UNDERSTANDING ADHESION TO MONITOR AND CONTROL ENVIRONMENT
PROCESSES: FROM VIRUS DETECTION TO MEMBRANE FOULING MITIGATION
                                        By
                                  Xunhao Wang
                               A DISSERTATION
                                    Submitted to
                           Michigan State University
                   in partial fulfillment of the requirements
                                for the degree of
             Environmental Engineering – Doctor of Philosophy
                                       2022


                                          ABSTRACT
    UNDERSTANDING ADHESION TO MONITOR AND CONTROL ENVIRONMENT
  PROCESSES: FROM VIRUS DETECTION TO MEMBRANE FOULING MITIGATION
                                              By
                                         Xunhao Wang
Adhesion is a physicochemical process of great importance for various environmental
engineering technologies including those that are employed in water treatment facilities
Understanding adhesion is the key to elucidating separation mechanisms in unit
processes and operations such as flocculation, granular media filtration, membrane
separation. The likelihood of adhesion can be quantified in terms of the interfacial
energy of interaction between two objects.
The first part of this dissertation is devoted to the study of virus adhesion to surfaces
commonly encountered in various indoor settings. Fomites are inanimate surfaces,
which can transfer the pathogens to a new human host. Fomite-based transfer is an
important pathway of virus transmission, along with direct contact and transmission
through aerosols. Adhesion of two viruses – one enveloped (human respiratory
syncytial virus, HRSV) and one non-enveloped (human adenovirus 5, HAdV5) – to four
fomites (silica, nylon, stainless steel, polypropylene) was quantified and interpreted
based on physicochemical properties of viruses and fomites. Quartz Crystal
Microbalance with Dissipation monitoring (QCM-D) is employed to quantify virus
attachment to fomites. XDLVO modeling is applied to predict the virus-fomites
interactions. It is found that for both HAdV5 and HRSV, the areal mass density of


deposited viruses correlated with the free energy of virus-fomite interfacial interaction in
water, ∆𝐺𝑣𝑤𝑓 .
The second part of this dissertation describes the study of membrane filter aging as a
result of membrane’s intermittent exposure to foulants and cleaning agents. This study
explores how the surface chemistry of polyvinylidene fluoride (PVDF) membranes
evolves in challenge tests with humic acid (HA) fouling and sodium hypochlorite (NaClO)
employed as a model foulant and cleaning agent, respectively. The evolution of
physicochemical properties of the ageing membranes is characterized based on surface
energy calculations. The results point to the formation of a chemically irreversible layer
of foulants that is conditioned by consecutive exposures to foulants and is comprised of
the adsorbed foulant fraction that is hard to oxidize further.


Copyright by
XUNHAO WANG
2022


                                   ACKNOWLEDGEMENTS
First, I would like to express my sincere gratitude to my advisor Dr. Volodymyr Tarabara
for the support of my PhD study, and for his patience and encouragement. I am
particularly impressed by his enthusiasm and devotion to scientific research. Dr.
Tarabara leads me to the world of scientific research. I am really grateful for his guide
and support over this long journey.
Besides my advisor, I would like to thank the rest of my thesis committee: Dr. Irene
Xagoraraki, Dr. Wei Zhang, and Dr. Rafael Auras, for their insightful comments and help.
I am also grateful to all my former and current lab members for their support and help. I
also would like to thank Brijen Miyani for their help with biological experiments. I would
like to thank all the specialists in Center of Advanced Microscopy Amy Albin, Carol
Flegler, and Stanley Flegler for helping me with the SEM.
I also want to acknowledge all great staff members in the Department of Civil and
Environmental Engineering. Many thanks to Lori Larner, Laura Post, Bailey Weber, the
research administrator team in our department, and Joseph Nguyen, lab technician in
our department.
Lastly, I would like to thank my parents for your unconditional love, support and care. I
would like to thank all my friends who accompanied me walk through this long journey.
Your support and encouragement are very precious to me.
                                                v


                                            TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................ix
LIST OF FIGURES ..........................................................................................................xi
CHAPTER ONE .............................................................................................................. 1
Introduction and dissertation overview ............................................................................ 1
  1.1 The environmental importance of adhesion ........................................................... 1
  1.2 Virus attachment to fomites.................................................................................... 1
  1.3 Virus adhesion to personal care products: PCPs as an important type of
       fomites ................................................................................................................... 5
  1.4 Membrane separation processes ........................................................................... 6
  1.5 Overview of adhesion measurements and data interpretation ............................... 8
     1.5.1 Classical method ............................................................................................. 8
     1.5.2 Graphical method ............................................................................................ 9
  1.6 Dissertation overview ........................................................................................... 10
  REFERENCES .......................................................................................................... 12
CHAPTER TWO ............................................................................................................ 20
Lip balm drying promotes virus attachment: Characterization of lip balm coatings and
XDLVO modeling .......................................................................................................... 20
  2.1 Abstract ................................................................................................................ 20
  2.2 Introduction .......................................................................................................... 21
  2.3 Materials and Methods ......................................................................................... 25
     2.3.1 Reagents ....................................................................................................... 25
     2.3.2 Preparation of lip balm surfaces .................................................................... 26
     2.3.3 Characterization of the lip balm surface: Surface charge and morphology .... 27
     2.3.4 Characterization of HAdV5 virions and SiO2 colloids: Size and charge ......... 28
     2.3.5 Quantifying hydrophobicity of HAdV5 and lip balm surfaces.......................... 29
     2.3.6 QCM-D studies of SiO2 and HAdV5 attachment to lip balm surfaces ............ 31
     2.3.7 XDLVO modeling of colloid interactions with lip balm .................................... 32
  2.4 Results and Discussion ........................................................................................ 34
     2.4.1 Optimal coating protocol and coating morphology ......................................... 34
     2.4.2 Hydrophobicity and surface charge of lip balm-coated surfaces .................... 36
     2.4.3 Hydrodynamic size and surface charge of HAdV5 and SiO2 colloids ............ 38
     2.4.4 Attachment of viruses and SiO2 particles to lip balm-coated surfaces:
           XDLVO predictions ........................................................................................ 43
     2.4.5 Attachment of SiO2 colloids and HAdV5 virions to lip balm-coated surfaces:
           Preliminary QCM-D study .............................................................................. 49
     2.4.6 Implications for virus control and public health protection .............................. 55
  2.5 Conclusions ......................................................................................................... 56
  APPENDIX ................................................................................................................. 59
  REFERENCES .......................................................................................................... 93
                                                              vi


CHAPTER THREE ...................................................................................................... 100
Virus adhesion to archetypal fomites: A study with human adenovirus and human
respiratory syncytial virus ............................................................................................ 100
  3.1 Abstract .............................................................................................................. 100
  3.2 Introduction ........................................................................................................ 101
  3.3 Materials and Methods ....................................................................................... 107
    3.3.1 Reagents, fomites, viruses........................................................................... 107
    3.3.2 Virion characterization: Hydrodynamic size, 𝜁-potential, concentration ....... 108
    3.3.3 Quantifying surface energy of viruses and fomites ...................................... 109
       3.3.3.1 Approach ............................................................................................... 109
       3.3.3.2 Experiments........................................................................................... 110
    3.3.4 Modeling virus-fomite interactions. Extended Derjaguin-Landau-Verwey-
           Overbeek theory .......................................................................................... 111
    3.3.5 QCM-D studies of HAdV5 and HRSV attachment to fomites ....................... 114
       3.3.5.1 Approach: Quantifying virus-fomite attachment efficiency ..................... 114
       3.3.5.2 QCM-D experiments .............................................................................. 115
  3.4 Results and Discussion ...................................................................................... 117
    3.4.1 Virus concentration, size, charge and hydrophobicity .................................. 117
    3.4.2 Fomite hydrophobicity .................................................................................. 122
    3.4.3 Quantifying virus-fomite interactions. Four archetypal fomites..................... 123
    3.4.4 Virus adhesion to fomites: QCM-D measurements ...................................... 129
       3.4.4.1 Model of mass transfer in QCM-D chamber. Assumptions .................... 129
       3.4.4.2 Quantifying mass transfer in QCM-D chamber: Mass transfer
               coefficients for HAdV5 and HRSV and size of depositing virions. ......... 130
       3.4.4.3 Virus deposition onto fomites. Virus-fomite attachment efficiency,
               𝛼𝑣𝑓. ....................................................................................................... 131
    3.4.5 XDLVO predictions ...................................................................................... 137
  3.5 Conclusions ....................................................................................................... 141
APPENDIX .................................................................................................................. 147
REFERENCES ............................................................................................................ 178
CHAPTER FOUR ........................................................................................................ 189
PVDF membrane ageing due to chemical cleaning: Understanding the evolution of
membrane-foulant interactions .................................................................................... 189
  4.1 Abstract .............................................................................................................. 189
  4.2 Introduction ........................................................................................................ 190
  4.3 Experimental ...................................................................................................... 192
    4.3.1 Reagents and materials ............................................................................... 192
    4.3.2 Chemical cleaning study of membranes ...................................................... 192
    4.3.3 Static fouling-cleaning test of membranes ................................................... 193
    4.3.4 Quantifying surface energy of PVDF membrane ......................................... 193
    4.3.5 Quantification of HA-membrane interactions. .............................................. 194
  4.4 Results ............................................................................................................... 195
    4.4.1 Impact of NaClO cleaning on membrane characteristics ............................. 195
    4.4.2 Static fouling and cleaning cycles ................................................................ 197
  4.5 Discussion.......................................................................................................... 198
                                                            vii


 4.6 Conclusions ....................................................................................................... 199
APPENDIX .................................................................................................................. 200
REFERENCES ............................................................................................................ 203
CHAPTER FIVE .......................................................................................................... 206
Conclusions and future work ....................................................................................... 206
REFERENCES ............................................................................................................ 209
                                                          viii


                                              LIST OF TABLES
Table 1. Size, charge and hydrophobicity of colloids considered in this work.
         The three properties are quantified in terms of hydrodynamic diameter 𝑑ℎ ),
         zeta-potential (𝜁), and free energy of interfacial interaction in water (∆𝐺𝑠𝑤𝑠 ),
         respectively. Notes: A This study (see SM, Fig. );B Shi et al. [21]; C Shi and
         Tarabara [30]; D Calculated based on contact angle data reported by
         Zdziennicka et al. [60]. The error is not provided because the original contact
         angle values are reported as averages only; E This study (see SM, Table 5);
         F Dang and Tarabara [25]; G This study (see SM, Fig. 13); H This study
         (see Fig. 5) ..................................................................................................... 44
Table 2. Values (in units of kT) of the primary maximum (𝐸𝑚𝑎𝑥 ) and secondary
         minimum (𝐸𝑚𝑖𝑛 ) in the XDLVO energy profile for colloid interaction with
         hydrated lip balms. ......................................................................................... 46
Table 3. Concentrations of various electrolytes in human saliva [9]. ............................. 64
Table 4. Composition of the four lip balms evaluated in this work. ................................ 65
Table 5. Contact angles (measured), surface energy parameters (calculated), and
         free interfacial energy of surface-surface interaction in water (calculated) for
         dry and hydrated lip balm samples and for human adenovirus 5. Lip balm
         coatings were prepared according to Protocol A (see Fig. 1). ........................ 66
Table 6. Contact angles (measured), surface energy parameters (calculated), and
         free interfacial energy of surface-surface interaction in water (calculated) for
         three types of lip balm (dry) with four different preparation protocols (A, B. C,
         and D.) The protocols are described in Fig. 1. Values for protocol A duplicate
         corresponding values (protocol A) given for dry lip balms in Table 5. ............ 67
Table 7. Guide to Figures with results of XDLVO modeling. Two of the 20 Figures
         are shown in the main manuscript while the other 18 are shown in the
         SM file. ........................................................................................................... 68
Table 8. Contact angle of probe liquids, surface energy parameters (𝛾 𝐿𝑊 , 𝛾 + , 𝛾 −
         , 𝛾 𝐴𝐵 , 𝛾 𝑡𝑜𝑡 ) and the free energy of interfacial interaction in water ((∆𝐺𝑠𝑤𝑠 ) ) of
         four fomites (clean, before use) and two viruses. Error estimates were
         obtained by propagating experimental errors in measured contact angles
         through the calculation of surface energy components (eqs (1) and (2)) and
         (∆𝐺𝑠𝑤𝑠 ) (eq. (3)). Additional contact angle measurements were performed
         for the four fomites using 150 mM NaCl solution at pH 5.8; the contact angle
         data and calculated surface energy values are given in SM (Table 11). ...... 119
Table 9. Free energy of virus-fomite interfacial interaction in water, ∆𝐺𝑣𝑤𝑓 , and
          its dispersive (∆𝐺𝑑𝐿𝑊    0
                                      ) and polar (∆𝐺𝑑𝐴𝐵       0
                                                                 ) components, for eight different
                                                           ix


