Journal of Colloid and Interface Science 600 (2021) 1–13

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

journal homepage: www.elsevier .com/locate / jc is

Regular Article

Precipitation dynamics of surrogate respiratory sessile droplets leadingto possible fomites

https://doi.org/10.1016/j.jcis.2021.04.1280021-9797/� 2021 Elsevier Inc. All rights reserved.

⇑ Corresponding author at: Department of Mechanical Engineering, Indian Institute of Science, Bengaluru, KA 560012, India.E-mail address: sbasu@iisc.ac.in (S. Basu).

Abdur Rasheed a, Shubham Sharma a, Prasenjit Kabi b, Abhishek Saha c, Swetaprovo Chaudhuri d,Saptarshi Basu a,b,⇑aDepartment of Mechanical Engineering, Indian Institute of Science, Bengaluru, KA 560012, Indiab Interdisciplinary Centre for Energy Research, Indian Institute of Science, Bengaluru, KA 560012, IndiacDepartment of Mechanical and Aerospace Engineering, University of California San Diego, La Jolla, CA 92093, USAd Institute for Aerospace Studies, University of Toronto, Toronto, Ontario M3H 5T6, Canada

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

a r t i c l e i n f o

Article history:Received 28 February 2021Revised 23 April 2021Accepted 24 April 2021Available online 28 April 2021

Keywords:FomitesRespiratory dropletDroplet evaporationParticle depositionProtein-salt solutionCrystallizationVirus-emulating particle

a b s t r a c t

Hypothesis: The droplets ejected from an infected host during expiratory events can get deposited asfomites on everyday use surfaces. Recognizing that these fomites can be a secondary route for diseasetransmission, exploring the deposition pattern of such sessile respiratory droplets on daily-use substratesthus becomes crucial.Experiments: The used surrogate respiratory fluid is composed of a water-based salt-protein solution, andits precipitation dynamics is studied on four different substrates (glass, ceramic, steel, and PET). For track-ing the final deposition of viruses in these droplets, 100 nm virus emulating particles (VEP) are used andtheir distribution in dried-out patterns is identified using fluorescence and SEM imaging techniques.Findings: The final precipitation pattern and VEP deposition strongly depend on the interfacial transportprocesses, edge evaporation, and crystallization dynamics. A constant contact radius mode of evaporationwith a mixture of capillary and Marangoni flows results in spatio-temporally varying edge deposits.Dendritic and cruciform-shaped crystals are majorly seen in all substrates except on steel, where regularcubical crystals are formed. The VEP deposition is higher near the three-phase contact line and crystalsurfaces. The results showed the role of interfacial processes in determining the initiation of fomite-type infection pathways in the context of COVID-19.

� 2021 Elsevier Inc. All rights reserved.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

1. Introduction

The ongoing COVID-19 pandemic has disrupted global travel,healthcare systems, social interactions, and business activitiesworldwide. Primary transmission of the virus occurs at the micro-scale level, where respiratory droplets rapidly spread the SARS-CoV-2 amongst human beings [1,2]. To arrest the transmission ofthe virus, wearing a facemask and maintaining social distanceshas been advised by the scientific and medical community world-wide [3–5]. The ejected droplets are in size range of 1–2000 lm [6]and creates two possible scenarios of infection. Smaller dropletscan evaporate, precipitate [7], travel far [8], and stay airborne fora sufficiently long time before being directly inspired by anotherhealthy human being [9]. On the other hand, the larger dropletsmay settle under gravity or impinge on a material surface, formingfomites [10,11]. In either scenario, infection mechanics, whichinvolve virus survivability [12], remains elusive. In this article,we shall limit our discussion to the physicochemical transforma-tions within a VEP (virus emulating particles) [3,7,13] loaded sur-rogate respiratory droplet drying on different commonly availablereal life surfaces.

The pattern formation during the evaporation of suspendedparticles in a sessile droplet was first explained by Deegan et al.[14]. The capillary flow inside the droplet led to the deposition ofthe solute particles near the pinned contact line. The pinning ofthe contact line was attributed to the inherent roughness of thesubstrate. Ring deposition morphs to dome shape for moderateto high particle concentration (5–20 mg=mL) mainly due to inter-particle interactions [15]. The pattern was independent ofhydrophobicity at lower particle concentrations (<5 mg=mL) forwhich ring deposition was observed [15]. Further, Shmuylovichet al. [16] and Maheswari et al. [17] observed the formation ofconcentric-rings pattern during pinning and de-pinning of the con-tact line, while a single monolith pattern was observed by Baldwinet al. [18]. The thick-edge deposit near the contact line can bealtered by surface patterning [19–22], choice of particles shape[23], solvent-vapor-substrate interaction [24], convection currents[25], surface roughness and addition of other soluble components[26]. The flow patterns inside the evaporating droplet influencethe final deposition of the particles. The evaporation flux in a ses-sile droplet was triggered by diffusion of solvent at the liquid–airinterface, and it diverged near the contact line of a pinned droplet[27,28]. The divergent evaporative-flux and pinning of the dropletcontact line led to the capillary flow towards the droplet contactline [14,29].

In contrast to the capillary flow, Marangoni flow is observeddue to surface tension changes at the liquid–air interface. This sur-face tension difference can arise due to temperature changes, i.e.,thermocapillary flow [30] or solute concentration changes, i.e.,solutal Marangoni flow [31–33]. The thermo-capillary flow wascreated either by the temperature difference between the liquid–air interface and underlying substrate [34] or due to the non-uniform cooling (due to non-uniform evaporation) along with theliquid–air interface [35,36]. Barmi et al. [37] numerically modeledthe evaporation rate and internal circulation in a thermo-capillaryflow of a sessile droplet with a pinned contact line. The thermalgradient at the air–liquid interface was shown to be the drivingforce for the flow, and the flow magnitude decreases during thelater stages of evaporation due to the reduction in a thermal gradi-ent. Solutal Marangoni flows were prompted by the surface tensiondifference between the edge and the apex of the evaporating dro-plet, along the liquid–air interface [38,31]. The surface tension dif-ference arises due to non-uniform evaporation at the liquid–airinterface, which increases salt concentration near the droplet con-tact line. The Marangoni-induced flow can be directed to or awayfrom the contact line depending on the solution constituents

2

[31,32]. The Marangoni circulation was generally observed duringthe evaporation of sessile saline-droplets [39,37,40–43]. Shahidza-deh et al. [41] found that the deposition stains of saline dropletswere different from colloidal droplets. The formation of salt crystalstains depended on wetting properties, nucleation pathways, andgrowth rates. On a super-hydrophobic substrate, Shin et al. [40]showed the formation of ‘‘ring” or ‘‘igloo” shaped salt depositionsdepending on the initial salt concentrations and evaporation rates.The resulting precipitate shapes depended on the combined effectof solubility, evaporation rate, and contact line hysteresis. The flowpattern inside the evaporating salt droplet was experimentallyshown by Kuznetsov et al. [44] and Efstratiou et al. [43]. Efstratiouet al. [43] showed a uniform outward flow towards the contact lineduring most of the evaporation stage. However, during the crystal-lization stage, a jet flow towards the nucleation site was observed.A detailed overview of a drying droplet on a substrate was pro-vided by Larson et al. [45] and Parsa et al. [46].

Evaporation of bio-fluid droplets has been applied as a medicaldiagnosis tool in earlier works [47–49]. Devineau et al. [50] ana-lyzed the pattern formation of evaporating protein droplets con-taining suspended polystyrene particles. The suppression of thecoffee ring effect due to protein adsorption on the surface of theparticles was observed. In another study [51], the formation ofedge deposition or uniform deposition of particles was shown tobe dependent on the protein charge. The bio-fluid compounds usu-ally contain dissolved salt in them, and the deposition patterns ofthese salt-protein biofluid droplets were reported [52–54]. The for-mation of dendritic patterns was shown by Darwich et al. [55] for abi-component solution of oppositely charged alginate polysaccha-rides and gold nanoparticles. The dendritic structure formationwas dependent on the salt concentration, drying mode, and parti-cle size. Choudhury et al. [56] studied the drying of colloidal gelcontaining dissolved salt and showed that the final depositionmorphology depended on the type of host gel, salt concentration,and substrate used. Carreon et al. [57] discussed the effectivenessof first-order statistics (FOS) and grey level-co-occurrence matrix(GLCM) methods to characterize the protein deposition patternsand their complex texture. Pathak et al. [54] studied the precipita-tion dynamics of two components (salt and proteins) bio-fluids atdifferent constituent compositions. The final deposition patterns(either crystalline or dendritic) were distinctive and dependenton the initial ratio of the two components.

