcome the above limitations of the paper disc mediated glucose biosensor [31] and explore the development of glucose biosensor by immobilizing GOx within porous structure of paper discs placed on top of PB-modified SPCEs, as shown in Figure (1). With this biosensor, only 0.5 µL of analyte
Page 3 of 31
was required for the detection of glucose. This results in cost saving due to a reduction in reagent consumption and prevents incessant exposure to bulk solution, as in a typical mediated glucose biosensor [31]. Therefore, leaching of PB from the SPCE, resulting in loss of catalytic activity, is significantly reduced. The various parameters for the optimization of
ip t
the biosensor development such as applied potential, pH and GOx loading were investigated. 2. Experimental
cr
2.1 Reagents and Instrumentation
All reagents used were of analytical grade. GOx from Aspergillus Niger (type X-S,
us
EC 232-601-0; 234,900 U g-1 solid) were purchased from Sigma-Aldrich (St Louis, MO, USA). β-D-glucose was purchased from Nacalai Tesque (Kyoto, Japan). For validation
an
studies, glucolin glucose powder (420g) and hydralyte were purchased from a local
M
supermarket and pharmacy store respectively. All solutions were prepared with 18 mΩ ultrapure water obtained from Millipore Alpha-Q water system (Bedford, MA, USA). All
d
electrochemical characterizations and measurements were performed using a four-channel
te
system (eDAQ QuadStat, e-Corder 8 and Echem software, eDAQ Europe, Poland). PB-SPCEs (DRP-710) and the boxed connector for SPEs (DRP-DSC) were purchased
Ac ce p
from DropSens (Asturias, Spain). The working electrode (4 mm diameter) consisted of Carbon/Prussian Blue, while Ag/AgCl and a carbon ring were the reference and counter electrodes, respectively. Paper discs were cut from Grade 1 filter papers [Whatman Asia Pacific Pte Ltd (Singapore)]. Data points were plotted using Microsoft Excel (USA) and ORIGIN (Northampton, MA, USA). 2.2 Preparation of GOx Discs and Electrochemical Characterization Grade 1 filter paper was cut into round discs with ca. 9 mm diameter using a paper punch. Then, 5 µL of GOx solution (150 U mL-1 in 0.1 M PBS, pH 7.0) was carefully added to each paper disc and allowed to dry at room temperature (25 °C). These paper discs laden
Page 4 of 31
with GOx were used for glucose analyses and stored at 4 °C for stability study of the biosensor. Before conducting any electrochemical measurements, each new PB-SPCE was preconditioned by successive cyclic scanning at 25 mV s-1 from +0.30 V to −0.50 V for 5 cycles.
ip t
The PB-SPCE was slotted into the boxed connector that was connected to the potentiostat. Typically, the paper disc with immobilized GOx was placed on top of the PB-SPCE to
cr
completely cover the working, counter and reference electrodes before each measurement.
As shown in Figure (2), 12 µL of 0.1 M pH 7.0 PBS buffer solution was added onto
us
the paper disc and good contact was formed with the PB-SPCE with this volume of buffer.
an
Amperometric experiments were carried out at a potential of −0.30 V while cyclic voltammetry scans were performed from +0.30 V to −0.50 V at a scan rate of 10 mV s-1.
M
Unless otherwise indicated, all measurements were performed in triplicates on paper discs prepared using 5 μL from a 150 U ml-1 GOx solution (0.75 U disc-1), unless stated otherwise,
d
and glucose detection was performed using 0.5 μL from a 1 mM glucose solution.
te
Furthermore, all different concentrations of GOx and glucose solutions were prepared using 0.1 M pH 7.0 PBS buffer solution, unless stated otherwise.
Ac ce p
FIGURE 1 IS PLACED HERE.
