Colloids and Surfaces A: Physicochem. Eng.

Aspects 429 (2013) 112–120

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces A: Physicochemical and

Engineering Aspects
journal homepage: www.elsevier.com/locate/colsurfa

Influence of relative humidity and nano-particle concentration on

pattern formation and evaporation rate of pinned drying drops of
nanofluids
D. Brutin ∗
Aix-Marseille University, IUSTI UMR 7343 CNRS, 13453 Marseille, France

h i g h l i g h t s g r a p h i c a l a b s t r a c t

• Relative humidity changes the evap-

oration rate but not the pattern
formation.
• Transition from the coffee ring
effect to homogenous deposition is
observed.
• The evaporation rate increase with
decreasing humidity more than pre-
dicted.

a r t i c l e i n f o a b s t r a c t

Article history: We report the pattern formation of nano-particles of iso-density water-based nanofluid drops. During
Received 16 November 2012 the spontaneous evaporation of a droplet in air under atmospheric conditions, the solvent evaporates and
Received in revised form 18 February 2013 nano-particles are deposited onto the substrate. Depending on the concentration, two different patterns
Accepted 6 March 2013
are observed: an o-ring pattern and a continuous nano-particle flower pattern. We observe a critical
Available online 11 April 2013
concentration that corresponds to the transition between the two patterns. Humidity is controlled and
modified to investigate its effect on the evaporation dynamics and on pattern formation. Surprisingly, the
Keywords:
leading pattern is identical regardless of the evaporation time, and the evaporative mass fluxes measured
Nano-particles
Humidity
are slightly higher than the expected flux, following a classical purely diffusive model for pure fluids.
Diffusion Two explanations are proposed: the first is that the diffusion coefficient of the nanofluid in humid air is
Evaporation different from that of pure water in dry air. The second is based on the experimental observation of an
increase in the spreading of the nanofluid, leading to an increase in the drop perimeter, a decrease in the
contact angle and thus an improvement in the evaporation rate.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction Using metallic nano-particles, inkjet printer technology is evolv-

ing to allow for the printing of metallic inks on soft substrates for
While nanofluids are widely studied in several scientific com- photovoltaic and several others applications. These metallic inks
munities (physics, biology, chemistry, etc.), several questions are mostly made of silver, and consequently, they are very expen-
regarding their heat transfer efficiency, stability and toxicity to sive (about $500 per milliliter). To create controlled patterns using
human health currently remain open problems. One important inkjet printing technology, a chain of droplets, whose volume has
application for nanofluids is inkjet printing [1], though nanoflu- been calculated, must create lines of constant diameter following
ids are very promising for tow-phase heat transfer as well [2]. evaporation. The homogeneity of the line thickness is important
because it will affect the electrical resistivity of the resulting pat-
tern. If the concentration of nano-particles is too low, an o-ring
∗ Tel.: +33 4 91 10 68 86; fax: +33 4 91 10 69 69. pattern will appear, which means that the matter will not have been
E-mail address: david.brutin@univ-amu.fr uniformly deposited on the substrate. If the concentration is too

0927-7757/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.colsurfa.2013.03.012
 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120 113