The above discussion provides an overview of the evaporationof sessile droplets containing suspended particles, salts, and bio-logical compounds. However, respiratory fluid is complex anddemands a more critical approach to study. Sefiane et al. [58]investigated the pattern formation from drying droplets of variousfluids, including biofluid. Researchers have shown that complexbio-fluid evaporation involving blood [59] and the synovial fluid[60] agrees well with observations and theoretical models of sim-ple fluid droplets. Mukhopadhyay et al. [61] showed recently dif-ferent pattern formation driven by interfacial energy on dryingblood droplets. Vejerano et al. [62] have presented a detailedexperimental exposition of respiratory droplets (surrogate andreal), drying on a surface, and their impact on virus survivability.However, several questions remain. What is the role of the sub-strate beyond wettability and thermal conductivity? What doesthe flow inside a droplet of complex fluid look like? If virus parti-cles are dispersed inside such a droplet, how do they aggregate inthe final residue? We shall discuss experimental findings that arenot intuitively apparent and present a heuristic understanding inthis article.

The details of the materials and methods used in the presentwork are given in Section 2. Section 3 provides the results and dis-cussion of this paper. Here we discuss the variation in evaporationlifetime of a surrogate droplet resting on these substrates with dif-

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

ferent thermal conductivity and wettability. The internal flowdynamics are discussed next. We establish the link between theonset of crystallization and the dynamics of evaporation and flow.We discuss the qualitative as well as quantitative features of thecrystallization pattern. Further, we explain how the crystallizationpattern distributes the VEP within the precipitate. Finally, the sum-mary and conclusions from the present work are shown inSection 4.

2. Materials and Methods

The stringent requirements of bio-safety during COVID-19 andour focus on the fluid mechanics of virus transmission necessitateusing appropriate surrogate materials for the experiment. The res-piratory fluid consists of dissolved salt ions and alveolar surfac-tants [63] with a typical pH value of 6 or greater. We use thesame artificial respiratory fluid composition as reported by Vejer-ano et al. [62]. Similar to the other works in recent literature[3,7,13], 100 nm polystyrene nanoparticles were used as a modelmaterial for the COVID-19 virus based on similarity in geometricalproperties, i.e. spherical shape and size (although they differ in bio-logical and chemical properties). Such nanoparticles can predictthe hydrodynamics of COVID-19 viruses (which are nonmotile[64]) during droplet evaporation which majorly defines the finaldeposition pattern. However, it must not be generalised that geo-metrically similar particles can be used for emulating hydrody-namics of all types of pathogens (e.g. bacteria). Earlier studieshad shown a considerable difference in the deposition patternbetween inert microparticles and motile pathogens [65–67]. Thesestudies have used bacteria with self-propelling motor mechanismswhich can alter the final deposition patterns. In contrast, virionslike COVID-19 don’t have self-motile mechanisms [64]. They travelthrough the host organism supply chain or from one cell to theneighbouring cell. Therefore, hydrodynamics of non-motile virusescan be predicted using emulating nanoparticles as done in the pre-sent work. At the same time, the differences in surface chemistryand interfacial properties of virus emulating particles (VEP) andactual viruses might still lead to some alterations in the final depo-sition; however, these analyses are beyond the scope of the presentwork.

2.1. Sample preparation

The surrogate model of the respiratory fluid was created from0.9 % by wt. of NaCl, 0.3% by wt. of gastric mucin (Type III, SigmaAldrich), and 0.05 % wt. of di-palmitoyl-phosphatidyl-choline(DPPC (Avanti Polar Lipids)) in deionized water [62]. The final for-mulation was sonicated for 15 min and centrifuged at 5000 rpm for15 min [68]. The nanoparticle properties which make them a suit-able candidate for VEPs are as follows. (1) The nanoparticles shouldbe geometrically similar to the COVID-19 virus in shape (spherical)and size (�100 nm [69]). Figure S1 shows the size and shape of thenanoparticles used in the study. (2) Nanoparticles should be chem-ically and biologically inert. The nanoparticles used in presentwork (R100, Thermofisher) are generally used as tracers for lPIVmeasurements; as such these particles are both inert and chargestabilized against sedimentation and agglomeration. (3) Thenanoparticles should exhibit fluorescence when excited by thegreen light generally used with microscope. This help in identifyingthe distribution pattern of particles in dried precipitate. Based onthe above criterion, 100 nm polystyrene nanoparticles (R100, Ther-mofisher) are used as a viable candidate at an initial viral load [70]of � 106 particles/mL of the surrogate liquid. At very low particleconcentration as used in the current study (� 10�3 mg/mL), theparticles cannot be influenced by DLVO forces. However, the

3

non-DLVO forces like solvation and hydration forces could play adominant role in particle dispersion stability and distribution.The nanoparticles used in present work are slightly hydrophobicand readily adsorbs proteins through non-DLVO hydrophobicinteractions (as per the specification sheet of particles and litera-ture [71,72]), which in turn prevents particle agglomeration. Fur-ther, we confirmed the nanoparticles to be stable for theduration of the experiment.

2.2. Evaporation experiments

Common surfaces encountered in daily life such as glass (micro-scope slides, Bluestar), steel (surgical scalpel blade), plastic (com-mercial Polyethylene Terephthalate (PET)), and ceramic (commonfloor tile) are used as substrates. Droplets of 0.46 ll �10 % volumewere gently dispensed on the substrates and allowed to evaporateat an ambient temperature of 25� 1 oC and relative humidity46� 1%. Backlit images (using Karl Storz LED source) of the dryingdroplet side profile were imaged every 0.5 s using a Nikon D7200camera fitted with a Navitar 6.5x zoom lens (spatial resolution2.5 lm/pixel) till the end of evaporation. The captured imageswere binarized, and the ‘‘Analyze Particles” plugin of ImageJ soft-ware (open source) was used to calculate the major axis (contactdiameter; 2a) and minor axis (twice of the droplet apex height;h) of the droplet as it evaporates. The contact angle h ¼ 2tan�1ðh=aÞ 180=p was extracted from the images.

The drying droplet top-view was imaged at 2.19 fps using ahigh-resolution CCD camera (pco2000) mounted on a BX51 Olym-pus frame. Halogen-based illumination (TH4 200, Olympus) andmicroscope objectives (10x and 20x) were used here.

For top-view fluorescent imaging of the drying droplets, lightfrom a halogen lamp was replaced by a mercury lamp (URFL-T,Olympus) using inbuilt prism assembly inside the microscope. Adichroic arrangement was used to excite the nanoparticles at532 nm, and the pco2000 camera acquired the resulting emissionat 590 nm.

2.3. Optical micro-characterization

The samemicroscope arrangement described above was used tocharacterize the microstructure of the deposits. 100x objective wasused for both bright-field and fluorescent imaging. ImageJ software(open source, JAVA platform) was used for image analysis. Detailsare provided in appropriate sections.

2.4. Flow measurement

Micro-Particle Image Velocimetry (lPIV) measurements of thedrying droplet on a glass and PET substrates show the flow patterninside the droplet. Fluorescent particles with a mean diameter of860 nm and a density of 1.05 g/cm3 are seeded into the surrogatesolution. A 532 nm Nd: Yag laser illuminates the sessile dropletfrom the bottom, necessitating the use of transparent surfaces suchas glass and PET. Therefore, flow measurement results of ceramicand steel could not be obtained. Fluorescence emission from theneutrally buoyant tracer particles is directed to an Imager Intensecamera (Lavision) connected to a motorized microscope (Flowmas-ter MITAS, Lavision). A 10x objective is used, which provides a fieldof view of 600 lm � 450 lm, depth of field of 8 lm, and pixel res-olution of 1.72 lm/pixel. The PIV is done at different horizontalplanes ðz=h0 � 0;0:5Þ. Here ’z’ is the plane distance measured fromthe bottom, and ’h0’ is the droplet initial apex height. Single frameimages are taken every 1 s, which corresponds to a particle move-ment of 4–5 pixels between two consecutive frames. The imagesare captured and processed using commercially available Davis

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

8.4 software. The captured images are pre-processed for intensitynormalization. A multi-pass vector cross-correlation from a win-dow size of 128 � 128 pixels to 64 � 64 pixels (with a 50% overlapbetween them) is done between two consecutive images. Theobtained vector field is then post-processed, where de-noising ofvector-field is done.

