Cellulose (2022) 29:777–798

https://doi.org/10.1007/s10570-021-04368-7 (0123456789().,-volV)
( 01234567
89().,-volV)

ORIGINAL RESEARCH

Supercritical CO2-assisted atomization for deposition

of cellulose nanocrystals: an experimental
and computational study
Shadi Shariatnia . Prajesh Jangale . Rohit Mishra . Amir Asadi .
Dorrin Jarrahbashi

Received: 29 July 2021 / Accepted: 9 December 2021 / Published online: 21 January 2022
Ó The Author(s), under exclusive licence to Springer Nature B.V. 2021

Abstract Nanoparticle spray deposition finds investigate the effect of different process parameters,
numerous applications in pharmaceutical, electronics, such as injection pressure, gas-to-liquid ratio, the axial
manufacturing, and energy industries and has shown distance between the nozzle and substrate, and CNC
great promises in engineering the functional properties concentration on the final patterns left on the substrate
of the coated parts. However, current spray deposition upon evaporation of water droplets. To this end, we
systems either lack the required precision in control- show how tuning process parameters control the size
ling the morphology of the deposited nanostructures or of carrier droplets, dynamics of evaporation, and self-
do not have the capacity for large-scale deposition assembly of CNCs, which in turn dictate the final
applications. In this study, we introduce a novel spray architecture of the deposited nanostructures. We will
system that uses supercritical CO2 to assist the particularly investigate the morphology of the nanos-
atomization process and create uniform micron-size tructures deposited after evaporation of micron-size
water droplets that are used as cellulose nanocrystal droplets that has not been fully disclosed to date.
(CNC) carriers. CNCs are selected in this study as they Different characterization techniques such as laser
are abundant, possess superior mechanical properties, diffraction, polarized microscopy, and high-resolution
and contain hydroxyl groups that facilitate interaction profilometry are employed to visualize and quantify
with neighboring materials. We fundamentally the effect of each process parameter. Numerical
simulations are employed to inform the design of
experiments. Finally, it is shown that the fabricated
Supplementary Information The online version contains nanostructures can be engineered based on the size of
supplementary material available at https://doi.org/10.1007/ the carrier droplets controlled by adjusting spray
s10570-021-04368-7.
parameters and the concentration of nanoparticles in
S. Shariatnia  P. Jangale  R. Mishra  the injected mixture. Process parameters can be
A. Asadi  D. Jarrahbashi (&) selected such that nanoparticles form a ring, disk, or
J. Mike Walker ’66 Department of Mechanical dome-shaped structure. Moderate operational condi-
Engineering, Texas A&M University, MEOB 309,
tions, simplicity, and time efficiency of the process,
College Station, TX 77843, USA
e-mail: djarrahbashi@tamu.edu and use of abundant and biodegradable materials, i.e.,
water, CNCs, and CO2 promote the scalability and
A. Asadi sustainability of this method.
Manufacturing and Mechanical Engineering Technology,
Department of Engineering Technology and Industrial
Distribution, Texas A&M University, MEOB 309, Keywords Nanoparticle spray deposition 
College Station, TX 77843, USA Supercritical-assisted atomization  Evaporation

123
778 Cellulose (2022) 29:777–798

induced nanostructure fabrication  Cellulose and final morphology of the deposited nanostructures
nanocrystals (Pawlowski 2008).
Direct-write deposition techniques on the other
hand, e.g. inkjet and aerosol jet printing, precisely
control the deposition of colloidal droplets and
Introduction formation of nanostructures on a targeted location
(Le 1998; Calvert 2001). Upon deposition on the
Deposition of nanoparticles via liquid atomization substrate, the contact line of the ink droplets pins to the
finds several practical applications in food (Huang substrate. As the droplet evaporation proceeds, the
et al. 2017), drug delivery (Singh and Van den Mooter contact angle between the droplet and substrate
2016), manufacturing (Shariatnia et al. 2019), energy decreases. As a result, a capillary flow from the center
(Kuznetsov et al. 2011; Krebs 2009), electronics (Zhao of the droplet towards the pinned contact line initiates
et al. 2012), and surface coating (Phan et al. 2009). to compensate the liquid mass loss at the droplet’s
Liquid atomization is referred to as a hydrodynamic periphery (Deegan et al. 1997). This flow drags the
process through which a liquid jet injected via a small particles and accumulates them along the edge of the
nozzle breaks up into several micron-size droplets droplet leaving a ring-shaped trace of particles on the
upon exposure to the surrounding fluid and forms a substrate (Yunker et al. 2011). This phenomenon is
spray. Spray deposition is a simple one-step, safe and known as the coffee ring effect (CRE) that can be
low-cost method for coating large surface areas within exploited or suppressed to enforce a specific nanopar-
few seconds to promote the efficiency and scalability, ticle pattern on the substrate depending on the
while reducing materials usage. In the nanoparticle application (Deegan et al. 1997). Suppression of the
spray deposition process, a colloidal suspension is CRE in direct writing methods requires costly and
atomized to create droplets containing nanoparticles multi-step processes such as the use of flammable and
of interest that subsequently evaporate and leave the toxic surfactants to the solvent, (Anyfantakis et al.
particles on the target surface. 2015) physical and chemical modification of the
Various configurations of thermal sprays (Moridi substrate (Dicuangco et al. 2014; Cui et al. 2012),
et al. 2014; Barbezat 2005), electrical sprays (Fau- and imposing external forces (Mampallil et al. 2015).
chais, Heberlein, and Boulos 2014), and direct-write Inkjet printing is limited to deposition of a single or a
deposition (Dinh et al. 2016; Bugakova et al. 2019) are few droplets at a time that covers a small surface area
among the most common spray deposition methods limiting the scalability of this technique. In addition,
that have been extensively studied. Thermal spray since the process parameters are set prior to print,
systems such as warm (Kuroda et al. 2015), plasma achieving thickness and material variability, requires
(Ke et al. 2019), and electro sprays (Pawlowski 2008) multiple rounds of deposition or using multiple print
use a heat source to melt the feedstock material and heads (Seifert et al. 2015). Aerosol jet printing utilizes
spray it on a substrate using a high-speed jet. Cold an air-assisted atomization technique for breakup of
spray systems where solid powders are accelerated the liquid jet stream and a specific directed nozzle to
towards the substrate, fall under the thermal spray facilitate targeted deposition (Wilkinson et al. 2019).
category. Due to the harsh conditions in these sprays, This technique is faster than inkjet printing and
coatings and substrates are limited to materials that capable of handling a wide range of materials in
can withstand large impact forces and are compatible moderate operating conditions (Lu et al. 2020; Paulsen
with high temperature (Fauchais et al. 2015). Thermal et al. 2012). It also provides precise control over the
sprays are cost effective and can cover large surface thickness and profile of the material deposition with
areas in a short period of time with a thickness that can the first round of spray (Secor 2018; Mette et al. 2007).
range from * 20 microns to several millimeters. However, aerosol jet printing lacks the capability of
However, thermal methods lack the precision needed large-scale deposition (Azarova et al. 2010; Sarobol
for coating layers with a few nanometer thicknesses, et al. 2016). The internal design of the nozzle that
and they do not provide any control over the formation directly affects the quality and dimensional resolution
of the print is very complicated (Seifert et al. 2015;
Jabari and Toyserkani 2015). Other limitations of this