high, then the final thickness will also be too high, exceeding a cer-
tain thickness above which the electrical resistivity will no longer Digital camera
+ macro lens
change; thus, nano-particles, which are extremely expensive, will
USB to a computer
have been wasted. Thus, the search for the optimal concentration
that avoids o-ring formation and consequently enables a uniform
deposition by minimizing the amount of nano-particles is of great
interest. Several authors are currently looking at the driving param- Cold light
eters that lead to such pattern formation [3]. However, because the
evaporation of even a pure fluid droplet remains a complex topic,
there is much research to be done to understand the evaporation RS-232 to a computer
of complex fluids [4–6]. 0,00706 g
Electronic
O-ring formation during colloidal drop evaporation was first balance
explained by R.D. Deegan in 2000 [9]. By using sulfate-terminated Glovebox
polystyrene spheres of two sizes, 1 and 0.1 ␮m, at a maximum vol-
ume concentration of 2% dried on glass plates, the author observed Fig. 1. Schematic of the experiment.
the formation of different patterns depending on the concentration
and size of the spheres used. For spheres measuring 100 nm and
concentrations ranging from 0.25% to 0.063%, Deegan observed the deposited particle ring size can be precisely controlled by chang-
formation of multiple rings for initial droplets measuring 6 mm in ing the particle size. The segregation effect may consequently play
diameter. In all cases, particles remained at the drop center and an important role in polymer and biological fluids, which are com-
were not only located at the drop edge. In this study, we observed posed of poly-disperse particles. Using 100 nm particles, Marin et al.
two different final patterns depending on the concentration: an [8] demonstrated that the particles that settle first along the contact
o-ring pattern and a nearly uniform deposition pattern on the sub- line of a coffee ring pattern have an ordered, crystalline structure.
strate. Similar results has been observed by Denkov et al. [22] for Toward the center of the drop, a transition to a disordered particle
latex particles. In the case of mixtures of more complex fluids, at arrangement was observed.
low concentrations, the formation of nano-particle rings following Pattern formation has also been observed for biological fluids,
the evaporation of droplets was evidenced by Govor et al. in 2004 such as whole blood, very recently and shows similar patterns to
[11] for a liquid matrix of a binary mixture of nitrocellulose, amyl those of nanofluids [13]. The drying dynamics observed for this
acetate and hexane. The authors used 6 nm nano-particles of CoPt3 complex biological fluid show two typical regimes of drying, with
at concentrations below 1%, leading to rings measuring 0.6–1.5 ␮m the first one driven by diffusion and convection and the second
in diameter. They experimentally observed the phase separation driven only by diffusion through the dried deposit [14]. Finally,
that leads to the formation of a bilayer structure. The CoPt3 nano- pattern formation is influenced by the wettability of a droplet on a
particles located on the contact line assembled along the line. In substrate. The relationship between the spreading of a droplet and
this case, in which spheres smaller than our nano-particles were the evaporation dynamics has been previously discussed [15]. The
used, the accumulation at the drop edge was more pronounced. triple line dynamics and the pinning of a drop play a major role
Conway et al. [7] first investigated the influence of particles on the in the drying process. Consequently, the motion of flow inside a
evaporation dynamics of a drop. Using 200 and 750 nm polystyrene drop is also strongly affected, as we have evidenced using infrared
beads, the author observed the formation of an o-ring and a crater imaging [12].
for these bead diameters and monitored the drop height and mass In this study, we used 24-nm-diameter polystyrene sphere at
over time. Conway et al. provide linear normalized model of the concentrations ranging from 0.01% to 5.70% deposited on a 0.5 mm
drop mass and normalized model of the drop height using a root thick aluminum disk with a diameter of 10 mm coated with a Nuflon
square function. 16× coating to enable the pinning of the drops regardless of the
Chhasatia et al. [10] investigated the effect of relative humid- surrounding conditions. The humidity was controlled, and the tem-
ity on the contact angle and nano-particle deposition during the perature was observed to not affect the evaporation dynamics. Only
evaporation of a pico-liter droplet of a nanofluid. They observed, the effect of the nano-particle concentration on the pattern for-
for the pinned droplet mode, a spreading effect with the increase mation was monitored. The nano-particles used were coated with
in relative humidity. The observed variations in the contact angle carboxylate and an acid used as a surfactant to stabilize the par-
were linear but exhibited different slopes depending on the relative ticles. The experiments reveal similar drying regimes at different
humidity. The researchers conclude that due to the small volume of concentrations but the formation of two types of patterns.
the droplet, the relative humidity becomes an important parameter
driving the evaporation process and consequently the deposition. 2. Experimental setup
They also observed an increase in the initial spreading of the drop
with increasing humidity. Chhasatia and Sun [1] used bi-dispersed In the experiments, droplets were evaporated in a Jacomex
colloids made of micro- and nano-particles and observed the inter- 314 liter G-Box-T2 glovebox with the humidity regulated within
action between these particles and the contact line. For 100 nm and the range of 13–85% (Fig. 1). The nano-particles were iso-density
1.1 ␮m particles, they observed that, at the end of the evaporation with the fluid; hence, the volumetric and mass concentrations
process, the nano-particles were located near the triple line, fol- were identical. Nine different nanofluid concentrations were inves-
lowed by the micro-particles. This phenomenon is explained by a tigated: 0.01%, 0.05%, 0.11%, 0.23%, 0.47%, 1.15%, 2.30%, 4.80%
force balance. and 5.70%. Experiments were performed at atmospheric pressure
During the drying of droplets containing latex particles mea- Pa = 1014 ± 5 hPa. Room temperature and pressure are recorded by
suring 40 nm to 5 ␮m, Monteux and Lequeux [18] observed a a Lufft weather station (see Brutin et al. 2011 [13] for more details).
rearrangement that led to the formation of rings. The smallest nano- The nanofluids used were obtained from Invitrogen(R) for their
particles were deposited on the outer side of the ring, while the stability, homogeneity in size and chemical potential (sample
largest particles were located on the inner side. The authors con- F8787). The solutions were stable due to the presence of carboxyl-
clude that this deposition mechanism leads to a tight deposition but ate surface functional groups on the surface of the nano-particles.
leaves a thin liquid film of pure water at the edge of the drop. The The solvent matrix was water, with 2.30% ± 0.03% in mass, and
114 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120