A 520 nm continuous laser sheet (Schafter + Kirchoff) is used toilluminate the sessile droplet for side view flow visualization. The860 nm fluorescent nanoparticles are used as tracers, and imagesare captured for an exposure time of five seconds.

3. Results and discussion

3.1. Effect of wettability and thermal conductivity on evaporation

The maximum contact diameter of the surrogate respiratorydroplets on all substrates is less than the capillary length,lc ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðrw=ðqwgÞÞp � 2:7 mm, which allows the volume estimation

using the expression, V ¼ ðphð3a2 þ h2ÞÞ=6 for a spherical cap. Hererw is the air–liquid surface tension, qw is the water density, and g =9.81 m/s2. The three-phase contact line (TPCL) of the dropletdeposited on different substrates is pinned throughout the dropletlifetime duration, as shown in Fig. 1a. Consequently, the contactangle must reduce to account for the loss of height due to evapora-tion. The volume of the droplet is calculated at various instancesand shown in Fig. 1b. The thick deposit near the edge blocks theside profile view at an advanced stage of drying, making it difficultto track the droplet evolution. The final lifetime (tf ) is betterobserved from the top-view, and the results are shown in Fig. 1cfor different substrates.

Conduction of heat (under steady-state conditions) across thedroplet’s bulk is expressed as

_q ¼ kwADTh

ð1Þ

Where kw = 0.6 W/m-K is the thermal conductivity of water, A and hare the contact area ðpa2Þ and the droplet height. Where a is thecontact radius of the droplet. The droplet will compensate for thethermal loss by conducting from the substrate, ksubADT=d. However,thermal compensation is hindered by substrates with low thermal

Fig. 1. Dynamics of droplet evaporation. (a) Sequential image showing the side-view ofðV=VoÞ vs. time (t). (c) Variation of the onset of crystallization within the droplet (tc) and

4

conductivity [73], causing evaporative cooling in droplets, whichlowers the mass-loss rate. Bazargan et al. [74] define a critical con-tact radius ðrcÞ, below which evaporative cooling is significant. Thecritical radius is defined as

rc � kwdksubtanðh=2Þ ð2Þ

Where ksub; h, and d are the respective thermal conductivity, contactangle, and thickness of the different substrates. The initial value ofa < rc for all substrates, except steel, leading to evaporative cooling.Sefiane et al. [58] proposed the droplet evaporation rate as:

_mSB � _m

½1þ f ðhÞabð _mhfgh2pakwÞð1þ b kw

ksubh�

ð3Þ

Where, _m � paDwCsð1� RHÞf ðhÞ. Dw ¼ 2:54� 10�5 m2=s is the dif-fusion coefficient of water vapor into the air,Cs ¼ 2:3� 10�2kg=m3 is the saturation concentration of water vaporat 298 K. The polynomial [29] f ðhÞ ¼ 0:27hþ 1:3. The typical contactangles on glass, ceramic, PET, and steel are 30�, 37�, 48�, and 70�,respectively. Where a ¼ hmsinh

hh ;b ¼ @Cs@T =Cs � 0:063 for water,

hfg ¼ 2:44� 106 J/kg is the latent heat of vaporization of waterand b ¼ dh=hm . Using the value of ksub as 0.8 W/m-K, 0.12 W/m-K,and 50 W/m-K for glass, PET, and steel, the droplet’s evaporationtimes on these substrates are shown in Fig. 1c. The assumption ofconstant mass-loss rate while estimating the total evaporation timeand not considering the change in water properties due to additionof the solutes leads to the observed error. However, the correcttrend is established.

3.2. Evaporation and flow

When exposed to an ambient environment, a sessile dropletundergoes evaporation, and the flux of solvent has a spatial varia-tion across the droplet surface and is expressed as [28,29],

Jðx; tÞ ¼ _mð1� gðhÞÞpa2

½1� ðxaÞ2�ð�kðhÞÞ

ð4Þ

Where x is any location within the droplet,

gðhÞ ¼ 0:2239ðh� p=4Þ2 þ 0:3619 and kðhÞ ¼ 1=2� h=p. The evapo-

the evaporating droplet at different time instances. (b) Normalized droplet volumethe end of crystallization i.e., total lifetime of the droplet (tf ) on different substrates.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

ration rate increases from apex to edge of the droplet, along the dro-plet periphery and near the contact line ðx=a � 1Þ, the vapor flux ishighest. The liquid-loss due to evaporation triggers an internal flowwhose direction and magnitude depend on the dominant mode offlow. In the present case, the induced flow can occur through capil-lary or Marangoni routes. As discussed in Section 1, for a particle-laden water droplets capillary flow is observed, and a salt-solution droplet results in the recirculation flow. The dominant cap-illary flow is directed from apex to edge of evaporating dropletwhile the Marangoni flow is recirculating in nature. The final flowpattern inside the droplet is intricate and depends on the magni-tude of evaporative flux and the relative dominance of capillary orMarangoni flow, as will be discussed next. The results of PIV mea-surement on PET and glass substrate are shown in Fig. 2 and 3,respectively.

3.2.1. Flow dynamics on PET substrateFig. 2a shows the streaklines indicating circulation flow and are

obtained after side-view planner imaging of evaporating surrogaterespiratory droplet on a PET substrate at t=tf � 0:1. 860 nm fluores-cent nanoparticles are used as tracers, and images are captured forthe camera exposure time of 5 s. The curvature effects hinder theside view imaging at lower contact angles. Therefore, side-viewimaging could not be done for the remaining substrates. For PETsubstrate, the initial bulk flow is majorly dominated by the recircu-lation flow, directing the liquid flow outward along with the liq-uid–air interface and inward near the droplet base (region-II).

Fig. 2. Spatio-temporal flow-field inside droplet evaporating on a PET substrate. (a) Sidblended with x-direction velocity ðUxÞ contours at z/h� 0, (b) at the droplet edge (Regionof each sequence is shown as a fraction of the droplet lifetime.

5

Velocity vector maps for PET surface at plane z=h0 � 0 are pre-sented in Fig. 2b and 2c. The region I shown in Fig. 2b includesthe local area near the droplet edge and is slightly above the dro-plet base. Region-I shows the bifurcation of the incoming flow.Here, the x-component of the velocity vectors ðUxÞ directed fromright to left represent the capillary flow ðUx ¼ UcÞ compensatingfor the evaporative mass loss, while the vectors directed from leftto right represents the inward component of recirculation flowðUx ¼ UreÞ. Simultaneously, Fig. 2c at t=tf ¼ 0:1 shows the recircu-lation flow convergence at the droplet center near the droplet base(Region II). Thus, capillary flow ðUcÞ deposits the solute populationat the droplet edge while Ure homogenize them in the interiorregions. The lPIV results at higher plane ðz=h � 0:5Þ shows an out-ward flow for both glass and PET substrate throughout the evapo-ration stage (see Fig. S2b). The observed recirculation flow on PETsubstrate can be attributed to solutal Marangoni flow [26]. Thesolutal Marangoni flow reported by Marin et al. [26] is qualitativelysimilar to the flow pattern shown in Fig. 2. They also show that theinward flow does not extend until the periphery of the drying dro-plet, as shown in Fig. 2b at t=tf � 0:1. The flow near the contact lineshows a bifurcation into inward Marangoni circulation and out-ward capillary flow at the region I.