         virus-fomite pairs. Standard deviations were obtained by propagating
         experimental errors in measured contact angles (Table 3) through the
         calculation of surface energy parameters (eqs. (1) and (2)) and ∆𝐺𝑣𝑤𝑓
         (eqs. (8), (9), and (10)). ................................................................................ 125
Table 10. 𝜁-potential values computed using Ohshima’s expression (eq. (S1)) based
          on measured electrophoretic mobilities of HRSV as a function of pH. ........ 173
Table 11. Contact angle of probe liquids, surface energy parameters (𝛾 𝐿𝑊 , 𝛾 + , 𝛾 − ,
           𝛾 𝐴𝐵 , 𝛾 𝑡𝑜𝑡 ) and the free energy of interfacial interaction in water (∆𝐺𝑠𝑤𝑠 of
           four fomites (clean, before use). Surface energies are calculated based on
           contact angles of 150 mM NaCl instead of DI water. Corresponding surface
           energy values calculated based on contact angles of DI water are given in
           Table 3. Error estimates were obtained by propagating experimental errors in
           measured contact angles through the calculation of 𝛾 (eq. (1)) and ∆𝐺𝑠𝑤𝑠
           (eq. (3)). ..................................................................................................... 174
Table 12. Contact angle of probe liquids, surface energy parameters (𝛾 𝐿𝑊 , 𝛾 + , 𝛾 − ,
            𝛾 𝐴𝐵 , 𝛾 𝑡𝑜𝑡 ) and the free energy of interfacial interaction in water ((∆𝐺𝑠𝑤𝑠 ) ) of
           four fomites (after use and subsequent cleaning; see section S4 for the
           description of cleaning protocols). Error estimates were obtained by
           propagating experimental errors in measured contact angles through the
           calculation of 𝛾 (eq. (1)) and (∆𝐺𝑠𝑤𝑠 ) (eq. (3))............................................. 175
Table 13. Contact angle of probe liquids, surface energy parameters (𝛾 𝐿𝑊 , 𝛾 + , 𝛾 − , 𝛾 𝐴𝐵 ,
           𝛾 𝑡𝑜𝑡 ) and the free energy of interfacial interaction in water ((∆𝐺𝑠𝑤𝑠 ) ) of four
           fomites (after use and subsequent cleaning; see section S4 for the
           description of cleaning protocols). Surface energies are calculated based on
           contact angles of 150 mM NaCl instead of DI water. Error estimates were
           obtained by propagating experimental errors in measured contact angles
           through the calculation of 𝛾 (eq. (1)) and (∆𝐺𝑠𝑤𝑠 ) (eq. (3)). ....................... 176
Table 14. Contact angle of probe liquids, surface energy parameters (𝛾 𝐿𝑊 , 𝛾 + , 𝛾 − , 𝛾 𝐴𝐵 ,
           𝛾 𝑡𝑜𝑡 ) and the free energy of interfacial interaction in water ((∆𝐺𝑠𝑤𝑠 ) ) of
           polypropylene at pH 4.1 and 3.8. ............................................................... 177
Table 15. The amount of each ingredient to make 50 g 15% solution of PVDF and
           PVDF/PVP in DMFA .................................................................................. 201
                                                           x


                                        LIST OF FIGURES
Figure 1. Lip balm coating protocols evaluated in this work. Of the three protocols
          (B, C, D) for preparing ultrathin coatings suitable for QCM-D measurements,
          protocol B yielded lip balm surface with the surface energy most closely
          matching that of a minimally processed sample (protocol A, baseline)
          and was, therefore, adopted for studying adhesion of colloids
          to such coatings. ............................................................................................ 28
Figure 2. Representative SEM images of a) Carmex b) ChapStick, c) Burt's Bees,
          and d) Vaseline coatings on a QCM-D sensor. .............................................. 37
Figure 3. Free energy of interfacial interaction in water, ∆Gsws , of the Carmex
           lip balm as a function of the coating protocol (see Fig. 1). Measured
          contact angles of probe liquids and calculated surface energy parameters
          used to compute ∆Gsws values are given in Table 6. Each measurement
          was done in triplicate. Errors correspond to standard deviations. .................. 40
Figure 4. Free energy of interfacial interaction in water, ∆Gsws , of the four lip balms
          coated using protocol B (see Fig. 1) in dry and hydrated states. Measured
          contact angles of probe liquids and calculated surface energy parameters
          used to compute ∆Gsws values are given in Table 5. Each measurement
          was done in triplicate. Errors correspond to standard deviations. .................. 41
Figure 5. Surface charge of HAdV5 virions and lip balm surfaces as a function
          of pH. Vertical dashed lines indicate pH values (4.2 and 7.2) used in QCM-D
          measurements of HAdV5 attachment to lip balm surfaces. Lines connecting
          experimental data points are added to the guide the eye. Average and
          standard deviations for lip balm samples are based on four independent
          measurements for each lip balm. Average and standard deviations for
          HAdV5 are based on three independent measurements. HAdV5 data is
          shown in both graphs to add data interpretation. ........................................... 42
Figure 6. XDLVO energies of the interaction of human adenovirus 5 with
          Carmex lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated
          (b, d) states. DLVO and XDLVO energy profiles for HAdV5 interaction with
          the other three lip balms (Burt’s Bees, ChapStick and Vaseline) are given in
          SM. DLVO total energy is provided for reference. .......................................... 47
Figure 7. DLVO and XDLVO energies of the interaction of SiO2 particles with Burt’s
          Bees lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated
          (b, d) states. DLVO and XDLVO energy profiles for HAdV5 interaction with
          the other three lip balms (Carmex, ChapStick and Vaseline) are given in
          SM. DLVO total energy is provided for reference. .......................................... 48
                                                      xi


Figure 8. QCM-D measurements of the deposition of SiO2 colloids onto hydrated
          Burt’s Bees lip balm under conditions when SiO2 loading is either a) constant
          (1.05 mg/ml) or b) increases stepwise from 0.52 mg/ml to 1.05 mg/ml. In both
          tests, deposition occurs from 150 mM NaCl electrolyte at pH 7.2. The mass
          values are calculated based on Sauerbrey equation (eq. (3)) with 𝑛 = 3. The
          results for 𝑛 = 5 are shown in SM (Fig. 32). ................................................... 51
Figure 9. QCM-D measurements of the deposition of human adenovirus 5 onto
          hydrated Carmex lip balm from 150 mM NaCl electrolyte at pH 7.2. HAdV5
          concentration in the QCM-D feed is ~109 GC/mL (~ 0.76 µg/mL). The mass
          values are calculated based on Sauerbrey equation (eq. (3)) with 𝑛 = 3. The
          results for 𝑛 = 5 are shown in SM (Fig. 33). ................................................... 54
Figure 10. Particle size distribution in the stock HAdV5 and ST-ZL SiO2 suspensions.
            Lines are added to guide the eye. ................................................................ 69
Figure 11. Free energy of interaction in water for three types of lip balm as a function
            of surface preparation method (see Fig. 1). ................................................. 70
Figure 12. Free energy, ∆𝐆𝐬𝐰𝐬 , for four lip balms as a function of drying time: a)
            Carmex; b) Chapstick, c) Burt’s Bees, d) Vaseline. Lip balms coatings were
            prepared using protocol B. In each graph, the solid line corresponds to the
            surface prepared using protocol A (melted lip balm with no dissolution and
            drying steps involved), which was used as the baseline: a) 65.03 mJ/m 2 for
            Carmex, b) -72.18 mJ/m2 for Chapstick, c) -95.62 mJ/m2 for Burt’s Bees. d) -
            90.71 mJ/m2 for Vaseline. Measured contact angles used to calculate ∆Gsws
            values are given Table 6. ............................................................................. 71
Figure 13. Surface charge of ST-ZL SiO2 colloids as a function of pH. Vertical dashed
            lines indicate pH values (4.2 and 7.2) used in QCM-D measurements of SiO2
            attachment to lip balm surfaces. Lines connecting experimental data points
            are added to the guide the eye. ................................................................... 72
Figure 14. XDLVO energy of interaction of human adenovirus 5 with Burt’s Bees lip
            balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d)
            states. DLVO total energy is provided for reference..................................... 73
Figure 15. XDLVO energy of interaction of human adenovirus 5 with ChapStick lip
            balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d)
            states. DLVO total energy is provided for reference..................................... 74
Figure 16. XDLVO energy of interaction of human adenovirus 5 with Vaseline lip
            balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d)
            states. DLVO total energy is provided for reference..................................... 75
Figure 17. XDLVO energy of interaction of human adenovirus 40 with Carmex lip
            balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d)
            states. DLVO total energy is provided for reference..................................... 76
                                                xii


Figure 18. XDLVO energy of interaction of human adenovirus 40 with Chapstick lip
           balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d)
           states. DLVO total energy is provided for reference..................................... 77
Figure 19. XDLVO energy of interaction of human adenovirus 40 with Burt’s Bees lip
           balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d)
           states. DLVO total energy is provided for reference..................................... 78
Figure 20. XDLVO energy of interaction of human adenovirus 40 with Vaseline lip
           balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d)
           states. DLVO total energy is provided for reference..................................... 79
Figure 21. XDLVO energy of interaction of P22 with Carmex lip balm at pH 4.2 (a, b)
           and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO total
           energy is provided for reference................................................................... 80
Figure 22. XDLVO energy of interaction of P22 with Chapstick lip balm at pH 4.2 (a, b)
           and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO total
           energy is provided for reference................................................................... 81
Figure 23. XDLVO energy of interaction of P22 with Burt’s Bees lip balm at pH 4.2
           (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO total
           energy is provided for reference................................................................... 82
Figure 24. XDLVO energy of interaction of P22 with Vaseline lip balm at pH 4.2 (a, b)
           and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO total
           energy is provided for reference................................................................... 83
Figure 25. XDLVO energies of interaction of MS2 with Carmex lip balm at pH 4.2
           (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO
           total energy is provided for reference. .......................................................... 84
Figure 26. XDLVO energy of interaction of MS2 with Chapstick lip balm at pH 4.2
           (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO
           total energy is provided for reference. .......................................................... 85
Figure 27. XDLVO energies of interaction of MS2 with Burt’s Bees lip balm at
           pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states.
           DLVO total energy is provided for reference. ............................................... 86
Figure 28. XDLVO energy of interaction of MS2 with Vaseline lip balm at pH 4.2
           (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO
           total energy is provided for reference. .......................................................... 87
Figure 29. XDLVO energies of interaction of SiO2 with Carmex lip balm at pH 4.2
           (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO
           total energy is provided for reference. .......................................................... 88
                                              xiii


Figure 30. XDLVO energy of interaction of SiO2 with Chapstick lip balm at pH 4.2
           (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO
           total energy is provided for reference. .......................................................... 89
Figure 31. XDLVO energies of interaction of SiO2 with Vaseline lip balm at pH 4.2
           (a, b) and pH 7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO
           total energy is provided for reference. .......................................................... 90
Figure 32. Deposition of SiO2 colloids onto hydrated Burt’s Bees lip balm under
           conditions when SiO2 loading is either a) constant (1.05 mg/ml) or b)
           increases stepwise from 0.52 mg/ml to 1.05 mg/ml. In both tests deposition
           occurs from 150 mM NaCl electrolyte at pH 7.2. The mass values are
           calculated based on Sauerbrey equation (eq. (3)) with 𝑛 = 5. The results
           for 𝑛 = 3 are similar and shown in the manin manuscript (Fig. 8). ................ 91
Figure 33. QCM-D measurements of the deposition of human adenovirus 5 onto
           hydrated Carmex lip balm from 150 mM NaCl electrolyte at pH 7.2. HAdV5
           concentration in the QCM-D feed is ~109 GC/mL (~ 0.76 µg/mL). The mass
           values are calculated based on Sauerbrey equation (eq. (3)) with 𝑛 = 5.
           The results for 𝑛 = 3 are similar and shown in the manin manuscript
           (Fig. 9). ......................................................................................................... 92
Figure 34. ζ-potential as a function of pH (A, D), size distribution (B, E), and TEM
           images (C, F, G) of HRSV (A-C) and HAdV5 (D-G). In A, B, D and E, lines
           are added to guide the eye. Charge and size values for HAdV5 (D, E) are
           adopted fromour earlier report [36]. Additional TEM images are shown
           in SM, Figure 42. ........................................................................................ 118
Figure 35. Free energy of interfacial interaction in water (∆𝑮𝒇𝒘𝒇 ) of four fomites as a
           function of pH. Depending on the type of aqueous solution used as a probe
           liquid, values are shown using either empty symbols (150 mM NaCl at
           pH 5.8; Table 10) or gray symbols (DI water at pH 5.8; Table 3) or black
           symbols (1 mM NaCl water at pH 3.8 and 4.1; Table 14). .......................... 124
Figure 36. Free energy, ∆𝐺𝑣𝑤𝑓 , of interfacial interaction of HRSV with monopolar fomites
           (𝛾𝑓+ = 0) in DI water. The solid line corresponds to ∆𝐺𝑣𝑤𝑓 = 0 (i. e. solution
           of eq. (16)). ................................................................................................ 128
Figure 37. Example QCM-D data set: deposition of HRSV on stainless steel.
           Representative QCM-D data for all virus-fomite pairs and deposition
           conditions are given in SM (Figures 50 – 57). All calculations based on the
           QCM-D data were performed using the signal for the 5th overtone (𝑛 = 5
           in eq. (14)). The reason for choosing the 5th harmonic was that the signals
           for the 3rd and 1st harmonics were unstable, likely due to their high sensitivity
           to mounting stress caused by the O-rings holding the sensor within the
           QCM-D chamber [96]. ................................................................................ 132
                                                           xiv