123
Cellulose (2022) 29:777–798 779

method is low solubility of particles in the solvent and in SAA, this potentially important application has not
lack of control over droplet sizes (Goth, Putzo, and been fully explored in the literature to date. This
Franke 2011; Mahajan, Frisbie, and Francis 2013). knowledge gap motivated the current study where we
To overcome these limitations, supercritical-as- have designed and built a novel SAA system to
sisted atomization (SAA) has been introduced as a atomize aqueous Cellulose nanocrystal (CNC) sus-
method that utilizes a fluid above its thermodynamic pension and deposit the droplets containing CNCs on a
critical point to facilitate the atomization process by solid substrate to fabricate nanostructures with con-
enhancing the nanoparticles dissolution in solvents trolled size and morphology. In our previous study, we
(Reverchon 2002). SAA accelerates liquid atomiza- experimentally studied the effect of a wide range of
tion by exploiting the hybrid gas-like and liquid-like process parameters on the breakup and final droplet
properties of supercritical fluids. High density and size in the absence of nanoparticles (Shariatnia, Asadi,
high diffusivity along with low viscosity of supercrit- and Jarrahbashi 2021). In this study, we use SAA as a
ical fluids enhances the dissolution of gasses into the new large-scale delivery method for depositing CNCs
injected mixture and reduces the surface tension and controlling the fabrication of nanostructures on
between the injected liquid and the surrounding gas solid substrates. The proposed spray deposition tech-
that both improve atomization process (Shariati and nique can seamlessly be adopted for several practical
Peters 2003; Tom and Debenedetti 1991). Owing to applications involving CNCs and other nanoparticles.
their high diffusivity, supercritical fluids highly The main objective of this study is to (1) understand
dissolve in the liquid mixture prior to injection and the effect of spray parameters and colloidal suspension
separate from the mixture in the form of gas bubbles properties on the shape of the nanostructures formed
upon injection into the surrounding environment. The on the substrate after droplet evaporation, and (2) use
sudden expansion of these gas bubbles, triggers the this knowledge to effectively control and tailor the
breakup of the liquid jet and shatters it into very fine architecture of deposited nanostructures.
and highly uniform droplets with a narrow size We adopted CNCs in this study as they form a
distribution compared to other spray-based techniques stable dispersion in water and possesses unique
(Reverchon 2007). CO2 is an abundant, degradable, mechanical and chemical properties that make it
nontoxic, and nonflammable gas with moderate crit- appealing for a variety of applications, from 3D
ical temperature and pressure (31̊C and 7.4 MPa) printing (Shariatnia et al. 2019) and manufacturing
compared to other fluids, which makes it a viable (Palaganas et al. 2017; Tang et al. 2017; Shariatnia
option for several applications such as temperature- et al. 2020) to drug delivery (Roman et al. 2009) and
sensitive materials used in pharmaceutical and bio- electronics (Grishkewich et al. 2017). CNCs with
logical applications (Aguiar-Ricardo 2017). The high linear chain glucose units (C6H10O5) are abundant,
solubility of supercritical CO2 (SCO2) in most organic non-toxic, and biodegradable spindle-shaped nanopar-
and inorganic solvents has made it the supercritical ticles obtained from plants, algae, bacteria, and marine
fluid of choice in SAA systems (Costa et al. 2018). animals (Mariano et al. 2014). CNCs contain acces-
Sensitivity of mixture properties to the operational sible hydroxyl groups on its surface that makes them
conditions enables the regulation of droplet sizes and suitable for chemical modification (Moon et al. 2011).
the final morphology of the fabricated particles (Della In addition, CNCs possess unique features such as low
Porta, De Vittori, and Reverchon 2005). Although density (1.5 g/cm3), elastic modulus of 110–220 GPa,
SAA provides great control over process parameters, it tensile strength of 3–7.5 GPa, high aspect ratio
is limited to solely manufacturing micro/nanoparticles (10–100), and high surface area (Moon et al. 2011).
that are collected in a precipitator in the form of dry Evaporation-induced self-assembly of aqueous CNC
powder after atomization. droplets has been widely studied for optical sensing,
An important application of SAA is the direct security labeling, food, cosmetics, textiles, and art
deposition of nanoparticle-carrier droplets resulting applications (Parker et al. 2016, 2018; Gu et al. 2016).
from atomization of nano-colloidal suspensions However, these studies are focused on investigating
exposed to SCO2 and exploiting that process to the patterns in a liquid film or a single droplet and
engineer the nanostructures on a substrate. However, involve time and cost-inefficient lab-scale processes.
due to complex underlying atomization mechanisms To this end, we experimentally investigate the effect

123
780 Cellulose (2022) 29:777–798

of different spray parameters and the concentration of The internal geometry of the nozzle is a straight,
CNCs on the mean droplet sizes and the morphology circular cylinder with an actuator.
of the created nanostructure. We have leveraged
computational fluid dynamics (CFD) simulations to Materials
obtain the optimum spray parameters, achieve the
desired film thickness, and indicate the prime location Wood fiber based CNCs with an average diameter and
for delivery of droplets on the substrate where length of 3 and 75 nm, respectively were provided by
minimum droplet evaporation and spray bounce-back CelluForce (Quebec, Canada). The white CNC pow-
occurs. The computational results inform the exper- ders are produced via spray drying process and have a
imental system design, and thus reduce the trial-and- molecular formula of [(C6O5H10]22–28 SO3 Na]4–6.
error process to obtain the optimum deposition CNCs with concentration of 0.2, 0.5, 2 wt% were
outcome with minimum material waste and ensure dispersed in 500 mL of deionized water (DI-H2O)
system scalability. using probe sonication (Qsonica Q125 equipped with a
12 mm sonotrode) for 30 min at a frequency of
20 kHz and 75% intensity. Sonication was performed
Experimental method at room temperature and the colloidal suspension was
used within two hours to prevent sedimentation and
Spray setup and diagnostics ensure the quality of the dispersion. The zeta potential
and hydrodynamic diameter of dissolved CNCs in DI-
Figure 1 shows a schematic of the experimental setup water were measured as - 54 mV and 70 nm,
for SAA spray deposition of CNCs on a glass respectively.
substrate. The spray system has two feed lines that
deliver CO2 and the aqueous CNC suspension to a Test Conditions
custom-made pressure vessel. The ternary mixture,
i.e., CNCs, water, and CO2, mix and reside in the Table 1 represents different test cases and correspond-
pressure vessel. A pump feeds the colloidal suspension ing experimental conditions including the concentra-
into the pressure vessel, while another pump connects tion of CNCs in the injected suspension, injection
the CO2 tank to the pressure vessel. Pressure and pressure, and gas-to-liquid ratio (GLR). GLR is the
temperature are monitored in multiple locations along ratio of the CO2 mass flow rate to the liquid
the feeding lines and inside the vessel using several (nanoparticle suspension) mass flow rate measured
pressure gauges and thermometers. The mixture is upon feeding CO2 and the liquid separately into the
then injected into the ambient atmospheric air towards mixing chamber prior to injection. GLR is commonly
a glass substrate (VWR, micro cover glass No. 1.5). used to represent the gas content in the injection
mixture. The axial distance from the injection nozzle

Fig. 1 A schematic of the

experimental setup
indicating the feed gas
tanks, pumps, pressure
vessel

123
Cellulose (2022) 29:777–798 781

Table 1 Design of experiments

CNC concentration (wt%) Injection pressure (MPa) Axial distance (cm) GLR

0.2 3, 6, 7.5, 9 10, 15, 20 0.02, 0.05, 0.075, 0.1, 0.2, 0.5, 1, 2, 3, 4
0.5
2

where droplet Sauter Mean Diameter (SMD) is the spray. Sampling errors and the back-end algo-
measured is outlined in Table 1. The average size of rithms that are deployed in the system software to
droplets in a spray is commonly represented by Sauter convert the scattered light into meaningful particle
D3 size measurements are the main limitations of these
Mean Diameter (SMD) and is calculated as SMD ¼ Dv2 ,
s
systems (Andrews et al. 2011), yet the system has a
where the surface diameter and volume diameter are
qffiffiffi 1
1 Hz acquisition rate, 0.1 lm resolution, and 99%
defined by Ds = Ap, and Dv = ( 6V
p ) , respectively and
3
accuracy in size measurements. Measurements are
A and V represent the surface area and volume of the captured from the diffraction pattern of the superim-
droplet, respectively. All experiments were carried out posed laser beam and the spray. All reported droplet
at room conditions (25 °C and relative sizes are at least an average of six measurement
humidity * 40%). realizations.

Characterization techniques Profilometry

Microscopy Bruker DektakXT Surface Profiler (Bruker Corp.,

USA.), which is a stylus-based surface profilometer
A Leica DM6B (Leica Microsystems Inc., Germany) with a vertical resolution of 1 Å, is used to map the
motorized microscope equipped with 2x-40 9 objec- height of CNCs deposited on the surface of the glass
tives is used to portray the distribution of droplets on substrates after droplet evaporation. A stylus with a
the glass substrate. Polarized light mode is applied to 6.5 ml tip and 3 mg force is used for all measure-
visualize the distribution of the crystalline CNCs that ments. All experiments were carried out at least six
are otherwise transparent to brightfield lighting. samples and the average height profile is reported.

High-speed imaging Rheometer

A Fastcam SA5 (CA, USA) high-speed camera A cone-and-plate rheometer (Anton Paar-MCR 301,
equipped with a Nikon Nikkor (Tokyo, Japan) micro Austria) is used to measure the viscosity of aqueous
lens is used for diffuse back-illumination imaging of suspensions of CNCs with different concentrations at
the spray development. The resolving power of this room temperature. All experiments were repeated at
optical system correlated with the smallest feature that least six times and the average viscosity is reported.
it can accurately capture is * 20 l m. The images are
captured with a frame rate of 500,000 fps and have a Zetasizer
128  64 pixels field of view.
A Malvern Zetasizer Ultra (Malvern, UK) is used for
Laser diffraction measuring the diffusion coefficient of cellulose
nanoparticles in water using non-invasive light scat-
A Malvern Panalytical’s laser diffraction system tering. All experiments were repeated at least six times
(Malvern, UK) with a He–Ne laser source is used for and the average diffusion coefficient for each case is
real-time measurements of the average volume-based reported.
droplet size (SMD) at different axial locations across