The droplet height reduces due to continuous evaporation,which enhances the capillary flow. As a result, it dominates therecirculation flow during later stages of evaporation. The domi-nance of capillary flow is shown in Fig. 2b and 2c at t=tf � 0:6;by noting the reversal of the velocity vectors leading to an entirely

e-view planner imaging showing flow streaklines. (b-c) Bottom-view, flow vectorsI) and (c) at the droplet center (Region II) at different time instances. The time stamp

Fig. 3. Spatio-temporal flow-field inside droplet evaporating on a glass substrate. Bottom-view, flow vectors blended with x-direction velocity ðUxÞ contours at z=h � 0 (a) atthe droplet edge (Region I) and (b) at the center of the droplet (Region II) at different time instances. The timestamp of each sequence is presented as a fraction of the dropletlifetime.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

outward flow near the edge. Thus, the role of recirculating flow inhomogenizing the solute concentration is limited only up to half ofthe droplet lifetime ðt=tf � 0:5� 0:6Þ, when the droplet contactangle is 25–22 degrees; beyond which the capillary flow domi-nates. At later stages (Fig. 2b at t=tf � 0:8), the flow velocity is verylow due to high solutal concentrations near the edge and the for-mation of the gelatinous solution. Simultaneously a disorderedflow in region II is observed. Finally, near the end stages, excessivesolute concentration in the solution leads to the viscous dampingof fluid motion within the droplet.

In the present experiments, the flow velocity obtained from theside view and the top view measurements is � Oð10�5Þ m/s.Whereas the solutal Marangoni velocity scales as [75]UMa � ðDr:hÞ=ðl:aÞ. Therefore, from this expression, for producinga flow velocity of 10�5 m/s, the surface tension change required is10�8 N/m, and the corresponding concentration change Dcrequired is [76] � 10 lmol/kg, which is very small and unphysicalfor a flow to be solely induced by surface tension gradient. Thiskind of quantitative mismatch between experimental and pre-dicted results on Marangoni velocity for a saline droplet evapora-tion was also observed by Marin et al. [26]. Similar observationswere made by Didens et al. [77] for multi-component droplets.Therefore, a clear physical explanation for this mismatch is stilllacking.

3.2.2. Flow dynamics on glass substrateThe droplet initial contact angle on a glass substrate is least

among all considered surfaces, and therefore, has a higher evapora-tive mass loss (see Eq. 4). To compensate for the evaporative loss,capillary flow increases and thereby dominates the recirculatingflow even during the initial stages of droplet evaporation, as shownin Fig. 3a. Similar outward flow is also observed from the center to

6

edge of the drying droplet (see Fig. 3b) during the entire dropletlifetime. In contrast to the PET surface, the recirculating flow Ure

on the glass surface does not dominate the capillary flow at anytime during the evaporation. The capillary flow velocity due tomass loss scales as Uc � a=tf � 10�5 m/s, which is of the sameorder as observed during this work.

3.3. Global features of crystallization

NaCl accounts for more than 70% of the solute mass in the dro-plet. As shown in Fig. 3, Uc will lead to solute growth near the con-tact line, and accumulated solute at these locations can act aspotential sites for heterogeneous nucleation of salt. Since thesolute concentration grows faster near the contact line, nucleationoccurs earlier here than at the center, at a time ’tc ’. The rate atwhich solute concentration increases near the contact line dependson the solute accumulation rate and the instantaneous volume ofthe wedge. The wedge’s instantaneous volume encloses the dro-plet’s liquid–vapor interface and the solid substrate near its con-tact line. The shorter drying time of the droplet on glass leads tohigher capillary flow rates, further leading to faster solute accumu-lation near the contact line and the earliest onset of nucleation (seeFig. 1c). The converse is true for PET substrates. The first column inFig. 4 shows the droplet state before nucleation. The onset ofnucleation in the droplet’s periphery is shown in the second col-umn of Fig. 4. Nucleation sites are also observed in the interiorregions ðt=tf � 0:95Þ on all surfaces. The crystallization fronts fromthe periphery and center appear to interconnect with each other inglass and ceramic substrate (Fig. 4, third column).

The onset of crystallization occurs when supersaturated condi-tions are achieved in the solution. NaCl crystals are expected to becuboid as mass accumulates on the crystal face. Goto et al. [78]demonstrated the dendrite formation during salt crystallization

Fig. 4. Sequential images of crystallization showing the top-view of the drying droplet on (a) Glass (b) Ceramic (c) Steel (d) PET substrate. The first column shows the onset ofnucleation near the droplet edge at t ¼ tc . Second and thirds column shows the evolution of crystal growth along the periphery and towards the center. Column IV show at anadvanced stage of crystallization, the merging of crystallization fronts from various locations. The last column shows the end of crystallization at tf . Inset windows show themagnified view of the major crystal shape at the end of crystallization on various substrates.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

when mass transfer across the supersaturated solution interface ishindered due to the presence of gelatinous agar. Dutta Chouduryet al. [79] attributed the formation of dendrites to the non- homo-geneous distribution of salt in the gelatinous solution. The sus-tained growth of regular crystals requires the initial nucleus tobe connected to a saturated solution. However, during theadvanced stages of drying within a gelatinous mixture, the saltsolution is segregated into disparate pockets. When these pocketsare supersaturated, crystallites form and interconnect directionallyto form dendrites. The role of mucin in the present study isobserved to be similar to agar in forming dendrites. Fig. 5 showsdendritic branches, both at the edge (bottom row) and center(top row) of drying droplets on glass, ceramic, and PET substrates.The dendrites observed near the droplet center are thinner thanthose near the edges. Dendrites are thinnest for glass, followedby ceramic and PET. The reason for variation in the dendritedimension is discussed in a subsequent section. Steel shows thicktriangular crystals at the edge and cubical shaped crystals at thefinal precipitate center.

3.4. Dendrites vs. cubical crystals

The crystallization sequence on steel and glass is shown in sup-plementary Fig. S3 and Fig. S4. The nucleation always starts at theedge due to high salt concentration at the edge and edge depositionacting as nucleation sites. The crystal nucleation on central regionoccurs, comparatively faster on steel. Thus when the supersatura-tion levels are low the crystals grow slowly by diffusion and formcubical crystals on steel. Regular cubical crystals are the character-istic shape of the NaCl crystal that form under normal conditions

7

when salt can diffuse at equilibrium conditions to the growing crys-tal front. NaCl crystals take a dendritic shapewhen non-equilibriumconditions hinder the salt diffusion. The schematic of these condi-tions is shown for NaCl-mucin solution in Fig. 6. As shown inFig. 6a, the mucin dispersed in the droplet hinders the salt diffusionand creates a non-equilibrium condition for salt diffusion. In con-trast, on a hydrophobic surface, mucin gets well adsorbed mainlythrough hydrophobic interactions [72] that could allow equilibriumsalt diffusion to form regular cubical crystals. Several studies,including Chandrasekaran et al. [80], have reported strong adsorp-tion of proteins on stainless steel surfaces. Even on the same sub-strate, the instantaneous concentration of non-salt solutes duringcrystal growth could affect salt diffusion and resulting crystalshape. The roughness of the substrate could play an essential rolein this variation. The glass and steel substrate average roughnessesaremeasured to be 0.011lmand 0.12lm, as shown in Supplemen-tary Fig. S5. The kinks on the ridges of the steel surface could act asnucleation sites, and crystal growth starts at a time t=tf much ear-lier than on other surfaces (tc=tf is least on steel, see Fig. 4). Thiscould result in the early initiation of crystallization when the dro-plet height is higher. Thus, the mucin concentration on steel duringnucleation would be lower than that on other substrates, thus pro-viding lesser resistance for salt diffusion. Therefore, the above dis-cussion explains regular crystal formation on steel substrate. Itmust be noted that the formation of cubic or dendritic crystals can-not be attributed to the flow inside droplet, as the advection getceased (see Fig. 2 and 3 at t=tf � 0:8) at the time of crystallizationðt=tf � 0:9Þ, and diffusion of salt (as explained earlier) particlesremains the dominant mechanism for crystal formation.

Fig. 5. Final morphology of the precipitate on glass, ceramic, steel and PET surface, at the (a) center and (b) edge. Scalebar represents 150 lm.

