1
Rational Design of Three-Dimensional Microfluidic Paper-Based Analytical
2
Devices (3D-μPADs)
3 4
Giorgio Gianini Morbioli,1,2,3 Thiago Mazzu-Nascimento,1,2 Amanda M. Stockton,3*
5
and Emanuel Carrilho1,2*
6 7
1
8
São-carlense, 400, 13566-590 São Carlos, SP, Brazil
9
2
Instituto Nacional de Ciência e Tecnologia de Bioanalítica, Campinas, SP, Brazil
10
3
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta,
11
Georgia 30332, USA
Instituto de Química de São Carlos, Universidade de São Paulo, Av. Trabalhador
12 13
*Corresponding authors: Carrilho, E. (
[email protected]); Stockton, A.M.
14
(
[email protected])
15 16
Keywords: Wax printing; point-of-care devices; enzymatic assays; magnetic
17
apparatus
18 19
Total number of words: 2067
20 21 22 23 24 1
25
Abstract
26
Paper-based devices are a portable, user-friendly and affordable technology that is
27
one of the best analytical tools for inexpensive diagnostic devices. Three-
28
dimensional microfluidic paper-based analytical devices (3D-μPADs) are an
29
evolution of single layer devices and they permit effective sample dispersion,
30
individual treatment of layers and multiplexed assays. Here, we present the rational
31
design of 3D-μPADs using wax printing technology, which enables more
32
homogeneous permeation of fluids along the cellulose matrix when compared with
33
the existing designs in the literature. Moreover, we show the importance of the
34
rational design of channels on these devices using a glucose oxidase, peroxidase
35
and ABTS assay. We present an alternative method for layer stacking using a
36
magnetic apparatus, which facilitates fluidic dispersion and improves the
37
reproducibility of tests performed on 3D-μPADs. We also provide the optimized
38
designs for printing, facilitating further studies using 3D-μPADs.
39 40 41 42 43 44 45 46 47 48 2
49
1.
Introduction
50
Microfluidic paper-based analytical devices (μPADs) are low-cost analytical
51
tools that are easily manufactured, manipulated, transported and stored,1 which
52
makes μPADs attractive for diagnostic applications and meets the requirements of
53
the World Health Organization (WHO) in the ASSURED Challenge.2 Three-
54
dimensional microfluidic paper-based analytical devices (3D-μPADs) consist of a
55
stack of single layer μPADs,3–9 and is a natural evolution of single layer systems10–
56
21
57
thousand-fold, which favors multiplexed assays;6 iii) enclosing of intermediate
58
layers, which protects the reagents stored on the device, without the need for an
59
additional toner layer22 and iv) sample preparation steps into the device, including
60
separation and washing.3
due to: i) individual treatment of layers; ii) sample dispersion from ten to
61
In order to allow fluid permeation through the device it is necessary to
62
maintain good contact between adjacent hydrophilic zones, which can be done
63
either irreversibly (Figure S1), in which the layers cannot be separated after
64
assembly without damaging the device;5,6,23,24 or reversibly (Figure S2), where it is
65
possible to separate layers after assembly.3,8,9
66
Among these methods of layer assembly, the use of tape and cellulose
67
powder is the most laborious one, requiring precise alignment and addition of
68
cellulose powder in each spot, which hinders mass production.6 The use of
69
adhesive spray, on the other hand, may be the most appropriate method for mass
70
production of irreversibly-bound paper-based devices.5
71
Origami μPADs are advantageous in comparison with the irreversibly bound
72
microchips, as they eliminate the need for extra layers of tape 6 or coating steps,5 3
73
and they are more useful for step-by-step studies3 or in fluidic dispersion studies
74
due to the ease of post-testing analysis.
75
Distinct patterns of fluidic distribution on the device depend on the design of
76
the paper-based microchip (Figure S3).5 However, it is relevant to note how the
77
design of the 3D-μPAD could affect fluidic dispersion on the cellulosic matrix and,
78
therefore, the use of the device itself.
79
Fluidic dispersion on paper-based devices is especially relevant in
80
enzymatic assays supported on the cellulosic matrix, because these reactions are
81
time-sensitive: the longer the enzymatic reaction proceeds, the higher amount of
82
product is generated, enhancing the detected signal in those areas, whether
83
colorimetric,1 electrochemical7 or fluorescent.9 If there is a preferential path on the
84
fluidic dispersion favoring some reaction zones, then a positive bias will be
85
observed in those zones, affecting the figures of merit of the analytical method.
86
Therefore, the design of paper-based microfluidic devices is critical for their
87
successful application.
88
In this paper we present a new, rational approach to paper-based
89
microchannel design, and demonstrate how channel design influences fluidic
90
permeation of the cellulosic matrix and the read-out measurements of the assay.