3. Results and Discussion
3.1 Cyclic Voltammetric Investigation and Electrocatalytic Properties of the PB-SPCE Typically, there are two major groups of peaks observed in cyclic voltammograms for
PB-SPCEs. The cathodic group corresponds to the Prussian Blue/Prussian White redox reaction whereas at very high anodic potentials, i.e. Prussian Blue is converted to Berlin Green [17]. Preliminary investigations on the catalytic redox behavior of PB on H2O2 were first conducted with PB-SPCEs using blank paper discs whereby wide cyclic voltammetry scans were performed from +1.00 V to −1.00 V at a scan rate of 50 mV s-1. It was observed
Page 5 of 31
that the cathodic group showed a significant increase in current when H2O2 was added whereas the anodic group showed a small increase in current. This observation has been reported in the literature [32]. Since Prussian White has been reported to catalyze the reduction of H2O2 at low applied potential, the Prussian Blue/Prussian White redox reaction
ip t
was more favorable as the effect of potential electrochemical interfering compounds can be avoided or greatly reduced. Additionally, it was also observed that the increase in reduction
cr
current was more significant as compared to the slight increase in oxidation current in the cyclic voltammograms of the Prussian Blue/Prussian White redox reaction with H2O2 (data
us
not shown). Therefore, glucose detection based on the electrocatlytic reduction of H2O2 was preferred and the cathodic peaks representing the Prussian Blue/Prussian White redox
an
reaction were focused on for this study.
M
The performance of SPCE and PB-SPCE for the detection of glucose using GOx disc was investigated via cyclic voltammetry. Figure (2) shows two sets of cyclic voltammograms
d
for GOx disc on the SPCE (I) and PB-SPCE (II). For set I, no redox peaks were observed
te
and addition of 5 mM glucose did not produce any change to the cyclic voltammogram. However, for set II, redox of PB was observed in (A) and addition of 5 mM glucose (B)
Ac ce p
caused an increase in the reduction current at ca. -0.24 V. Thus, it was demonstrated that PB was a good electrocatalyst for the reduction of H2O2. FIGURE 2 IS PLACED HERE.
Figure (3) shows the cyclic voltammograms of the GOx disc on PB-SPCE with buffer
alone (A), addition of 1 mM (B) and 5 mM of (C) glucose solutions, respectively. It can be noted that upon adding glucose, significantly well-defined reduction peaks were observed at a low applied potential of ca. −0.24 V. The reduction current increased with increasing concentration of glucose.
Thus, it could be deduced that with increasing glucose
concentration, more H2O2 was formed and thus resulting in increasing reduction current. On
Page 6 of 31
the other hand, the oxidation peaks did not significantly increase as much as in comparison to the reduction peaks. It can also be inferred that for a paper disc on the PB-SPCE, a small sample volume of 0.5 µL was sufficient in producing excellent signals. As with previous literature, these scans are characteristic of the enzymatic reaction of
ip t
GOx on glucose and H2O2 reduction was catalyzed by PB [32]. Moreover, the enhancement of reduction current in the scans clearly demonstrates that the close proximity of PB and the
cr
GOx disc allowed for its efficient electrocatalytic role for the H2O2 reduction reaction.
us
FIGURE 3 IS PLACED HERE.
3.2 Selection of the Applied Potential
an
The selection of the applied potential at the working electrode plays a significant role
M
in the response of the paper-based glucose biosensor. The effect of applied potential on the amperometric response of the paper-based biosensor to glucose is shown in Figure (4). The
te
−0.36 V.