Figure 38. Virus-fomite attachment efficiency as a function of the free energy of
           virus-fomite interaction in water. Empty symbols correspond to tests with
           150 mM NaCl electrolyte as the deposition solution. Filled symbols
           correspond to tests with virus deposition onto polypropylene from 1 mM
           NaCl at pH 4.1 for HAdV5 and at pH 3.8 for HRSV where virus-fomite
           electrostatic interaction is favorable and at its maximum absolute value.
           (See SM, section S6 for the algorithm used to select these pH values.) .... 134
Figure 39. Mass and dissipation QCM-D data for the deposition of a) HRSV on
           polypropylene and b) HAdV5 on silica. Insets illustrate areal mass density
           values computed based on 5th, 7th, 9th, and 11th harmonics of the QCM-D
           signal.......................................................................................................... 136
Figure 40. XDLVO total energy of interactions during virus deposition (a, c) and
           cleaning (b, d) stages of QCM-D experiments. The results are for HRSV
           interaction with polypropylene (a, b) and HAdV5 interaction with silica
           (c, d). .......................................................................................................... 139
Figure 41. Electrophoretic mobility of HRSV virions as a function of pH. Table 10 lists
           mobility and zeta potentials as functions of pH in a tabular format. ........... 154
Figure 42. TEM images of (a, b) HAdV5 virions and (c, d) HRSV virions .................... 156
                                                                                                            𝑑𝑚
Figure 43. Rate of HAdV5 deposition on stainless steel QCM-D sensors ( 𝑑𝑡𝑣) as a
           function of HAdV5 concentration in the feed suspension (𝐶𝑏 ). The
           intermediate value of 𝐶𝑏 and deposition conditions are the same as in all
           other QCM-D tests. Solution is 150 mM NaCl. The observed linearity
                                                                          𝑑𝑚
           validates the assumption behind eq. (3): 𝑑𝑡𝑣 ≈ 𝛼𝑣𝑓 𝑘𝐶𝑏 𝐴. .......................... 157
Figure 44. Contact angle of four fomites: a) silica, b) nylon, c) stainless steel, d)
           polypropylene, as a function of pH. Depending on the type of aqueous
           solution used, values are shown using either empty symbols (150 mM NaCl
           at pH 5.8) or gray symbols (DI water at pH 5.8) or black symbols (1 mM
           NaCl at pH 3.8 and 4.1). ............................................................................ 159
Figure 45. XDLVO total energy of interactions during virus deposition (a, c) and
           cleaning (b, d) stages of QCM-D experiments. The results are for HRSV
           interaction with polypropylene (a, b) and stainless steel (c, d). .................. 160
Figure 46. XDLVO total energy of interactions during virus deposition (a, c) and
           cleaning (b, d) stages of QCM-D experiments. The results are for HRSV
           interaction with nylon (a, b) and silica (c, d). .............................................. 161
Figure 47. XDLVO total energy of interactions during virus deposition (a, c) and
           cleaning (b, d) stages of QCM-D experiments. The results are for HAdV5
           interaction with polypropylene (a, b) and stainless steel (c, d). .................. 162
                                                            xv


Figure 48. XDLVO total energy of interactions during virus deposition (a, c) and
           cleaning (b, d) stages of QCM-D experiments. The results are for HAdV5
           interaction with nylon (a, b) and silica (c, d). .............................................. 163
Figure 49. XDLVO total energy of interactions during virus deposition (a, c) and
           cleaning (b, d) stages of QCM-D experiments. The results are for HRSV
           (a, b) and HAdV5 (c, d) interaction with polypropylene Conditions
           correspond to to 𝛼𝑣𝑓 = 1. ............................................................................ 164
Figure 50. Example QCM-D data set: deposition of HRSV on polypropylene. The
           data are based on the 5th harmonic of the QCM-D signal. ......................... 165
Figure 51. Example QCM-D data set: deposition of HRSV on stainless steel. The
           data are based on the 5th harmonic of the QCM-D signal. ......................... 166
Figure 52. Example QCM-D data set: deposition of HRSV on nylon. The data are
           based on the 5th harmonic of the QCM-D signal. ....................................... 167
Figure 53. Example QCM-D data set: deposition of HRSV on silica. The data are
           based on the 5th harmonic of the QCM-D signal. ....................................... 168
Figure 54. Example QCM-D data set: deposition of HAdV5 on polypropylene. The
            data are based on the 5th harmonic of the QCM-D signal ......................... 169
Figure 55. Example QCM-D data set: deposition of HAdV5 on stainless steel. The
            data are based on the 5th harmonic of the QCM-D signal. ........................ 170
Figure 56. Example QCM-D data set: deposition of HadV5 on nylon. The data are
           based on the 5th harmonic of the QCM-D signal ........................................ 171
Figure 57. Example QCM-D data set: deposition of HAdV5 on silica. The data are
            based on the 5th harmonic of the QCM-D signal. ...................................... 172
Figure 58. Free energy of interfacial interaction in water (∆𝐺𝑠𝑤𝑠 ) of membrane as a
           function of cleaning time (cleaning agent: 1% NaClO) ............................... 195
Figure 59. Free energy of interfacial interactions (∆𝐺𝑚𝑤𝑚 ) of PVDF (red) and
           PVDF/PVP (blue) membrane as a function of fouling and cleaning cycles.
           The solid and empty symbols refer to replicated tests. ............................. 197
                                                xvi


                                      CHAPTER ONE
                          Introduction and dissertation overview
1.1 The environmental importance of adhesion
Adhesion processes are of great significance in many treatment processes and
operations employed by engineers to cure contamination of water, air, and soil.
Common water treatment techniques, such as flocculation and sand filtration rely on
adhesion. The efficiency of membrane separations, among many other processes, is
often limited by fouling – a phenomenon that has adhesion as its basis. Understanding
adhesion is critical for both gaining better mechanistic understanding of environmental
separations and improving process performance. This dissertation considers several
specific contexts where adhesion is of critical importance - virus adhesion to fomites,
virus adhesion to personal care products, foulant adhesion to polymeric membranes –
and combines experimental measurements with modeling of interfacial interactions to
understand and predict adhesion in these contexts.
1.2 Virus attachment to fomites
Numerous studies have been conducted to explore virus adhesion to surfaces such as
minerals [1-11], dissolved and particulate natural organic matter [12-14], activated
carbon [15-18], granular media filters [19-21], ion exchange resins [22], membrane
filters [20, 21, 23-28], vegetables and fruits [29-32], as well as human skin [33-41].
                                               1


However, the relationship between the interaction forces and adhesion and how the
adhesion is controlled by the physicochemical properties of the viruses and surfaces
remain insufficiently understood [42].
Naturally, virus adhesion depends on the properties of the virus [43]. It has been
suggested that differences in virus surface charge and hydrophobicity play a significant
role in virus adhesion [27]. Most viruses have capsid composed of protein polypeptides
that contain amino acids such as glutamic acid, aspartic acid, histidine, and tyrosine.
These amino acids feature acidic and basic groups (i. e., carboxyl and amino groups)
that, if ionized, give the viral capsid an electrical charge. The isoelectric point (IP) is the
pH value at which a surface has a zero net charge; surfaces are positively charged at
pH < pI and negatively charged at pH> pI. Viruses with a lipid envelope tend to be
hydrophobic, whereas viruses without a lipid envelope are generally hydrophilic [44].
Individuals across the globe spend most of their lives indoors, where they are constantly
exposed to viruses residing on surfaces. The term “fomite” stands for any inanimate
object that, when contaminated with infectious microorganisms, can act as a vector for
transferring the disease-causing agents to a new human host. Most important fomites
are those that humans contact frequently; examples include countertops, indoor walls,
clothing, and furniture. Despite the already large and quickly growing knowledge base
on the persistence of viruses on various surfaces, physicochemical bases of virus
attachment to and removal from surfaces are not fully understood. can consider
personal care products that human applies on their body as fomites. The study of
                                                2


fomites has traditionally focused on determining whether there is presence of specific
pathogenic organisms. Understanding virus-fomite interaction at the mechanistic level,
in terms of specific interaction forces can help predict adhesion and identify most likely
accumulation loci (“hot spots”) for viruses. Potentially, such knowledge can guide the
selection and placement of materials and surface coatings as well as optimization of
protocols of their cleaning with the ultimate goal of protecting human health.
Most viruses are colloids, which making colloidal stability theories applicable to viruses.
Virus-surface interactions can be modeled using the XDLVO theory, which builds on the
DLVO theory and takes hydrophobic interactions into consideration. XDLVO theory has
been widely used to study virus-fomites interactions [26, 45-52]. XDLVO theory
                                                𝑋𝐷𝐿𝑉𝑂
describes the total energy of interaction 𝑈𝑣𝑤𝑓        between a spherical particle
(representing a virus (𝑣)) and a flat surface (representing a fomite (𝑓)) in water (𝑤) as a
                                                          𝐿𝑊
sum (see eq. (7)) of Lifshitz-van der Waals (LW), 𝑈𝑣𝑤𝑓         , electrostatic double layer (EL),
  𝐸𝐿                                    𝐴𝐵
𝑈𝑣𝑤𝑓  , and Lewis acid-base (AB), 𝑈𝑣𝑤𝑓     , energies expressed as follows:
  𝐿𝑊
               𝐴𝑎           𝑎                                                                   (4)
𝑈𝑣𝑤𝑓  (𝑑) = −       = 2𝜋𝑑02 ∆𝐺𝑑𝐿𝑊
               6𝑑           𝑑       0
  𝐸𝐿
                                   1 + 𝑒 −𝜅𝑑
𝑈𝑣𝑤𝑓   (𝑑) = 𝜋𝜀𝑟 𝜀0 𝑎 [2𝜓𝑣 𝜓𝑓 𝑙𝑛 (           ) + (𝜓𝑣2 + 𝜓𝑓2 ) 𝑙𝑛(1 − 𝑒 −2𝜅𝑑 )]                  (5)
                                   1 − 𝑒 −𝜅𝑑
  𝐴𝐵
                               𝑑0 − 𝑑
𝑈𝑣𝑤𝑓  (𝑑) = 2𝜋𝑎𝜆∆𝐺𝑑𝐴𝐵    𝑒𝑥𝑝 (        )                                                         (6)
                       0
                                  𝜆
  𝑋𝐷𝐿𝑉𝑂      𝐿𝑊       𝐸𝐿      𝐴𝐵
𝑈𝑣𝑤𝑓     = 𝑈𝑣𝑤𝑓  + 𝑈𝑣𝑤𝑓   + 𝑈𝑣𝑤𝑓                                                                (7)
                                                  3