91
We use colorimetric glucose determination by the glucose oxidase enzymatic
92
assay as a model.15 Moreover, we also present a new method of layer assembly
93
using an external magnetic apparatus, facilitating fluidic dispersion studies and
94
improving the reproducibility of tests performed on 3D-μPADs.
95 96
2.
Materials & Methods 4
97
Reagents
98
Glucose oxidase from Aspergillus niger (E.C. 232-601-0), peroxidase from
99
horseradish (E.C. 232-668-6), 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)
100
diammonium salt (ABTS) and D-(+)-trehalose dehydrate from Saccharomyces
101
cerevisiae were purchased from Sigma-Aldrich (St. Louis, MO). D-(+)-glucose was
102
purchased from VWR (Solon, OH), red fountain pen ink was purchased from
103
Waterman (Paris, France) and sodium hydrogen phosphate anhydrous (ACS,
104
99.0% min.) was purchased from Alfa Aesar (Ward Hill, MA). All reagents were
105
used as received.
106 107
Paper-based devices fabrication
108
Origami paper-based microchip devices were designed using CorelDraw®
109
X7 software (Figure 1a), and printed on Whatman No. 1 chromatography paper
110
(letter size) using a Xerox Phaser® 8580 wax printer (Figure 1b). The patterned
111
sheets were submitted to thermal treatment in an oven (150 oC for 2 min, Figure
112
1c) to melt the wax through the entire thickness of the paper in order to create
113
hydrophobic barriers.13 The design files (Figure 1d) and specifications are available
114
in the Electronic Supporting Information (E.S.I.).
115 116
Magnetic external apparatus
117
Two stainless steel sheets were cut in the following dimensions: 39 x 48 mm, as
118
presented in Figure 1f. The top sheet was perforated in the center, and a cut
119
pipette tip was glued onto it (Figure 1f). The bottom part was perforated 16 times,
120
in a 4x4 configuration, according to the bottom layer of the paper-based 5
121
microdevices. The paper-based devices were introduced between the top and
122
bottom metal sheets, aligning the holes of the apparatus with the pattern of the
123
paper-based device (Figure 1g). To keep the whole apparatus together a pair of
124
neodymium magnets were used (curved pieces in Figure 1f).
125 126
Figure 1. μPADs step-by-step fabrication and use. (a) Microdevice design. (b) Wax
127
printing. (c) Melting and permeation of wax on paper. (d) Cut microchip. (e) Folding
128
of the device. (f) Insertion into the magnetic apparatus. (g) Sample introduction. (h)
129
Bottom of the device after the assay. (i) Device digitalization. (j) Image analysis.
130 131
Incremental permeation study
132
A red ink solution (0.5 mL ink: 10 mL DI water) was applied at the top of the
133
magnetic apparatus containing one paper-based device, in 5 μL increments (0 to
134
70 μL). This experiment was performed at least in triplicate for each one of the
135
volumes and for each one of the 5 studied designs.
136 6
137
Permeation study recording
138
A Logitech® HD Pro Webcam C920 was fixed in the base of a universal support to
139
record the bottom of the device during fluid permeation. The magnetic apparatus
140
containing the paper-based device was positioned above the camera, and 70 μL of
141
the red ink solution was applied at the top of the magnetic apparatus, recording
142
each device for 15 min. At least 15 devices for each one of the 5 studied designs
143
were recorded. Video recordings of the fluid permeation on each studied design
144
are available in the Electronic Supporting Information (E.S.I.).
145 146
Enzymatic assay
147
A volume of 1 µL of a 25 mmol L−1 ABTS redox indicator diluted in water was
148
pipetted onto each one of the 16 spots in the bottom layer of the 3D-μPAD. After
149
complete dryness of the spots (~30 min), 1 µL of a solution containing glucose
150
oxidase (120 U mL−1), horseradish peroxidase (300 U mL−1) and trehalose (0.6 mol
151
L-1) in PBS buffer (0.1 mol L−1; pH 6.0) was added on each spot in the bottom layer
152
of the 3D-μPAD, and allowed to dry ~30 min for complete dryness of the spots.15
153
The 3D-μPAD containing reagents was folded and placed in the external magnetic
154
apparatus, and 65 µL of a 2 mmol L-1 glucose standard solution diluted in water
155
was applied at the top of the magnetic apparatus. After 10 min the device was
156
removed from the magnetic apparatus and after complete dryness (~30 min) the
157
device read-off was digitalized in a flatbed scanner. This experiment was
158
performed at least in triplicate for the original and optimized designs. The average
159
color intensity in the RGB channel was obtained using the Adobe Photoshop ® CC
7
160
2015 software (but other open source software could be also used, such as
161
ImageJ (http://imagej.nih.gov/ij)).