contained 24-nm-diameter nano-particles. Solutions of various

concentration were obtained by the dilution (lower concentrations)
or evaporation (higher concentrations) of a 2.30% original sample.
Dilution was performed using distilled water (Bioblock) by weigh-
ing the samples. The accuracy of the electronic balance was 10 ␮g.
For 10 mL samples, this accuracy led to an uncertainty due to the
dilution or evaporation of less than 0.01%. Thus, the uncertainty in
the concentrations was assumed to be equal to the original sam-
ple concentration: 0.03%. Droplets with a constant volume of 7 ␮L
were created using an electronic syringe (Eppendorf X-Stream) in
all experiments. The droplets were then gently laid down on the
substrate.
The variation in drop mass was recorded using a Mettler Toledo
XS 105 balance. The balance resolution was 10 ␮g, which enabled
us to weigh samples with masses of up to 81 g. The drop masses
investigated were typically 7 mg. The mass signal was recorded
on a computer at 10 Hz. Image analysis was performed using a Fig. 2. Surface roughness using a confocal microscope (STILL Micromesure 2). Aver-
digital camera Canon EOS 7D coupled with a 1–5× macro lens age standard roughness of 1.736 ␮m.

Average concentration in nano-particles low

during most of the evaporation

Evaporation mass flux decrease

linked to the contact angle decrease

No influence of Nano-particles
the nano-particles influences the
on the evaporation evaporation
dynamic’s dynamic’s

Fig. 3. Mass evolution (bottom figure), evaporation mass flux (middle figure) and mean solid concentration (top figure) as a function of time for a nanofluid concentration
of 4.8% at a room humidity of 50%. Two stages of evaporation can be noticed depending on the mass flow rate of evaporation: one when the evaporation mass flux decreases
linearly and the other when the evaporation mass flux sharply decrease.
 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120 115

at 0.1 Hz. This camera allowed us to obtain images measuring

5184 × 3456 pixels over an area measuring 22.3 mm ×14.9 mm.
At the 1× magnification, the resolution was 4.30 ␮m, and at the
greatest magnification of 5×, the resolution was 0.86 ␮m. This opti-
cal set-up allowed for the visualization of drops from above with
enough detail to further analyze specific areas. Illumination was
provided by a cold cathode back light at 5000 ± 270 K, without any
surrounding light to minimize the reflection at the drop interface.
The substrate was an aluminum disk measuring 10 mm in diam-
eter and 0.5 mm in thickness. A 20–30 ␮m layer of Nuflon 16× was
deposited on the substrate by chemical vapor deposition (APS Coat-
ing, Bordeaux, France). The surface roughness was characterized
using a confocal microscope (STILL Micromesure 2); an average
standard roughness of 1.736 ␮m was measured (Fig. 2).
The interaction between the Nuflon 16× layer and the water Fig. 4. Nano-particle deposition in substrate grooves.