where 𝑎 is the virus radius, 𝑑 is the virus-fomite minimal interfacial separation distance,
𝐴 = −12𝜋𝑑02 ∆𝐺𝑑𝐿𝑊0
                    is the Hamaker constant, 𝜀𝑟 is the relative dielectric permittivity of
water (for water at 25 °C, 𝜀𝑟 ≈ 78.3), 𝜀0 is the dielectric permittivity of vacuum (𝜀0 =
8.854×1012 C·V-1·m-1), 𝜓𝑣 and 𝜓𝑠 are surface potentials of the virus and fomite,
respectively, 𝜅 is the inverse Debye screening length, λ is the characteristic delay length
of AB interactions in water (λ = 0.6 nm), and 𝑑0 is the minimum separation distance (𝑑0
= 0.158 nm) due to Born repulsion. Surface potentials 𝜓𝑣 and 𝜓𝑓 are commonly
approximated by 𝜁-potentials (𝜁𝑣 and 𝜁𝑓 ). Values of ∆𝐺𝑑𝐿𝑊  0
                                                              and 𝐺𝑑𝐴𝐵 0
                                                                         in eq. (4) and eq. (6)
are given by:
                                                                                             (8)
∆𝐺𝑑𝐿𝑊
    0
      = 2 (√𝛾𝑣𝐿𝑊 − √𝛾𝑤𝐿𝑊 ) (√𝛾𝑤𝐿𝑊 − √𝛾𝑓𝐿𝑊 )
                                                                                             (9)
∆𝐺𝑑𝐴𝐵
    0
      = 2 √𝛾𝑤+ (√𝛾𝑓− + √𝛾𝑣− − √𝛾𝑤− ) + 2√𝛾𝑤− (√𝛾𝑓+ + √𝛾𝑣+ − √𝛾𝑤+ )
                                 −2√𝛾𝑓+ 𝛾𝑣− − 2√𝛾𝑓− 𝛾𝑣+
The free energy of virus-fomite interfacial interaction in water, ∆𝐺𝑣𝑤𝑓 , is
∆𝐺𝑣𝑤𝑓 = ∆𝐺𝑑𝐿𝑊0
                + ∆𝐺𝑑𝐴𝐵0
                                                                                           (10)
Quartz crystal microbalance with dissipation monitoring (QCM-D) is a measurement
technique that makes it uniquely suitable for studies of virus adhesion. First, QCM-D is
very sensitive and can measure deposited materials with mass density as small as 17.7
ng/cm2 [53]. Second, the technique allows for real-time monitoring of the deposited
                                               4


mass and, therefore, can be used to measure deposition kinetics. With long range
transport of species to the QCM-D sensor factored out, mass vs time data can offer
insights into adsorption (or, more generally, adhesion) kinetics. The deposition of virus
particles on the quartz surface can lead to a decrease in the frequency of vibration.
Since the wet mass deposited onto the sensor is proportional to the changes in
resonance and overtone frequencies, the rate of wet mass deposition can be
represented by the rate of frequency shift [13]. Third, this technique can also offer
information about the viscoelastic behavior of the adsorbed layer by measuring
dissipation shifts [54]. QCM-D has been broadly used to study attachment of viruses
such as bacteriophage MS2 [11, 12, 42], norovirus virus-like particle (VLPs) [55, 56],
adenovirus [57], and pathogenic plant viruses [58, 59] to various surfaces including bare
QCM crystals [11, 55, 60] (most often gold or silica but sensors with other coating are
available), natural organic matter [12, 13, 61], polyelectrolyte multilayers [46, 57, 58, 62,
63], or self-assembled monolayers [42, 56, 58, 59].
1.3 Virus adhesion to personal care products: PCPs as an important type of
    fomites
Microbial contamination of personal care products has been a significant problem for
researchers and manufacturers worldwide [64]. Personal care products often include
natural ingredients, including organic compounds. The raw material that these natural
ingredients are sourced from is what often determines the potential of microbial
contamination. Further, personal care products can be contaminated during use. Of
                                             5


particular concern are products such as lipstick and lip balm, which can facilitate
contagion through routes such as ingestion and inhalation.
1.4 Membrane separation processes
Membrane processes are being increasingly used in drinking water treatment and water
reuse due to their effectiveness in removing a broad range of contaminants and
relatively low cost. Microfiltration (MF) is suitable for reducing turbidity and removing
microorganisms (bacteria, protozoans) while ultrafiltration (UF) can remove various
waterborne viruses and most dissolved organic matter. Polyvinylidene fluoride (PVDF)
is widely used as the base polymer for MF and UF membranes due to its chemical
resistance and excellent mechanical strength. However, pure PVDF membranes are
hydrophobic, making them more susceptible to fouling [65]. To mitigate this practical
limitation, membranes made from hydrophobic polymers such as PVDF are often
blended with hydrophilic additives.
Indeed, membrane fouling is one of the critical challenges in the successful application
of membrane processes. Fouling arises from interactions between a membrane and
various components present in the raw water [66]. Natural organic matter (NOM), such
as humic acid, is ubiquitous in surface water – a common source of raw water for
drinking water treatment plants. NOM is one of the major membrane foulants as
documented in research literature and reports from professional practice community
                                                6


[66-69]. Fouling by humic acid is a complex phenomenon which involves both reversible
and irreversible processes [66].
Membrane cleaning is the primary approach to fouling mitigation and recovering the
membrane performance. Membrane cleaning is defined as the action of removing
substances that are not an integral part of the membrane surface or membrane bulk.
Cleaning can be conducted either by using chemicals (chemical cleaning) or water flow
(hydraulic cleaning) [70]. Other processes such as vibration and ultrasound [71, 72] are
often used to supplement chemical and hydraulic cleaning. There are several categories
of cleaning chemicals including surfactants, oxidants, acids, and chelating agents.
Among them, sodium hypochlorite (NaClO) is one of the most commonly used cleaning
reagents because of its chemical stability, low price, and effectiveness against organic
foulants [73]. However, over an extended time, NaClO can impact the physicochemical
characteristics of the membrane, affect membrane performance, and reduce the service
life of the membrane [74]. NaClO can cause damage to membrane structure [75] and
depth [76]. The membrane can also become more hydrophilic due to NaClO cleaning
[77], which can result in a flux increase [78]. The zeta potential and the contact angle of
the membrane can also be changed by NaClO.
                                              7


1.5 Overview of adhesion measurements and data interpretation
1.5.1 Classical method
Hydrophobicity of a solid (𝑠) can be quantified in terms of the free energy of its
interfacial interaction with an identical material when immersed in water (𝑤), ∆𝐺𝑠𝑤𝑠 [79].
We applied this approach to evaluate hydrophobicity of viruses (𝑣) and fomites (𝑓) by
computing ∆𝐺𝑣𝑤𝑣 and ∆𝐺𝑓𝑤𝑓 , respectively. The calculation relies on the knowledge of
the solid’s surface energy in terms of its three components: two Lewis acid-base
(electron acceptor, 𝛾𝑠+ , and electron donor, 𝛾𝑠− ) components and the Lifshitz-van der
Waals component, 𝛾𝑠𝐿𝑊 . The Lewis acid-base components are also described as polar
while the Lifshitz-van der Waals component is often referred to as dispersive or apolar.
The components can be determined by measuring contact angles (𝜃) of three probe
liquids (𝑙) with known 𝛾𝑙𝐿𝑊 , 𝛾𝑙+ and 𝛾𝑙− on the surface of the solid and substituting these
values into the Young-Dupré equation [80, 81]
(1 + 𝑐𝑜𝑠𝜃)𝛾𝑙𝑇𝑂𝑇 = 2 (√𝛾𝑠𝐿𝑊 𝛾𝑙𝐿𝑊 + √𝛾𝑠+ 𝛾𝑙− + √𝛾𝑠− 𝛾𝑙+ ),                                   (1)
where 𝛾𝑙𝑇𝑂𝑇 is the total surface energy of the probe liquid:
𝛾 𝑡𝑜𝑡 = 𝛾 𝐿𝑊 + 𝛾 𝐴𝐵 = 𝛾 𝐿𝑊 + 2√𝛾 − 𝛾 + .                                                   (2)
The free energy of solid-solid interfacial interaction in water is given by
                                2
∆𝐺𝑠𝑤𝑠 =   −2 (√𝛾𝑠𝐿𝑊   −    𝐿𝑊
                        √𝛾𝑤 )     − 4 (√𝛾𝑠+ 𝛾𝑠− + √𝛾𝑤+ 𝛾𝑤− − √𝛾𝑠+ 𝛾𝑤− − √𝛾𝑠− 𝛾𝑤+ )         (3)
                                                 8


where 𝛾𝑤𝐿𝑊 , 𝛾𝑤+ and 𝛾𝑤− are surface energy components of water. A positive value of
∆𝐺𝑠𝑤𝑠 indicates a hydrophilic surface, while negative ∆𝐺𝑠𝑤𝑠 corresponds to a
hydrophobic surface. The absolute value of ∆𝐺𝑠𝑤𝑠 indicates the degree of hydrophilicity
(or hydrophobicity, when ∆𝐺𝑠𝑤𝑠 < 0) of the surface. Contact angles of three probe liquids
– DI water, glycerol, and diiodomethane (DID) – on target solids are measured using
sessile drop method. Additional contact angle measurements with ethylene glycol (EG)
as the fourth probe liquid are often performed to obtain additional data and improve the
accuracy of the prediction; in such cases, the graphical method is used.
1.5.2 Graphical method
For some of surfaces, the “classical method” may give a small negative value of √𝛾𝑠+ ;
this is attributed to a limited accuracy of the method wherein the standard error around
a small positive “true” average result in a prediction of a small negative average.
Therefore, a new approach – described herein as the “graphical method”, is used to
solve √𝛾𝑠+ , √𝛾𝑠− , and √𝛾𝑠𝐿𝑊 of these types of surfaces.
Since 𝛾𝑙− and 𝛾𝑙+ of DID are both equal to zero, Eq. (1) can be simplified:
(1 + 𝑐𝑜𝑠𝜃)𝛾𝐷𝐼𝐷𝑇𝑂𝑇
                  = 2√𝛾𝑠𝐿𝑊 𝛾𝐷𝐼𝐷
                              𝐿𝑊
                                  ,                                                   (4)
          𝑇𝑂𝑇       𝐿𝑊
where 𝛾𝐷𝐼𝐷    and 𝛾𝐷𝐼𝐷   are the surface tension components of DID. Using Eq. (4), one can
calculate 𝛾𝑠𝐿𝑊 of solids by measuring their contact angles with DID only.
                                               9


Eq. (1) can be recast as
(1 + 𝑐𝑜𝑠𝜃)𝛾𝑙𝑇𝑂𝑇 − 2√𝛾𝑠𝐿𝑊 𝛾𝑙𝐿𝑊              𝛾𝑙+                                             (5)
                                = √𝛾𝑠+ + √     √𝛾 −
            2√𝛾𝑙−                          𝛾𝑙− 𝑠
and 𝛾𝑠− can be determined from the slope of a linear dependence of the left-hand side of
eq. (5) on √𝛾𝑙+ ⁄𝛾𝑙− with the intercept (√𝛾𝑠+) set to zero. To improve the reliability of the
prediction, additional contact angle measurements with additional probe liquids can be
performed. The calculation can be done using the LINEST function in Excel; other than
the common availability of the software, the approach is also very useful in that LINEST
calculation returns errors for each both parameters (slope and intercept) of the linear fit.
The methodology introduced in this section provides a brief introduction. A detailed
discussion of the technique as applied in different adhesion contexts is provided in
individual chapters of this dissertation.
1.6 Dissertation overview
The material presented in this dissertation is divided into Chapters where each Chapter
is a manuscript either already published or in preparation for submission.
Chapter 2 presents the study of virus adhesion to archetypal fomites. In this chapter,
adhesion of two viruses – one enveloped (human respiratory syncytial virus, HRSV) and
one non-enveloped (human adenovirus 5, HAdV5) – to four fomites (silica, nylon,
                                               10


stainless steel, polypropylene) was quantified and interpreted based on
physicochemical properties of viruses and fomites. Virus-fomite interactions are
predicted using the extended Derjaguin-Landau-Verwey-Overbeek (XDLVO) theory and
are experimentally assessed in tests with quartz crystal microbalance with dissipation
(QCM-D).
Chapter 3 tests the hypothesis that drying-induced decrease in lip balm surface energy
can enhance virus adhesion due to the strong hydrophobic colloid-surface interactions.
The surface properties of lip balm and human adenovirus 5 (HAdV5) were measured.
Extended Derjaguin-Landau-Verwey-Overbeek (XDLVO) energy of interactions
between lip balm coatings and HAdV5 as well as four other colloids: HAdV40, MS2 and
P22 bacteriophages, and SiO2, were calculated.
Chapter 4 describes a model of virus recovery by tangential flow filtration to guide the
design of sample concentration process.
Chapter 5 focuses on the PVDF membrane ageing due to chemical cleaning. In this
study, the effect of humic acid fouling and sodium hypochlorite cleaning on the
physicochemical properties and surface energy of PVDF is investigated. The home-
made PVDF membranes, with and without PVP additives, were treated by several
cycles of humic acid fouling and sodium hypochlorite cleaning. The evolution of surface
energy over this process is recorded. This study helps understand the evolution of
membrane-foulant interactions when cleaned with sodium hypochlorite.
                                             11