162 163
3.
Results & Discussion
164
Layer assembling method
165
The reversible method for layer assembly using an aluminum housing with
166
screws9 can be problematic, as the torque applied to each one of the 4 screws can
167
introduce a force imbalance into the system that can induce a bias in fluidic
168
dispersion in the 3D-μPAD. Using 2 flat stainless steel plates united by strong Nd
169
magnets alleviates this issue, as the force applied to keep the layers of the paper-
170
based device together is more uniform. As this method minimizes variation in
171
fluidic permeation, one should also expect improvement in the figures of merit for
172
the analytical method.
173 174
Permeation study
175
As seen in Figure 2, the use of the original design presented in literature5
176
creates a preferential path for fluidic permeation to the central spots in the bottom
177
layers of the device. This bias can influence the results of enzymatic assays, which
178
are time-dependent. This behavior is due to the smaller hydrodynamic resistance
179
of the path traveled by the fluid,25 as shown in Figure 3.
8
180 181
Figure 2. Incremental fluid dispersion study using as model a design provided in
182
literature.5 Distinct volumes of testing dye solution were applied at the top of the
183
device (in 5 μL increments) and permeated from top to the bottom, and the devices
184
were digitalized after 30 min. There is a preferential dispersion through the central
185
spots, as can be observed from 5 to 20 μL.
186
9
187
188 189
Figure 3. Hydrodynamic resistance in each layer of the original model design.5 The
190
central spots present a smaller hydrodynamic resistance than the peripheral spots,
191
explaining the observed bias.
192 193
In a first optimization attempt, an additional layer was added to the original
194
design5 in order to ameliorate preferential fluidic dispersion (Figure S4). The
10
195
dispersion of fluid was more homogeneous in this design, with no apparent
196
preferential paths. However, an extra layer was introduced in the device, which
197
adds a minor additional degree of fabrication complexity.5
198
Two other optimization designs were tested (Figure S5 and S6) that have
199
the same number of layers as the original design. The combination of these
200
designs led to an optimized design (Figure 4), which incorporates just 4 layers and
201
displays homogeneous fluidic dispersion.
202
203 204
Figure 4. Optimized paper-based microchip design. This design contains only 4
205
layers, and permits a more homogeneous dispersion of fluid in the device. It can
206
comport 100 μL of sample without leaking. 11
207 208
As can be seen in Figure 4, volumes as small as 40 μL can reach all the
209
spots in the bottom layer of the optimized device. However, experience has shown
210
that a little excess of liquid (65 μL) provides better results with the enzymatic
211
assay. The improved performance appears to be due to the prolonged hydration of
212
the bottom spots (containing enzymes and redox indicator), which enables the
213
reactions to proceed further to completion. In order to evaluate the liquid holding
214
capacity of the devices we have added an excess of fluid (until 100 μL) at the top
215
of the magnetic apparatus. We did not observe any leaking when excess fluid was
216
applied to the devices, indicating that the system can hold larger volumes of fluid.
217
This characteristic can be exploited further to improve the LOD and LOQ of
218
enzymatic assays supported on these devices.
219 220
Enzymatic assay
221
When the original design is used in conjunction with the enzymatic assay
222
there is a difference in the mean pixel intensity of the 4 central spots and the 12
223
peripheral ones (Figure 5a). This difference is statistically significant (test-t, C.I.
224
95%, results in S.I.), which proves our initial hypothesis. The optimized design
225
(Figure 5b), however, shows no statistical difference between the spots (test-t, C.I.
226
95%, results in E.S.I.). These results prove that assays can be unbiased or biased
227
based on μPADs design, and that the rational μPADs design presented here
228
enables unbiased μPADs assays.
12
229 230
Figure 5. Statistical comparison between colorimetric enzymatic assay for glucose
231
determination. (a) Original design.5 There is a higher signal developed in the 4
232
central spots (inside the green square) in comparison with the coloration developed
233
at the peripheral spots. (b) Optimized design. There is no significant statistical
234
difference between the central and peripheral spots coloration. Original digitalized
235
images are provided in the E.S.I.
236
4.
Conclusion 13
237
The 3D-μPADs systems provide multiple advantages, including sample
238
distribution, multiplexed assays and individualized treatment of layers. Here, we
239
have shown that the method of assembling layers and the 3-dimensional design of
240
paper-based devices play a critical role in assay performance. We have introduced
241
a device that optimizes the production of 3D-μPADs and assay performance. This
242
work
243
reproducibility and other figures of merit using a 3D paper-based platform.
furthers
the
development
of
low-cost
diagnostic
tools,
improving
244 245
Acknowledgements
246
The authors would like to thank the funding agencies FAPESP (Grant No.