123
782 Cellulose (2022) 29:777–798

Post-processing methods transport equations employed for solving the continu-

ity (Eq. 1) and momentum for the liquid film deposited
ImageJ (NIH) is utilized for post-processing the on the substrate (Eq. 2). For the details on Lagran-
microscopy images to measure the diameter of gian–Eulerian spray simulations, the reader is referred
droplets. In order to measure the surface area that is to the authors’ earlier works (Jarrahbashi et al.
coated with CNCs, we have binarized images by 2017a, b; Jarrahbashi et al. 2017a, b). The heat
imposing a global intensity threshold above which the transfer on the solid substrate has been neglected here
intensity was set to one and the remaining pixel to isolate the behavior of the splashing droplets upon
intensities were set to zero. We have then used ImageJ reaching the substrate from evaporation effects. It is
to measure the area covered with pixels that have the noted that small droplets may have a non-Newtonian
intensity of one. The jet development simulation is behavior due to the presence of the nanoparticles;
visualized using the EnSight software package from however, as the CNC concentration is low, this study
ANSYS. In addition, the Matplotlib library in Python assumed a Newtonian behavior.
is utilized to plot and analyze the data from laser
diffraction and profilometry techniques. od 1 X N side
! Sd
þ ð V film :b
n Þ i di l i ¼ ; ð1Þ
ot Awall i¼1 ql Awall

!
Computational method oðd: V film Þ 1 X N side
! !
þ V film ð V film :b
n Þi di li £i
ot Awall i¼1
Governing equations PN side PN side !
 i¼1 ðPb n Þ i di l i M tang s Ai
¼ þ þ i¼1 ; ð2Þ
ql Awall ql Awall ql Awall
Three-dimensional Computational Fluid Dynamics
(CFD) simulation is carried out in the open-source The continuity and momentum equations are pre-
C??-based CFD package OpenFOAM-2.2.x (Weller sented in (1) and (2), respectively and Awall is the area
et al. 1998) to model the atomization and breakup of !
of the wall cell, V film is the film velocity, li is the
the aqueous suspension, formation of the liquid film substrate length at side i,ql is the film density, di is the
deposited on the solid substrate, and the behavior of film thickness at side i, £i is the impingement angle,
the droplets on the substrate that include stick, and Sd is the source term. The following equations are
rebound, spread, and splash (Tsang et al. 2014). A used to calculate the pressure as P ¼ Pcell þ Pd; where
two-dimensional domain is considered for simulating Pcell is the free stream pressure. Pd is the dynamic
the liquid film on the solid substrate. Dynamic pressure due to impingement and splashing of the
structure Large Eddy Simulation (LES) is imple- droplets defined as follows:
mented in this study (Mishra and Rutland 2019) to
N
X drop NXsplash
incorporate the turbulence effect of the fluid phase Adi Aj
Pd ¼ ql V 2nd þ ql V 2nj ð3Þ
(liquid suspension and the surrounding gas) and Awall A wall
i¼1 j¼1
Stanton-Rutland model is employed to model film
formation on the solid substrate (Stanton and Rutland where V nd is the normal component of velocity of the
1996a, b). A Lagrangian–Eulerian approach (Markt incoming droplets and V nj is the normal component of
et al. 2018) is used for the spray simulation that treats velocity of the jth secondary droplet due to splashing.
the gas phase as a continuum for which a complete set Adi and Aj are projected areas of the ith incoming
of transport equations are solved while the liquid droplet and jth splashed droplet, respectively. M tang is
phase is considered as a discrete phase transported the tangential momentum due to the impingement and
with the gas medium. The sub-models used for the splashing of the droplets defined as follows:
spray simulation include the dispersion model,
splash 
! 
N
X drop NX
Kelvin–Helmholtz Rayleigh–Taylor (KH-RT) !
M tang ¼ ðmi V sdi Þ  m j V sj : ð4Þ
breakup model (Reitz 1987), vaporization model i¼1 j¼1
(Zuo 2000), Ranz-Marshall (Ranz 1952) heat transfer
model, and dynamic structure turbulence model. For Finally, the shear force acting on the substrate due
the sake of brevity, we only discuss the modified to the droplet splashing is defined as

123
Cellulose (2022) 29:777–798 783

N
X edge NX
splash Results and discussion
ðAj !
s jÞ ¼ ð!
s Þedge;i di li þ ð!
s Þwall Awall
j¼1 i¼1 Spray formation
þ ð!
s Þliq=air Awall; ð5Þ
It is crucial to understand the breakup mechanisms of
where ð! s Þedge;i is the shear stress along the edges of
the supercritical CO2-assisted atomization of the
the film, ð!s Þwall is the wall shear stress, and ð!
s Þliq=air aqueous CNC suspension as it creates droplets that
is the shear stress at the interface between the gas and carry and deposit the nanoparticles on the substrate. In
the liquid. order to fully understand the effect of different
parameters on the deposition process and the created
Droplet behaviour on the substrate micro/nano structure, a wide range of test cases
outlined in Table 1 are studied. The experiments are
The droplet behavior upon reaching the substrate is designed to encompass different phases (i.e., subcrit-
detected via the droplet splashing criteria suggested by ical, critical, and supercritical phases) of the hybrid
Stanton (Stanton 1996) as outlined in Table 2. It CO2-water mixture. Supercritical CO2 (SC-CO2) has a
indicates whether the droplets stick, rebound, or high density and is highly soluble in water around the
spread on the solid substrate depending on the critical point of the CO2-water mixture, since the
frequency of the incoming impinging droplets. This diffusion coefficient of CO2 significantly increases
criterion is based on the Weber number (We) that is close to this critical pressure (i.e., * 7.5 MPa). The
defined as the ratio of the drag force to surface tension high solubility of CO2 in water reduces the interfacial
force acting on the droplets. The parameters given in tension of the injection mixture that is shown to
Table 2 include dd, the diameter of impinging droplet; facilitate the atomization process (Jarrahbashi and
f, the frequency of droplets impinging on the wall, v, Sirignano 2014; Jarrahbashi et al. 2016). Table 3
the velocity, and r is the surface tension coefficient. presents the thermophysical properties of CO2-water
We will calculate the velocity of the droplets after mixture at sub-critical, critical, and supercritical
rebound and the angle at which the droplets bounce off states. Comparing these values shows that that opti-
from the substrate. We will identify the position of the mum condition (high diffusion coefficient and low
droplets based on their velocity upon impact with the interfacial tension) is achieved at the critical pressure
substrate to predict the liquid film growth towards the of the CO2-water mixture and supercritical pressure.
edges of the substrate as the droplets stick to the Increasing the pressure beyond 9 MPa does not further
substrate. The number density of the droplets bounc- change the solubility and the interfacial tension. The
ing off the substrate will be calculated. Finally, a translational diffusion coefficient of CNCs in water is
Weibull distribution (1951) is used to calculate the measured using dynamic light scattering (DLS) and is
diameter of the droplets which break down and bounce plotted as a function of CNC concentration in Fig. 2.
back from the surface of the liquid film on the By increasing the concentration of CNCs from 0.2 to
substrate. The equations used for calculating the 2wt%, the measured diffusion coefficient decreases,
above-mentioned parameters are summarized in the and viscosity increases.
supplementary information document. High mole fraction of dissolved SCO2 in aqueous
CNC suspension results in the formation of a bulged
core filled with CO2 in the liquid jet very close to the
nozzle. The emergence of gas bubbles and depressur-
Table 2 Droplet splashing criteria (Stanton 1996) ization into the atmospheric pressure causes bubble
expansion and eventually bubble burst. The force
Stick We\5
produced by the bubble burst shatters the liquid into
Rebound 5\We\10 micron-size long and slender ligaments that eventually
Spread 10\We\324v1=4 f 3=4 ðq=rÞ1=2 breakup and form small droplets (Shariatnia, Asadi,
Splash We [ 324dd v1=4 f 3=4 ðq=rÞ1=2 and Jarrahbashi 2021). The temporal development of
the bubbles and ligaments in a region close to the
nozzle (* 300 l m downstream of the orifice exit) is

123
784 Cellulose (2022) 29:777–798

Table 3 Thermophysical properties of subcritical, critical, and supercritical CO2-H2O. (Diamond and Akinfiev 2003; Bachu and
Bennion 2009)
Pressure (MPa) CO2 solubility (mol%) CO2 diffusion coefficient (m2/s) 9 10–10 Interfacial tension (mN/m) Density (kg/m3)

3 1.35 6.7 56.5 1015.2

6 2.15 15.1 40.8 1016.1
7.5 2.35 18.5 36.2 1018.5
9 2.41 13.9 33.5 1020.3

Fig. 2 Diffusion coefficient

of CNCs in water (left axis,
dashed blue line), and
viscosity of aqueous CNC
suspension (right axis, solid
green line) as a function of
CNC concentration. nozzle,
substrate, and the laser
diffraction system

portrayed in Fig. 3. The lower interfacial tension of parameters (i.e., injection pressure, gas-to-liquid ratio,
CO2-water mixture at supercritical conditions facili- axial distance from injection orifice) and injection
tates the ligament breakup and the combined effects mixture properties (i.e., the concentration of nanopar-
result in enhanced primary breakup and formation of ticles) on the droplet average size. These analyses will
fine droplets with homogenous size distribution that aid in designing the process parameters of the spray
ensures a uniform distribution of the deposited system. The next section discusses the effect of
nanostructures on the substrate upon evaporation of various process parameters on the SMD of the droplets
the solvent. We have detailed the breakup mechanism containing nanoparticles.
of the liquid jet in the same SAA system in our earlier
paper (Shariatnia, Asadi, and Jarrahbashi 2021). Droplet size distribution
The droplets created through the atomization
process carry the nanoparticles and place them on The effect of different spray parameters and physical
the substrate. The assembly of the nanoparticles and properties of the injection mixture on the average size
the pattern of the deposited nanoparticles are highly of carrier droplets that is especially crucial for
affected by the droplet evaporation dynamics on the designing the nanoparticle delivery system is dis-
substrate and the size of the droplets. As a result, it is cussed in this section. The laser diffraction system is
important to understand the effect of different spray used for real-time measurement of SMD at 10, 15, and

