Content uploaded by Mohammad Awashra
Author content
All content in this area was uploaded by Mohammad Awashra on Jan 05, 2024
Content may be subject to copyright.
RESEARCH ARTICLE
www.advmatinterfaces.de
Superhydrophilic/Superhydrophobic Droplet Microarrays of
Low Surface Tension Biofluids for Nucleic Acid Detection
Mohammad Awashra,* Pinja Elomaa, Tuomas Ojalehto, Päivi Saavalainen,
and Ville Jokinen*
Superhydrophilic/superhydrophobic patterned surfaces can be used to create
droplet microarrays. A specific challenge with the liquids needed for various
biomedical applications, as compared to pure water, is their lower surface
tension and potential for contaminating the surfaces through adsorption.
Here, a method is shown to create biofluid droplet microarrays using
discontinuous dewetting of pure water, an oil protective layer, and finally
biofluid exchange with the water droplet array. With this method, a droplet
array of a viscous nucleic acid amplification solution can be formed with a low
surface tension of 34 mN m−1and a contact angle of only 76°with the used
hydrophobic coating. This droplet array is applied for nucleic acid detection of
SARS-CoV-2 virus using strand invasion-based amplification (SIBA)
technology. It is shown that by using an array of 10 000 droplets of 50 μm
diameter the limit of detection is 1 RNA copy μL−1. The results demonstrate
that SIBA on droplet microarrays may be a quantitative technology.
1. Introduction
Sample partitioning to tiny droplets is required to perform many
bioassays such as digital nucleic acid detection. There are mainly
three approaches to partitioning aliquots into droplets of con-
trolled volume. First, droplet microfluidics, where a microfluidic
M. Awashra, V. Jokinen
Department of Chemistry and Materials Science
School of Chemical Engineering
Aalto University
Tietotie , Espoo , Finland
E-mail: mohammad.awashra@aalto.fi; ville.p.jokinen@aalto.fi
P. Elomaa, P. Saavalainen
Translational Immunology Research Program
Folkhälsan Research Center
University of Helsinki
Haartmaninkatu , Helsinki , Finland
T. Ojalehto
Aidian Oy
Koivu-Mankkaan tie B, Espoo , Finland
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/./admi.
© The Authors. Advanced Materials Interfaces published by
Wiley-VCH GmbH. This is an open access article under the terms of the
Creative Commons Attribution License, which permits use, distribution
and reproduction in any medium, provided the original work is properly
cited.
DOI: 10.1002/admi.202300596
pump is used to serially generate water
droplets confined in a second immiscible
continuous phase. The drawbacks of this
method are its requirement for constant
flow and proper surfactants, and the
possibility of droplets destabilization
during processing, resulting in cross-
contamination and droplets merging.
Second, solid microchambers/wells,
where the droplets are confined in a
solid container. This method does not
need surfactants and unlike microflu-
idics droplets, the formed droplets have
a defined location, enabling simple
indexing of each droplet. The solid
walls prevent droplet merging and cross-
contamination and enable a very uniform
volume distribution of the droplets.[1]
Third, 2D open droplet microarray
(also called surface droplets) where the
droplets are loaded on a flat surface by a printing method,[2–4] or
by discontinuous dewetting using a patterned surface with selec-
tive hydrophilic/hydrophobic modification.[5] While microwells
and water-in-oil droplets are frequently used for sample parti-
tioning, surface droplet microarrays can be considered a combi-
nation of both, where there is a microarray of accessible droplets
(as in water-in-oil droplets), and at the same time, the droplets
are fixed in position and more controllable (as in microwells).
Moreover, instead of the oil phase or physical wall separation, the
droplets can be separated using a patterned superhydrophobic
barrier.[1,6–8]
Nucleic acid testing is extensively applied in biomedical fields.
The detection of pathogens, such as SARS-CoV-2, is normally
realized by nucleic acid amplification. Reverse transcription-
quantitative polymer chain reaction (RT-qPCR) is currently the
gold standard for nucleic acid detection.[9,10 ] However, RT-qPCR
assays are based on relative quantification, which necessitates an
external calibration using genetic standards or internal reference
DNA templates.[11] On the other hand, digital PCR (dPCR) relies
solely on the count of positive partitions rendering the standard
curve requirement unnecessary. Consequently, dPCR achieves
superior sensitivity and accuracy in comparison to qPCR.[3] One
of the requirements of dPCR is the partition of the diluted nucleic
acid sample into a huge number of separate microreactors.[12,13 ]
Performing dPCR in commercial equipment is significantly
more expensive than conventional PCR.[14] Developing a point of
care (POC) device that makes dPCR not only more sensitive but
also cheaper, is a crucial matter. Microstructured devices are used
Adv. Mater. Interfaces 2024,11, 2300596 (1 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
www.advancedsciencenews.com www.advmatinterfaces.de
for dPCR and integrated into POC devices.[15] Several studies
used water-in-oil droplet microfluidics for dPCR analysis.[16] Oth-
ers reported the use of solid microchambers/wells.[17–25 ] While
these methods make it hard to access the droplet and can have
complex droplet loading methods, surface microarrays make the
droplet very accessible and addressable. This approach simplifies
the manipulation of droplets, enabling direct analysis on the chip
using various optical and other techniques.[1] When the PCR so-
lution gets partitioned into a droplet microarray, following Pois-
son distribution, most of the droplets theoretically will contain
only one or zero target molecules.[13] Each copy is then individu-
ally amplified by PCR.