REFERENCES
    12


                                       REFERENCES
[1] S. Chattopadhyay, R.W. Puls, Adsorption of bacteriophages on clay minerals,
Environmental science & technology 33(20) (1999) 3609-3614.
[2] G.W. Fuhs, M. Chen, L.S. Sturman, R.S. Moore, Virus adsorption to mineral
surfaces is reduced by microbial overgrowth and organic coatings, Microbial ecology
11(1) (1985) 25-39.
[3] S.M. Lipson, G. Stotzky, Adsorption of reovirus to clay minerals: effects of cation-
exchange capacity, cation saturation, and surface area, Appl. Environ. Microbiol. 46(3)
(1983) 673-682.
[4] S.M. Lipson, G. Stotzky, Effect of proteins on reovirus adsorption to clay minerals,
Appl. Environ. Microbiol. 48(3) (1984) 525-530.
[5] S.H. Lo, O.J. Sproul, Polio-virus adsorption from water onto silicate minerals, Water
Research 11(8) (1977) 653-658.
[6] J.P. Loveland, J.N. Ryan, G.L. Amy, R.W. Harvey, The reversibility of virus
attachment to mineral surfaces, Colloids and Surfaces A: Physicochemical and
Engineering Aspects 107 (1996) 205-221.
[7] R.S. Moore, D.H. Taylor, M.M. Reddy, L.S. Sturman, Adsorption of reovirus by
minerals and soils, Appl. Environ. Microbiol. 44(4) (1982) 852-859.
[8] R.S. Moore, D.H. Taylor, L.S. Sturman, M.M. Reddy, G.W. Fuhs, Poliovirus
adsorption by 34 minerals and soils, Appl. Environ. Microbiol. 42(6) (1981) 963-975.
[9] M. Schiffenbauer, G. Stotzky, Adsorption of coliphages T1 and T7 to clay minerals,
Appl. Environ. Microbiol. 43(3) (1982) 590-596.
[10] D.H. Taylor, R.S. Moore, L.S. Sturman, Influence of pH and electrolyte composition
on adsorption of poliovirus by soils and minerals, Appl. Environ. Microbiol. 42(6) (1981)
976-984.
[11] M. Tong, Y. Shen, H. Yang, H. Kim, Deposition kinetics of MS2 bacteriophages on
clay mineral surfaces, Colloids and Surfaces B: Biointerfaces 92 (2012) 340-347.
[12] B. Yuan, M. Pham, T.H. Nguyen, Deposition kinetics of bacteriophage MS2 on a
silica surface coated with natural organic matter in a radial stagnation point flow cell,
Environmental science & technology 42(20) (2008) 7628-7633.
[13] M. Pham, E.A. Mintz, T.H. Nguyen, Deposition kinetics of bacteriophage MS2 to
natural organic matter: Role of divalent cations, Journal of colloid and interface science
338(1) (2009) 1-9.
                                             13


[14] L. Gutierrez, X. Li, J. Wang, G. Nangmenyi, J. Economy, T.B. Kuhlenschmidt, M.S.
Kuhlenschmidt, T.H. Nguyen, Adsorption of rotavirus and bacteriophage MS2 using
glass fiber coated with hematite nanoparticles, Water Research 43(20) (2009) 5198-
5208.
[15] J.T. Cookson Jr, Mechanism of virus adsorption on activated carbon, Journal‐
American Water Works Association 61(1) (1969) 52-56.
[16] J.T. Cookson, W.J. North, Adsorption of viruses on activated carbon. Equilibriums
and kinetics of the attachment of Escherichia coli bacteriophage T4 on activated carbon,
Environmental science & technology 1(1) (1967) 46-52.
[17] T. Matsushita, H. Suzuki, N. Shirasaki, Y. Matsui, K. Ohno, Adsorptive virus
removal with super-powdered activated carbon, Separation and Purification Technology
107 (2013) 79-84.
[18] Q.L. Shimabuku, F.S. Arakawa, M. Fernandes Silva, P. Ferri Coldebella, T. Ueda-
Nakamura, M.R. Fagundes-Klen, R. Bergamasco, Water treatment with exceptional
virus inactivation using activated carbon modified with silver (Ag) and copper oxide
(CuO) nanoparticles, Environmental technology 38(16) (2017) 2058-2069.
[19] A.A. Torkelson, A.K. da Silva, D.C. Love, J.Y. Kim, J.P. Alper, B. Coox, J. Dahm, P.
Kozodoy, R. Maboudian, K.L. Nelson, Investigation of quaternary ammonium silane‐
coated sand filter for the removal of bacteria and viruses from drinking water, Journal of
applied microbiology 113(5) (2012) 1196-1207.
[20] H. Wang, M. Li, K. Brockman, T.H. Nguyen, Reduction of MS2 bacteriophage and
rotavirus in biosand filters, Environmental Science: Water Research & Technology 2(3)
(2016) 483-491.
[21] H. Wang, T. Narihiro, A.P. Straub, C.R. Pugh, H. Tamaki, J.F. Moor, I.M. Bradley, Y.
Kamagata, W.-T. Liu, T.H. Nguyen, MS2 bacteriophage reduction and microbial
communities in biosand filters, Environmental science & technology 48(12) (2014) 6702-
6709.
[22] P.S. Porter, M.J. Semmens, A review of virus and protein sorption by ion exchange
resins, Environment International 3(4) (1980) 311-319.
[23] A.M. ElHadidy, S. Peldszus, M.I. Van Dyke, An evaluation of virus removal
mechanisms by ultrafiltration membranes using MS2 and φX174 bacteriophage,
Separation and Purification Technology 120 (2013) 215-223.
[24] S.R. Farrah, Chemical factors influencing adsorption of bacteriophage MS2 to
membrane filters, Appl. Environ. Microbiol. 43(3) (1982) 659-663.
[25] J. Lukasik, T.M. Scott, D. Andryshak, S.R. Farrah, Influence of salts on virus
adsorption to microporous filters, Appl. Environ. Microbiol. 66(7) (2000) 2914-2920.
                                             14


[26] B. Michen, F. Meder, A. Rust, J. Fritsch, C. Aneziris, T. Graule, Virus removal in
ceramic depth filters based on diatomaceous earth, Environmental science &
technology 46(2) (2012) 1170-1177.
[27] P.A. Shields, S.R. Farrah, Influence of salts on electrostatic interactions between
poliovirus and membrane filters, Appl. Environ. Microbiol. 45(2) (1983) 526-531.
[28] K. Zodrow, L. Brunet, S. Mahendra, D. Li, A. Zhang, Q. Li, P.J.J. Alvarez,
Polysulfone ultrafiltration membranes impregnated with silver nanoparticles show
improved biofouling resistance and virus removal, Water research 43(3) (2009) 715-723.
[29] M.A. Esseili, Q. Wang, L.J. Saif, Binding of human GII. 4 norovirus virus-like
particles to carbohydrates of romaine lettuce leaf cell wall materials, Appl. Environ.
Microbiol. 78(3) (2012) 786-794.
[30] B.K. Ward, C.M. Chenoweth, L.G. Irving, Recovery of viruses from vegetable
surfaces, Appl. Environ. Microbiol. 44(6) (1982) 1389-1394.
[31] K. Verhaelen, M. Bouwknegt, A. Carratalà, F. Lodder-Verschoor, M. Diez-Valcarce,
D. Rodríguez-Lázaro, A.M. de Roda Husman, S.A. Rutjes, Virus transfer proportions
between gloved fingertips, soft berries, and lettuce, and associated health risks,
International journal of food microbiology 166(3) (2013) 419-425.
[32] E. Tuladhar, W.C. Hazeleger, M. Koopmans, M.H. Zwietering, E. Duizer, R.R.
Beumer, Transfer of noroviruses between fingers and fomites and food products,
International journal of food microbiology 167(3) (2013) 346-352.
[33] S.A. Ansari, S.A. Sattar, V.S. Springthorpe, G.A. Wells, W. Tostowaryk, Rotavirus
survival on human hands and transfer of infectious virus to animate and nonporous
inanimate surfaces, Journal of clinical microbiology 26(8) (1988) 1513-1518.
[34] S.A. Ansari, V.S. Springthorpe, S.A. Sattar, S. Rivard, M. Rahman, Potential role of
hands in the spread of respiratory viral infections: studies with human parainfluenza
virus 3 and rhinovirus 14, Journal of Clinical Microbiology 29(10) (1991) 2115-2119.
[35] A. Elkhyat, P. Agache, H. Zahouani, P. Humbert, A new method to measure in vivo
human skin hydrophobia, International journal of cosmetic science 23(6) (2001) 347-352.
[36] M.L. Grayson, S. Melvani, J. Druce, I.G. Barr, S.A. Ballard, P.D.R. Johnson, T.
Mastorakos, C. Birch, Efficacy of soap and water and alcohol-based hand-rub
preparations against live H1N1 influenza virus on the hands of human volunteers,
Clinical Infectious Diseases 48(3) (2009) 285-291.
[37] T.R. Julian, J.O. Leckie, A.B. Boehm, Virus transfer between fingerpads and
fomites, Journal of applied microbiology 109(6) (2010) 1868-1874.
                                             15


[38] J.N. Mbithi, V.S. Springthorpe, J.R. Boulet, S.A. Sattar, Survival of hepatitis A virus
on human hands and its transfer on contact with animate and inanimate surfaces,
Journal of Clinical Microbiology 30(4) (1992) 757-763.
[39] A.K. Pitol, H.N. Bischel, A.B. Boehm, T. Kohn, T.R. Julian, Transfer of enteric
viruses adenovirus and coxsackievirus and bacteriophage MS2 from liquid to human
skin, Appl. Environ. Microbiol. 84(22) (2018) e01809-18.
[40] A.K. Pitol, H.N. Bischel, T. Kohn, T.R. Julian, Virus Transfer at the Skin–Liquid
Interface, Environmental science & technology 51(24) (2017) 14417-14425.
[41] A.M. Wilson, K.A. Reynolds, M.P. Verhougstraete, R.A. Canales, Validation of a
stochastic discrete event model predicting virus concentration on nurse hands, Risk
Analysis (2019).
[42] A. Armanious, M. Aeppli, R. Jacak, D. Refardt, T.r.s. Sigstam, T. Kohn, M. Sander,
Viruses at solid–water interfaces: a systematic assessment of interactions driving
adsorption, Environmental science & technology 50(2) (2015) 732-743.
[43] C.P. Gerba, Applied and theoretical aspects of virus adsorption to surfaces,
Advances in applied microbiology, Elsevier1984, pp. 133-168.
[44] A.K. Vidaver, R.K. Koski, J.L. Van Etten, Bacteriophage φ6: a lipid-containing virus
of Pseudomonas phaseolicola, Journal of Virology 11(5) (1973) 799-805.
[45] C.V. Chrysikopoulos, I.D. Manariotis, V.I. Syngouna, Virus inactivation by high
frequency ultrasound in combination with visible light, Colloids and Surfaces B:
Biointerfaces 107 (2013) 174-179.
[46] H.T.T. Dang, V.V. Tarabara, Virus deposition onto polyelectrolyte-coated surfaces:
A study with bacteriophage MS2, Journal of colloid and interface science 540 (2019)
155-166.
[47] Z. Feng, R. Lu, B. Yuan, Z. Zhou, Q. Wu, T.H. Nguyen, Influence of solution
chemistry on the inactivation of particle-associated viruses by UV irradiation, Colloids
and Surfaces B: Biointerfaces 148 (2016) 622-628.
[48] J.-A. Park, S.-B. Kim, DLVO and XDLVO calculations for bacteriophage MS2
adhesion to iron oxide particles, Journal of contaminant hydrology 181 (2015) 131-140.
[49] V.I. Syngouna, C.V. Chrysikopoulos, Bacteriophage MS2 and titanium dioxide
heteroaggregation: Effects of ambient light and the presence of quartz sand, Colloids
and Surfaces B: Biointerfaces 180 (2019) 281-288.
[50] K. Wong, B. Mukherjee, A.M. Kahler, R. Zepp, M. Molina, Influence of inorganic
ions on aggregation and adsorption behaviors of human adenovirus, Environmental
science & technology 46(20) (2012) 11145-11153.
                                             16