Fig. 3 High-speed images (500,000 fps) capturing the early development of the spray. GLR = 0.2, Pinj = 9 MPa developing with time
from left to right with a 2 sl time interval between the frames

123
Cellulose (2022) 29:777–798 785

20 cm axially located downstream of the nozzle. with smaller sizes. This is owed to the higher solubility
These points are selected to fully represent the whole of CO2 in water and lower interfacial tension of CO2-
spray plume. Figure 4 plots the measured SMD as a water mixture at higher pressures as was indicated in
function of GLR for different injection pressures and Table 3. The combined effects enhance the primary
axial locations. It is observed in Fig. 4a–c that for each breakup of the liquid jet due to the burst of dissolved
injection pressure, the mean droplet size decreases as gas bubbles and surface capillary breakup. As a result,
GLR increases and the rate of SMD reduction the variation of droplet sizes by changing injection
decreases with an increase in GLR and reaches a pressure is more evident in cases where measurement
plateau at the GLR of 0.2. At this point, increasing is performed closer to the nozzle (i.e., 10 cm axial
GLR does not have a noticeable effect on the SMD and distance in Fig. 4(a)) compared to measurements
hence this value (i.e., GLR = 0.2) is selected for spray further away from the injection orifice (i.e., 15 and
deposition experiments. 20 cm from the orifice in Fig. 4(b, c)). It is also
At each GLR and axial distance, increasing the evident that for each injection pressure, increasing the
injection pressure results in the formation of droplets axial distance between the injection orifice and SMD

Fig. 4 SMD measurements

as a function of GLR for
different injection pressures
and an axial distance of
a 10 cm, b 15 cm, and
c 20 cm from injection
orifice

123
786 Cellulose (2022) 29:777–798

probe from 10 cm in Fig. 4a to 20 cm in Fig. 4c, of the injection mixture leads to an average of * 54%
results in smaller mean droplet sizes and their size growth in droplet sizes of the spray. It is well
does not vary significantly with GLR. This can be established that an increase in liquid viscosity results
attributed to the ‘‘secondary breakup’’ of droplets that in the formation of larger droplets as it suppresses the
occurs at locations further away from the nozzle. The breakup process by dampening the interfacial pertur-
secondary breakup is referred to a process in which the bations between the liquid and gas upon injection that
droplets exposed to high shear forces breakup into eventually break it up to multiple droplets (Bouse et al.
multiple smaller droplets. 1990; Dayal et al. 2004). The direct effect of droplet
Figure 5(a) shows the measured SMD as a function sizes on the dynamics of solvent evaporation which in
of injection pressure at 1 cm axial distance with turn influences the assembly of nanoparticles and
respect to the nozzle and for three different concen- architecture of nanostructures formed on the substrate
trations of CNCs in the suspension (i.e., 0.2, 0.5, and is discussed in the next section.
2wt%). Similar to Fig. 4, increasing the injection
pressure reduces the droplet sizes. The droplet sizes Nanostructure patterns
breakup into smaller droplets at locations further away
from the orifice due to the secondary breakup. The In this section, we discuss the CNC patterns that form
sharpest decrease in the droplet is achieved at 7.5 MPa after the evaporation of the liquid droplets generated
injection pressure that is close to the critical pressure through atomization of the aqueous CNC suspension.
of the CO2-water mixture. The proximity to the critical We study the effects of different process parameters
pressure enhances the diffusivity of CO2 in water and on the created nanostructure on a glass substrate.
the creation of more bubbles inside water upon Figures 6 visualizes the polarized micrographs of
injection that enhances the atomization process and otherwise transparent CNC nanostructures that are
reduces the droplet size. Further increasing the formed on the substrate upon droplet evaporation for
injection pressure to 9 MPa has a negligible effect different injection pressures. The glass substrates are
on the droplet size. This can be attributed to the 1 cm by 1 cm. Figure 6 illustrates the architecture of
maximum diffusion coefficient of CO2 in water that nanostructures for various injection pressures for
occurs at 7.5 MPa (Tewes and Boury 2005), which in 0.2wt% (left column) and 2wt% (right column) CNC
turn results in minimum surface tension value for the concentration. The main pattern of assembled nanos-
water-CO2 mixture at this pressure (Bachu and tructures in these top-view micrographs can be
Bennion 2009). In addition, increasing the concentra- categorized in one of the three shapes: (1) ring-shape,
tion of nanoparticles in the injection mixture, from 0.2 where the majority of nanoparticles accumulate along
to 2wt%, increases the overall size of the carrier the edge of the evaporating droplet, (2) homogenous
droplets. The viscosity of the aqueous suspension distribution, where particles scatter across the surface
increases from 1.2 to 4 mPa.s by increasing the area of the evaporating droplet, and (3) transition
concentration from 0.2 to 2wt% as indicated in stage, where there is still a distinct ring-shape structure
Fig. 2. This enhancement (* three-fold) in viscosity and some particles are also scattered within the center

Fig. 5 SMD as a function of

injection pressure for
different CNC
concentrations measured at
an axial distance of 15 cm
and GLR = 0.2

123
Cellulose (2022) 29:777–798 787

0.2 wt% CNC 2 wt% CNC

a-1 10 m a-2

Rings

(a) Pinj=3 MPa

Rings

b-1 b-2 Ring

Ring
(b) Pinj=6 MPa Transition Transition
Rings
Ring

c-1 c-2

Homogenous Homogenous
(c) Pinj=7.5 MPa

d-1 d-2

(d) Pinj=9 MPa Homogenous

Homogenous

Fig. 6 Polarized microscopy of CNC patterns after evaporation 9 MPa injection pressures and 2 wt% CNC concentration for a-2
of water in CNC aqueous suspension droplets on a glass 3 MPa, b-2 6 MPa, c-2 7.5 MPa, and d-2 9 MPa injection
substrate located at 15 cm axial distance for 0.2 wt% CNC pressures. The 10 ml scale bar is identical in all images
concentration for a-1 3 MPa, b-1 6 MPa, c-1 7.5 MPa, and d-1

of the evaporating droplet. It is illustrated in Figure 7 demonstrates the profile/height measure-

Fig. 6(a1–d1) that regardless of the injection pressure, ments of assembled nanostructures upon evaporation
the droplets with diameters smaller than * 5.5 ml for droplet sizes varying from 5 to 13 ml and different
exhibit a homogenous distribution, while droplets concentrations (0.2wt%, 0.5wt%, 2wt%). Combined
larger than * 7.5 ml have generated a ring-shaped with top-view micrographs presented in Fig. 6, they
structure, and droplets with diameter sizes in between provide a 3D realization of the shape of assembled
the two thresholds (i.e., between 5.5 and 7.5 ml) CNC nanostructures. In Fig. 7, the droplets have been
represent a transition between the two identified injected at 9 MPa. We discussed the effect of injection
regimes. All three patterns were observed for all pressure on the nanoparticle patterns in Fig. 6 and
injection pressures as the droplet size distribution showed that droplets with the same size and concen-
envelopes the detected thresholds. In Fig. 6(a2-d2) tration shared the same pattern regardless of the
that illustrate droplets with a higher concentration of injection pressure. The profilometry height measure-
CNC particles (i.e., 2wt%), the homogenous distribu- ments indicate that nanostructures represent a ring,
tion, transition, and ring structure occurs for \ 9.5 disk, or dome shape. A ring pattern that is identified
ml, * 9.5–11.5 ml, and [ 11.5 ml droplet sizes, with two peaks on the height profile is referred to the
respectively. It is noted that at least 6 images were accumulation of nanoparticles along the edge of the
taken at different locations of the same substrate; all of droplet (labeled as ‘‘ring’’ in the top view in Fig. 6). A
which were in great agreement with the threshold dome forms when nanoparticles are captured at the
detected in these figures. interface during evaporation and mainly remained in
the center after droplet evaporation i.e., only one peak

123
788 Cellulose (2022) 29:777–798

Fig. 7 Profilometer height measurement of CNC nanostructures created on substrate after evaporation of water in droplets as a function
of droplet diameter. CNC concentrations are 0.2, 0.5, and 2wt% and the injection pressure is 9 MPa

is observed on the height profile. The dome structure surface area of the droplet compared to the dome and
was identified as ‘‘transition’’ in the top view depicted ring and. The disk pattern was identified as ‘‘homoge-
in Fig. 6. Finally, a disk pattern forms when the height nous’’ in the top view Fig. 6. Figure 7 shows that by
profile is nearly flat at the center. This indicates decreasing the droplet size from 13 lm (red) down to
nanoparticles are scattered more uniformly across the 5 lm (blue), the assembly of particles transits from