The amplification techniques of nucleic acid fall into one of
the two categories: thermocycling amplification or isothermal
amplification.[15] The thermocycling technique requires complex
and time-consuming rapid temperature changes during amplifi-
cation. In contrast, isothermal amplification is a one-step heat-
ing process, making it simpler and implementable in portable
and POC devices. Isothermal amplification systems include nu-
cleic acid sequence-based amplification (NASBA), loop-mediated
isothermal amplification (LAMP), and many other reactions.[15]
Another isothermal amplification technique is Strand Invasion
Based Amplification (SIBA) that was first described in 2014.[26]
SIBA works at relatively low temperatures, which lowers the
evaporation rate of tiny droplets. SIBA takes advantage of target
template denaturation that is mediated by recombinase enzyme
and invasion oligonucleotide (IO). Local disruption of double-
stranded DNA (dsDNA) is created to target the template allow-
ing amplification primers to anneal. A single-stranded binding
protein stabilizes the dsDNA and prevents re-annealing between
the two separated DNA stands. Due to the enzymatic denatura-
tion step, amplification can be carried out at a constant temper-
ature (≈44 °C). Target-specific IO has 3′end 2′-O-methyl RNA
modification that prevents it from acting as a template for the
polymerase. This makes SIBA highly specific and resistant to any
non-specific amplification, enabling high sensitivity for a single
target molecule.[26,27 ]
In this study, a facile droplet microarray was developed to
spatially separate aliquots of a controlled volume of SIBA re-
action mixture. The array consists of circular superhydrophilic
(SHL) patterns of varying size (50–500 μm) separated by super-
hydrophobic (SHB) black silicon barriers. The reagent loading
method was optimized to solve complications arising from the
adsorption and surface tension of the SIBA mixture. The SIBA
mixture was successfully partitioned into droplet microarray us-
ing the developed protective oil method. The low temperature
used in isothermal SIBA technology (44 °C for 20 min) is a clear
advantage for dPCR experiments where such exceedingly small
droplet volumes are used. We show the detection and quantifica-
tion of SARS-CoV-2 RNA down to 1 copy μL−1.
2. Results and Discussion
2.1. SHL/SHB Black Silicon Surface Characterization
The black silicon (bSi) nanograss structure is shown in Figure 1a.
The bSi surface was patterned into superhydrophilic circular
spots with superhydrophobic borders using a fluoropolymer coat-
ing. This process was used to fabricate 40 different combina-
tions of SHL spot sizes and spacings. The interface of the coated
and oxidized bSi and their static contact angles (CAs) are shown
in Figure 1b. Water droplet microarray (DMA) spontaneously
formed on the patterned bSi by discontinuous dewetting. The
nanograss structure increases the solid area for the Wenzel state
on the SHL spot and minimizes it for the Cassie state on the SHB
regions (nano tips) making an extreme wettability difference be-
tween the two regions. Figure 1c shows the trapped air plastron
(silver-like color) beneath the water film indicating a Cassie wet-
ting state, while the SHL spots appear in black indicating full wet-
ting of the surface (Wenzel state). Figure 1d,e shows top and side
views of the water DMA, respectively.
To understand the DMA formation and select the optimal pa-
rameters for carrying out quantitative nucleic acid detection, dif-
ferent SHL spot sizes and spacings were fabricated (see Table S1,
Supporting Information). The aim was to fit in as many spots
as possible without merging nearby droplets. Droplet formation
and evaporation in ambient air were monitored for chips with
SHL spot sizes of 50–500 μm with different spacings under op-
tical microscopy (Figure S1, Supporting Information). Droplet
evaporation on different chips was studied in ambient air at
room temperature (21 °C) and room relative humidity (40%)
(Figure 1f). The evaporation time of the DMA was found to in-
crease with SHL spot size and decrease with spacing for all the
combinations of spot sizes and spacings (Figure S2a, Support-
ing Information). The larger droplets have a higher volume-to-
surface ratio, and therefore, slower evaporation rate. With in-
creasing droplet density, the atmospheric surrounding of the
droplets is more saturated in water vapor, which also slows down
the evaporation rate.
The volume of the water droplet for the different arrays was
measured using goniometry. The volume distribution of the
droplets in an array is narrow, making the array suitable for ap-
plications that require uniform droplet volumes. The droplet vol-
ume is increasing from 0.1–10 nL with spot diameter at fixed
50 μm spacing (Figure 1g). This increase is exponential as a cube
of spot diameter [V(nL)=10−7d2.92(μm), R2=0.997] following
the behavior of semi-spherical water droplets, which is also con-
firmed by other studies.[6,28,29 ] Xu et al.[28] obtained the equa-
tion V(nL)=32.9d2.44(mm), which is more deviated than our
results compared to the spherical droplet volume (d3). However,
the range of the spot diameter they used was 0.3–2 mm making
the effect of the gravity a major factor in determining the shape of
the droplet, while we used 50–500 μm spots that are small enough
to avoid high gravitational forces. Moreover, this group used the
beads counting method to measure droplet volume while in our
case goniometry was used. The droplet volume also increased
from 5.4–8.9 nL as the spacing between the 400 μm sized spots
increased from 50–800 μm (Figure 1g). A study by Chang et al.[30]
showed that the droplet volume of hydrophilic squared spots with
a diameter of 500 μm and different spacing of 0.2–0.5 mm gives a
droplet volume of ≈12–14 nL, which is in perfect agreement with
our values for this spot size (10–12 nL, see Figure S2b, Support-
ing Information). The droplet evaporation rate for the array with
larger spacing is faster, even though the droplet volume is big-
ger, indicating that the vapor saturation factor had the advantage
over the small increase of droplet volume. According to Mands-
berg et al.[29] The volume of the droplets can be affected by the
withdrawing velocity and angle of the chips from the water. The
Adv. Mater. Interfaces 2024,11, 2300596 (2 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmatinterfaces.de
Figure 1. Water droplet array formation on SHL/SHB bSi surface. a) Scanning electron microscopy (SEM) image of the nanograss structure of bSi.