[51] X. Wu, Z. Feng, B. Yuan, Z. Zhou, F. Li, W. Sun, Effects of solution chemistry on
the sunlight inactivation of particles-associated viruses MS2, Colloids and Surfaces B:
Biointerfaces 162 (2018) 179-185.
[52] S. Xu, R. Attinti, E. Adams, J. Wei, K. Kniel, J. Zhuang, Y. Jin, Mutually facilitated
co-transport of two different viruses through reactive porous media, Water research 123
(2017) 40-48.
[53] B. Becker, M.A. Cooper, A survey of the 2006–2009 quartz crystal microbalance
biosensor literature, Journal of Molecular Recognition 24(5) (2011) 754-787.
[54] M.C. Dixon, Quartz crystal microbalance with dissipation monitoring: enabling real-
time characterization of biological materials and their interactions, Journal of
biomolecular techniques: JBT 19(3) (2008) 151.
[55] A.K. da Silva, O.V. Kavanagh, M.K. Estes, M. Elimelech, Adsorption and
aggregation properties of norovirus GI and GII virus-like particles demonstrate differing
responses to solution chemistry, Environmental science & technology 45(2) (2010) 520-
526.
[56] G.E. Rydell, A.B. Dahlin, F. Höök, G. Larson, QCM-D studies of human norovirus
VLPs binding to glycosphingolipids in supported lipid bilayers reveal strain-specific
characteristics, Glycobiology 19(11) (2009) 1176-1184.
[57] M. Dimitrova, Y. Arntz, P. Lavalle, F. Meyer, M. Wolf, C. Schuster, Y. Haikel, J.C.
Voegel, J. Ogier, Adenoviral gene delivery from multilayered polyelectrolyte
architectures, Advanced functional materials 17(2) (2007) 233-245.
[58] N.F. Steinmetz, E. Bock, R.P. Richter, J.P. Spatz, G.P. Lomonossoff, D.J. Evans,
Assembly of multilayer arrays of viral nanoparticles via biospecific recognition: a quartz
crystal microbalance with dissipation monitoring study, Biomacromolecules 9(2) (2008)
456-462.
[59] X. Huang, J. Xu, H.-F. Ji, G. Li, H. Chen, Quartz crystal microbalance based
biosensor for rapid and sensitive detection of maize chlorotic mottle virus, Analytical
Methods 6(13) (2014) 4530-4536.
[60] J. Lee, J. Jang, D. Akin, C.A. Savran, R. Bashir, Real-time detection of airborne
viruses on a mass-sensitive device, Applied physics letters 93(1) (2008) 013901.
[61] L. Gutierrez, S.E. Mylon, B. Nash, T.H. Nguyen, Deposition and aggregation
kinetics of rotavirus in divalent cation solutions, Environmental science & technology
44(12) (2010) 4552-4557.
[62] G. Tiliket, G. Ladam, Q.T. Nguyen, L. Lebrun, Polyethylenimine surface layer for
enhanced virus immobilization on cellulose, Applied Surface Science 370 (2016) 193-
200.
                                               17


[63] V. Curtis, S. Cairncross, R. Yonli, Domestic hygiene and diarrhoea–pinpointing the
problem, Tropical medicine & international health 5(1) (2000) 22-32.
[64] S.E. Stewart, M.D. Parker, A. Amézquita, T.L. Pitt, Microbiological risk assessment
for personal care products, Int. J. Cosmet. Sci. 38(6) (2016) 634-645.
[65] S.Z. Abdullah, P.R. Bérubé, Assessing the effects of sodium hypochlorite exposure
on the characteristics of PVDF based membranes, Water research 47(14) (2013) 5392-
5399.
[66] W. Yuan, A.L. Zydney, Humic acid fouling during ultrafiltration, Environmental
Science & Technology 34(23) (2000) 5043-5050.
[67] I.H. Suffet, M. Patrick, Aquatic humic substances: influence on fate and treatment
of pollutants, 1988.
[68] V. Lahoussine-Turcaud, M.R. Wiesner, J.-Y. Bottero, Fouling in tangential-flow
ultrafiltration: the effect of colloid size and coagulation pretreatment, Journal of
Membrane Science 52(2) (1990) 173-190.
[69] Y. Kaiya, Y. Itoh, K. Fujita, S. Takizawa, Study on fouling materials in the
membrane treatment process for potable water, Desalination 106(1-3) (1996) 71-77.
[70] H. Liang, W. Gong, J. Chen, G. Li, Cleaning of fouled ultrafiltration (UF) membrane
by algae during reservoir water treatment, Desalination 220(1-3) (2008) 267-272.
[71] S. Muthukumaran, K. Yang, A. Seuren, S. Kentish, M. Ashokkumar, G.W. Stevens,
F. Grieser, The use of ultrasonic cleaning for ultrafiltration membranes in the dairy
industry, Separation and Purification Technology 39(1-2) (2004) 99-107.
[72] J. Wei, G. Wei, L. Xiaoping, H. Pingfang, W. Yanru, Effect of the ultrasound
generated by flat plate transducer cleaning on polluted polyvinylidenefluoride hollow
fiber ultrafiltration membrane, Chinese Journal of Chemical Engineering 16(5) (2008)
801-804.
[73] Q. Wang, H. Zeng, Z. Wu, J. Cao, Impact of sodium hypochlorite cleaning on the
surface properties and performance of PVDF membranes, Applied Surface Science 428
(2018) 289-295.
[74] C. Regula, E. Carretier, Y. Wyart, G. Gésan-Guiziou, A. Vincent, D. Boudot, P.
Moulin, Chemical cleaning/disinfection and ageing of organic UF membranes: A review,
Water research 56 (2014) 325-365.
[75] V. Gitis, R.C. Haught, R.M. Clark, J. Gun, O. Lev, Application of nanoscale probes
for the evaluation of the integrity of ultrafiltration membranes, Journal of membrane
science 276(1-2) (2006) 185-192.
                                                 18


[76] S.H. Wolff, A.L. Zydney, Effect of bleach on the transport characteristics of
polysulfone hemodialyzers, Journal of Membrane Science 243(1-2) (2004) 389-399.
[77] T. Mohammadi, S. Madaeni, M. Moghadam, Investigation of membrane fouling,
Desalination 153(1-3) (2003) 155-160.
[78] I. Wienk, E. Meuleman, Z. Borneman, T. Van Den Boomgaard, C. Smolders,
Chemical treatment of membranes of a polymer blend: mechanism of the reaction of
hypochlorite with poly (vinyl pyrrolidone), Journal of Polymer Science Part A: Polymer
Chemistry 33(1) (1995) 49-54.
[79] C.J. Van Oss, Interfacial forces in aqueous media, CRC Press2006.
[80] C.J. van Oss, M.K. Chaudhury, R.J. Good, Interfacial Lifshitz-van der Waals and
polar interactions in macroscopic systems, Chem. Rev. 88(6) (1988) 927-941.
https://doi.org/https://doi.org/10.1021/cr00088a006.
[81] M.E. Schrader, Young-Dupre revisited, Langmuir 11 (1995) 3585-3589.
https://doi.org/https://doi.org/10.1021/la00009a049.
                                             19


                                    CHAPTER TWO
                     Lip balm drying promotes virus attachment:
             Characterization of lip balm coatings and XDLVO modeling
2.1 Abstract
Hypothesis
Drying-induced decrease in lip balm surface energy enhances virus adhesion due to the
emergence of strong hydrophobic colloid-surface interactions.
Experiments
A protocol was developed for preparing lip balm coatings to enable physicochemical
characterization and adhesion studies. Surface charge and hydrophobicity of four
brands of lip balm (dry and hydrated) and human adenovirus 5 (HAdV5) were measured
and used to calculate the extended Derjaguin-Landau-Verwey-Overbeek (XDLVO)
energy of interactions between lip balm coatings and HAdV5 as well as four other
colloids: HAdV40, MS2 and P22 bacteriophages, and SiO2. Quartz crystal microbalance
with dissipation monitoring (QCM-D) tests employed SiO2 colloids, HAdV5 and hydrated
lip balms.
                                           20


Findings
Drying of lip balms results in a dramatic decrease of surface energy (𝛿(∆𝐺𝑠𝑤𝑠 ) ≥ 83.0
mJ/m2) making the surfaces highly hydrophobic. For dry lip balms, the interaction with
all five colloids is attractive. For lip balms hydrated in 150 mM NaCl (ionic strength of
human saliva), XDLVO calculations predict that hydrophilic colloids (MS2, P22, SiO 2)
may attach into shallow secondary minima. Due to the relative hydrophobicity of human
adenoviruses, primary maxima in XDLVO profiles are low or non-existent making
irreversible deposition into primary energy minima possible. Preliminary QCM-D tests
with SiO2 colloids and HAdV5 confirm deposition on a hydrated lip balm.
Keywords: adenovirus, personal care products, lip balm, lipstick, adhesion, XDLVO,
QCM-D, fomites, virus transfer, public health
2.2 Introduction1
Environmental transmission via contaminated surfaces is an important pathway for
pathogen transfer that can complement transfer with contaminated food or water, or via
direct person-to-person contact. There is a growing body of literature on the pathogen
adhesion to and resuspension from fomites [1]. Due to their smaller size, high infectivity,
and resistance to disinfection, viruses are of special concern. Despite a clear need for
1
  Abbreviations: Extended Derjaguin-Landau-Verwey-Overbeek (XDLVO); quartz crystal microbalance
with dissipation monitoring (QCM-D); human adenovirus 5 (HAdV5); human adenovirus 40 (HAdV40);
American Type Culture Collection (ATCC); scanning electron microscopy (SEM). isoelectric point (IEP).
                                                21


such data, relatively little is known about the propensity of viruses to attach to common
surfaces. A number of studies have explored virus transfer to and from human body. [2-
8]. Most but not all of these studies focused on unmodified human skin Julian et al. [5]
showed that washing fingerpads reduced virus transfer; the trend was tentatively
attributed to changes in the moisture level or pH of the skin [9] or to the presence of
soap residuals pointing to the possible importance of skin coatings. In the two studies
by Pitol et al. [6, 7], Vaseline was used to delimit the human skin area where
bacteriophages were allowed to adhere to human skin. The approach was based on a
viability assay, which showed that MS2 adsorption to Vaseline was at least one order of
magnitude lower than to the skin. However, the duration of the longest test was 10 min
and thus one can assume that Vaseline did not have time to dry.
Microbial contamination of personal care products has been a significant problem for
researchers as well as manufacturers worldwide [10]. The presence of high numbers of
pathogens poses a serious health threat to consumers, especially those who are
already ill or in a weakened state [11]. According to the Rapid Alert System database,
from January 2008 until week 26 of 2014 sixty-two cosmetic products were recalled
because they were contaminated with pathogenic or potentially pathogenic
microorganisms [12]. Although the use of preservatives, good manufacturing practices
and quality control programs have improved product quality, cases of contaminated
cosmetic products have been reported. For example, on December 7, 2010, there was
an Import Alert regarding Alexia Lip gloss cosmetic contaminated by Sphingomonas
paucimobilis and P. aeruginosa, manufactured by Maesa, Jinwan Zhuhai [13].
                                              22


Legislation and introduction of good manufacturing practices have improved the
microbiological standards, but contaminated cosmetics are still found [14] and
consequences for end users can be grave. As an example, contaminated ocular
cosmetics caused a Pseudomonas corneal ulcer after a woman sustained minor corneal
trauma with a mascara applicator [15].
Cosmetics are designed and manufactured to ensure product stability and
microbiological safety during normal and reasonably foreseeable product use. However,
according to the U.S. Food and Drug Administration, personal care products intended
for the general population are not intended to be sterile [10]. Personal care products
often include natural ingredients, including organic compounds; the raw materials that
these natural ingredients are sourced from are one of the main reasons for microbial
contamination [16]. Further, cosmetic products can be contaminated during use. Of
particular concerns are products such as lipstick and lip balm, which can facilitate
contagion through common routes such as ingestion and inhalation. For example,
sharing lipstick may increase the chance of infection of Epstein–Barr virus, an agent
linked to systemic lupus erythematosus [17]. Lipstick-associated pathogens are most
likely to enter the human body through the mouth when eating or drinking. This
hypothesized route is of particular importance given the scale of lipstick usage.
According to a study conducted by the Scientific Committee on Consumer Safety, the
estimated daily amount applied lipstick is 0.057 g and the frequency of application is
twice a day [18].
                                             23


There are multiple studies on virus attachment to surfaces such as organic matter [19],
soil [20], membrane filters [21], sand [22], polymers (e.g. polyvinylidene fluoride [23] and
anion exchange resins [24]), and polyelectrolyte-coated surfaces [25]. Most of this work
employed bacteriophages as human virus surrogates [19, 20]. To our knowledge, there
has been no published work on the attachment of viruses to personal care products
including lip balm and lipsticks. More broadly, to our knowledge, there have been no
prior reports on the adhesion of particles of any kind to lip balm or lipstick surfaces. This
is likely because of the absence of standard protocols for such measurements and
difficulties of sample preparation. The present work aims at filling this knowledge gap.
This study began with developing a methodology for coating lip balms to prepare
surfaces suitable for physicochemical characterization of lip balm as well as for testing
adhesion of colloids to such materials. Both dry and hydrated lip balms were
characterized in terms of charge and hydrophobicity. The data was used to predict the
XDLVO energy of interfacial interactions between these surfaces and five colloids:
human adenoviruses (serotypes 5 and 40) and two bacteriophages (P22 and MS2), and
SiO2 particles. The XDLVO modeling study was complemented with a preliminary
experimental study of SiO2 and HAdV5 deposition onto hydrated lip balm.
                                              24