123
Cellulose (2022) 29:777–798 789

ring-shape to a dome-shape structure for all CNC concentration, 90% of droplets are smaller than
concentrations. The 9 lm-droplet (green) represents 4 ml in size. Based on the microscopic images and
the transition between ring to a disk-shape structure. height profilometry, droplet sizes smaller than 4 ml
By increasing the concentration of CNCs, the transi- will represent a homogenous nanoparticle distribution.
tion from a ring structure to disk occurs at larger As a result, most of the deposited nanostructures will
droplet sizes. As will be discussed in the next section, exhibit a disk-shape structure. The information from
the droplet size and concentartion directly affects the this measurement combined with detailed discussions
evaporation rate of the solvent, which in turn influ- on the 3D architecture of fabricated micro/nanostruc-
ences the particle advection and diffusion and ulti- tures have important implications in designing prac-
mately the nanoparticle patterns. tical deposition systems to ensure that majority
In summary, by controlling the droplet sizes we can ([ 90%) of droplets fall under a certain category
engineer the desired pattern (ring, dome, disk) for (i.e., ring versus homogenous distribution).
different concentrations. The use of SCO2 enables
achieving a very uniform distribution of droplet sizes Evaporation-induced nanoparticle assembly
within the spray that facilitates achieving a uniform
distribution of CNCs with the desired pattern on the In this section, we will explore the CNC assembly in
substrate. The injection pressure can directly control micron-size evaporating droplets (Govor et al. 2004;
the overall size of droplets within the spray plum and Lu et al. 2013). In order to find the link between the
can be adjusted to the size requirements of the specific dynamics of droplet evaporation and the formation of
application where the spray deposition system is being a specific pattern upon evaporation, two main param-
used. To quantitatively demonstrate the control over eters are identified: (1) droplet evaporation rate that is
the nanostructure patterns with the injection pressure linked to the convective transport of CNCs as the
Dv90 measurement using laser diffraction method is droplet edge recedes back during droplet evaporation;
plotted vs. injection pressure for variable concentra- and (2) the Brownian diffusion rate of CNCs in water.
tions plotted in Fig. 8. Dv90 indicates the mean It has been shown that in the absence of other
diameter size that represents 90% of the total volume competing mechanisms e.g., external forces, special
of the existing liquid droplets Fig. 8 shows that at treatment of the substrate or solvent, the competition
7.5 MPa injection pressure and 0.2wt% CNC between the convective and diffusive transport of

Fig. 8 Dv90 as a function of injection pressure for different CNC concentrations measured at an axial distance of 15 cm and
GLR = 0.2

123
790 Cellulose (2022) 29:777–798

particles dictates the final pattern after droplet evap- where R, D, H, Cv, and h are the droplet radius (1 mm),
2
oration (Vehring 2008). The ratio of the convective to water vapor diffusivity (2.42 9 10–5 ms ), relative
diffusive transport of particles during water evapora- humidity (40%), saturated water vapor concentration
tion is represented by the non-dimensional Péclet (Weast and Astle 1981) (23.2 mm g
3 ), and droplet contact
(Pe = r2/Dte) number, where ‘r’ is the droplet radius, angle (0.369 rad), respectively. The contact angle is
‘D’ is the particle mass diffusivity in the liquid phase measured on an image that is taken normal to a back-
and ‘te’ is the droplet evaporation time. For millime- illuminated droplet deposited on a solid substrate.
ter-sized droplets, it has been shown (Wei 2015) that Larson’s model predicts 709 s for a 1 mm droplet to
the ring pattern is typically favored for Pe [ 1 as the evaporate and our experimental measurement indi-
convective rate surpasses the diffusive rate. A reduc- cated 718 s, which is in close agreement with the
tion in the Pe, which implies a diffusion-dominated model prediction. The translational diffusion coeffi-
transport, is known to mitigate the ring formation cient of CNC in DI-water measured by DLS for 0.2,
toward a more uniform particle distribution (Mam- 0.5, and 2wt% concentration (Fig. 2) is 7.18 9 10–12,
pallil and Eral 2018). 6.7 9 10–12, and 4.3 9 10–12 m2/s, respectively. The
Our SMD measurements of the spray suggest that diffusion coefficient reduces with concentration due to
the droplet sizes are below 20 microns for which the packed space hindering the freedom of particles to
measuring the droplet evaporation rate is experimen- transport (Van Rie et al. 2019). We use this data along
tally very challenging. As an alternative approach, with the evaporation rate obtained from Larson’s
there are various mathematical and analytical models model and droplet sizes captured by the laser diffrac-
to calculate the evaporation rate of a sessile droplet tion measurements to calculate Pe.
(Hu and Larson 2002, 2005; Nguyen and Nguyen Figure 9 shows the calculated Pe as a function of
2012). Larson’s model (Eq. 6), which is applicable for droplet size for different CNC concentrations. By
semispherical sessile droplets, is commonly used as increasing the droplet size for each concentration, the
one of the most accurate models that has been verified evaporation time is also increased while the diffusion
empirically (Hu and Larson 2002). This model is more coefficient is constant for the same concentration. This
accurate when the Bond number is smaller than 0.1 results in higher Pe at higher concentrations. Accord-
(Bo ¼ qgRh lur
r ) and the capillary number (Ca ¼ r ) is
0
ing to Fig. 9, the corresponding Pe for a droplet size of
smaller than 1. Bo is the ratio of the gravitational to 13 l m is 1.27, 1.37, and 2.13, for 0.2, 0.5, and 2.0 wt%
surface tension forces and accounts for the initial concentration, respectively. Pe [ 1 indicates the
shape of the droplet whereas Ca is the ratio of viscous domination of the convective transport of CNC
to capillary forces and accounts for deformation of the particles towards the edge of the droplet induced by
droplet during evaporation. Here q, g, R, h0, r, l, and the evaporation of DI-water and formation of a ring-
ur are the density, gravitational acceleration, contact shape structure as was depicted in Fig. 7. Droplets
line radius, initial droplet height, liquid–air surface within the 6–8 l m diameter range have an average Pe
tension, liquid viscosity, and average radial velocity of 0.7, 0.8, and 1 for 0.2, 0.5, and 2.0 wt% concen-
due to evaporation, respectively. We first compare the tration, respectively. These cases where convective
experimentally measured evaporation rate for a 1 ll and diffusion rates are almost equal were identified as
droplet (* 1 mm in radius) deposited on a glass the transition between ring and dome shape structures
substrate with the predictions of the Larson’s model. in Fig. 6.
The evaporation rate has been measured with a timer It is noted that pure CNCs are almost hydrophilic,
at room condition (i.e., 25  C temperature and * 40% and thus tend to form a ring. However, various
relative humidity). Comparing the Bo (* 0.07) and observed patterns for different concentrations and
Ca (O(10–8)) for the largest droplet (i.e., 1 mm radius) different droplet sizes imply that in addition to the
indicates that the droplet has a spherical cap shape and particle shape, level of hydrophilicity, the droplet size,
satisfies the requirement for using the Larson’s model and particle mass concentration also play a role in
(Eq. 6): determining the final pattern. For instance, Fig. 7
m_ ðtÞ ¼ pRDð1  HÞC v ð0:27h þ 1:3Þ ð6Þ showed that for a 9 lm-droplet, the pattern changed
from a ring at 0.2 and 0.5wt% to a disk at 2wt%. Our

123
Cellulose (2022) 29:777–798 791

Fig. 9 Peclet number as a function of droplet sizes for different CNC concentrations

results showed that increasing the ratio of mass resulting droplets deposited on the substrate. The
concentration to droplet size tends to change the black color represents the water, and the gray shows
pattern from ring to dome as the particles are captured the ambient air. The droplets are observed around the
at the interface between the evaporating liquid and the core of the liquid jet due to the progression of the
surrounding air before they get a chance to accumulate atomization process. The distance between the nozzle
at the droplet periphery and dry as a dome or disk after and the substrate varies from 5 to 30 cm with 5 cm
liquid evaporation. increments and the tested injection pressure is 3, 6, 7.5,
and 9 MPa, consistent with the experiments. The
Evaluating the effectiveness of the nanoparticle nozzle geometry selected for the simulations is
spray deposition method consistent with the experiment (diameter of 125 lm
and cone-angle of 6°). For the sake of consistency, a
Droplets created by injecting a high-pressure liquid jet baseline liquid film thickness of 1 lm is set and when
toward a substrate in ambient temperature and pres- this thickness is achieved at any point on the substrate,
sure can evaporate before reaching the substrate, it is assumed that the spray has reached the substrate.
deposit on the substrate, or bounce back from the The simulations are conducted using pure water at
substrate before depositing their nanoparticle content room temperature without considering nanoparticles.
on the substrate. Optimizing the distance between the Since the concentration of nanoparticles in water is
injection orifice and the substrate at different injection very low in experiments the nanoparticles do not
conditions is necessary from two perspectives: (1) to interfere with the spray behavior.
predict and control the evaporation of the droplets to The spray development is visualized in Fig. 10 for
ensure most of the droplets containing nanoparticles the cases where the substrate is located at 10 cm axial
reach the substrate before evaporation. This is impor- distance from the nozzle and injection pressures are 3,
tant because if most of the droplets evaporate before 6, 7.5 and, 9 MPa. As Fig. 10 shows, increasing the
reaching the substrate the nanoparticle content will be injection pressure and increasing the jet momentum
dispersed in the surrounding air and wasted; (2) to results in longer liquid penetration length, and thus the
predict and control the splashing of the droplets that time taken to reach the substrate and forming the
will affect the nanoparticle content and patterns left on liquid film decreases from 4.8 to 2.8 ms by increasing
the substrate. Figure 11 depicts a series of simulations the injection pressure from 3 to 9 MPa. The time
performed to study the behavior of the spray and the required for the spray to reach the substrate is an