b) SEM image of the interface between the fluoropolymer-coated SHB (left) and the oxidized non-coated SHL (right) bSi nanograss and corresponding
contact angles. c) bSi chip immersed in water. A water film covers the surface and fully wets the SHL spots (black circles) while the SHB regions have a
Cassie wetting state with air plastron (silver-like color). d,e) Top and side views of the water droplet microarray, respectively. The eect of SHL spot size
(blue square) and spacing (red triangle) on the water droplet f) evaporation time and g) volume.
withdrawal of the chip in this study was done manually but we
kept a fixed protocol using the maximum possible velocity and an
angle in the range of 10 –30°. It is worthy to mention here that
the camera resolution of the goniometer did not allow an accu-
rate volume measurement of droplets on smaller spot sizes like
50 μm.
The array chip measuring 50 ×50 μm in size and spac-
ing was selected for subsequent experiments. The spot size
50 μm allows an easy formation of the droplets besides hav-
ing the highest droplet density with this size. While the spac-
ing 50 μm is optimal for this size since it gives the highest
possible droplet density without having any droplets merging
or film formation. The droplet density of this array is 10 000
droplets cm−2.
2.2. Wettability and Surface Tension Characterization
Unlike pure water, the SIBA reaction mixture did not form dis-
crete droplets spontaneously by simply immersing the array in
the liquid. This complication is shown in Figure 2a that shows
that there is no plastron (no Cassie state) when this mixture is
Adv. Mater. Interfaces 2024,11, 2300596 (3 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmatinterfaces.de
Figure 2. Wettability characterization for developing a reliable method for creating droplet microarrays out of SIBA reaction mixture. a) While water
(left) has two dierent wettability states on the SHL and SHB regions of the bSi surface, the SIBA mixture (right) has the same wettability state on both,
preventing the formation of MDA. b) Surface tension measurement of SIBA ( mN m−) using pendant droplet method. c) The advancing contact
angle of SIBA mixture on the HB surface in the air (°). The advancing contact angle of SIBA under hexadecane on d) HB surface (°)ande)HL
surface (°). The needle diameter in (b–e) is . mm. f) The workflow of the SIBA mixture DMA formation. g) SIBA droplet volume of dierent SHL
spot sizes (spacing is fixed at μm) before (blue) and after (red) isothermal heating at °Cformin.
used. The key differences between the two liquids can be: the
surface tension, the viscosity, and the potential of molecule ad-
sorption from the SIBA reaction mixture. Ramalingam et al.[31]
claimed that the fluid properties of PCR solution are different
from that of pure water, where they found that Promega PCR
static CA on polydimethylsiloxane (PDMS) surface is 46°(as com-
pared to water that has a roughly 100°static CA with PDMS) and
its surface tension is ≈31 mN m−1(72 mN m−1for water). Some
other studies that used PCR solution often approximate its prop-
erties to that of pure water.[31]
Therefore, to develop a method for the reliable formation of
droplet arrays of PCR solutions, we characterized the surface
tensions and wetting properties of water, the nucleic acid am-
plification solution (SIBA mixture), and hexadecane oil on two
types of surfaces. First, the surface tension of the SIBA mixture
was measured using the pendant droplet method and found to
Adv. Mater. Interfaces 2024,11, 2300596 (4 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmatinterfaces.de
Tabl e 1 . Advancing and receding contact angles of a system formed of
water, SIBA, and hexadecane on HL and HB surfaces.
Experiment Adv. CA
[°]
SD
[°]
Rec. CA
[°]
SD
[°]
Water droplet in air vs HL , * -
Water droplet in air vs HB , ,
Hexadecane droplet in air vs HL - -
Hexadecane droplet in air vs HB , ,
SIBA droplet in air vs HL , -
SIBA droplet in air vs HB -
Hexadecane droplet under SIBA vs HB - - -
Hexadecane droplet under SIBA vs HL - - -
Hexadecane droplet under water vs HB ,
Hexadecane droplet under water vs HL ,
SIBA droplet under hexadecane vs HB -
SIBA droplet under hexadecane vs HL , -
HL: hydrophilic; HB: hydrophobic; Adv. CA: advancing contact angle; Rec. CA: reced-
ing contact angle; SD: standard deviation; SIBA: strand invasion-based isothermal
amplification technology for nucleic acid detection. *Any contact angle lower than °
(what the goniometer can measure) was recorded as °.
be 34 mN m−1(±0.2 mN m−1) (Figure 2b). While pure water
and pure hexadecane surface tensions were found to be 72 and
27 mN m−1, respectively. The contact angle measurements were
performed with planar silicon surfaces which have the same two
surface chemistries as we use on the bSi arrays. The reason for
utilizing planar surfaces is that we were interested in the ther-
modynamic question of whether liquid Acan replace liquid B
on a given surface. This depends only on the chemistry and not
on the topography. Table 1 summarizes the CA measurement re-
sults. First, the advancing (adv.) and receding (rec.) CAs for wa-
ter, SIBA, and hexadecane were measured on both surfaces in
ambient air. Focusing on the hydrophobic (HB) surface, the adv.