Fig. 6. Schematic diagram showing the effect of mucin presence on different types of substrates. (a-b) Effect of surface wettability. (c-d) Effect of surface roughness. (a) Showslesser adsorption of mucin on hydrophilic substrate. (b) Shows the high affinity of mucin adsorption on hydrophobic surface. The square at the center of all schematics depictsthe crystal nucleate with the arrows indicating the salt diffusion under non-equilibrium conditions (a) and equilibrium conditions (b). (c) Shows the crystal nucleate on arelatively smooth substrate. At a later stage and nonequilibrium salt diffusion due to high non-salt solute concentration. (d) Shows the early crystal nucleation on a roughsubstrate and uniform salt diffusion at relatively low non-salt solute concentration. The length scales of the mucin monomer, crystal nuclei and the substrate roughness areexaggerated for easier visualization.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

3.5. Growth dynamics during crystallization

The dendritic network encompasses the precipitate footprint.Thus, their growth dynamics is crucial to understand the particledistribution within the precipitate. For each drying droplet, theinstantaneous length of the dendrite ðlÞ is considered from thepoint of nucleation until the tip of the growing front, as shown inFig. 7a. For the steel substrate, the side length of the cubical crystalis taken as ðlÞ. The typical growth of l on different substrates isshown in Fig. 7b. Cubical crystals on steel grow slower than thedendrites. Dendrites grow slowest on PET and fastest on glass.The final dendrite length lf , shows a scattered trend with a maxi-mum dendritic length not more than 700 lm (Fig. 7b). Fig. 7cshows a scatter plot between the lf and the corresponding averagegrowth rate ðdl=dtÞavg . The growth rate of a bulk crystal in a super-

saturated solution can be expressed as [81] dl=dt ¼ klðS� 1Þg ,

8

where S ¼ C=C;C, and C are the supersaturation and equilibriumsaturation concentration of salt in the solution, g is the order of thereaction and equals unity for NaCl. The rate constant kl ¼ Cleð�Ea=RTÞ

. Where Cl ¼ 1:14x10�4 m/s, Ea is the activation energy and equals58.18 kJ/mol, R is the universal gas constant, and T is the temper-ature of the solution in Kelvin. For a supersaturation (S) of 1.6 (ef-florescence limit), the crystal growth rate is � 1 lm/s. Insupersaturated solution, both regular cubical and bulk crystalgrowth rate is of the same order. However, the regular cubical crys-tal’s length scale is three to four times higher than bulk crystal, asthe supersaturation level increases continuously due to evapora-tion. The growth rate of dendrites, as shown in Fig. 7c, is aboutan order of magnitude higher than the growth of regular cubicalcrystals and varies across different substrates. The rate of dendriticgrowth should be driven by the variation in concentration gradientof salt around the leading tip of the dendrite. This involves a com-

Fig. 7. Growth dynamics of the crystal. (a) Sequence of dendrite growth showing the initial nucleus and the dendrite length ðlÞ. (b) The plot of instantaneous dendrite lengthðlÞ with time ðtÞ. (c) Variation in final length of the dendrite ðlf Þ for a different average rate of dendrite growth ðdl=dtÞavg corresponding to different substrates. Centroidscorresponding to each substrate is shown with purple colour of corresponding symbol. Scale bar equals 100 lm.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

petition between solvent loss (which increases the concentrationof salt) and diffusion of salt ions towards the leading edge of thetip (which reduces the concentration of salt in solution). The aver-age growth rate of dendrites is fastest on the substrate with ahigher evaporation rate and slowest on the substrate with slowerevaporation, as can be seen from Fig. 7c. The centroids data indi-cates that the final crystal length doesn’t increase much at a highergrowth rate due to more number of nucleation sites.

3.6. Distribution of virus emulating particles within the precipitate

Fluorescent signal from the virus emulating particles (VEP) isused to track their distribution during droplet precipitation[7,13]. The particles are neutrally buoyant, charge stabilized anddid not show noticeable surface affinity. The precipitate patternof the same solution without VEP does not show any fluorescencesignal. Fig. 8 is obtained after superimposing a fluorescent-baseddistribution of VEP onto a bright-field image. The procedure ofimage superimposition is shown in Fig. S6 (SupplementaryInformation).

The amount of particle deposition on edge and near the centerdepends on the flow dynamics during the evaporation and crystal-lization stage. Fig. 8a shows the fluorescence signal from the over-all droplet deposition. The particles are deposited majorly near theperiphery of the droplet and at the edge of the crystals. Fig. 8b and8c show the zoomed-in images at the central and peripheralregions. The peripheral region on all substrates shows high fluores-cence signal indicating a higher deposition of particles. The capil-lary driven flow drives the particles and other solutes towardsthe edge, leading to edge deposition. In the central region, PETshows higher fluorescence signal than other substrates (Fig. 8b).Some deposits also occur during the recirculation of the fluid inthe interior region. Apart from edge deposition, particles are alsofound near the edge of the crystal for all substrates (see Fig. 8). Thisdeposition can be mainly attributed to the jet flow during the crys-

9

tallization stage [43]. During crystal formation, the solution super-saturation level near the crystal surface reduces, which derives ajet flow towards the crystal surface. We suspect that particlesdeposited near the crystal periphery are due to this jet flow.Fig. 8d shows the ratio of fluorescence signal intensity (particledeposition) acquired over the same spatial window at the centraland the peripheral regions, Ic=Ie. The Ic=Ie is found to be propor-tional to the evaporation rate. Thus, the glass/PET showed mini-mum/maximum values of intensity ratios. Even though dropletsevaporate slightly slower on ceramic than on steel, the Ic=Ie is com-paratively higher on steel. This could be due to the crystal shapedifference between steel (cuboidal) and ceramic (dendritic).

3.7. Exposure of viruses on dried out patterns.

Fig. 9 shows the SEM images of the crystal shapes, patterns, andzoomed-in images used to identify particle exposure on precipi-tated deposits. The formation of dendrites on glass and cubicalcrystals on steel is also evident from the SEM images Fig. 9a and9b. Fig. 9c shows the rosette pattern of dendritic crystals fromthe nucleation site in the central region. The crystal pattern froman inner nucleation site in cruciform dendritic shape with branchesradiating out is shown in Fig. 9d. A zoomed-in image over the den-drite is shown in Fig. 9(e). Here, we can see a spherical protrusionon the dendrites, possibly due to the nanoparticles infused on thesurface. Similarly, Fig. 9(f) shows the deposits near the dropletperiphery region. Here a uniform layer of mucin is visible withcracks on the surface. No exact particle could be seen on the crystalor the edge deposits. Very few spherical protrusions are visible onthe dendrite’s surface, which indicates that particles are mostlyinfused inside the dendrites and are not directly exposed on thesubstrate. The SEM images of these substrates indicate that duringthe evaporation of cough droplets on a different substrate, virusesmay not get directly exposed either on the substrate or the crystalsformed over it, assuming overall similarity in the flow and interfa-

Fig. 8. Superimposed fluorescent and bright-field images of the droplet deposition pattern on the glass, ceramic, steel, and PET substrate from left to right. (a) Zoomed-outview of deposition. (b) Central region. (c) Peripheral region. (d) Histogram of fluorescent intensity ratios at the center to the edge ðIc=IeÞ of droplet deposition on differentsubstrates. The fluorescent intensity depicts the extent of particle deposition at the end of crystallization.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

cial characteristics. They are mostly infused inside the cubic ordendritic crystals (depending on the substrate) or beneath themucin layer formed on the substrate’s surface.

4. Conclusions

We have explored the crystallisation dynamics of a complexsystem which involves colloids (nano-particles) suspended in awater based protein-salt solution. Optics based diagnosis showsthat the surrogate respiratory fluid [62] on evaporation exhibitsdifferent flow and crystallization dynamics on different daily-usesubstrates (glass, steel, ceramic, and PET) considered in this study.