used as the matrix of the nanofluid allowed for the formation of

a pinned triple line during evaporation. A typical contact angle of tetrahedral pyramids measuring 1.736 ␮m on each side. Thus, each
70◦ was obtained. Detailed values are provided in Tables 2 and 3. cavity has a volume of 5.79 × 10−19 m3 and can store 1.21 × 105
The radii of the all of the droplets were below the capillary length nano-particles. The fluid density is 1050 kg/m3 . The tetrahedral
(2.7 mm for pure water) to ensure the formation of a spherical cap cavities are uniformly and continuously distributed over the sub-
at the drop interface (Eq. (1) where R is the wetting radius,  is the strate; thus, we compute the number of nano-particles stored in a
contact angle and V is the sessile droplet volume). Thus, the initial cavity based on a compact arrangement (66%). Table 1 shows the
contact angles were calculated from the volume and wetting diam- percentage of nano-particles stored in the cavities compared to the
eter. In this study, a quasi-steady-state assumption could be made total number of nano-particles in the droplet.
because Cv(1 − H)  ; thus, the diffusion time was far shorter than The number of nano-particles stored inside the cavities of the
the characteristic time of evaporation. rough surface is extremely small compared to the total num-
ber of nano-particles. For the lowest concentration, only 0.81%
R3 (1 − cos )2 (2 + cos ) of the nano-particles are deposited in the cavities, while 99.19%
V (R, ) = (1)
3 sin3  is deposited above the cavities. Consequently, we can conclude
that immediately after drop deposition, the external periphery of
the substrate beneath the drop is covered with nano-particles.
3. Drying dynamics of a drop of nanofluid
Thus, pattern formation is not affected by a surface roughness of
1.736 ␮m.
For all experiments, geometrical and physical characteristics
were measured, including the initial drop mass m0 , the final drop
5. Influence of nanofluid concentration on drying and
mass mF , the initial wetting diameter d and the total time of dry-
pattern formation
ing tF . The characteristic masses and times were obtained from the
variation in the droplet mass recorded by the electronic balance,
5.1. Drying and pattern formation
as illustrated in Fig. 3. The total time of evaporation tF is defined
by mF (t) = 0 as dm/dt tends to 0. The uncertainty in the initial drop
Experiments were performed by varying the concentration of
mass m0 is 0.1 mg, that in the total time of evaporation tF is 50 s and
the nanofluids at a constant glovebox humidity of 50%. Drops of a
that in the radii RE and RI is 2.2% based on the image analysis.
given mass were weighed and controlled initially; the initial disper-
From the mass evolution m(t), we can derive the evaporation
sion in drops with a mass of 7 mg ± was 6.0%. The measured room
mass flux and the average solid concentration. The evaporation
temperature was a nearly constant 25.0 ◦ C ± 0.6 ◦ C. Because the
mass flux clearly shows a drying transition regime immediately
experiments were carried out over several days, the atmospheric
below 2000 s, with a transition from a linearly decreasing evapora-
pressure slightly varied, 1013.4 hPa ± 4.9 hPa. All of the drops had
tion mass flux to a sharply decreasing mass flux. At the same time,
an initial contact angle of 67.5◦ ± 8.1%, as illustrated in Table 2
the mean solid concentration sharply increases. The first regime,
where m0 is the drop mass, tF is the time of drying, R are the external
with a mean nano-particle concentration below 20%, is very simi-
lar to that of the evaporation of a pure fluid drop with a receding
triple line (thus a decreasing evaporation mass flux). The transition Table 1
regime observed for concentrations above 20% is clearly affected by Amount of nano-particles in a 7 ␮L droplet, depending on the concentration,
deposited on a substrate cavities. Percentage of nano-particles in the cavities and
the presence of nano-particles, which slow down the evaporation
apparent layer above the cavities. The three highest concentrations show a uniform
kinetics. deposition of the nano-particles, while the others show the formation of an o-ring;
this last area of deposition is consequently much smaller [NP: nano-particles].
4. Influence of the surface roughness on the patterns  (%) Number of NP Area covered Number of NP in cavities %
(−) by cavities cavities below of total
To compare the total volume of nano-particles in a droplet with (mm2 ) the deposit
the volume of the cavities in the substrate based on a roughness of 5.7 5.25E+15 10.752 8.59E+06 0.02
1.736 ␮m, we assume that the substrate roughness can be modeled 4.8 4.42E+15 12.316 9.84E+06 0.03
by tetrahedral cavities. We show that the rough substrate rapidly 2.3 2.12E+15 9.186 7.34E+06 0.04
1.15 1.06E+15 2.852 2.28E+06 0.03
becomes smooth due to the deposition of nano-particles in the sub- 0.47 4.33E+14 2.212 1.77E+06 0.05
strate cavities, which fill the empty spaces, using only a few percent 0.23 2.12E+14 2.187 1.75E+06 0.10
of the nano-particles available (see Fig. 4). 0.11 1.01E+14 1.575 1.26E+06 0.15
We assume that the spherical nano-particles measure 24 nm in 0.05 4.61E+13 1.185 9.47E+05 0.25
0.01 9.21E+12 0.767 6.13E+05 0.81
diameter and that the initial drop volume is 7 ␮L. The cavities are
116 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120

Table 2
∗
Set of experiments conducted at constant humidity of 50% for different nanofluid concentrations ranging from 0.01% to 5.70%. Jmodel was obtained based on a purely diffusive
model [15].
∗ ∗
 (%) m0 (mg) tF (s) RE (mm) RI (mm) w/rE (−) Jexp (mg s−1 m−1 ) Jmodel (mg s−1 m−1 )  (◦ )

5.70 7.4 2350 1.85 – – 1.70 1.30 66.2

4.80 7.7 2450 1.98 – – 1.70 1.27 59.2
2.30 7.4 2600 1.71 – – 1.66 1.34 75.9
1.15 7.0 2250 1.71 1.42 0.169 1.82 1.06 73.6
0.47 6.4 2600 1.71 1.49 0.129 1.44 1.32 70.4
0.23 7.0 2600 1.84 1.64 0.104 1.50 1.30 65.0
0.11 6.7 2350 1.86 1.72 0.075 1.50 1.28 61.1
0.05 6.6 2200 1.77 1.66 0.064 1.73 1.30 66.9
0.01 7.1 2350 1.78 1.71 0.039 1.74 1.32 69.6