CA of water, SIBA, and hexadecane were 114°,76°,and50°,re-
spectively. Water is clearly having a non-wetting behavior on the
fluoropolymer-coated HB surface. Alternatively, SIBA and hex-
adecane have a wetting behavior (CA <90°). The lower CA of
SIBA (Figure 2c) can be explained by its lower surface tension
(cos(𝜃)=𝛾sv −𝛾sl
𝛾lv ).
For water, as the CAs are >90°on the HB surface and <90°on
the HL one, this leads to the possibility of spontaneous discrete
droplet formation (i.e., by discontinuous dewetting) on the bSi
array (in fact this happened with all spot sizes and spacings we
tested). In contrast, for the SIBA mixture, both the HL and HB
surfaces have CAs that are <90°, which means that there is no
possibility for discrete droplet formation since the Cassie state is
not obtained on the bSi samples and the Wenzel state only en-
hances the wetting further. As a result of that, we observed the
formation of a stable continuous film over SHL and SHB regions
on the bSi surface without spontaneously splitting into droplets.
To complicate further, the rec. CA of SIBA on both HL and HB
surfaces was 0°indicating strong adhesive forces with both sur-
faces. This has led to SIBA not receding from both SHL and SHB
regions of the bSi surface confirming that SIBA droplet array
formation is non-spontaneous. However, SIBA wettability was
also investigated under hexadecane, and it was found that SIBA
cannot replace hexadecane on the HB surface (adv. CA >90°,
Figure 2d), nonetheless, on the HL surface, it can (adv. CA <90°,
Figure 2e).
To make the SIBA mixture array formation more reliable, we
developed the protective oil method shown in Figure 2f. First,
(1and 2) a water droplet microarray is formed on the SHL/SHB
pattern and then covered with a thin layer of hexadecane. Next, (3
and 4) the SIBA mixture was added on top with the assistance of
a PDMS chamber and was kept for 5 s, which was found to be suf-
ficient exchange time but was not optimized downward further.
The key idea of this method is applying the protective oil coating
on the SHB region to prevent the adsorption of the biofluid con-
tents while, at the same time, protecting the SHL spots using wa-
ter. As shown by the CA measurements, the biofluid can replace
water from the SHL spots, but it cannot replace oil in the SHB
region. We assume that the exchange process between SIBA and
water is most probably a hydrodynamic process where the thin oil
layer on top of the protruding water droplets gets sheared away
so that the SIBA droplet makes direct contact with the chip form-
ing SIBA droplet. The SIBA droplet microarray was successfully
formed and can be seen by the naked eye if 200 μm spots are used
as shown in Figure S3 (Supporting Information). The isothermal
SIBA amplification is performed at 44 °C for 20 min, and there-
fore, possible evaporation of the droplets under these conditions
was studied. Figure 2g shows that the droplet’s volume remained
nearly unchanged before and after undergoing the heating step
while the droplets were covered with hexadecane oil.
For the validation of our method, two other biofluids were ap-
plied on the chip using the steps in Figure 2f. First, Evagreen dye
with high concentration of DNA solution was used to confirm
the exchange process between the biofluid and water droplets.
The results indicated high fluorescence for the imaged droplets
as shown in Figure S4a (Supporting Information). Second, a non-
fluorescent protein-rich cell media was used to form the droplet
microarrays. The microarray was successfully formed as shown
in Figure S4b (Supporting Information). As droplet microarrays
were formed on our chip for three biofluids, we concluded that
the protective oil method could be used to overcome the chal-
lenges caused by the different properties of biofluids.
Patterned SHL/SHB surfaces have been employed to perform
different bioassays. Our previous study used SHL/SHB bSi array
for single-cell trapping by splitting cell media into tiny droplets.[8]
Popova et al.[6] used a SHL/SHB droplet microarray for high-
throughput cell-based screening. Qian et al.[32] formed 106fem-
toliter droplets array using a patterned perfluorinated surface to
perform bead-free digital immunoassay for human interleukin-
6. Nevertheless, some bioassays employ biofluids with properties
that differ greatly from water, especially surface tension, viscosity,
and adsorption capabilities.[33] Peethanetal.
[34] studied the effect
of surface tension of a CTAB (cetyl trimethyl ammonium bro-
mide) solution on droplet splitting to a droplet microarray on a
candle-soot-coated patterned SHL/SHB surface. The surface ten-
sion of the solution varied from 61 to 38.5 mN m−1with varying
concentrations of CTAB, and the authors found that the droplet
is still splitting into arrays and the daughter droplet volume re-
mains nearly the same. They concluded that the substrate can be
used for making bioassays using body fluids like blood. However,
surface tension is not the only factor that can affect the splitting
behavior of the droplets. Protein and other biomolecules can get
Adv. Mater. Interfaces 2024,11, 2300596 (5 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmatinterfaces.de
Figure 3. a) Fluorescent images of the dSIBA quantification experiment performed with SARS-CoV- RNA concentrations ranging from to
copies μL−using SHL/SHB bSi droplet microarray platform. b) Shows a strong linear correlation between the counted positive droplets and expected
RNA concentration. cp μL−is not included in the curve.
adsorbed and contaminate the surface changing its wettability.
We noticed this effect by the rec. CA of 0°of SIBA mixture on
our surface indicating strong adhesion and Wenzel state. Wang
et al.[35] showed that bovine serum albumin (BSA) protein solu-
tion causes loss in air plastron of an SHB surface as soon as they
are immersed in the protein solution, indicating a Cassie to Wen-
zel state transition. They attributed this behavior to two factors:
the lower surface tension of the BSA solution (≈50 mN m−1), and
the protein adsorption on the SHB substrate.