10

The final deposition pattern in all the substrates showed a thickedge deposition as reported in earlier works [52–54,82], but theoverall crystallization pattern showed considerable differenceamong each substrate. The dendrites and cruciform crystals[54,68] are the primary crystal shapes formed on glass, ceramic,and PET substrates, while cubical crystals were formed on steelsurfaces. The plausible reason for the formation of cubical and den-dritic crystals was also discussed based on mucin adsorption affin-ity with the substrate. Particle deposition on the evaporatingprotein-salt sessile droplet is studied extensively for the first timeand the virus emulating nanoparticles (VEP) were found to bedeposited more on edge due to the capillary flow during droplet

Fig. 9. SEM images of the crystal shapes, patterns, and the infused particles in the crystal. (a) Dendritic crystal on the glass substrate. (b) Cubical crystal on steel substrate (c)Rosette crystal pattern glass substrate. (d) Crystal pattern showing cruciform dendritic crystal with branches (e) Zoomed-in image on the dendrite surface showing sphericalprotrusions (possibly the particles infused inside it). (f) Zoomed-in image on the peripheral region showing mucin deposition with cracks. Scale bars are included in thefigures.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

evaporation. Further, particles are also observed near the crystalsurface. From the perspective of COVID-19, we expect that virusesare either deposited near the contact line or the crystal surfacesduring drying of virus-laden sessile-droplet on the considered sub-strates under the assumption that the virus emulating particlesassume the overall flow and interfacial characteristics. The workexplicitly sheds new light on the pandemics by exploring the phy-sics behind the formation of fomites in surrogate cough droplets.As the COVID-19 virus is nonmotile, one can expect a direct corre-lation of observed hydrodynamic and flow behavior betweenactual viruses and VEPs. However, surface chemistry and interfa-cial property differences between actual viruses and VEPs mightstill alter the deposition pattern of the viruses to some extent.

As a future scope, one can investigate whether the depositionwill stick to the human skin (palm) when it comes into contact,which depends on the deposition moisture content and the bond-ing strength of the crystals and the mucin layer on the surface. Thewide variation in crystal size and shapes will vary the transfer effi-ciency of contact which can be explored using well-calibratedshear and compression contact based measurements. The water-sorption and desorption characteristics of mucin could play a cru-cial role. Water-sorption and desorption studies as done for mucinfilm on glass in earlier study [83] can be done for the present com-position of solution and on different substrates. Using suitablesafety precautions actual pathogens can be used to study thepathogen spread in deposits.

Declaration of Competing Interest

The authors declare that they have no known competing finan-cial interests or personal relationships that could have appearedto influence the work reported in this paper.

11

Acknowledgement

The authors thank O. Hegde for assistance with SEM. SBacknowledges funding received through DRDO Chair Professorship.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at https://doi.org/10.1016/j.jcis.2021.04.128.

References

[1] S. Chaudhuri, S. Basu, P. Kabi, V.R. Unni, A. Saha, Modeling the role ofrespiratory droplets in covid-19 type pandemics, Physics of Fluids 32 (6)(2020) 063309. doi: https://doi.org/10.1063/5.0015984

[2] S. Chaudhuri, S. Basu, A. Saha, Analyzing the dominant sars-cov-2 transmissionroutes toward an ab initio disease spread model, Physics of Fluids 32 (12)(2020) 123306, https://doi.org/10.1063/5.0034032.

[3] S. Sharma, R. Pinto, A. Saha, S. Chaudhuri, S. Basu, On secondary atomizationand blockage of surrogate cough droplets in single- and multilayer face masks,Science Advances 7 (10) (2021), https://doi.org/10.1126/sciadv.abf0452.

[4] S. Esposito, N. Principi, C.C. Leung, G.B. Migliori, Universal use of face masks forsuccess against covid-19: evidence and implications for prevention policies,European Respiratory Journaldoi:10.1183/13993003.01260-2020.

[5] N.H. Leung, D.K. Chu, E.Y. Shiu, K.H. Chan, J.J. McDevitt, B.J. Hau, H.L. Yen, Y. Li,D.K. Ip, J.S. Peiris, W.H. Seto, G.M. Leung, D.K. Milton, B.J. Cowling, Respiratoryvirus shedding in exhaled breath and efficacy of face masks, Nature Medicine26. doi:10.1038/s41591-020-0843-2.

[6] J.P. Duguid, The size and the duration of air-carriage of respiratory droplets anddroplet-nuclei, Journal of Hygiene 44 (6) (1946) 471–479, https://doi.org/10.1017/S0022172400019288.

[7] S. Basu, P. Kabi, S. Chaudhuri, A. Saha, Insights on drying and precipitationdynamics of respiratory droplets in the perspective of covid-19, Physics ofFluids (November). doi:10.1063/5.0037360.

[8] P. Prasanna Simha, P.S. Mohan Rao, Universal trends in human cough airflowsat large distances, Physics of Fluids (8). doi:10.1063/5.0021666.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

[9] T. Dbouk, D. Drikakis, On respiratory droplets and face masks, Physics of Fluids32 (6) (2020) 063303. arXiv:https://doi.org/10.1063/5.0015044, doi:10.1063/5.0015044.

[10] S.A. Boone, C.P. Gerba, Significance of fomites in the spread of respiratory andenteric viral disease, Appl. Environ. Microbiol. 73 (6) (2007) 1687–1696,https://doi.org/10.1128/AEM.02051-06.

[11] B. Stephens, P. Azimi, M.S. Thoemmes, M. Heidarinejad, J.G. Allen, J.A. Gilbert,Microbial exchange via fomites and implications for human health, CurrentPollution Reports 5 (4) (2019) 198–213, https://doi.org/10.1007/s40726-019-00123-6.

[12] E. Goldman, Exaggerated risk of transmission of covid-19 by fomites, Lancet.Infect. Dis 20 (8) (2020) 892–893, https://doi.org/10.1016/S1473-3099(20)30561-2.

[13] S.R. Lustig, J.J.H. Biswakarma, D. Rana, S.H. Tilford, W. Hu, M. Su, M.S.Rosenblatt, Effectiveness of common fabrics to block aqueous aerosols ofvirus-like nanoparticles, ACS Nano 14 (6) (2020) 7651–7658, https://doi.org/10.1021/acsnano.0c03972. pMID: 32438799.

[14] R.D. Deegan, O. Bakajin, T.F. Dupont, G. Huber, S.R. Nagel, T.A. Witten, Capillaryflow as the cause of ring stains from dried liquid drops, Nature 389 (6653)(1997) 827–829, https://doi.org/10.1038/39827.

[15] M. Anyfantakis, D. Baigl, B.P. Binks, Evaporation of drops containing silicananoparticles of varying hydrophobicities: Exploiting particle-particleinteractions for additive-free tunable deposit morphology, Langmuir 33.doi:10.1021/acs.langmuir.7b00807.

[16] L. Shmuylovich, A.Q. Shen, H.A. Stone, Surface morphology of drying latexfilms: Multiple ring formation, Langmuir 18 (9) (2002) 3441–3445, https://doi.org/10.1021/la011484v.

[17] S. Maheshwari, L. Zhang, Y. Zhu, H.C. Chang, Coupling between precipitationand contact-line dynamics: Multiring stains and stick-slip motion, Phys. Rev.Lett. 100 (4) (2008) 1–4, https://doi.org/10.1103/PhysRevLett.100.044503.

[18] K.A. Baldwin, S. Roest, D.J. Fairhurst, K. Sefiane, M.E. Shanahan, Monolithformation and ring-stain suppression in low-pressure evaporation of poly(ethylene oxide) droplets, J. Fluid Mech. 695 (2012) 321–329, https://doi.org/10.1017/jfm.2012.22.

[19] K.M. Chen, X. Jiang, L.C. Kimerling, P.T. Hammond, Selective self-organizationof colloids on patterned polyelectrolyte templates, Langmuir 16 (20) (2000)7825–7834, https://doi.org/10.1021/la000277c.

[20] F. Fan, K.J. Stebe, Assembly of Colloidal Particles by Evaporation on Surfaceswith Patterned Hydrophobicity, Langmuir 20 (8) (2004) 3062–3067, https://doi.org/10.1021/la030380c.

[21] A. Mathur, A.D. Brown, J. Erlebacher, Self-ordering of colloidal particles inshallow nanoscale surface corrugations, Langmuir 22 (2) (2006) 582–589,https://doi.org/10.1021/la0520379.

[22] D. Debuisson, A. Merlen, V. Senez, S. Arscott, Stick-Jump (SJ) Evaporation ofStrongly Pinned Nanoliter Volume Sessile Water Droplets on Quick Drying,Micropatterned Surfaces, Langmuir 32 (11) (2016) 2679–2686, https://doi.org/10.1021/acs.langmuir.6b00070.

[23] P.J. Yunker, T. Still, M.A. Lohr, A.G. Yodh, Suppression of the coffee-ring effectby shape-dependent capillary interactions, Nature 476 (7360) (2011) 308–311, https://doi.org/10.1038/nature10344.