d
signals produced in response to glucose were measured in the potential range of −0.24 V to
It was found that the sensitivity of the paper-based glucose biosensor increased
Ac ce p
progressively with as potential was changed from −0.24 V to −0.30 V, with a maximum at −0.30 V and declined sharply when the potential was brought more negative from −0.33 V to −0.36 V. This increased sensitivity with applied potential from −0.24 V to −0.30 V was attributed to the upsurge in the driving force for the reduction of H2O2. On the basis of the above results, a potential of −0.30 V was selected thereafter for the subsequent experiments in order to attain paramount sensitivity. FIGURE 4 IS PLACED HERE. 3.3 Selection of pH An optimum pH range is vital to the sensitivity of glucose biosensors as it influences both the bioactivity of the GOx and electrochemical behavior of PB. It has been reported that
Page 7 of 31
extreme pH may possibly modify the kinetics of glucose measurements as a result of disturbing the redox state of the GOx reaction [33]. The effect of pH on the biosensor response was investigated in the range of 5.0 to 9.0 and the trend is shown in Figure (5). From Figure (5), it can be deduced that the paper-based glucose biosensor exhibited
ip t
an optimum response at pH 7.0 over the experimented pH range of 5.0 to 9.0 in 0.1M PBS buffer solutions. It can also be noted that the electrochemical response of the paper-based
cr
glucose biosensor was very poor when exposed to strong acidic or alkaline environments. This is because high acidity causes a decrease in GOx activity and strong alkalinity causes a
us
drop in PB stability and decrease of GOx activity too [34]. On the basis of the above results, a pH of 7.0 was selected thereafter for the subsequent experiments in order to attain paramount
an
sensitivity.
M
FIGURE 5 IS PLACED HERE.
d
3.4 GOx Loading
te
Typically enzyme loading affects the amperometric response of the paper-based glucose biosensor [31]. Furthermore, to improve the biosensor performance, the amount of
Ac ce p
GOx should be sufficient so as to achieve a broad linear response range. The trend is such that with increasing GOx loading, more H2O2 is formed and hence leading to an increase in current. Different concentrations of GOx (0.25 to 1 U disc-1) immobilized on paper discs were used to determine the optimum response for the paper-based glucose biosensor. As shown in Figure (6), it can be observed that the sensitivity of the biosensor
increased steadily with increasing concentration of glucose oxidase. GOx concentrations of 0.75 and 1.0 U disc-1 on the PB-SPCE displayed maximum sensitivity with a wide linear range in the concentration range of 0.5 to 2 mM glucose. Figure (6), showed that for various GOx concentrations, the reduction current increased rapidly and linearly initially at low substrate concentrations but gradually levelled off towards high glucose concentrations. This
Page 8 of 31
trend is in accordance with enzyme kinetics whereby the rate becomes zero order because all the enzyme molecules are bound to the substrate molecules. On the basis of the above results, an optimized enzyme concentration of 0.75 U disc-1 was selected in order to attain an optimum sensitivity. However, a drop in current was
ip t
observed in this graph when compared to the other graphs obtained previously. For instance in Figures (4) and (5), a 0.75 U disc-1 with 1 mM glucose concentration produced an average
cr
current of 2.35 μA at a potential of −0.3 V. Whereas in Figure (6), a 0.75 U disc-1 with 1 mM glucose concentration produced a current response of 1.16 μA at the same potential. This
us
lower response can be accounted for the continuous leaching of PB from the PB-SPCE due to
an
the increased length of time needed when investigating this particular parameter [24,29].
3.5 Biosensor Response Characteristics
M
FIGURE 6 IS PLACED HERE.
d
By employing the optimum conditions from the above parameters investigated, the
te
calibration graph of the sensor was obtained as shown in Figure (7). The PB-modified screenprinted glucose biosensors exhibited a linear range between 0.25 to 2.00 mM. The linear
Ac ce p
regression equation was y=2.1387x + 0.489, where y signifies the current in μA and x is the glucose concentration in mM with R2 value was found to be 0.996. In addition, the relative standard deviation of the biosensor at 1 mM glucose was 4.3% (n=6) thereby showing good reproducibility and the limit of detection (LOD), calculated based on IUPAC definition was 0.01 mM glucose, with n = 5 at 90% confidence level [35]. FIGURE 7 IS PLACED HERE.