2.3. Digital SIBA – Isothermal Amplification on the Droplet
Microarray
In this study, SARS-CoV-2 SIBA assay is used to test the perfor-
mance of the designed SHL/SHB bSi 2D- open droplet microar-
ray platform. SIBA is an isothermal amplification technology for
nucleic acid testing[26] that works at low temperatures (44 °C). It
has been used before in a droplet microfluidics system (Elomaa
et al 2023, unpublished). Hexadecane oil was not found to affect
the SIBA amplification based on the real-time amplification ex-
periment conducted using RT-qPCR machine.
Before the use of the patterned chips for digital SIBA (dSIBA)
reaction, they were sterilized using a UV lamp. The sterilization
was found to not affect the hydrophobic coating and the wettabil-
ity of the chip. Several fabricated arrays with different spot sizes
and spacings were used to perform dSIBA, and the depicted re-
sults showed a successful formation of SIBA droplet array using
our method (Figure 2f), and the fluorescent positive droplets were
imaged and are shown in Figure S5 (Supporting Information).
The 50 ×50 μm chip was used to perform a quantification test of
dSIBA assay. The quantification experiment was performed us-
ing a serial dilution of the target RNA and for negative control, a
sample buffer was added without any target RNA. By partitioning
the SIBA mixture into droplets on the platform, the dissolved tar-
get RNA copies are getting separately trapped in one of the 10 000
droplets formed on the 1 ×1 cm chip following Poisson statistics,
which means that with low concentrations the spots contain only
one or zero target.[17] Isothermal amplification of SIBA was per-
formed at 44 °C for 20 min to amplify each trapped RNA copy in
its droplet. The prepared concentrations were 1, 10, 20, 50, 100,
and 10 000 copies μL−1. The quantification results and a represen-
tative section of each experiment are shown in Figure 3. The dif-
ferentiation between droplets lacking the target (negatives) and
those containing it (positives) was accomplished by implement-
ing a fluorescence amplitude threshold. The number of positive
spots is clearly increasing with RNA concentration and the nega-
tive control has no signal (Figure 3a). Moreover, Figure 3b shows
a strong linear relationship, with a correlation factor of 0.9986,
between the expected target RNA concentration and measured
positive spot fraction. This correlation indicates that dSIBA could
be quantitative.[13,36 ]
We tested analyzing the positive and negative spots with Aifo-
ria AI in addition to ImageJ thresholding. The results of both
methods were very similar (See Figure S6, Supporting Informa-
tion). One advantage of the surface array method is that it is very
easy to analyze for several reasons: 1) The hydrophobic back-
ground is systematically completely black, 2) The spots are circles
with a known diameter, and 3) The spots are in a known array.
Because of these features, we concluded that the simpler ImageJ
analysis was sufficient and reliable for our assay.
There is a trade-off in choosing a certain droplet size for the
assay. The 50 ×50 μm chip that was characterized has a high
density of droplets enabling it to detect high concentrations up
to 10 000 copies μL−1if 1 ×1cmchipisused(Figure3a),and
its limit of detection is limited by the small total volume of the
partitioned sample (1 copy μL−1). Moreover, this spot size has the
highest droplet density at which the droplet evaporation and im-
age acquisition are well handled. At lower spot size like 25 μm
the evaporation is extremely fast even under oil and the droplet
cannot survive the isothermal heating step. On the other hand,
arrays with much larger spots would have lower droplet density
but higher single droplet volume. This would lead to worse quan-
tification at higher concentrations due to a higher likelihood of
Adv. Mater. Interfaces 2024,11, 2300596 (6 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmatinterfaces.de
having multiple target RNA copies in a single spot. However, the
total volume would be larger that would enable a lower limit of
detection.
There are two previous studies that used SHL/SHB arrays
for performing digital nucleic acid detection, both with a differ-
ent nucleic acid amplification mixture. Chi et al.[37] fabricated
hydrophilic/superhydrophobic pseudopaper photonic nitrocellu-
lose substrate to perform digital LAMP (dLAMP). They used a
spot diameter of 500 μm limiting their droplet array density.
Mao et al.[38] used a polymethyl methacrylate (PMMA)-based
SHL/SHB microfluidic chip to perform dLAMP. These groups
did not clarify how they overcame the lower surface tension and
adsorption complication of the LAMP mixture, compared to wa-
ter. A major advantage of our method is that due to the protective
hexadecane, the contents and properties of the biofluid should
not matter as long as they cannot replace hexadecane (or other
oil) from the SHB regions. This could enable the use of the same
method for many types of biomedical droplet array applications.
3. Conclusion
We have reported in this study a superhydrophilic/superhydr
ophobic patterned surface that can form a droplet microarray
of a biofluid with different surface tension and molecular con-
tent compared to water. By applying a protective coating (hexade-
cane oil) to the superhydrophobic regions, the biofluid is pre-
vented from reaching there and it has access only to the super-
hydrophilic spots forming a droplet array. Our array gives the
user the possibility to gain the same information as working with
water-in-oil droplets without the need for external pumping units
and with fixed and indexed droplets. We have used this array as
a nucleic acid detection platform using SIBA technology and our
depicted results showed that this assay is quantitative with a limit
of detection down to 1 copy μL−1for 50 μm spot size. In the mar-
ket leader Bio-Rad ddPCR system, 20 000 droplets are counted us-
ing 20 μL PCR-reaction mix. Here, in a chip as small as 2 ×1cm,
we can form 20 000 droplets using reaction mix volume as low as
1μL.