123
792 Cellulose (2022) 29:777–798

Pinj= 3 MPa Pinj= 6 MPa

Pinj= 7.5 MPa Pinj= 9 MPa

t= 2 ms 3 ms 5 ms 2 ms 3 ms 5 ms

Fig. 10 Computational simulation of spatio-temporal evolution gray background represents the surrounding ambient air. The
of the spray with different injection pressures at 10 cm distance scale bar is identical in all images
from nozzle. The black color represents the injected water and

important factor in designing experiments and setting a relatively constant rates of evaporation. It is seen that
up the optimum location of the substrate. Repeating for 5 and 10 cm positions, the mass loss due to
the simulations for the cases that the substrate was evaporation of the droplets is in the range of 10–15%
located at 15 and 20 cm with the same injection and 20–25%, respectively. However, the mass loss
pressures of Fig. 10 revealed that at 20 cm, less than reaches 25–30% range for 15 and 20 cm from the
1% of the droplets reached the substrate which did not nozzle. These observations suggest that placing the
result in formation of a film with 1 lm thickness. substrate at a distance lower than 20 cm below the
These simulations suggest that the substrate should be injection orifice minimizes the mass loss due to
placed at an axial distance less than 20 cm from the evaporation for the explored pressure range of 3 to
nozzle to ensure a liquid film of 1 lm thickness is 9 MPa.
formed on the nozzle between 3 ms (for 6, 7.5, and Figure 12 shows the variation in the percentage of
9 MPa) to 4 ms (for 3 MPa) after the start of injection. the water mass deposited on the substrate for different
Microscopy images of the assembled CNC structures injection pressures for substrate positioned at 5, 10,
left on the substrate located at 15 cm are presented in and 15 cm from the orifice. The mass of water
Fig. 6. deposited on a substrate located 5 cm below the
Since droplet evaporation before the spray reaches injection point is * 65% of the total initial mass upon
the substrate is essential in efficiently delivering the injection, whereas for a 10 cm distance, 25–30% of its
nanomaterial to the substrate it is imperative to initial mass is deposited. This suggests that the
calculate the mass of evaporated droplet as a percent- remaining droplets either evaporated or scattered in
age of the total injected mass as shown in Fig. 11 for the surrounding air without reaching the substrate. For
different injection pressures and nozzle-substrate 20 and 30 cm nozzle-substrate distance not shown in
distance. For each injection pressure, increasing the Fig. 12, the simulations predict negligible droplet
distance between the nozzle and the substrate deposition on the substrate implying that most of the
increases the evaporated mass. The plots reach a droplets have evaporated before reaching the sub-
plateau once the spray reaches the substrate indicating strate. At each axial location, increasing the injection

123
Cellulose (2022) 29:777–798 793

Fig. 11 Temporal variation

of the percentage of the mass
of evaporated droplets with
changes in the distance from
the nozzle to the substrate at
different injection pressures

Fig. 12 Temporal variations of the mass of water added to the substrate with the change in injection pressures for 5, 10 and 15 cm
distance between the nozzle and the substrate

pressure results in the delivery of higher portions of nanoparticles. In addition, comparing the experimen-
the initial mass of droplets to the substrate. However, tal measurements of SMD at 7.5 MPa (Figs. 4 and 5)
there is a slight decrease in the deposited mass at demonstrated a more uniform and smaller droplet size
9 MPa injection pressure for the 10 and 15 cm cases. distribution.
This is because 9 MPa injection pressure generates The droplet bounce-back upon contact with the
smaller droplets (as is shown in Fig. 4 and 5) which are substrate is another important factor that affects the
more prone to evaporation before reaching the effectiveness of the nanoparticle deposition on the
substrate. These results suggest that 6 and 7 MPa substrate. Figure 13 represents the water loss due to
injection pressures are more appropriate for optimum splashing from the substrate for axial distances of 5
spray deposition as droplets have large enough and 10 cm. As expected, the percentage of the liquid
momentum to reach the substrate yet the mass of mass bounced back into the surroundings (* 0.5–1%)
evaporated droplets is smaller than the 9 MPa case is higher at 5 cm where the droplets have a higher
before reaching the target substrate. Satisfying these momentum upon interacting with the substrate. The
two conditions ensures more efficient delivery of the mass of bounced droplets at 15 and 20 cm distance are

123
794 Cellulose (2022) 29:777–798

Fig. 13 Temporal variations of the percentage of the mass of bounced droplets from the substrate with injection pressure for 5 and
10 cm distance between the nozzle and the substrate

negligible, hence not depicted here. Increasing the splashing and subsequent deposition. This eventually
injection pressure to 9 MPa results in higher splashing leads to higher deposition of nanoparticles and a larger
at the film interface as the droplets gain a higher film front where the nanoparticles accumulate. It can
momentum upon injection. While positioning the also be seen that there is no deposition for 10 cm case
substrate closer to the nozzle may not seem ideal for a at 2 ms due to larger distance between nozzle and
stable film formation due to increased splashing, the substrate as compared to the 5 cm case. Considering
percentage of droplets that are lost due to the bounce- different behaviors of droplets (evaporation, deposi-
back effect is much lower than the evaporation mass tion, bounce-back) relative to the axial distance of the
loss (1% versus 20%). substrate observed from simulations, it is concluded
Moreover, the higher number of bounced droplets that a location between 10 and 15 cm below the nozzle
will lead to higher film spread. This happens when the and 7.5 MPa injection pressure results in a more
droplets that have already bounced from the substrate effective and uniform nanoparticle deposition.
lose momentum and fall back on the substrate. The
temporal variation of the surface coverage and iso-
scales of the covered area by the spray are depicted Summary and conclusions
Fig. 14. The film area calculated for 5 cm distance
between the nozzle and substrate (colored with a film In this study, we designed and built a novel nanopar-
thickness of 0.01 lm) is larger due to enhanced ticle spray deposition system that utilizes supercritical

Fig. 14 Temporal variation of the surface covered on the injection pressure: a percentage of the substrate surface covered,
substrate by the film formed due to droplet deposition for 5 and b iso-scales depicting droplet deposition of over 0.01 lm film
10 cm distance between the nozzle and the substrate at 7.5 MPa thickness