and internal droplet radii when applicable, w is the corona width, using particle concentrations below 0.01% were not performed
∗ is the evaporation mass flux per drop radius length 2m /(t d ).
Jexp because the uncertainty in the concentration was too significant
0 F E
The evaporation of the nanofluid drops led to the formation of and because the measured width was not accurate.
certain patterns, as shown in Fig. 5. For concentrations below 1.15%, Two main types of patterns could be observed as a function of
the drops formed an o-ring shape whose width decreased with the concentration:
concentration. This o-ring formation was previously observed by
Deegan [9] using a 2% mass concentration of sulfate-terminated
polystyrene microspheres dyed yellow–green with diameters of 1 • From 0.01% to 0.47%, the drops formed an o-ring pattern with no
and 0.1 ␮m. However, the author did not investigate the effect of particles at the center. The width of the deposits was observed
the concentration on the ring formation. In our study, experiments to decrease. No nano-particles were detected at the drop center.

Fig. 5. Influence of nanofluid concentrations ranging from 0.01% to 5.70% on pattern formation (glovebox humidity of 50%, room temperature of 25.0 ◦ C ± 0.6 ◦ C).
 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120 117

Fig. 6. Normalized mass variation as a function of normalized time for nanofluid concentrations ranging from 0.01% to 5.70% (room humidity of 50%). The inset is a
magnification of t/tD from 0.85 to 1.10 that illustrates the effect of the concentration on the final stage of drying. The dashed line is Eq. (2).

All nano-particles were deposited along the o-ring, whose width between 1.7 and 1.3. This induces a slightly nonlinear variation in
decreased with the nanofluid concentration. the normalized mass over time.
• For concentrations of 1.15% and above, particles were deposited All experimental mass curves collapse, indicating that the con-
at the center of the drop and the deposition thickness increased centration does not disturb the evaporation dynamics, except
with concentration. The formation of an axisymmetric flower during the last step of the evaporation process. Indeed, as shown
petal pattern was clearly observed. Several patterns were noted. in the inset of Fig. 6, the concentration plays a role in the evolution
The number of flower petals decrease with increasing concentra- of the drops only during the last few stages of drying. This result
tion. The final diameter also decreased due to a clear retraction indicates that, for all concentrations, the main stages of droplet
of the drying edge. evaporation are not affected by the nanofluid concentration, while
at the end of evaporation, when the local concentration is impor-
At constant humidity, the evaporation mass flux was tant, the concentration plays an important role. Depending on the
1.60 ␮g/(s mm) ± 8.0%, with an average experimental uncer- initial concentration, either an o-ring will form or nano-particles
tainty of 5.6%, for concentrations ranging from 0.01% to 5.70%. will be deposited at the center of a drop (see Fig. 5).
The overall evaporation rate was higher than that obtained with
a purely diffusive model, following Eq. (2), for which an approxi-
5.3. Variation in o-ring width
mation of f() is provided by Hu and Larson [17]. The experimental
droplet radii and initial contact angles were substituted into the
For our 24-nm-diameter nano-particles, a critical concentration
model to determine the total time of evaporation and consequently
occurs between 0.47% and 1.15%. Below 0.47%, only o-ring pat-
the evaporation rate. The quasi-steady diffusion evaporation mass
terns are formed, while above 1.15%, nano-particles deposit at the
flux was 1.28 ␮g/(s mm) ± 6.6%, which is within the uncertainty of
center of drying drops (see Fig. 5). Both patterns can be quanti-
the experimental measurements. The experimental evaporation
tatively characterized. The o-ring pattern is characterized by its
rates were 25% higher than those predicted by the model. These
width, which decreases with the concentration following a power
parameters were also studied at a fixed concentration and different
law, as evidenced in Fig. 7. This power law behavior is explained
levels of humidity for verification.
by the effect of the nano-particles on the triple line behavior dur-
dm ing the last stages of evaporation. Nano-particles accumulate near
− = RDcv (1 − H)fHL () (2)
dt the triple line during the first 80% of the evaporation time; then, the
triple line recedes due to the deepening effect reported by previous
fHL () = 1.3 + 0.27 2 (3) authors for such cases [1].
The ring width can be related to the concentration by the
power law correlation: w/r = ˛C ˇ with ˛ = 1.72 ± 1.92% and
5.2. Drying dynamics ˇ = 0.346 ± 4.23%. In 2000, R. Deegan obtained similar patterns
with microspheres measuring 0.1 and 1 ␮m and also observed
Fig. 6 presents the normalized mass [m/m0 ] as a function of power law behavior. However, his power law exponent ˇ was
the normalized time [t/tD ] using the diffusion times and initial 0.78 ± 12.8% for 0.1 ␮m-diameter microspheres and 0.86 ± 11.6%
drop mass. The dashed line represents a constant evaporation mass for 1 ␮m-diameter microspheres. In our case, using nano-particles
flux based on Eq. (2). Because the drop contact angle decreased measuring 24 nm in diameter, the value ˇ = 0.346 ± 4.23% is consis-
from approximately 70◦ to 0, the contact angle function fHL varies tent even if no quantitative link can be made between these values
118 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120