2.3 Materials and Methods
2.3.1 Reagents
All chemical reagents were of high purity (>99%). Petroleum distillate (Penetrol, PPG
Industries) as well as ethyl acetate, isopropanol, methanol, pentane, and hexane (all –
Sigma Aldrich) were used as solvents for lip balms. Aqueous solutions of KCl (Sigma
Aldrich) were employed in measurements of particle size and electrophoretic mobility of
SiO2 colloids and HAdV5, streaming potential of lip balm surfaces, and colloid adhesion
to lip balms. Eight different brands of over-the-counter lip balm, all purchased in the
local supermarket (Meijer, Okemos, MI) were initially characterized in water contact
angle tests. Based on contact angle values and manufacturer-supplied information on
balm composition, four representative types of lip balm were selected for further
characterization and colloid adhesion tests. While nanoparticles such as ZnO and TiO 2
are sometimes included in lip balm formulations for added UV protection, neither of the
four lip balms selected for this study contained nanoparticles. The detailed composition
of each lip balm is given in SM (Table 4). The stock of viable HAdV5 was purchased
from American Type Culture Collection (ATCC) [26] (see SM, section S3). As described
by the manufacturer, HAdV5 (ATCC VR-1516) was propagated by culturing on HEK 293
cells from the Working Cell Bank in Cell Cubes® and purified using a single-column
chromatography with an anion exchange resin. The method was shown to give virus
stock of the purity comparable to that achieved with CsCl gradient purification [27]. SiO2
suspension (Snowtex ZL, 40-41 wt%, , SiO2 density 2.3 g/cm3 [28]) was obtained from
                                              25


Nissan Chemical Industries (Tokyo, Japan).
2.3.2 Preparation of lip balm surfaces
For contact angle and surface charge measurements, lip balms samples were melted
and then spin coated into a relatively thick film on a surface of a glass slide (Protocol A,
Fig. 1). Briefly, a solid lip balm sample was first melted on the heater and then pipetted
on a glass slide positioned on the chuck of the spin coater operated at 3000 rpm. The
deposition procedure was sufficiently fast (~ 5 s) to avoid premature on-contact
solidification of the sample. The duration of spin-coating was 15 s.
For QCM-D measurements, the coating layer should be not only continuous but also
very thin (a few microns at most) to avoid overloading the QCM-D sensor. To form such
thin layers, melted lip balm samples were mixed with a solvent prior to being spin-
coated on the QCM-D sensor surface. First, several solvents - ethyl acetate,
isopropanol, methanol, pentane, hexane, and petroleum distillate were evaluated as
alternatives for liquefying lip balm. The selection of solvents to test was partly based on
the results of a study by Dasari and Goud, who extracted castor seed oil using polar
and non-polar solvents [29]. Of the five solvents, petroleum distillate yielded the most
homogenous lip balm solution. Then, three different protocols (B, C and D; Fig. 1) were
assessed to select a method that gives a sufficiently thin coating with the surface
energy most closely matching that of a solvent-free sample. Briefly, solid lip balm
samples were melted in a glass vial using a heating plate and then dissolved in
                                              26


petroleum distillate. The liquid sample was stirred (Multistirrer, Thomas Scientific) under
800 rpm and autoreverse (10 s) mode for 1 day and then sonicated overnight at 37 °C
to break up larger, undissolved particles. The processed sample (~ 0.5 ml) was pipetted
onto a clean support surface (glass slide or QCM-D sensor), spin-coated at 7500 rpm
for 50 s, and left in the fume hood at room temperature for 5 days. After drying, the
sample was heated on a heater at the minimum temperature of 100 °C for 2 min to
evaporate residual solvents.
2.3.3 Characterization of the lip balm surface: Surface charge and morphology
Surface charge was determined using streaming current measurements (SurPASS
electrokinetic analyzer, Anton Paar GmbH) performed on 20 mm x 10 mm lip balm-
coated glass slides. Prior to measurements, each sample was immersed in 1 mM KCl
solution overnight. Samples were fixed on sample holders and inserted into an
adjustable gap cell with the gap height set at 145 µm. Measurements were done using
KCl as the electrolyte and repeated four times for each sample. The homogeneity of the
coating on the gold sensors was evaluated using scanning electron microscopy (SEM,
JEOL 6610 LV microscope). Because of the concern that volatile components of lip
balm samples may interfere with SEM imaging, microscopy was performed in the low
vacuum mode (see SM, section S1).
                                            27


Figure 1. Lip balm coating protocols evaluated in this work. Of the three protocols (B, C,
D) for preparing ultrathin coatings suitable for QCM-D measurements, protocol B
yielded lip balm surface with the surface energy most closely matching that of a
minimally processed sample (protocol A, baseline) and was, therefore, adopted for
studying adhesion of colloids to such coatings.
2.3.4 Characterization of HAdV5 virions and SiO2 colloids: Size and charge
Hydrodynamic diameter and electrophoretic mobility were measured by dynamic light
scattering (ZetaPALS, Brookhaven Instruments) and laser doppler micro-
electrophoresis (Zetasizer Nano ZS, Malvern), respectively. Zeta potential values were
                                             28


calculated based on electrophoretic mobilities determined as a function of pH, which
was adjusted using NaOH and HCl [30]. The particle size distribution in the SiO2 stock
(40.5 wt%) could not be measured because of high turbidity. Thus, the samples used to
measure the particle size and 𝜁-potential were both diluted to 0.08 wt% in 1 mM KCl.
Prior to use, the stock suspensions were sonicated (VWR ultrasonic cleaner, 35 kHz, 40
W, VWR International) for 20 min to ensure complete dispersion. Prior to measuring the
hydrodynamic size and charge of colloids (virus or SiO2), the suspensions were filtered
through 0.22 μm filter.
2.3.5 Quantifying hydrophobicity of HAdV5 and lip balm surfaces
While the hydrophobicity of a surface can be roughly evaluated based on its contact
angle with water, a more accurate measure of surface hydrophobicity is given by the
free energy of interfacial interaction (∆𝐺𝑠𝑤𝑠 ) of two surfaces, identical to the one in
question, when immersed in water [31]. The negative sign of ∆𝐺𝑠𝑤𝑠 indicates that the
surface is hydrophobic [32]. The absolute value of ∆𝐺𝑆𝑊𝑆 indicates the degree of
hydrophobicity (or hydrophilicity, when ∆𝐺𝑆𝑊𝑆 > 0) of the surface. To determine ∆𝐺𝑠𝑤𝑠 for
HAdV5 and lip balms (both dry and hydrated), contact angles of three probe liquids - DI
water, glycerol, and diiodomethane - on virus lawn and on lip balm surface were
determined using the sessile drop method (goniometer/tensiometer model 250, ramé-
hart). Contact angle values were calculated by DROPimage Advanced software based
on recorded droplets shapes. The droplet volume was in the 8 µl to 10 µl range. To
prepare a virus lawn, purified virus stock was filtered through a 50 kDa ultrafiltration
                                               29


membrane. The membrane coated with a multilayer cake of virions (> 4 monolayers)
was dried until the water contact angle on the membrane stabilized (≥ 6 h across all
samples) [21, 30, 33]. Lip balm surfaces were prepared by spin coating as described in
section 2.2. To hydrate lip balm, coated glass slides were immersed for 30 min in NaCl
solution with the ionic strength (150 mM) matching that of human saliva. The contact
angle tests in air were performed at the ambient temperature of 22 °C and the relative
humidity of 47 %.
Every contact angle measurement was repeated three times. Surface tension
components of the surface (𝛾𝑠𝐿𝑊 , 𝛾𝑠+ , 𝛾𝑠− ) were obtained by substituting measured
contact angles and known surface tensions of probe liquids into the Young-Dupré
equation (Eq. (1)) where 𝜃 is the contact angle of the probe liquid on the surface, 𝛾𝑙𝑇𝑂𝑇 is
the total surface energy, while 𝛾 𝐿𝑊 , 𝛾 + and 𝛾 − are Lifshitz-van der Waals (i.e. apolar),
electron acceptor, and the electron donor components of surface energy. Subscripts 𝑙
and 𝑠 refer to the probe liquid and the surface, respectively [31].
(1 + 𝑐𝑜𝑠𝜃)𝛾𝑙𝑇𝑂𝑇 = 2(√𝛾𝑠𝐿𝑊 𝛾𝑙𝐿𝑊 + √𝛾𝑠+ 𝛾𝑙− + √𝛾𝑠− 𝛾𝑙+ )                            (1)
The free energy of interfacial interaction in water, ∆𝐺𝑠𝑤𝑠 , was calculated based on the
surface energy component of the solid (virus or lip balm) and the tabulated values of the
surface energy components of water [31]:
                              2
∆𝐺𝑠𝑤𝑠 = −2 (√𝛾𝑠𝐿𝑊 − √𝛾𝑤𝐿𝑊 ) − 4(√𝛾𝑠+ 𝛾𝑠− + √𝛾𝑤+ 𝛾𝑤− − √𝛾𝑠+ 𝛾𝑤− − √𝛾𝑠+ 𝛾𝑤− )       (2)
                                               30


2.3.6 QCM-D studies of SiO2 and HAdV5 attachment to lip balm surfaces
Colloidal deposition onto a surface can be quantified using quartz crystal microbalance
with dissipation monitoring (QCM-D). This system allows real-time monitoring of the
changes in vibrational frequency due to mass deposition onto the quartz crystal sensor.
With the deposited mass proportional to the changes in resonance and overtone
frequencies, the amount of mass deposited can be computed based on the frequency
shift [34]. This technique can also offer information on the viscoelastic behavior of the
adsorbed layer by measuring dissipation [35]. QCM-D has been used to study
attachment of MS2 bacteriophage [19, 20, 36], human adenovirus [37] and pathogenic
plant viruses [38, 39] onto various surfaces such as clays [20], natural organic matter
[19, 34, 40], polyelectrolyte multilayers [25, 41], and household paints [37].
The QCM-D E4 system (Biolin Scientific) was used to quantify the deposition of colloids
onto the lip balm-coated QCM sensor surfaces. Prior to measurement, gold QCM-D
sensors were cleaned following the procedure recommended by the manufacturer (see
SM, section S2) and then mounted into the flow chamber to determine their resonance
frequency in air. This was followed by a 5-min measurement of resonance frequency to
establish a stable baseline. QCM-D tests were carried out at 25 °C in a continuous flow
mode (0.15 ml/min) using a digital peristaltic pump (IPC, four channels, ISMATEC). To
acquire QCM resonances, lip balm-coated sensors were first contacted with 150 mM
NaCl solution for at least 20 min until the baseline of the frequency signals was
                                              31


                 ∆𝑓
stabilized using    ≤ 0.025 Hz/min as the baseline criterion. The 150 mM solution
                 ∆𝑡
matched the continuous phase of the colloidal suspension in the QCM-D measurement
and had the ionic strength approximating that of human saliva (~ 136 mM; Table 3) [42].
In tests with SiO2 colloids, the sensors were challenged with silica suspensions of one
of two concentrations: 0.52 mg(SiO2)/ml or 1.05 mg(SiO2)/ml. In tests with HAdV5, the
concentration of HAdV5 in the feed was ~ 109 GC/mL. Based on the measured value of
HAdV5 hydrodynamic size (103 nm, see section 3.3) and the approximate virion density
(assumed 1.33 g/cm3 [43, 44]) the corresponding mass concentration of HAdV5 was
estimated to be 0.76 µg/mL. QCM frequency and dissipation were recorded every 1
min. The frequency shifts were fitted into the Sauerbrey equation [45] to compute the
mass change:
∆𝑚 = −𝐶∆𝑓/𝑛                                                                     (3)
where 𝐶 = 17.7 ng·Hz-1·cm-2 is the mass sensitivity constant, 𝑛 is the overtone number
and Δ𝑓 is the frequency shift (Hz). Mass data were calculated based on 3rd 5th
overtones.
2.3.7 XDLVO modeling of colloid interactions with lip balm
The interactions particles and surfaces include repulsive electrostatic interactions and
attractive van der Waals forces, which can be described by the Derjaguin-Landau-
Verwey-Overbeek (DLVO) theory of colloid stability [46, 47]. The classic DLVO theory,
however, showed only limited ability to predict nanoparticle adhesion to surfaces [19, 34,
36, 48]. Extended DVLO (XDLVO) model [49] builds on the DLVO theory by taking
                                             32