123
Cellulose (2022) 29:777–798 795

CO2 to assist the atomization process and create visualization of the assembled nanostructure
uniform micron-size CNC-carrier aqueous droplets and show that they either form a ring, disk, or
and deposit them onto the substrate to form tailored dome-shaped architecture. Increasing the mass
nanostructures upon evaporation of water. The effect concentration to droplet size ratio shifts the
of spray parameters on formation of droplets were morphology of assembled nanoparticles from
studied numerically and experimentally. The main ring to dome as the particles are trapped at the
conclusions from this work are summarized below. liquid–air interface before they get a chance to
The 1st, 4th, and 8th conclusion points provide new move towards the edge of the droplet.
insights into the main breakup mechanisms of SAA 6. For each CNC concentration and regardless of
system, pattern of nanostructures left on the substrate the injection pressure, there is a droplet size
upon evaporation of water, and relationship between threshold range above which the assembled
the size of carrier droplets and the pattern of deposited nanostructures exhibit a ring pattern and below
CNCs, respectively. that they exhibit a homogenous distribution. For
concentration of 0.2wt%, the lower and upper
1. Supercritical CO2-assisted atomization sparks
bounds of the threshold are 5.5 and 7.5 lm,
two concurrent mechanisms to boost liquid
respectively while for 2wt% CNC concentration
atomization: reducing the liquid surface tension
these values increase to 9.5 lm and 11.5 lm,
and enhancing CO2 dissolution in water. The
respectively.
combined effect results in the formation of fine
7. The injection pressure on the other hand dictates
droplets with a narrow size distribution that can
the size of the majority of droplets within the
be used as nanoparticle-carrier droplets.
spray plume and can be used to design a system
2. Laser diffraction measurement of SMD shows
where the bulk of droplets fall under one of the
that in general, increasing the injection pressure,
identified nanostructure patterns.
GLR, and axial distance from the injection
8. The size of the carrier droplets strictly influ-
orifice results in the creation of smaller droplets.
ences the evaporation rate of solvent in particle-
In addition, increasing the concentration of
carrier droplets upon deposition on the sub-
CNCs in the injection mixture increases the
strate. The evaporation rate in turn, affects the
overall size of the carrier droplets.
prevalence of convective to diffusive transport
3. However, increasing the injection pressure
of particles that is represented by Peclet number.
above the critical pressure of the CO2-water
9. The evaporation time is prolonged by increasing
mixture (i.e., 7.5 MPa) and increasing the GLR
the droplet size for each concentration, which
do not noticeably decrease the droplet sizes.
results in Pe [ 1 that indicates the higher rate of
This means there is no need for extremely high
convective transport of particles to diffusive
pressures or excessive amount of assisting gas to
transport leading to accumulation of particles
create micron-sized droplets.
along the periphery of the droplet and formation
4. Microscopic visualization of the assembled
of a ring-shaped structure. At Pe \ 1 where the
nanoparticles on the substrates illustrates that
diffusive movement of CNCs is dominant, a
morphology of nanostructures falls into three
dome-shaped structure is formed for all tested
main categories: (1) ring-shape pattern, where
concentrations. At Pe * 1, droplets fall in the
the majority of nanoparticles accumulate along
transitional region where both ring and dome-
the edge of the evaporating droplet, (2) homoge-
shaped structures are observed.
nous distribution or disk pattern, where particles
10. Computational simulations show that consider-
scatter more uniformly across the surface area
ing different droplet behavior interacting with
of the evaporating droplet, and (3) transition
the substrate (evaporation, deposition, bounce-
stage, where there is still a distinct ring-shape
back) versus the injection pressure and axial
structure yet some particles are scattered within
distance between the substrate and nozzle,
the edges of the evaporating droplet.
positioning the substrate at 10 or 15 cm below
5. The profilometry height measurements com-
the nozzle and 7.5 MPa injection pressure result
bined with micrographs provide a 3D

123
796 Cellulose (2022) 29:777–798

in a more effective and uniform nanoparticle sprays using the Malvern SpraytecÒ. J Pharm Sci
deposition. 93:1725–1742
Deegan RD, Bakajin O, Dupont TF, Huber G, Nagel SR, Witten
TA (1997) Capillary flow as the cause of ring stains from
Data availability All the data and materials supporting the dried liquid drops. Nature 389:827
claims herein are included in the manuscript and comply with Della Porta G, De Vittori C, Reverchon E (2005) Supercritical
the field standards. assisted atomization: a novel technology for microparticles
preparation of an asthma-controlling drug. AAPS
Declarations PharmSciTech 6(E421):E428
Diamond L, Akinfiev N (2003) Solubility of CO2 in water from
Conflict of interest The authors declare that they no conflict -1.5 to 100 °C and from 0.1 to 100 MPa: evaluation of
of interest. literature data and thermodynamic modelling. Fluid Phase
Equilib 208:265–290
Dicuangco M, Dash S, Weibel JA, Garimella SV (2014) Effect
of superhydrophobic surface morphology on evaporative
deposition patterns. Appl Phys Lett 104:201604
References Dinh NT, Sowade E, Blaudeck T, Hermann S, Rodriguez RD,
Zahn DRT, Schulz SE, Baumann RR, Kanoun O (2016)
Aguiar-Ricardo A (2017) Building dry powder formulations High-resolution inkjet printing of conductive carbon nan-
using supercritical CO2 spray drying. Current Opini Green otube twin lines utilizing evaporation-driven self-assem-
Sustain Chem 5:12–16 bly. Carbon 96:382–393
Andrews SW, Nover DM, Reuter JE, Schladow SG (2011) Fauchais P, Vardelle M, Vardelle A, Goutier S (2015) What do
Limitations of laser diffraction for measuring fine particles we know, what are the current limitations of suspension
in oligotrophic systems: pitfalls and potential solutions. plasma spraying? J Therm Spray Technol 24:1120–1129
Water Resourc Res. https://doi.org/10.1029/ Fauchais PL, Joachim VR, Heberlein, Maher IB (2014) ’Wire
2010WR009841 arc spraying’ In: Thermal Spray Fundamentals. Springer,
Anyfantakis M, Geng Z, Morel M, Rudiuk S, Baigl D (2015) Berlin
Modulation of the coffee-ring effect in particle/surfactant Goth C, Sonja P, Joerg F (2011) ‘‘Aerosol Jet printing on rapid
mixtures: the importance of particle–interface interactions. prototyping materials for fine pitch electronic applica-
Langmuir 31:4113–4120 tions.’’ In 2011 IEEE 61st Electronic components and
Azarova NA, Owen JW, McLellan CA, Grimminger MA, technology conference (ECTC), 1211–16. IEEE
Chapman EK, Anthony JE, Jurchescu OD (2010) Fabri- Govor LV, Reiter G, Bauer GH, Parisi J (2004) Nanoparticle
cation of organic thin-film transistors by spray-deposition ring formation in evaporating micron-size droplets. Appl
for low-cost, large-area electronics. Org Electron Phys Lett 84:4774–4776
11:1960–1965 Grishkewich N, Mohammed N, Tang J, Tam KC (2017) Recent
Bachu S, Brant Bennion D (2009) Interfacial tension between advances in the application of cellulose nanocrystals.
CO2, freshwater, and brine in the range of pressure from (2 Current Opinion in Colloid Interface Science 29:32–45
to 27) MPa, temperature from (20 to 125) C, and water Gu M, Jiang C, Liu D, Prempeh N, Smalyukh II (2016) Cellu-
salinity from (0 to 334 000) mg L- 1. J Chem Eng Data lose nanocrystal/poly (ethylene glycol) composite as an
54:765–775 iridescent coating on polymer substrates: structure-color
Barbezat G (2005) Advanced thermal spray technology and and interface adhesion. ACS Applied Materials Interfaces
coating for lightweight engine blocks for the automotive 8:32565–32573
industry. Surface Coatings Tech 200:1990–1993 Hu H, Larson RG (2002) Evaporation of a sessile droplet on a
Bouse LF, Kirk IW, Bode LE (1990) Effect of spray mixture on substrate. J Phys Chem B 106:1334–1344
droplet size. Trans ASAE 33:783–0788 Hu H, Larson RG (2005) Analysis of the microfluid flow in an
Bugakova D, Slabov V, Sergeeva E, Zhukov M, Vinogradov AV evaporating sessile droplet. Langmuir 21:3963–3971
(2019) Comprehensive characterization of TiO2 inks and Huang S, Vignolles M-L, Chen XD, Le Loir Y, Jan G, Schuck P,
their application for inkjet printing of microstructures. Jeantet R (2017) Spray drying of probiotics and other food-
Colloids Surf a: Physicochem Eng Aspects 586:124146 grade bacteria: A review. Trends in Food Sci Tech 63:1–17
Calvert P (2001) Inkjet printing for materials and devices. Chem Jabari E, Ehsan %J Carbon Toyserkani. (2015) Micro-scale
Mater 13:3299–3305 aerosol-jet printing of graphene interconnects. Carbon
Costa C, Casimiro T, Aguiar-Ricardo A (2018) Optimization of 91:321–329
supercritical CO2-assisted atomization: phase behavior and Jarrahbashi D, Sirignano WA (2014) Vorticity dynamics for
design of experiments. J Chem Eng Data 63:885–896 transient high-pressure liquid injection. Phys Fluids
Cui L, Zhang J, Zhang X, Li Y, Wang Z, Gao H, Wang T, Zhu S, 26:101304
Hailing Yu, Yang B (2012) Avoiding coffee ring structure Jarrahbashi D, Sirignano WA, Popov PP, Hussain F (2016)
based on hydrophobic silicon pillar arrays during single- Early spray development at high gas density: hole, liga-
drop evaporation. Soft Matter 8:10448–10456 ment and bridge formations. J Fluid Mech 792:186–231
Dayal P, Shaik MS, Singh M (2004) Evaluation of different Jarrahbashi D, Kim S, Genzale CL (2017a) Simulation of
parameters that affect droplet-size distribution from nasal combustion recession after end-of-injection at diesel