247
2011/13997-8), CNPq (Grant No. 131306/2013-8 and 205453/2014-7) by the
248
scholarships and financial support to the Instituto Nacional de Ciência e Tecnologia
249
de Bioanalítica – INCTBio (FAPESP Grant Nr. 2008/57805-2 / CNPq Grant Nr.
250
573672/2008-3), the Georgia Institute of Technology (Georgia Tech) and the State
251
of Georgia, USA. We gratefully acknowledge valuable discussion with Dr. Luis
252
Aparecido Milan (UFSCar, SP, Brazil). The authors declare having no competing
253
financial interests.
254 255
References
256
1
Whitesides, Anal. Chem., 2008, 80, 3699–707.
257 258
2
R. W. Peeling, K. K. Holmes, D. Mabey and a Ronald, Sex. Transm. Infect., 2006, 82 Suppl 5, v1-6.
259 260
A. W. Martinez, S. T. Phillips, E. Carrilho, S. W. Thomas, H. Sindi and G. M.
3
L. Ge, S. Wang, X. Song, S. Ge and J. Yu, Lab Chip, 2012, 12, 3150–8. 14
261
4
2010, 10, 2622–7.
262 263
5
6
7
D. Zang, L. Ge, M. Yan, X. Song and J. Yu, Chem. Commun. (Camb)., 2012, 48, 4683–5.
268 269
A. W. Martinez, S. T. Phillips and G. M. Whitesides, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 19606–11.
266 267
G. G. Lewis, M. J. DiTucci, M. S. Baker and S. T. Phillips, Lab Chip, 2012, 12, 2630–3.
264 265
X. Gong, X. Yi, K. Xiao, S. Li, R. Kodzius, J. Qin and W. Wen, Lab Chip,
8
M. Zhang, L. Ge, S. Ge, M. Yan, J. Yu, J. Huang and S. Liu, Biosens. Bioelectron., 2013, 41, 544–50.
270 271
9
H. Liu and R. M. Crooks, J. Am. Chem. Soc., 2011, 133, 17564–17566.
272
10
A. Abbas, A. Brimer, J. M. Slocik, L. Tian, R. R. Naik and S. Singamaneni, Anal. Chem., 2013, 85, 3977–83.
273 274
11
K. Abe, K. Suzuki and D. Citterio, Anal. Chem., 2008, 80, 6928–6934.
275
12
R. S. J. Alkasir, M. Ornatska and S. Andreescu, Anal. Chem., 2012, 84, 9729–37.
276 277
13
7091–5.
278 279
14
15
284
C. A. Chaplan, H. T. Mitchell and A. W. Martinez, Anal. Methods, 2014, 6, 1296.
282 283
D. M. Cate, W. Dungchai, J. C. Cunningham, J. Volckens and C. S. Henry, Lab Chip, 2013, 13, 2397–404.
280 281
E. Carrilho, A. W. Martinez and G. M. Whitesides, Anal. Chem., 2009, 81,
16
C.-M. Cheng, A. W. Martinez, J. Gong, C. R. Mace, S. T. Phillips, E. Carrilho, K. A. Mirica and G. M. Whitesides, Angew. Chem. Int. Ed. Engl., 2010, 49, 15
4771–4.
285 286
17
2010, 82, 4158–64.
287 288
18
19
20
21
T. S. Park, W. Li, K. E. McCracken and J.-Y. Yoon, Lab Chip, 2013, 13, 4832–40.
295 296
M. M. Mentele, J. Cunningham, K. Koehler, J. Volckens and C. S. Henry, Anal. Chem., 2012, 84, 4474–80.
293 294
A. W. Martinez, S. T. Phillips, G. M. Whitesides and E. Carrilho, Anal. Chem., 2010, 82, 3–10.
291 292
B. R. Lutz, P. Trinh, C. Ball, E. Fu and P. Yager, Lab Chip, 2011, 11, 4274– 8.
289 290
M. S. Khan, G. Thouas, W. Shen, G. Whyte and G. Garnier, Anal. Chem.,
22
K. M. Schilling, A. L. Lepore, J. a Kurian and A. W. Martinez, Anal. Chem., 2012, 84, 1579–85.
297 298
23
H. Noh and S. T. Phillips, Anal. Chem., 2010, 82, 4181–7.
299
24
K. M. Schilling, D. Jauregui and A. W. Martinez, Lab Chip, 2013, 13, 628–31.
300
25
K. W. Oh, K. Lee, B. Ahn and E. P. Furlani, Lab Chip, 2012, 12, 515–45.
301
16