The analytical performance of the present biosensor was compared against other types of PB-based glucose biosensors as shown in Table (1). It can be noted that the LOD for the present biosensor is lower than our earlier paper-based mediated glucose biosensor [31] and other PB-modified glucose biosensors [26,28,36,]. Despite the simplicity of the present
Page 9 of 31
biosensor, it produced a lower LOD compared to the highly complex paper-based electrochemical device fabricated with an integral battery and PB electrochromic read-out, i.e. colour change of PB due to different redox state in the presence of glucose [37]. Moreover, the linear range of this simple biosensor is comparable with other more complex
ip t
PB-modified glucose biosensors [26,28,36] and is wider compared to our previous paperbased mediated glucose biosensor [31].
cr
TABLE 1 IS PLACED HERE
The storage stability of the biosensors in solution was investigated at 4°C. The
us
response of the paper-based glucose biosensor as a function of the storage time was shown in Figure (8). The response to 1 mM glucose of the paper-based glucose biosensors, stored at
an
4 °C when not in use and tested for 45 days retained ca. 72% of the initial current value. The
M
lifetime of the present paper-based biosensor was compared with other types of PB-based glucose biosensors as shown in Table (1). Our shelf life study showed that the present low
d
cost, simple biosensor was comparable to other more complex PB-modified biosensors based
te
on carbon paste electrodes with solid paraffin, i.e. 0% drop in the signal after 30 days of storage [28]. Also, the sensitivity of a glucose biosensor based on PB-modified screen-
Ac ce p
printed electrodes and GOx immobilized on nafion dropped by 50% within half a day [26]. In this work, the immobilization technique has been extremely low cost and straightforward as no additional chemical reagents, apart from pH 7.0 PBS buffer solution, and thus the fiber matrix in paper proved to be an effective stable microenvironment for GOx. FIGURE 8 IS PLACED HERE.
3.6 Selectivity against Interferences Electroactive interferences have been a problem when biological or industrial samples are assayed by amperometric biosensors. The effect of electroactive interferents that might affect the response of the paper-based glucose biosensor was investigated. Nine possible
Page 10 of 31
interferents were used to evaluate the selectivity of the glucose biosensor. The current obtained for each interfering substance at a concentration of 1 mM (unless otherwise stated) in the presence of 1 mM glucose (Iθ) is compared to the current obtained in the presence of 1mM glucose alone (I) as a criterion for the selectivity of the biosensor, shown in Table (2).
ip t
Acetic acid, ascorbic acid, citric acid, fructose, oxalic acid, sucrose, urea and uric acid do not cause any observable interference. Only L-cysteine interfered significantly.
cr
TABLE 2 IS PLACED HERE.
us
3.7 Validation Study TABLE 3 IS PLACED HERE.
an
To ascertain the applicability of the paper-based glucose biosensor, glucose concentrations in two commercial beverages were measured by the paper-based biosensor
M
and the results were compared quantitatively with the results obtained from Benedict’s Test along with UV/Vis spectrophotometry, shown in Table (3). Beverage A was a pure glucose
d
energy supplement (Glucolin) and beverage B was an oral rehydration solution comprising of
te
electrolytes and glucose only (Hydralyte). The original concentrations were obtained from
Ac ce p
the nutritional information stated on container of beverages. The glucose concentration determined by the two methods corroborates rather well and it can be inferred that the results obtained by the biosensor had lower RSD compared to the Benedict’s test.
4. Conclusions
In the present study, an amperometric glucose biosensor based on PB-SPCEs was
developed by using paper as a matrix for GOx immobilization via a simple physical adsorption method. This immobilization method on a green matrix is extremely beneficial as complex immobilization methods involving the use of multiple chemicals can be avoided and this also ensures that the original enzyme configuration is not compromised thus allowing optimal activity. It was demonstrated that the integration of paper disc as a matrix for enzyme
Page 11 of 31
immobilization with SPEs was highly advantageous due to the simplicity and costeffectiveness of its preparation procedures, easy control of enzyme loading with good shelflife. This novel paper-based glucose biosensor provides a robust and convenient platform with promising prospects for applications which require miniturization, small volumes of
ip t
sample, high sensitivity and selectivity such as point-of-care medical diagnostics in resourcelimited settings as well as field-deployable environmental analysis and monitoring.
cr
Sensitivity could be enhanced by incorporating nano-electrocatalysts into paper discs.
us
5. Acknowledgements
The team acknowledges the generous funding provided by National Institute of Education-
Ac ce p
te
d
M
an
Nanyang Technological University (NIE-NTU) (RI7/12 TSN).