[24] S.M. Rowan, M.I. Newton, F.W. Driewer, G. McHale, Evaporation ofmicrodroplets of azeotropic liquids, J. Phys. Chem. B 104 (34) (2000) 8217–8220, https://doi.org/10.1021/jp000938e.

[25] H. Hu, R.G. Larson, Marangoni effect reverses coffee-ring depositions, J. Phys.Chem. B 110 (14) (2006) 7090–7094, https://doi.org/10.1021/jp0609232.

[26] A. Marin, S. Karpitschka, D. Noguera-Marín, M.A. Cabrerizo-Vílchez, M. Rossi,C.J. Kähler, M.A. Rodríguez Valverde, Solutal Marangoni flow as the cause ofring stains from drying salty colloidal drops, Physical Review Fluids 4 (4)(2019) 41601. arXiv:1809.07722, doi:10.1103/PhysRevFluids.4.041601.

[27] R.G. Picknett, R. Bexon, The evaporation of sessile or pendant drops in still air,Journal of Colloid And Interface Science 61 (2) (1977) 336–350, https://doi.org/10.1016/0021-9797(77)90396-4.

[28] H. Hu, R.G. Larson, Evaporation of a sessile droplet on a substrate, J. Phys.Chem. B 106 (6) (2002) 1334–1344, https://doi.org/10.1021/jp0118322.

[29] H. Hu, R.G. Larson, Analysis of the effects of marangoni stresses on themicroflow in an evaporating sessile droplet, Langmuir 21 (9) (2005) 3972–3980, https://doi.org/10.1021/la0475270.

[30] S. Masoudi, H.C. Kuhlmann, Axisymmetric buoyant-thermocapillary flow insessile and hanging droplets, J. Fluid Mech. 826 (2017) 1066–1095, https://doi.org/10.1017/jfm.2017.479.

[31] M. Majumder, C.S. Rendall, J.A. Eukel, J.Y. Wang, N. Behabtu, C.L. Pint, T.Y. Liu,A.W. Orbaek, F. Mirri, J. Nam, A.R. Barron, R.H. Hauge, H.K. Schmidt, M.Pasquali, Overcoming the coffee-stain effect by compositional marangoni-flow-assisted drop-drying, J. Phys. Chem. B 116 (22) (2012) 6536–6542,https://doi.org/10.1021/jp3009628.

[32] T. Still, P.J. Yunker, A.G. Yodh, Surfactant-induced Marangoni eddies alter thecoffee-rings of evaporating colloidal drops, Langmuir 28 (11) (2012) 4984–4988, https://doi.org/10.1021/la204928m.

[33] O. Hegde, S. Chakraborty, P. Kabi, S. Basu, Vapor mediated control ofmicroscale flow in sessile droplets, Physics of Fluids (12). doi:10.1063/1.5054632.

[34] F. Girard, M. Antoni, K. Sefiane, On the effect of Marangoni flow on evaporationrates of heated water drops, Langmuir 24 (17) (2008) 9207–9210, https://doi.org/10.1021/la801294x.

12

[35] W.D. Ristenpart, P.G. Kim, C. Domingues, J. Wan, H.A. Stone, Influence ofsubstrate conductivity on circulation reversal in evaporating drops, Phys. Rev.Lett. 99 (23) (2007) 1–4, https://doi.org/10.1103/PhysRevLett.99.234502.

[36] J.R. Trantum, Z.E. Eagleton, C.A. Patil, J.M. Tucker-Schwartz, M.L. Baglia, M.C.Skala, F.R. Haselton, Cross-sectional tracking of particle motion in evaporatingdrops: Flow fields and interfacial accumulation, Langmuir 29 (21) (2013)6221–6231, https://doi.org/10.1021/la400542x.

[37] M.R. Barmi, C.D. Meinhart, Convective flows in evaporating sessile droplets, J.Phys. Chem. B 118 (9) (2014) 2414–2421, https://doi.org/10.1021/jp408241f.

[38] E. Bormashenko, Y. Bormashenko, R. Pogreb, O. Stanevsky, G. Whyman,Droplet behavior on flat and textured surfaces: Co-occurrence of Deeganoutward flow with Marangoni solute instability, J. Colloid Interface Sci. 306 (1)(2007) 128–132, https://doi.org/10.1016/j.jcis.2006.10.043.

[39] N. Shahidzadeh-Bonn, S. Rafai, D. Bonn, G. Wegdam, Salt crystallization duringevaporation: Impact of interfacial properties, Langmuir 24 (16) (2008) 8599–8605, https://doi.org/10.1021/la8005629.

[40] B. Shin, M.W. Moon, H.Y. Kim, Rings, igloos, and pebbles of salt formed bydrying saline drops, Langmuir 30 (43) (2014) 12837–12842, https://doi.org/10.1021/la503095t.

[41] N. Shahidzadeh, M.F. Schut, J. Desarnaud, M. Prat, D. Bonn, Salt stains fromevaporating droplets, Scientific Reports 5 (2015) 1–9, https://doi.org/10.1038/srep10335.

[42] V. Soulié, S. Karpitschka, F. Lequien, P. Prené, T. Zemb, H. Moehwald, H. Riegler,The evaporation behavior of sessile droplets from aqueous saline solutions,PCCP 17 (34) (2015) 22296–22303, https://doi.org/10.1039/c5cp02444g.

[43] M. Efstratiou, J. Christy, K. Sefiane, Crystallization-driven flows withinevaporating aqueous saline droplets, Langmuir 36 (18) (2020) 4995–5002,pMID: 32319295. doi:10.1021/acs.langmuir.0c00576.

[44] G.V. Kuznetsov, S.Y. Misyura, R.S. Volkov, V.S. Morozov, Marangoni flow andfree convection during crystallization of a salt solution droplet, Colloids Surf.,A 572 (March) (2019) 37–46, https://doi.org/10.1016/j.colsurfa.2019.03.051.

[45] R.G. Larson, Transport and deposition patterns in drying sessile droplets, AIChEJ. 60 (5) (2014) 1538–1571, https://doi.org/10.1002/aic.14338.

[46] M. Parsa, S. Harmand, K. Sefiane, Mechanisms of pattern formation from driedsessile drops, Adv. Colloid Interface Sci. 254 (2018) 22–47, https://doi.org/10.1016/j.cis.2018.03.007.

[47] E. Rapis, A change in the physical state of a nonequilibrium blood plasmaprotein film in patients with carcinoma, Tech. Phys. 47 (4) (2002) 510–512,https://doi.org/10.1134/1.1470608.

[48] H. Ariffin, S.S. Daud, Z. Mohamed, K. Ibrahim, T.F. Lee, L.A. Chong, Evaluation oftwo short tandem repeat multiplex systems for posthaematopoieticstem cell transplantation chimerism analysis, Singapore Med. J. 48 (4)(2007) 333–337.

[49] T.A. Yakhno, Complex Pattern Formation in Sessile Droplets of Protein-SaltSolutions with Low Protein Content. What Substance Fabricates ThesePatterns?, Physical, Chemistry 1 (1) (2012) 10–13, https://doi.org/10.5923/j.pc.20110101.02.

[50] S. Devineau, M. Anyfantakis, L. Marichal, L. Kiger, M. Morel, S. Rudiuk, D. Baigl,Protein Adsorption and Reorganization on Nanoparticles Probed by the Coffee-Ring Effect: Application to Single Point Mutation Detection, J. Am. Chem. Soc.138 (36) (2016) 11623–11632, https://doi.org/10.1021/jacs.6b04833.

[51] A. Sett, M. Ayushman, S. Desgupta, S. Dasgupta, Analysis of the Distinct PatternFormation of Globular Proteins in the Presence of Micro- and Nanoparticles, J.Phys. Chem. B 122 (38) (2018) 8972–8984, https://doi.org/10.1021/acs.jpcb.8b05325.

[52] G. Chen, G.J. Mohamed, Complex protein patterns formation via salt-inducedself-assembly and droplet evaporation, European Physical Journal E 33 (1)(2010) 19–26, https://doi.org/10.1140/epje/i2010-10649-4.

[53] H.M. Gorr, J.M. Zueger, D.R. McAdams, J.A. Barnard, Salt-induced patternformation in evaporating droplets of lysozyme solutions, Colloids Surf., B 103(2013) 59–66, https://doi.org/10.1016/j.colsurfb.2012.09.043.