4. Experimental Section
Black Silicon Chips Fabrication and Characterization:The SHL/SHB
black silicon (bSi) microarray fabrication process is a modified version of
the earlier work.[] Figure 4 shows the fabrication steps of SHL/SHB pat-
terned bSi. First, bSi was fabricated using a maskless cryogenic deep reac-
tive ion etching (Oxford PlasmaPro Estrelas ICP-DRIE) process on a
″silicon wafer (Figure a,b). The process parameters were − °Ctem-
perature, mTorr pressure, ICP power W, forward power W, and
the gas flows were sccm for Oand sccm for SF. The wafer was
then covered with a fluoropolymer hydrophobic coating using a PECVD
process (Oxford Plasmalab Plus) (Figure c). The process parameters
were mTorr pressure, W power, and sccm CHFflow. After
that, photolithography was performed as follows: AZ thick photore-
sist (MicroChemicals) was spin-coated on the bSi surface ( rpm for
s), soft baked ( °C, min), exposed for s with dierent film pho-
tomasks (Süss MicroTec MA- with nm wavelength), and developed
for min in AZ B (Merck) to define hydrophilic areas (Figure d,e).
Next, the fluoropolymer was etched away from the non-protected areas us-
ing oxygen plasma reactive ion etching (RIE) (Oxford Plasmalab Plus)
(Figure f). The parameters were mTorr pressure, W power, sccm
Figure 4. The workflow of patterning the black silicon surface to superhy-
drophilic/superhydrophobic regions.
O, and sccm Ar. The photoresist was then removed by ultrasonication
in acetone (Figure g). The wafer was then diced into small chips with
the needed sizes using a dicing saw (DAD, Disco). The black silicon
structure was confirmed using scanning electron microscopy (SEM). The
wetting properties of the surface were characterized using goniometry and
fluorescent microscopy.
SU-8 Mold Fabrication:The chamber mold was fabricated using SU-
photolithography. First, a silicon wafer was dipped in buered hydrofluo-
ric acid for min. Then, SU- (Kayaku) was spin-coated on the silicon
wafer ( rpm, s), and soft baked first at °C for min and then
at °C for min (UniTemp GmbH – HP-). Then, it was exposed for
s with a film photomask (Süss MicroTec MA- with nm wavelength)
and developed for min in MR-DEV (PGMEA) to define the chamber
borders. Then, post-exposure baking was performed at °Cformin.Fi-
nally,a hydrophobic coating was applied using the PECVD process (Oxford
Plasmalab Plus).
PDMS Casting:Polydimethylsiloxane (PDMS) was prepared by mixing
the prepolymer and curing agent at a ratio of :. The degassed mixture
was poured over the SU- master mold and baked at °Cforhtoso-
lidify. The PDMS chambers were peeled o the mold and cut to ×cm
with a height of μm. All chambers were punctured to make an inlet in
Adv. Mater. Interfaces 2024,11, 2300596 (7 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmatinterfaces.de
the center using a . mm Miltex Biopsy Punch with Plunder (TED PELLA,
Inc).
Water Droplet Microarray:The water droplet microarray was formed
by dipping the bSi patterned chip in water and then withdrawing it. After
this point, hexadecane oil was used to prevent the droplet’s evaporation.
The evaporation time of the droplets was measured once in air at room
temperature ( °C) and room relative humidity (%) and another time
under oil at °C using optical microscopy.
Droplet Volume Measurement:Water and SIBA DMA volume on dif-
ferent chips were measured under hexadecane oil using the camera of a
contact angle goniometer (THETA, Biolin Scientific). For each chip with a
specific SHL spot size and spacing, ten microdroplets from dierent lo-
cations were used to measure static contact angles and baselines. Then,
the volume of the droplets was computed using OneAttension software
(Biolin Scientific).
Contact Angle and Surface Tension Measurements:Two reference sur-
faces were used for studying the wettability of a system formed of water,
hexadecane, and SIBA. A blank Si wafer exposed to oxygen plasma reac-
tive ion etching (Oxford Plasmalab Plus) was used as the hydrophilic
(HL) reference surface, and a Si wafer coated with fluoropolymer was used
as the hydrophobic (HB) reference surface. The dynamic advancing and
receding contact angles were measured using the needle-in-sessile drop
technique (THETA, Biolin Scientific). Advancing contact angles were mea-
sured from to μL droplet size and receding angles from to μLwith
a droplet rate of . μLs
−. The surface tension measurement of SIBA
was performed optically using the pendant drop method with a droplet
size of μL (THETA, Biolin Scientific). All experiments were performed in
triplicates and the reported value is the mean.
SIBA:SIBA SARS-CoV- assay from Aidian Oy was performed on the
SHL/SHB DMA. The oligos in the assay were designed to amplify the
SARS-CoV- RNA-dependent RNA-polymerase gene (RdRp). Aidan opti-
mized the SIBA SARS-CoV- assay to be used in droplets by creating an op-
timal fluorescence signal for droplet detection. The assay was SYGRGreen-
based.
SIBA Droplet Microarray Formation:All chips, chambers, and equip-
ment used in the RNA detection experiments were sterilized using a UV
lamp ( nm) for min. Figure f shows a schematic illustration of
SIBA microarray formation. SIBA DMA was formed by first forming a wa-
ter DMA covered with a thin layer of hexadecane. This was done by im-
mersing the SHL/SHB patterned chip in a water-oil layered system, where
the chip was withdrawn from this system starting with the water phase
and subsequently passing the oil phase (Figure f, 1and 2). The excess oil
was removed and the PDMS chamber was introduced on the chip. Then,
the SIBA mix was flushed in the sealed system using the inlet and the
contact time of SIBA mix with the chip was s. The chamber was then
disconnected, and the chip was placed in a glass petri dish containing
hexadecane (Figure f, and ).