Page 12 of 31
6. References [1]
S.M. Hossain, R.E. Luckham, A.M. Smith, J.M. Lebert, L.M. Davies, R.H. Pelton, C.D. Filipe, J.D. Brennan, Development of a bioactive paper sensor for detection of neurotoxins using piezoelectric inkjet printing of sol-gel-derived bioinks, Anal. Chem.
ip t
81 (2009) 5474-5483. W.A. Zhao, A. Van Den Berg, Lab on paper, Lab Chip. 8 (2008) 1988–1991.
[3]
D.D. Liana, B. Raguse, J.J. Gooding, E. Chow, Recent advances in paper-based
cr
[2]
sensors, Sensors (Basel). 12 (2012) 11505-11526.
A.W. Martinez, S.T. Phillips, G.M. Whitesides, E. Carrilho, Diagnostics for the
us
[4]
developing world: Microfluidic paper-based analytical devices, Anal. Chem. 82 (2010)
W. Lewis, E. Keshavarz-Moore, J. Windust, D. Bushell, N. Parry, Construction and
M
[5]
an
3–10.
evaluation of novel fusion proteins for targeted delivery of micro particles to cellulose
S. Su, R. Nutiu, C.D. Filipe, Y. Li, R. Pelton, Adsorption and covalent coupling of
te
[6]
d
surfaces, Biotechnol. Bioeng. 94 (2006) 625-632.
ATP-binding DNA aptamers onto cellulose, Langmuir. 23 (2007) 1300-1302. M. Tolba, L.Y. Brovko, M.W. Griffiths, Engineering of bacteriophages displaying
Ac ce p
[7]
affinity tags on its head for biosensor applications, Nanotechnology: Life Sciences, Medicine & Bio Materials. 2 (2008) 449-452.
[8]
S.J. Craig, A. Shu, Y. Xu, F.C. Foong, R. Nordon, Chimeric protein for selective cell
attachment onto cellulosic substrates, Protein Engineering, Design & Selection. 20 (2007) 235-241.
[9]
R. Pelton, Bioactive paper provides a low-cost platform for diagnostics, Trac-Trends in Anal. Chem. 28 (2009) 925-942.
Page 13 of 31
[10] P. Yager, T. Edwards, E. Fu, K. Helton, K. Nelson, M.R. Tam, B.H. Weigl, Microfluidic diagnostic technologies for global public health, Nature. 442 (2006) 412418. [11] A.W. Martinez, S.T. Phillips, M.J. Butte, G.M. Whitesides, Patterned paper as a
ip t
platform for inexpensive, low-volume, portable bioassays, Angew. Chem. Int. Ed. Engl. 46 (2007) 1318–1320.
cr
[12] A.W. Martinez, S.T. Phillips, E. Carrilho, S.W. Thomas, H. Sindi, G.M. Whitesides, Simple telemedicine for developing regions: camera phones and paper-based
us
microfluidic devices for real-time, off-site diagnosis, Anal. Chem. 80 (2008) 3699– 3707.
an
[13] A.W. Martinez, S.T. Phillips, G.M. Whitesides, Three-dimensional microfluidic
M
devices fabricated in layered paper and tape, Proc. Natl. Acad. Sci. USA. 105 (2008) 19606–19611.
d
[14] A.W. Martinez, S.T. Phillips, B.J. Wiley, M. Gupta, G.M. Whitesides, FLASH: A
(2008) 2146-2150.
te
Rapid Method for Prototyping Paper-Based Microfluidic Devices, Lab on a Chip. 8
Ac ce p
[15] J. Wang, Electrochemical glucose biosensors, Chem. Rev. 108 (2008) 814-825. [16] J.D. Newman, S.J. Setford, Enzymatic biosensors, Mol. Biotechnol. 32 (2006) 249268.