w/r (%) evaporated with different evaporation mass fluxes, which could be
predicted using Eq. (2), but the final pattern was almost exactly
O-ring Flower
pattern pattern
the same. The o-rings that formed were 23.1 % ± 5.6% of the initial
drop radii for all humidities. This result can be explained by the fact
that the mechanism by which the nano-particles are transported
to the drop edge is internal convection. The internal flow motion is
w/r directly proportional to the evaporative mass flux; consequently,
the intensity of evaporation scales with that of internal convection,
which transports the nano-particles to the drop perimeter, where
they accumulate.
The nano-particles were transported to the droplet edge by
internal flow motion. When the humidity was decreased, the global
evaporation mass flux increased; thus, at the triple line, the local
evaporation mass flux also increased. Therefore, the amount of liq-
uid transported to the triple line consequently increased due to the
C (%)
proportional increase in the internal flow motion. The internal flow
motion inside the droplets adapted to the external flow of vapor in
Fig. 7. Dimensionless o-ring width [w/r] as a function of the nanofluid concentra- the air. Because the nano-particles were carried out by the liquid,
tion (constant room humidity of 50%). w/r = ˛C ˇ with ˛ = 1.72 ± 1.92% and ˇ = 0.346 the increase in the internal flow velocities enhanced the transporta-
± 4.23%. tion mechanism pushing nano-particles to the drop edge. Thus, the
pattern formation was still the same but was realized more quickly
for micro- and nano-particles. Our nano-particles were coated with because the mechanism of formation was the same.
carboxylate (Invitrogen F8787), while the microspheres used in Similar flower pattern formation have been observed in high
Deegan’s study were not. In a future study, we will vary the nano- concentration colloidal suspensions such as polymer [20] or even
particle size to investigate the effect on ˇ. blood [14,19]. The different behaviors are analyzed by the authors
using a stability diagram obtained for colloidal suspensions. The
same mechanisms could be applied for drops of blood composed
6. Influence of humidity on drying and pattern formation
of high concentration particles of 8 ␮m. For these mechanisms
encountered in blood, the radial cracks formation is due to a stress
6.1. Drying and pattern formation
release in the dry-out. Due to the solvent evaporation, a closed
packed array of colloid is formed and stuck in between two inter-
Experiments were performed by varying the humidity in the
faces: substrate–colloids interface and colloids air interface. The
glovebox between 13% and 85%. The same initial mass of 7 mg
solvent evaporation create a pressure onto the colloids–air inter-
± 4.8% was weighed over time. See Table 3 for a summary of
face. This stress release periodically depending on the competition
the relevant droplet characteristics. T is room temperature, which
between the colloids adhesion onto the substrate and the evapo-
was a nearly constant 24.9 ◦ C ± 0.6 ◦ C, and P is the atmospheric
ration rate. The same mechanisms can explain the flower pattern
pressure, which was a nearly constant 1013.5 hPa. Because the
formation due to a stress release in the dry-out. This mechanisms is
experiments were carried out over several days, the atmospheric
not possible at low concentration since there is not enough material
pressure slightly varied, 1013.6 hPa ± 4.8 hPa.
deposited.
Experiments were performed at a controlled concentration of
1.15%, for which a very thin layer of nano-particles was deposited
at the center of the droplets. The initial drop diameter was 3.54 mm 6.2. Drying dynamics
± 3.7%; thus, the initial drop radius was 1.77 mm ± 3.7%. For all
drops, the wetting contact angle was 69.6◦ ± 8%, according to Eq. The mass flow rate of evaporation per drop radius is plotted in
(1). Fig. 9 and shows an increasing linear trend with decreasing humid-
Surprisingly, all of the dried drops of the nanofluids led to the ity (so increasing [1 − H]). A purely diffusive model using 2 also
same pattern, as shown in Fig. 8. Changing the glovebox humid- shows a linear trend but with a lower slope compared to that of the
ity greatly altered the total time of evaporation from 23 min at experimental results.
15% to 112 min at 87%. However, even if the internal flow inside For pure fluids, such as water, under the same conditions of sub-
the drops was strongly enhanced, we observed, using Fig. 8, that strate wettability, temperature and humidity without the heating
the final pattern was not modified. The same corona with a thin of the substrate, a purely diffusive model was used to fit the data.
layer of nano-particles in the central area was observed. The o-ring We observe in the slopes ˛ a difference of 24% with a higher exper-
width of all nine drops was measured to be 23.1% ± 5.6% of the ini- imental evaporative mass flux compared to that of the pure fluid.
tial diameter. The drops held under different glovebox humidities When the evaporation was slow for (1 − H) small, the difference