Other Biofluids Used to Validate the Protective Oil Method:The same
steps used in the previous subsection (Figure f) were repeated using
other low-surface tension biofluids to validate the method. First, Evagreen
dye (Biotium x) was used together with a genomic DNA control in early
droplet experiments to demonstrate the fluid exchange in the spots. Eva-
green binds to double-stranded DNA and therefore there was no need for
an amplification step at this stage. Second, protein-rich cell media was also
used to validate the method for droplet microarray formation. The cell me-
dia contained high glucose DMEM (Euroclone), % FBS (Thermofisher),
Penicillin Streptomycin (Gibco), and Glutamax (Gibco) and therefore was
high in protein.
dSIBA Experiments:The functionality of the chip was tested using
SARS-CoV- SIBA from Aidian Oy. SIBA reaction was constructed of SIBA A
mixture (substrates), SIBA Bmixture (enzymes), SIBA oligomixture ( n
oligonucleotides and SYBRGreen), and nuclease-free water. SARS-CoV-
RNA sample was added to the reaction mixture in a preferable concen-
tration diluted in sample buer (containing m magnesium acetate).
The total volume of one reaction was μL. The negative control contained
only a sample buer. The reaction was initiated by the addition of magne-
sium acetate and heating the sample at °C for min. All experiments
were performed in triplicates.
Real-Time Detection of SIBA:SIBA reaction was tested together with
the hexadecane oil to evaluate whether the oil was interacting with the
SIBA reaction. SIBA reaction mixture was prepared as described earlier and
loaded into a -well plate together with μL of hexadecane oil. RT-qPCR
instrument (Bio-Rad) was programmed to keep a constant temperature of
°C for min. One cycle was designed to last for s (total of cycles)
with a fluorescence readout at the end of each cycle.
Fluorescence Detection and Image Analysis:SIBA assay is based on
SYBRGreen chemistry that allows detection via fluorescence microscopy.
Pictures were taken using Zeiss Axio imager microscope with x magni-
fication and HE GFP filter (Biomedicum imaging unit, Helsinki). Tiles-
function was used for imaging large areas ( × cm). For quantification,
ImageJ was employed by counting droplets using an automated
particle counter. The software first converted the picture to black and white
and the fluorescence background was determined based on signal inten-
sity in the negative control and used for determining positive and negative
spots. Then, size and circularity filters were applied to exclude any defects.
AI was also used by implementing Aiforia Create Version . (Aiforia Tech-
nologies Plc, Helsinki, Finland), and the results were compared to ImageJ
analysis. The deep convolutional neural network was taught to recognize
positive, negative, and empty spots from the fluorescence pictures (see
Figure S, Supporting Information) Aiforia showed a total object error of
.% (consisting of false positive .% and false negative .% errors)
between all of the analyzed pictures.
Supporting Information
Supporting Information is available from the Wiley Online Library or from
the author.
Acknowledgements
This work utilized the cleanroom facilities of Micronova, which is part of
the OtaNano national research infrastructure. Imaging was done at the
Biomedicum Imaging Unit, Helsinki University, Finland, with the support
of Biocenter Finland. This work was funded by the Academy of Finland
(#). P.E. would also like to acknowledge the Doctoral Programme
in Clinical Research for her studentship.
Conflict of Interest
All SIBA patents/patent applications are owned by Aidian Oy. T.O. is an
employee of Aidian Oy.
Author Contributions
M.A. and P.E. contributed equally to this work. M.A., P.E., and V.J. proposed
and designed the research. M.A. did the surface fabrication and characteri-
zation, performed the wettability studies, and analyzed the data. P.E. devel-
oped and validated SIBA reaction in droplets. M.A. and P.E. performed the
nucleic acid detection experiments on the patterned surface, did fluores-
cence microscopy, and analyzed the data. T.O. provided SIBA mixture. P.S.
provided valuable advice for digital nucleic acid detection. V.J. developed
the concept of using protective oil coating for the formation of droplet mi-
croarray of low surface tension biofluids. M.A. and P.E. wrote the paper.
V.J. carried out the review and editing. T.O. and P.S. reviewed the article.
Data Availability Statement
The data that support the findings of this study are available from the cor-
responding author upon reasonable request.
Adv. Mater. Interfaces 2024,11, 2300596 (8 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmatinterfaces.de
Keywords
biofluids, black silicon, digital nucleic acid amplification, microfluidics,
wettability patterning
Received: July ,
Revised: August ,
Published online: October ,
[] W. Feng, E. Ueda, P. A. Levkin, Adv. Mater. 2018,30, .
[] Y. Zhu, Y.-X. Zhang, W.-W. Liu, Y. Ma, Q. Fang, B.o Yao, Sci. Rep. 2015,
5, .
[] W.-W. Liu, Y. Zhu, Y.i-M. Feng, J. Fang, Q. Fang, Anal. Chem. 2017,89,
.
[] Y. Sun, X. Zhou, Y. Yu, Lab Chip 2014,14, .
[] Z. Dong, Q. Fang, Trends Analyt. Chem. 2020,124, .
[] A. A. Popova, K. Demir, T. G. Hartanto, E. Schmitt, P. A. Levkin, RSC
Adv. 2016,6, .
[] A.A.Popova,S.M.Schillo,K.Demir,E.Ueda,A.Nesterov-Mueller,
P. A. Levkin, Adv. Mater. 2015,27, .
[] P. Raittinen, P. Elomaa, P. Saavalainen, V. Jokinen, Adv. Mater. Inter-
faces 2021,8, .