[17] A.A. Karyakin, E.E. Karyakina, L. Gorton, On the mechanism of H2O2 reduction at Prussian Blue modified electrodes, Electrochem. Commun. 1 (1999) 78–82.
[18]
A.A. Karyakin, O.V. Gitelmacher, E.E. Karyakina, Prussian Blue-based first-
generation biosensor. A sensitive amperometric electrode for glucose, Anal. Chem. 67 (1995) 2419–2423.
Page 14 of 31
[19] A.V. Mokrushina, M. Heim, E.E. Karyakina, A. Kuhn, A.A. Karyakin, Enhanced hydrogen peroxide sensing based on Prussian Blue modified macroporous microelectrodes. Electrochem. Commun. 29 (2013) 78-80. [20] A.A. Karyakin, E.E. Karyakina, L. Gorton, Prussian-Blue based amperometric
ip t
biosensors in flow-injection analysis, Talanta. 43 (1996) 1597–1606. [21] A.A. Karyakin, E.E. Karyakina, L. Gorton, The electrocatalytic activity of Prussian
cr
Blue in hydrogen peroxide reduction studied using a wall-jet cell with continuous flow, J. Electroanal. Chem. 456 (1998) 97-104.
applications, Electroanalysis. 13 (2001) 813-819.
us
[22] A.A. Karyakin, Prussian Blue and its analogues: electrochemistry and analytical
an
[23] A. Boyer, K. Kalcher, R. Pietsch, Voltammetric behavior of perborate on Prussian
M
Blue modified carbon paste electrodes, Electroanalysis. 2 (1990) 155–161. [24] A.A. Karyakin, M. Chaplin, Polypyrrole Prussian Blue films with controlled level of
te
(1994) 85–92.
d
doping: codeposition of polypyrrole and Prussian Blue, J. Electroanal. Chem. 56
[25] F. Ricci, A. Amine, G. Palleschi, D. Moscone, Prussian Blue based screen printed
Ac ce p
biosensors with improved characteristics of long-term lifetime and pH stability, Biosens. Bioelectron. 18 (2003) 165– 174.
[26] M.P. O’Halloran, M. Pravda, G.G. Guilbault, Prussian Blue bulk modified screenprinted electrodes for H2O2 detection and for biosensors, Talanta. 55 (2001) 605–611.
[27] A. M. Vinu Mohan, K. K. Aswini, A. Maria Starvin, M. Biju, Amperometric detection of glucose using Prussian blue-graphene oxide modified platinum electrode, Anal. Methods. 5 (2013) 1764-1770. [28] D. Moscone, D. D’Ottavi, D. Compagnone, G. Palleschi, Construction and Analytical Characterization of Prussian Blue-Based Carbon Paste Electrodes and Their Assembly as Oxidase Enzyme Sensors, Anal. Chem. 73 (2001) 2529-2535.
Page 15 of 31
[29] Q. Deng, B. Li, S. Dong, Self-gelatinizable copolymer immobilized glucose biosensor based on Prussian Blue modified graphite electrode, Analyst. 123 (1998) 1995-1999. [30] J. Wang, X. Zhang, Screen-printed cupric-hexacyanoferrate modified carbon enzyme electrode for single-use glucose measurements, Anal. Lett. 32 (1999) 1739-1749.
glucose biosensor, Sensor Actuat. B-Chem. 193 (2014) 536-541.
ip t
[31] C.S.L. Kuek, S.N. Tan, C.Z. Floresca, A “green” cellulose paper based amperometric
cr
[32] Q. Chi, S. Dong, Amperometric biosensors based on the immobilization of oxidases in a Prussian blue film by electrochemical codeposition, Anal. Chim. Acta. 310 (1995)
us
429-436.
[33] Z. Tang, X. Du, R.F. Louie, G.J. Kost, Effects of pH on Glucose Measurements With
an
Handheld Glucose Meters and a Portable Glucose Analyzer for Point-of-Care Testing,
M
Arch. Pathol. Lab. Med. 124 (2000) 577-582.