[54] B. Pathak, J. Christy, K. Sefiane, D. Gozuacik, Complex pattern formation insolutions of protein and mixed salts using dehydrating sessile droplets,Langmuir 36 (33) (2020) 9728–9737, https://doi.org/10.1021/acs.langmuir.0c01122.

[55] S. Darwich, K. Mougin, H. Haidara, From highly ramified, large scale dendritepatterns of drying alginate/Au NPs solutions to capillary fabrication of lab-scale composite hydrogel microfibers, Soft Matter 8 (4) (2012) 1155–1162,https://doi.org/10.1039/c1sm06623d.

[56] M.D. Choudhury, T. Dutta, S. Tarafdar, Pattern formation in droplets of starchgels containing NaCl dried on different surfaces, Colloids Surf., A 432 (2013)110–118, https://doi.org/10.1016/j.colsurfa.2013.04.064.

[57] Y.J. Carreón, M. Ríos-Ramírez, R.E. Moctezuma, J. González-Gutiérrez, Textureanalysis of protein deposits produced by droplet evaporation, ScientificReports 8 (1) (2018) 1–12, https://doi.org/10.1038/s41598-018-27959-0.

[58] K. Sefiane, R. Bennacer, An expression for droplet evaporation incorporatingthermal effects, Journal of Fluid Mechanics 667. doi:10.1017/S0022112010005446.

[59] D. Brutin, B. Sobac, B. Loquet, J. Sampol, Pattern formation in drying drops ofblood, J. Fluid Mech. 667 (2011) 85–95, https://doi.org/10.1017/S0022112010005070.

[60] K.A. Esmonde-White, F.W. Esmonde-White, M.D. Morris, B.J. Roessler,Characterization of biofluids prepared by sessile drop formation, Analyst 139(11) (2014) 2734–2741, https://doi.org/10.1039/c3an02175k.

A. Rasheed, S. Sharma, P. Kabi et al. Journal of Colloid and Interface Science 600 (2021) 1–13

[61] M. Mukhopadhyay, R. Ray, M. Ayushman, P. Sood, M. Bhattacharyya, D. Sarkar,S. DasGupta, Interfacial energy driven distinctive pattern formation during thedrying of blood droplets, J. Colloid Interface Sci. 573 (2020) 307–316, https://doi.org/10.1016/j.jcis.2020.04.008.

[62] E.P. Vejerano, L.C. Marr, Physico-chemical characteristics of evaporatingrespiratory fluid droplets, Journal of The Royal Society Interface 15 (139)(2018) 20170939, https://doi.org/10.1098/rsif.2017.0939.

[63] R. Effros, K. Hoagland, M. Bosbous, D. Castillo, B. Foss, M. Dunning, M. Gare, W.Lin, F. Sun, Dilution of respiratory solutes in exhaled condensates, Am. J.Respir. Crit. Care Med. 165 (5) (2002) 663–669, https://doi.org/10.1164/ajrccm.165.5.2101018. pMID: 11874811. arXiv:https://doi.org/10.1164/ajrccm.165.5.2101018.

[64] M.A. Kanso, J.H. Piette, J.A. Hanna, A.J. Giacomin, Coronavirus rotationaldiffusivity, Phys. Fluids 32 (11) (2020) 113101, https://doi.org/10.1063/5.0031875.

[65] A. Agrawal, S. Sinha, R. Mukherjee, D. Mampallil, Dynamics of bacterialdeposition in evaporating drops, Phys. Fluids 32 (9) (2020) 093308, https://doi.org/10.1063/5.0024078.

[66] T.T. Nellimoottil, P.N. Rao, S.S. Ghosh, A. Chattopadhyay, Evaporation-inducedpatterns from droplets containing motile and nonmotile bacteria, Langmuir23. doi:10.1021/la7006205.

[67] M.K. Raj, S. Misra, S.K. Mitra, Microparticle suspensions and bacteria-ladendroplets: Are they the same in terms of wetting signature?, Langmuir 37.doi:10.1021/acs.langmuir.0c03365.

[68] V.G. Kulichikhin, A.A. Elenskii, A.E. Kharlov, M.P. Shabanov, G.P. Yampol’skaya,Peculiarities of the surface crystallization of sodium chloride on mucin films,Colloid J. 74 (2) (2012) 207–214, https://doi.org/10.1134/S1061933X12010097.

[69] M.A. Shereen, S. Khan, A. Kazmi, N. Bashir, R. Siddique, COVID-19 infection:Origin, transmission, and characteristics of human coronaviruses, Journal ofAdvanced Research 24 (2020) 91–98, https://doi.org/10.1016/j.jare.2020.03.005.

[70] Y. Wang, G. Xu, Y.-W. Huang, Modeling the load of sars-cov-2 virus in humanexpelled particles during coughing and speaking, PLOS ONE 15 (10) (2020) 1–12, https://doi.org/10.1371/journal.pone.0241539.

[71] W. Norde, F.G. Gonzalez, C.A. Haynes, Protein adsorption on polystyrene latexparticles, Polymers for Advanced Technologies 6.doi:10.1002/pat.1995.220060713.

13

[72] L. Shi, K.D. Caldwell, Mucin adsorption to hydrophobic surfaces, J. ColloidInterface Sci. 224 (2) (2000) 372–381, https://doi.org/10.1006/jcis.2000.6724.

[73] G.J. Dunn, S.K. Wilson, B.R. Duffy, S. David, K. Sefiane, The strong influence ofsubstrate conductivity on droplet evaporation, J. Fluid Mech. 623 (2009) 329–351, https://doi.org/10.1017/S0022112008005004.

[74] V. Bazargan, B. Stoeber, Effect of substrate conductivity on the evaporation ofsmall sessile droplets, Phys. Rev. E 94 (3) (2016) 1–5, https://doi.org/10.1103/PhysRevE.94.033103.

[75] J. Park, J. Ryu, H.J. Sung, H. Kim, Control of solutal Marangoni-driven vorticalflows and enhancement of mixing efficiency, J. Colloid Interface Sci. 561 (2020)408–415, https://doi.org/10.1016/j.jcis.2019.11.006.

[76] L. Wang, M.T. Harris, Stagnation point of surface flow during drop evaporation,Langmuir 34 (20) (2018) 5918–5925, pMID: 29718675. doi:10.1021/acs.langmuir.8b00627.

[77] C. Diddens, Detailed finite element method modeling of evaporating multi-component droplets, J. Comput. Phys. 340 (2017) 670–687, https://doi.org/10.1016/j.jcp.2017.03.049.

[78] M. Goto, Y. Oaki, H. Imai, Dendritic growth of NaCl crystals in a gel matrix:Variation of branching and control of bending, Crystal Growth and Design 16(8) (2016) 4278–4284, https://doi.org/10.1021/acs.cgd.6b00323.

[79] M. Dutta Choudhury, T. Dutta, S. Tarafdar, Growth kinetics of NaCl crystals in adrying drop of gelatin: Transition from faceted to dendritic growth, Soft Matter11 (35) (2015) 6938–6947, https://doi.org/10.1039/c5sm00742a.arXiv:1506.06269.

[80] N. Chandrasekaran, S. Dimartino, C.J. Fee, Study of the adsorption of proteinson stainless steel surfaces using QCM-D, Chem. Eng. Res. Des. 91 (9) (2013)1674–1683, https://doi.org/10.1016/j.cherd.2013.07.017.

[81] A. Naillon, P. Duru, M. Marcoux, M. Prat, Evaporation with sodium chloridecrystallization in a capillary tube, J. Cryst. Growth 422 (2015) 52–61, https://doi.org/10.1016/j.jcrysgro.2015.04.010.

[82] T. Yakhno, Salt-induced protein phase transitions in drying drops, J. ColloidInterface Sci. 318 (2) (2008) 225–230, https://doi.org/10.1016/j.jcis.2007.10.020.

[83] S. Bjorklund, V. Kocherbitov, Water vapor sorption-desorption hysteresis inglassy surface films of mucins investigated by humidity scanning qcm-d, J.Colloid Interface Sci. 545 (2019) 289–300, https://doi.org/10.1016/j.jcis.2019.03.037.