Table 3
Set of experiments performed at a constant nanofluid concentration of 1.15% at different humidities ranging from 13% to 85%.
∗ ∗
H (%) m0 (mg) tF (s) RE (mm) RI (mm) w/RE (−) Jexp (mg s−1 m−1 ) Jmodel (mg s−1 m−1 )  (◦ ) T (◦ C) P (hPa)

13.0 7.0 1400 1.77 1.39 0.214 2.82 2.26 69.4 24.9 1019.0
20.0 7.2 1600 1.74 1.35 0.225 2.59 2.20 72.9 25.6 1018.6
30.0 6.7 1800 1.67 1.26 0.248 2.23 1.95 74.9 25.8 1018.6
41.0 7.8 2250 1.68 1.29 0.237 2.05 1.57 80.0 24.5 1008.7
50.0 7.0 2400 1.78 1.38 0.227 1.64 1.31 68.7 25.0 1013.1
60.5 6.9 2700 1.79 1.40 0.220 1.42 0.96 67.2 23.8 1004.7
70.0 7.3 3750 1.85 1.45 0.218 1.04 0.78 65.1 25.0 1013.1
78.0 6.7 4800 1.78 1.35 0.243 0.78 0.57 66.9 25.0 1013.1
85.0 6.8 6750 1.87 1.41 0.246 0.54 0.38 61.4 24.9 1012.2
 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120 119

Fig. 8. Influence of humidity ranging from 15 to 87% on pattern formation (nanofluid concentration of 1.15%).

was small in terms of mg s−1 m−1 . However, when the evaporation

was five times faster (ratio of evaporation rates at 13% humidity
compared to that at 85% humidity), a gap was clearly observed.
The experiments reported in the literature have rarely been per-
formed at high relative humidity, considering the longer period
over which such experiments were carried out. Additionally, in
most research laboratories, air conditioning controls the relative
humidity to be approximately 50% to limit bacteria propagation at
high RH and to prevent throat pains at lower RH. It is very likely
that, in the literature, such experimental results are assumed to be
in agreement with a purely diffusive model if the authors cannot
provide results with small uncertainty. With our relatively very low
uncertainties, we are able to show that such a difference seems to
be nearly constant, approximately 24% ± 2.62%. The explanation
for this difference should be investigated by studying the diffusion
coefficient of our nanofluids in humid air, which is based on the
evaporation of pure water in dry air. Complementary experiments
should confirm our assumptions.
Another possible explanation proposed is based on the effect
of relative humidity on spreading. Chhasatia and Sun [1] reported
an increase in spreading and a decrease in the initial contact angle
with increasing relative humidity for pico-liter droplets of a water-
based nanofluid. The nanofluids used in this study are the same as
Fig. 9. Evaporation mass flow rate per drop radius as a function of the [1 − H] for the
same nanofluids with a concentration of 1.15%. The experimental results were com- those used by Chhasatia et al. If we assume the same behavior, then
pared to those of a purely diffusive model based on Eq. (2). Linear evolution of both J* even a slight increase in spreading coupled with a slight decrease
with J* = ˛(1 − H) with ˛exp = 3297 mg s−1 m−1 ± 1.35% and ˛model = 2664 mg s−1 m−1 in the contact angle would induce a higher evaporation rate. Our
± 1.27%. diffusive model presented in Fig. 9 is based on the initial droplet
120 D. Brutin / Colloids and Surfaces A: Physicochem. Eng. Aspects 429 (2013) 112–120