[34] S. Banerjee, P. Sarkar, A.P.F. Turner, Amperometric biosensor based on Prussian
d
Blue nanoparticle-modified screen-printed electrode for estimation of glucose-6-
te
phosphate, Analytical Biochemistry. 439 (2013) 194-200. [35] J. Mocak, A.M. Bond, S. Mitchell, G. Scollary, A statistical overview of standard
Ac ce p
(IUPAC AND ACS) and new procedures for determining the limits of detection and quantification: application to voltammetric and stripping techniques, Pure Appl. Chem. 69 (1997) 297–328.
[36] F. Ricci, C. Gonçalves, A. Amine, L. Gorton, G. Palleshi, D. Moscone, Electroanalytical study of prussian blue modified glassy carbon paste electrodes, Electroanal. 15 (2003) 1204-1211.
[37] H. Liu, R.M. Crooks, Paper-based electrochemical sensing platform with integral battery and electrochromic read-out, Anal. Chem. 84 (2012) 2528-2532.
Page 16 of 31
Authors’ Biography Nadia Chandra Sekar received her BSc (Hons) in Biomedical Sciences from the University of Bradford and is currently a Master of Science (Research) student from the Nanyang Technological University (NTU), Singapore. Dr Tan Swee Ngin is an Associate Professor at NTU, Singapore. Her current research areas include electroanalytical chemistry and separation science.
ip t
the Singapore Institute of Singapore with focus on activities are focused on fuel cells; and design and
cr
Dr Seyed Ali Mousavi Shaegh is a Research Scientist at Technology (SIMTech). He completed his PhD at NTU, Microfluidic Fuel Cells at 2012. Currently, his research electrochemical microdevices including biosensors and micro fabrication of microfluidic devices for biomedical applications.
an
us
Dr Sum Huan Ng received his BEng and MEng in mechanical engineering in 1997 and 2000 respectively from the National University of Singapore. He obtained his PhD from the Georgia Institute of Technology. He is currently a research scientist with SIMTech working on the following areas: microfluidics, microfabrication techniques and abrasive removal processes.
Ac ce p
te
d
M
Dr Liya Ge obtained her PhD from NTU, Singapore. She has been a research fellow working at NTU since 2005.
Page 17 of 31
Figure Captions: Figure 1: Schematic illustration of GOx paper disc preparation and integration with the PBSPCE. 5 μL GOx solution was dropped on the surface of the paper disc with subsequent drying in air at room temperature; the paper disc with immobilized GOx was placed on top of the PB-SPCE surface for complete coverage of working, counter and reference electrodes; 12 µL PBS buffer solution with 0.5 μL glucose solution was dropped on the paper disc for electrochemical detection.
cr
ip t
Figure 2: Two sets of cyclic voltammograms for GOx disc on the SPCE (I) and PB-SPCE (II) in which each set contains one voltammogram with 12 μL of 0.1 M PBS (pH 7.0) buffer solution only (A) and another with the addition of 5 mM glucose solution (B). The cyclic voltammetry scans were performed from +0.30 V to -0.50 V at a scan rate of 10 mV s-1, [GOx]=0.25 U disc-1.
us
Figure 3: Cyclic Voltammograms for GOx disc on the PB-SPCE with 12 µL of 0.1 M PBS (pH 7.0) buffer solution (A), 1 mM (B) and 5 mM (C) glucose solutions. The cyclic voltammetry scans were performed from +0.30 V to -0.50 V at a scan rate of 10 mV s-1, [GOx]=0.75 U disc-1.
M
an
Figure 4: Effect of applied potential on the amperometric response of the GOx paper-based glucose biosensor on PB-SPCE with 1 mM glucose solution (pH 7), [GOx]=0.75 U disc-1. The error bars show standard deviations for n=3~4. Figure 5: Effect of pH on the amperometric response of the GOx paper-based glucose biosensor on PB-SPCE with 1 mM glucose solution at an applied potential of -0.30 V (vs Ag/AgCl), [GOx]=0.75 U disc-1. The error bars show standard deviations for n=3~4.