[] O. Vandenberg, D. Martiny, O. Rochas, A. Van Belkum, Z. Kozlakidis,
Nat. Rev. Microbiol. 2021,19, .
[] L. Benevides Lima, F. P. Mesquita, L. L. Brasil De Oliveira, F. Andréa
Da Silva Oliveira, M. Elisabete Amaral De Moraes, P. F. N. Souza, R.
C. Montenegro, Expert Rev. Mol. Diagn. 2022,22, .
[] L. Dong, J. Zhou, C. Niu, Q. Wang, Y. Pan, S. Sheng, X. Wang, Y.
Zhang, J. Yang, M. Liu, Y. Zhao, X. Zhang, T. Zhu, T. Peng, J. Xie, Y.
Gao, D.i Wang, X. Dai, X. Fang, Tal a nta 2021,224, .
[] L.i-P. Xu, Y. Chen, G. Yang, W. Shi, B. Dai, G. Li, Y. Cao, Y. Wen, X.
Zhang, S. Wang, Adv. Mater. 2015,27, .
[] M. Baker, Nat. Methods 2012,9, .
[] A. S. Whale, J. F. Huggett, S. Cowen, V. Speirs, J. Shaw, S. Ellison, C.
A. Foy, D. J. Scott, Nucleic Acids Res. 2012, e, 40.
[] L. Zanoli, G. Spoto, Biosensors 2012,3, .
[] L. Chen, V. Yadav, C. Zhang, X. Huo, C. Wang, S. Senapati, H.-C.
Chang, Anal. Chem. 2021,93, .
[] K. A. Heyries, C. Tropini, M. Vaninsberghe, C. Doolin, O. I. Petriv,
A. Singhal, K. Leung, C. B. Hughesman, C. L. Hansen, Nat. Methods
2011,8, .
[] Y. Men, Y. Fu, Z. Chen, P. A. Sims, W. J. Greenleaf, Y. Huang, Anal.
Chem. 2012,84, .
[] L. Cao, X. Guo, P. Mao, Y. Ren, Z. Li, M. You, J. Hu, M. Tian, C. Yao,
F. Li, F. X u , ACS Sens. 2021,6, .
[] Q.Zhu,Y.Xu,L.Qiu,C.Ma,B.Yu,Q.iSong,W.Jin,Q.Jin,J.Liu,Y.
Mu, Lab Chip 2017,17, .
[] C. D. Ahrberg, J. M. Lee, B. G. Chung, BioChip J. 2019,13, .
[] Y. Seo, S. Jeong, J. Lee, H. S. Choi, J. Kim, H. Lee, Nano Converg. 2018,
5,.
[] C. D. Ahrberg, J.i W. Choi, J. M. Lee, K. G. Lee, S. J. Lee, A. Manz, B.
G. Chung, Lab Chip 2020,20, .
[] S. Bhat, J. Herrmann, P. Armishaw, P. Corbisier, K. R. Emslie, Anal.
Bioanal. Chem. 2009,394, .
[] L. Du, Y. Li, X. Zhang, Z. Zhou, Y. Wang, D. Jing, J. Zhou, ACS Appl.
Mater. Interfaces 2023,15, .
[] M. J. Hoser, H. K. Mansukoski, S. W. Morrical, K. E. Eboigbodin, PLoS
One 2014,9, e.
[] K. E. Eboigbodin, K. Moilanen, S. Elf, M. Hoser, BMC Infect. Dis. 2017,
17, .
[] K. Xu, X. Wang, R. M. Ford, J. P. Landers, Anal. Chem. 2016,88,
.
[] N. K. Mandsberg, O. Hansen, R. Taboryski, Sci. Rep. 2017,7, .
[] B.o Chang, Q. Zhou, R. H. A. Ras, A. Shah, Z. Wu, K. Hjort, Appl. Phys.
Lett. 2016,108, .
[] N. Ramalingam, M. E. Warkiani, N. Ramalingam, G. Keshavarzi, L.
Hao-Bing, T. G. Hai-Qing, Biomed. Microdevices 2016,18, .
[] S. Qian, H. Wu, B. Huang, Q.i Liu, Y. Chen, B.o Zheng, Sens Actuators
BChem2021,345, .
[] Y. Sun, in Nucleic Acid Amplification Strategies for Biosensing, Bioimag-
ing and Biomedicine (Eds.: S. Zhang, S. Bi, X. Song), Springer, Berlin,
Germany 2019.
[] A. Peethan, M. Aravind, V. K. Unnikrishnan, S. Chidangil, S. D.
George, Appl. Surf. Sci. 2022, , 571.
[] Y. Wang, B. Zhang, H. Dodiuk, S. Kenig, C. Barry, J. Ratto, J. Mead,
Z. Jia, S. Turkoglu, J. Zhang, ACS Appl. Mater. Interfaces 2021,13,
.
[] Y. Xia, S. Yan, X. Zhang, P. Ma, W. Du, X. Feng, B.i-F. Liu, Anal. Chem.
2017,89, .
[] J. Chi, C. Shao, X. Du, H. Liu, Z. Gu, ACS Appl. Mater. Interfaces 2018,
10, .
[] P. Mao, L. Cao, Z. Li, M. You, B. Gao, X. Xie, Z. Xue, P. Peng, C. Yao,
F. Xu, Analyst 2021,146, .
[] V. Jokinen, L. Sainiemi, S. Franssila, Adv. Mater. 2008,20, .
Adv. Mater. Interfaces 2024,11, 2300596 (9 of 9) © The Authors. Advanced Materials Interfaces published by Wiley-VCH GmbH
21967350, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/admi.202300596 by Aalto University, Wiley Online Library on [05/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