contact angle and does not take into account the change in the grateful to F. Martinez and B. Sobac for their fruitful discussions
droplet contact angle during the evaporation process. The effects and participation.
of the increase in spreading could be more significant throughout
the entire evaporation process. This assumption is currently being References
tested by monitoring the entire evaporation process to compute
the drop diameter and contact angle at all times. [1] V.H. Chhasatia, Y. Sun, Interaction of bi-dispersed particles with contact line in
an evaporating colloidal drop, Soft Matter 7 (2011) 10135–10143.
[2] J. Barber, D. Brutin, L. Tadrist, A review on boiling heat transfer enhancement
7. Conclusions with nanofluids, Nanoscale Res. Lett. 6 (2011) 1–16.
[3] A. Askounis, K. Sefiane, V. Koutsos, M.E.R. Shanahan, The effect of evaporation
kinetics on nanoparticle structuring within contact line deposits of volatile
Polystyrene nano-particles measuring 24 nm were suspended drops, Colloids Surf. A, http://dx.doi.org/10.1016/j.colsurfa.2012.10.017, in
in deionized water to study the effect of nano-particle concen- press.
tration and surrounding humidity on the pattern formation and [4] S. David, K. Sefiane, L. Tadrist, Experimental investigation of the effect of ther-
mal properties of the substrate in the wetting and evaporation of sessile drops,
drying dynamics of sessile droplets. Droplets were deposited on Colloids Surf. A 298 (2007) 108–114.
rough Nuflon 16×-coated substrates. The patterns formed were [5] P.L. Kelly-Zion, C.J. Pursell, S. Vaidya, J. Batra, Evaporation of sessile drops under
observed to have been influenced mainly by the concentration of combined diffusion and natural convection, Colloids Surf. A 381 (2011) 31–36.
[6] V. Starov, K. Sefiane, On evaporation rate and interfacial temperature of volatile
nano-particles but not at all by the drying time. Two patterns were
sessile drops, Colloids Surf. A 333 (2009) 170–174.
observed: an o-ring formation for concentrations below 0.47% and [7] J. Conway, H. Korns, M.R. Fish, Evaporation kinematics of polystyrene bead
a flower pattern above 1.15%. The effect of humidity was propor- suspensions, Langmuir 13 (1997) 426–431.
tional to the drying rate, but no effect on pattern formation was [8] A. Marin, H. Gelderblom, D. Lohse, J.H. Snoeijer, Order-to-disorder transition in
ring-shaped colloidal stains, Phys. Rev. L (2011) 085502.
observed. The experimentally measured evaporative mass flux was [9] R.D. Deegan, Pattern formation in drying drops, Phys. Rev. E 61 (2000) 475–485.
slightly higher compared to that predicted by a purely diffusive [10] V.H. Chhasatia, A.S. Joshi, Y. Sun, Effect of relative humidity on contact angle and
model, which may indicate a slight enhancement in heat transfer particle deposition morphology of an evaporating colloidal drop, Appl. Phys.
Lett. 97 (2010) 231909.
at the triple line due to the local improvement in the substrate heat [11] L.V. Govor, G. Reiter, G.H. Bauer, J. Parisi, Nanoparticles ring formation in evap-
capacity by nano-particle deposition. orating micro-size droplets, App. Phys. Lett. 84 (23) (2004) 4774–4776.
For pure fluids, the effect of the substrates thermal properties [12] D. Brutin, B. Sobac, F. Rigollet, C. Le Niliot, Infrared visualization of thermal
motion inside a sessile drop deposited onto a heated surface, Exp. Therm. Fluid
was investigated, even at ambient temperature, because the latent Sci. 35 (2011) 521–530.
heat of evaporation is mostly affected by element with the high- [13] D. Brutin, B. Sobac, B. Loquet, J. Sampol, Pattern formation in drying drops of
est heat capacity, which is the substrate [21]. One explanation is blood, J. Fluid Mech. 667 (2011) 85–95.
[14] B. Sobac, D. Brutin, Structural and evaporative evolutions in desiccating sessile
that the diffusion coefficient of our water-based nanofluid, even drops of blood, Phys. Rev. E 84 (2011) 011603.
at low concentration, in humid air is different from that of pure [15] B. Sobac, D. Brutin, Triple-line behavior and wettability controlled by
water in dry air. Experiments are currently underway to address nanocoated substrates: influence on sessile drop evaporation, Langmuir 27
(2011) 14999–15007.
this discrepancy. Another explanation is based on the experimen-
[17] H. Hu, R. Larson, Evaporation of a sessile droplet on a substrate, J. Phys. Chem.
tal observation in the literature of an increase in the spreading B 106 (2002) 1334–1344.
of nanofluids leading to an increase in the drop perimeter and a [18] C. Monteux, F. Lequeux, Packing and sorting colloids at the contact line of a
simultaneous decrease in the contact angle. The consequence is an drying drop, Langmuir 27 (2011) 2917–2922.
[19] D. Brutin, B. Sobac, C. Nicloux, Influence of substrate nature on the evaporation
improvement in the evaporation rate. This assumption is also under of a sessile drop of blood, J. Heat Transfer 134 (2012) 061101.
investigation at this time. [20] L. Pauchard, C. Allain, Stable and unstable surface evolution during the drying
of a polymer solution drop, Phys. Rev. E 68 (2003) 052801.
[21] B. Sobac, D. Brutin, Thermal effects of the substrate on water droplet evapora-
Acknowledgements tion, Phys. Rev. E 86 (2012) 021602.
[22] N. Denkov, O. Velev, P. Kralchevski, I. Ivanov, H. Yoshimura, K. Nagayama,
The author thanks the Carnot STAR for the financial support pro- Mechanism of formation of two-dimensional crystals from latex particles on
substrates, Langmuir 8 (12) (1992).
vided through the NanoPhaseChange Project. The author is also