Ac ce p
te
d
Figure 6: Effect of GOx concentration (U disc-1) on the amperometric response of the paperbased glucose biosensor on PB-SPCE with various glucose concentrations at an applied potential of -0.30 V (vs Ag/AgCl). Paper discs with GOx concentrations of 0.25 U disc-1 (A), 0.50 U disc-1 (B), 0.75 U disc-1 (C) and 1 U disc-1 (D), n=3~4. Figure 7: Calibration curve of the amperometric response of the paper-based glucose biosensor on PB-SPCE with various glucose concentrations at an applied potential of -0.30 V (vs Ag/AgCl), pH=7.0, [GOx]=0.75 U disc-1. The error bars show standard deviations for n=3~4. Figure 8: Stability of the amperometric response of the paper-based glucose biosensor on PB-SPCE during 30 days determined with 1 mM glucose in PBS buffer (pH 7) at an applied potential of -0.30 V (vs Ag/AgCl), [GOx]=0.75 U disc-1. The error bars show standard deviations for n=6.
Page 18 of 31
Linear Range (mM)
Detection Limit (mM)
Lifetime
Paper-based/PB-SPCE
0.25 to 2.0
0.01
72% for 45 days
−
PB-based carbon paste electrodes with solid paraffin
0.1 to 20
0.10
cr
Table 1: Comparison of the analytical performance of the present biosensor against other related biosensors.
[28]
an
0.22
ip t
Lost 50% sensitivity after 4 hours
[26]
0.05 to 0.8
0.05
−
[36]
M
PB-glassy carbon paste electrodes Paper-based/SPCE with ferrocene carboxylic acid as mediator Paper-based electrochemical sensing platform with integral battery and PB electrochromic readout
Until 3
100% for 30 days
0.18
98% for 120 days
[31]
0.10
−
[37]
1.0 to 5.0
−
Ac ce p
Paperbased glucose biosensors
PB-modified screenprinted electrodes with GOx immobilized on nafion
References
d
PBmodified electrodes for glucose analysis
te
Current Work
us
Types of PB-based glucose biosensors
Page 19 of 31
Table 2: Possible Interferences Tested With The Paper-Based Glucose Biosensor
Acetic Acid
0.9
Citric Acid
1.0
Fructose
1.0
L-Cysteine
0.6
Oxalic Acid
1.1
Sucrose
0.9
cr
Ascorbic Acid
ip t
Current Ratioa (Iθ/I) 1.0
Possible Interfering Substance
Urea
1.0
1.2
us
Uric Acid
Current ratios (Iθ/I) for mixtures of 1mM interfering substance and 1mM glucose (C*) in comparison to that of 1mM glucose alone (C), [GOx]=0.75 U disc-1, n=3~4.
Ac ce p
te
d
M
an
a
Page 20 of 31
Table 3: Glucose Beverage Analysis Based On Amperometry and Benedict’s Tests
Benedict's Testb (g/L)
Paper-Based Glucose Biosensorc (g/L)
A
89.5
105.0±0.0003
90.4±0.1
B
14.4
11.9±0.001
14.1±0.1
cr
ip t
Beverage
Original Concentrationa (g/L)
us
a
Ac ce p
te
d
M
an
The original glucose concentrations were calculated based on the nutritional information stated on the beverage packaging. b.c Values are the mean (±standard error) of n=6~8, c [GOx]=0.75 U disc-1.
Page 21 of 31
Graphical Abstract (for review)
Ac ce
pt
ed
M
an
us
cr
ip t
Graphical Abstract
Page 22 of 31
Research Highlights
Highlights •
Ac ce
pt
ed
M
an
us
cr
ip t
• •
Glucose oxidase (GOx) was physically adsorbed on paper without use of chemicals. This biosensor configuration was cost-effective and showed good stability. Simple platform for applications that require small sample volumes.
Page 23 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 1
Page 24 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 2
Page 25 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 3
Page 26 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 4
Page 27 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 5
Page 28 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 6
Page 29 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 7
Page 30 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure 8
Page 31 of 31