123
Cellulose (2022) 29:777–798 797

engine conditions. J Eng Gas Turbines Power 139:102804- Palaganas NB, Mangadlao JD, de Leon AC, Palaganas JO,
04–102808 Pangilinan KD, Lee YJ, Advincula RC (2017) 3D printing
Jarrahbashi D, Kim S, Knox BW, Genzale CL (2017b) Com- of photocurable cellulose nanocrystal composite for fab-
putational analysis of end-of-injection transients and rication of complex architectures via stereolithography.
combustion recession. Int J Engine Res 18:1088–1110 ACS Appl Mater Interf 9:34314–34324
Ke D, Vu AA, Bandyopadhyay A, Bose S (2019) Composi- Parker RM, Frka-Petesic B, Guidetti G, Kamita G, Consani G,
tionally graded doped hydroxyapatite coating on titanium Abell C, Vignolini S (2016) Hierarchical self-assembly of
using laser and plasma spray deposition for bone implants. cellulose nanocrystals in a confined geometry. ACS Nano
Acta Biomater 84:414–423 10:8443–8449
Krebs FC (2009) Fabrication and processing of polymer solar Parker RM, Guidetti G, Williams CA, Zhao T, Narkevicius A,
cells: a review of printing and coating techniques. Solar Vignolini S, Frka-Petesic B (2018) The self-assembly of
Energy Mater Solar Cells 93:394–412 cellulose nanocrystals: hierarchical design of visual
Kuroda S, Kawakita J, Watanabe M, Kim KH, Molak R, appearance. Adv Mater 30:1704477
Katanoda H (2015) ’Current status and future prospects of Paulsen JA, Michael R, Kurt C, Richard P (2012) ‘‘Printing
warm spray technology.’ In: Future development of ther- conformal electronics on 3D structures with Aerosol Jet
mal spray coatings. Elsevier, Armsterdam technology.’’ In 2012 Future of Instrumentation Interna-
Kuznetsov IA, Greenfield MJ, Mehta YU, Merchan-Merchan tional Workshop (FIIW) Proceedings, 1–4. IEEE
W, Salkar G, Saveliev AV (2011) Increasing the solar cell Pawlowski L (2008) The science and engineering of thermal
power output by coating with transition metal-oxide spray coatings. Wiley, New Jersey
nanorods. Appl Energy 88:4218–4221 Phan HT, Caney N, Marty P, Colasson S, Gavillet J (2009)
Le Hue P (1998) Progress and trends in ink-jet printing tech- Surface wettability control by nanocoating: the effects on
nology. J Imaging Sci Tech 42:49–62 pool boiling heat transfer and nucleation mechanism. Int J
Lu G, Han H, Duan Y, Sun Y (2013) Wetting kinetics of water Heat Mass Transf 52:5459–5471
nano-droplet containing non-surfactant nanoparticles: a Ranz WE, Marshall WR (1952) Evaporation from drops. Chem
molecular dynamics study. Appl Phys Lett 103:253104 Eng Prog 48:141–146
Lu S, Zheng J, Cardenas JA, Williams NX, Lin Y-C, Franklin Reitz RD (1987) Modeling atomization processes in high-
AD (2020) Uniform and stable aerosol jet printing of car- pressure vaporizing sprays. Atomizat Spray Tech
bon nanotube thin-film transistors by ink temperature 3:309–337
control. ACS Appl Mater Interf 12:43083–43089 Reverchon E (2002) Supercritical-assisted atomization to pro-
Mahajan A, Daniel Frisbie C, Francis LF (2013) Optimization of duce micro-and/or nanoparticles of controlled size and
aerosol jet printing for high-resolution, high-aspect ratio distribution. Ind Eng Chem Res 41:2405–2411
silver lines. ACS Appl Mater Interf 5:4856–4864 Reverchon, Ernesto. 2007. ‘‘Process for the production of micro
Mampallil D, Eral HB (2018) A review on suppression and and/or nano particles.’’ In.: Google Patents.
utilization of the coffee-ring effect. Adv Colloid Interface Rie V, Jonas CS, Gençer A, Lombardo S, Gasser U, Kumar S,
Sci 252:38–54 Salazar-Alvarez G, Kang K, Thielemans W (2019) Ani-
Mampallil D, Reboud J, Wilson R, Wylie D, Klug DR, Cooper sotropic diffusion and phase behavior of cellulose
JM (2015) Acoustic suppression of the coffee-ring effect. nanocrystal suspensions. Langmuir 35:2289–2302
Soft Matter 11:7207–7213 Roman, Maren, Shuping Dong, Anjali Hirani, and Yong Woo
Mariano M, El Kissi N, Dufresne A (2014) Cellulose Lee. 2009. ’Cellulose nanocrystals for drug delivery.’ in
nanocrystals and related nanocomposites: review of some (ACS Publications).
properties and challenges. J Polym Sci, Part b: Polym Phys Sarobol P, Cook A, Clem PG, Keicher D, Hirschfeld D, Hall
52:791–806 AC, Bell NS (2016) Additive manufacturing of hybrid
Markt DP, Roberto T, Ashish P, Mehdi R, Sibendu S, Riccardo circuits. Annu Rev Mater Res 46:41–62
S, Seong-Young L, Jeffrey N (2018) ‘‘Using a DNS Secor EB (2018) Principles of aerosol jet printing. Flex Print
framework to test a splashed mass sub-model for Lagran- Electron 3:35002
gian spray simulations.’’ In: SAE Technical Paper Seifert T, Sowade E, Roscher F, Wiemer M, Gessner T, Bau-
Mette A, Richter PL, Hörteis M, Glunz SW (2007) Metal aerosol mann RR (2015) Additive manufacturing technologies
jet printing for solar cell metallization. Prog Photovolt: Res compared: morphology of deposits of silver ink using
Appl 15:621–627 inkjet and aerosol jet printing. Ind Eng Chem Res
Mishra R, Christopher R (2019) ‘‘Evaluating surface film 54:769–779
models for multi-dimensional modeling of spray-wall Shariati A, Peters CJ (2003) Recent developments in particle
interaction.’’ In.: SAE Technical Paper design using supercritical fluids. Current Opin Solid State
Moon RJ, Martini A, Nairn J, Simonsen J, Youngblood J (2011) Mater Sci 7:371–383
Cellulose nanomaterials review: structure, properties and Shariatnia S, Akshaj V, Suleiman O, Dorrin J, Amir A (2019)
nanocomposites. Chem Soc Rev 40:3941–3994 Atomization of cellulose nanocrystals aqueous suspensions
Moridi A, Hassani-Gangaraj SM, Guagliano M, Dao M (2014) in fused deposition modeling: A scalable technique to
Cold spray coating: review of material systems and future improve the strength of 3D printed polymers. Composit
perspectives. Surf Eng 30:369–395 Part B Eng 177:107291
Nguyen TAH, Nguyen AV (2012) Increased evaporation Shariatnia S, Annuatha VK, Ozge K, Amir A (2020) Hybrid
kinetics of sessile droplets by using nanoparticles. Lang- cellulose nanocrystals-bonded carbon nanotubes/carbon
muir 28:16725–16728

123
798 Cellulose (2022) 29:777–798

fiber polymer composites for structural applications. ACS Wei Y (2015) ’Effect of particles on evaporation of droplet
Appl Nano Mater 3(6):5421–5436 containing particles’
Shariatnia S, Amir A, Dorrin J (2021) Experimental analysis of Weibull W (1951) A statistical distribution function of wide
supercritical-assisted atomization. PhysFluId 33:013314 applicability. J Appl Mech 18(293):297
Singh A, den Guy VM (2016) Spray drying formulation of Weller HG, Tabor G, Jasak H, Fureby C (1998) A tensorial
amorphous solid dispersions. Adv Drug Deliver Review approach to computational continuum mechanics using
100:27–50 object-oriented techniques. Comput Phys 12:620–631
Stanton, Donald W, and Christopher J Rutland. 1996. ’Modeling Wilkinson NJ, Smith MAA, Kay RW, Harris RA (2019) A
fuel film formation and wall interaction in diesel engines’, review of aerosol jet printing—a non-traditional hybrid
SAE transactions: 808–24. process for micro-manufacturing. Int J Adv Manuf Tech
Stanton D, Rutland C (1996a) Modeling fuel film formation and 105:4599–4619
wall interaction in diesel engines. SAE Technical Paper. Yunker PJ, Still T, Lohr MA, Yodh AG (2011) Suppression of
https://doi.org/10.4271/960628 the coffee-ring effect by shape-dependent capillary inter-
Tang J, Sisler J, Grishkewich N, Tam KC (2017) Functional- actions. Nature 476:308–311
ization of cellulose nanocrystals for advanced applications. Zhao Da, Liu T, Park JG, Zhang M, Chen J-M, Wang B (2012)
J Coll Interf Sci 494:397–409 Conductivity enhancement of aerosol-jet printed elec-
Tewes F, Boury F (2005) Formation and rheological properties tronics by using silver nanoparticles ink with carbon nan-
of the supercritical CO2- water pure interface. J Phys otubes. Microelectron Eng 96:71–75
Chem B 109:3990–3997 Zuo B, Gomes AM, Rutland CJ (2000) Studies of superheated
Tom JW, Debenedetti PG (1991) Particle formation with fuel spray structures and vaporization in GDI engines. Int J
supercritical fluids—a review. J Aerosol Sci 22:555–584 Engine Res 1(4):321–336
Tsang C-W, Trujillo MF, Rutland CJ (2014) Large-eddy sim-
ulation of shear flows and high-speed vaporizing liquid fuel
Publisher’s Note Springer Nature remains neutral with
sprays. Comput Fluids 105:262–279
regard to jurisdictional claims in published maps and
Vehring R (2008) Pharmaceutical particle engineering via spray
institutional affiliations.
drying. Pharm Res 25:999–1022
Weast RC, MJ CRC, Astle, Boca Raton, Fla. 1981. ’CRC
Handbook of chemistry and physics 62nd ed’

123