hydrophobic interactions into consideration. The XDLVO theory can predict interactions
of dissolved and colloidal materials with various surfaces [50, 51] and has been applied
to describe virus-surface interactions [21, 25, 33, 52]. The XDLVO theory describes the
total energy of interaction 𝐸𝑠𝑋𝐷𝐿𝑉𝑂1 𝑤𝑠2
                                          between two surfaces in an aqueous medium as a
sum of the Lifshitz-van der Waals, 𝐸𝑠𝐿𝑊      1 𝑤𝑠2
                                                   , electrostatic double layer, 𝐸𝑠𝐸𝐿
                                                                                    1 𝑤𝑠2
                                                                                          , an acid-base,
𝐸𝑠𝐴𝐵
   1 𝑤𝑠2
         , energies. When one of the surfaces is a virus:
  𝑋𝐷𝐿𝑉𝑂         𝐿𝑊     𝐸𝐿       𝐴𝐵
𝐸𝑣𝑤𝑠       = 𝐸𝑣𝑤𝑠  + 𝐸𝑣𝑤𝑠  + 𝐸𝑣𝑤𝑠                                                            (4)
The XDLVO approach extends the DLVO theory by taking hydrophobic interactions
(𝐸𝑠𝐴𝐵
    1 𝑤𝑠2
          ) into consideration. In the expression above,
               𝐴𝑎      𝑎        𝐿𝑊                                                           (5)
  𝐿𝑊
𝐸𝑣𝑤𝑠    = − 6𝑑 = 2𝜋 𝑑 𝑑02 ∆𝐺𝑦0      ,
where 𝑎 is the virus radius, 𝑑 is the separation distance, 𝑑0 is the minimum separation
                                                                   𝐿𝑊
distance (𝑑0 = 0.158 nm) [31, 53], and 𝐴 = −12𝜋𝑦02 ∆𝐺𝑦0               is Hamaker constant. Further,
  𝐸𝐿
                                  1 + 𝑒 −𝑘𝐷𝑑                                                  (6)
𝐸𝑣𝑤𝑠    = 𝜋𝜀𝑟 𝜀0 𝑎 [2𝜓𝑣 𝜓𝑠 𝑙𝑛 (              ) ]
                                  1 − 𝑒 −𝑘𝐷𝑑
                  + (𝜓𝑣2 + 𝜓𝑠2 ) 𝑙𝑛(1 − 𝑒 −2𝑘𝐷𝑑 )
                            𝑑0 − 𝑑                                                            (7)
  𝐴𝐵
𝐸𝑣𝑤𝑠   = 2𝜋𝑎𝜆∆𝐺𝑑𝐴𝐵    𝑒𝑥𝑝 (           )
                    0
                                 𝜆
where 𝜀𝑟 is the dielectric constant of water (𝜀𝑟 = 79), 𝜀0 is the relative permittivity in
vacuum (𝜀0 = 8.8541012 C·V-1·m-1), 𝜓𝑣 and 𝜓𝑠 are the surface potentials of the colloid
and surface respectively, 𝑘𝐷 is the reverse Debye length, λ is the characteristic delay
length of the AB interaction (λ = 0.6 nm) [53].
                                                      33


∆𝐺𝑑𝐴𝐵
    0
       = 2 √𝛾𝑠+ (√𝛾𝑠− + √𝛾𝑣− − √𝛾𝑤− ) + 2√𝛾𝑠− (√𝛾𝑠+ + √𝛾𝑣+ − √𝛾𝑤+ )
                                                                                          (8)
                              −2(√𝛾𝑠+ 𝛾𝑣− + √𝛾𝑠− 𝛾𝑣+ )
                                                                                          (9)
∆𝐺𝑑𝐿𝑊
    0
       = 2(√𝛾𝑤𝐿𝑊 − √𝛾𝑠𝐿𝑊 )(√𝛾𝑣𝐿𝑊 − √𝛾𝑤𝐿𝑊 )
where 𝛾 + is the electron acceptor component, 𝛾 − is the electron donor component and
𝛾 𝐿𝑊 is the apolar surface energy component. The surface energy components of the
surface (𝛾𝑠+ , 𝛾𝑠− , 𝛾𝑠𝐿𝑊 ) and the virus (𝛾𝑣+ , 𝛾𝑣𝐿𝑊 , 𝛾𝑣− ) are calculated using the Young-Dupre
equation and contact angle values of the probe liquids.
2.4 Results and Discussion
2.4.1 Optimal coating protocol and coating morphology
At each processing step, lip balm samples remained homogenous with no apparent
changes other than in their flowability. While phase separation may occur in similar (e.g.
recrystallization of cocoa butter leading to “fat blooms” on the surface of lip balms [54]
and chocolate [55]), in our study no phase separation was observed at any point during
sample preparation (melting, coating, drying). We attribute this to the relatively simple
composition of the lip balms used in this work (see SM, Table 4). Detailed rheological
studies would be necessary to explore possible structural and compositional changes in
depth [56].
                                                    34


Lip balm film morphology was assessed using SEM (Fig. 2). The coating thickness was
relatively constant across the coated area and had a homogeneous internal structure for
all lip balms but Vaseline. The thickness was estimated to be 1.15 ± 0.19 µm, 2.03 ±
0.12 µm, 1.07 ± 0.16 µm, and 3.18 ± 0.10 µm for Carmex, ChapStick, Burt’s Bees, and
Vaseline, respectively. The surface roughness of the coatings was not measured more
accurately (e.g. with atomic force microscopy). However, how the roughness affected
surface energy was captured by the measured values of apparent contact angles.
(Surface roughness enhances apparent hydrophobicity of surfaces with water contact
angle > 90 ° [57].) We also note that there may be a difference in the surface roughness
(and, therefore, surface energy) between lip balm coatings on a relatively flat surface
such as a glass slide, an SEM stub or a QCM-D sensor and that of a human lip. While
outside of this study’s scope, the effects of the morphology of the underlying surface on
the adhesiveness of personal care products should be explored in future work.
Addition of petroleum distillate as a solvent enabled spin-coating of the lip balm sample
but could also alter the physicochemical properties that affect adhesion. In this study, a
change in the surface energy of lip balm was used as a composite indicator of solvent-
induced alterations to the sample during the preparation process. The underlying
assumption was that the deviation of the surface energy from its baseline value
(measured for dry samples) was indicative of the presence of residual solvent. By
extension, return of the surface energy value to that of the baseline (protocol A) was
accepted as evidence of the removal of residual solvent. The optimal sample
preparation procedure was selected in tests with Carmex lip balm coated onto a glass
                                              35


slide surface. Based on measured values of contact angles of probe liquids, surface
energies of Carmex coatings made using protocols B, C, and D were compared with
that for a coating made using the “solvent-free” protocol A. Protocol B has ∆𝐺𝑠𝑤𝑠 values
matching those obtained by protocol A and, therefore, was selected as the coating
method to prepare lip balm surfaces for adhesion studies (Fig. 3).
2.4.2 Hydrophobicity and surface charge of lip balm-coated surfaces
Lip balms hydrated in 150 mM NaCl (model human saliva) had the surface energy in the
15.2 mJ/m2 to 38.4 mJ/m2 range. The energy penalty associated with the replacement
of the layer of water molecules bound at such hydrophilic surfaces makes them less
adhesive. Drying had a dramatic effect on the surface energy of lip balms (Fig. 4)
                                           36


                                (a)                                                      (b)
                               (c)                                                     (d)
Figure 2. Representative SEM images of a) Carmex b) ChapStick, c) Burt's Bees, and d) Vaseline coatings on a QCM-D
sensor.
                                                       37


converting them into strongly hydrophobic surfaces with 𝐺𝑠𝑤𝑠 values ranging from - 65.0
mJ/m2 to - 90.7 mJ/m2. Such reversal of the surface energy from positive to highly
negative should translate into a significant change in the adhesive properties of these
surfaces.
All four lip balms had a pH-dependent charge indicating presence of ionizable surface
groups. The 𝜁-potential of Carmex and Vaseline showed a steady decrease with an
increase in pH (Fig. 5a). The isoelectric point (IEP) for Carmex and Vaseline lip balm
samples were ~ 4.15 and ~ 4.0, respectively. For ChapStick and Burt’s Bees lip balms,
the charge had a more complex dependence on pH (Fig. 5b) with two IEPs for each
surface: 4.0 and 7.7 for ChapStick and 4.1 and 7.7 for Burt’s Bees. The origin of the
positive slope in the 𝜁 vs pH dependence for ChapStick and Burt’s Bees is unclear. We
speculate that this non-monotonous nature of 𝜁(pH) function is due to the presence of
water-soluble compounds in the lip balms with a pH dependent charge and solubility.
The observed increase 𝜁 with an increase in pH can stem from leaching of negatively
charged compounds at higher pH values. The irregular behavior was observed for only
two out of four lip balms and should be explored further. This is particularly important
given that the second IPE is close to the pH range of saliva.
2.4.3 Hydrodynamic size and surface charge of HAdV5 and SiO2 colloids
The particle size distribution of HAdV5 suspension had a single narrow peak at ~103 ±
1 nm (see SM, Fig. 10) indicating high purity of the stock [26, 27]. The 𝜁-potential of
                                             38


HAdV5 continuously decreases with an increase in pH (Fig. 5). The IEP of HAdV 5 is ~
4.5 consistent with the result reported by Trilisky and Lenhoff [58]. The 𝜁-potentials
measured at pH < 4.5 and pH > 8 were different from the values reported in that earlier
study likely due to differences in the background electrolyte (see SM, section S5).
The particle size distribution of ST-ZL (see SM, Fig. 10) features a single narrow peak
at ~138 nm, which is above the manufacturer-provided size range (70 to 100 nm).
Earlier studies also reported larger sizes for ST-ZL silica (139 to 153 nm [28, 59]). The
𝜁-potential of ST-ZL decreased with an increase in pH (Fig. 13) and remained negative
over the pH range measured in this study, which is consistent with the result reported by
Kim et al. [59]. The 𝜁-potentials measured at pH < 4 and pH > 8 were different from the
result reported by Kim et al. likely due to different background electrolytes used.
                                              39


Figure 3. Free energy of interfacial interaction in water, ∆Gsws , of the Carmex lip balm as a function of the coating protocol
(see Fig. 1). Measured contact angles of probe liquids and calculated surface energy parameters used to compute ∆Gsws
values are given in Table 6. Each measurement was done in triplicate. Errors correspond to standard deviations.
                                                              40


Figure 4. Free energy of interfacial interaction in water, ∆𝐆𝐬𝐰𝐬 , of the four lip balms coated using protocol B (see Fig. 1) in
dry and hydrated states. Measured contact angles of probe liquids and calculated surface energy parameters used to
compute ∆𝐆𝐬𝐰𝐬 values are given in Table 5. Each measurement was done in triplicate. Errors correspond to standard
deviations.
                                                              41


                             (a)                                                         (b)
Figure 5. Surface charge of HAdV5 virions and lip balm surfaces as a function of pH. Vertical dashed lines indicate pH
values (4.2 and 7.2) used in QCM-D measurements of HAdV5 attachment to lip balm surfaces. Lines connecting
experimental data points are added to the guide the eye. Average and standard deviations for lip balm samples are based
on four independent measurements for each lip balm. Average and standard deviations for HAdV5 are based on three
independent measurements. HAdV5 data is shown in both graphs to add data interpretation.
                                                           42


2.4.4 Attachment of viruses and SiO2 particles to lip balm-coated surfaces:
XDLVO predictions
XDLVO modeling of colloid-lip balm interaction was performed for five colloids: two
human viruses (HAdV5, HAdV40), two bacteriophages (MS2, P22) and particulate SiO 2.
Colloid properties required as inputs to XDLVO model included particle size, ζ-potential
and surface energy. For HAdV5, all these characteristics were measured (section 4.3).
To our knowledge, this is the first report of surface energy (or any other metric of
hydrophobicity) for HAdV5. For SiO2, size and charge were also determined in this
study while the surface energy value was calculated based on contact angles of three
probe liquids as reported by Zdziennicka et al. [60]. For HAdV40, P22 and MS2, the
values were taken from the literature [21, 25, 30, 60] (Table 1).
                                              43


Table 1. Size, charge and hydrophobicity of colloids considered in this work. The three
properties are quantified in terms of hydrodynamic diameter (𝒅𝒉 ), zeta-potential (𝜻), and
free energy of interfacial interaction in water (∆𝑮𝒔𝒘𝒔 ), respectively.
Notes:
 A This study (see SM, Fig. );B Shi et al. [21]; C Shi and Tarabara [30]; D Calculated
based on contact angle data reported by Zdziennicka et al. [60]. The error is not
provided because the original contact angle values are reported as averages only; E
This study (see SM, Table 5); F Dang and Tarabara [25]; G This study (see SM, Fig. 13);
H This study (see Fig. 5)
                                                Colloid type
   Property
                   SiO2               Human viruses                   Bacteriophages
                 particles         HAdV5          HAdV40             P22           MS2
  𝑑ℎ , nm     137.9 ± 0.4 A      103 ± 1.3 A     98 ± 3.0 B      54 ± 1.3 C     27 ± 0.4 C
                  (n=10)           (n=10)          (n=10)          (n=10)         (n=10)
  𝜁, mV (at     -25 ± 2.9 G       7 ± 0.8 H      -8 ± 1.8 B      -19± 1.3 C     -31± 1.3 C
  pH 4.2)         (n=10)           (n=30)          (n=10)          (n=10)         (n=10)
  𝜁, mV (at     -36 ± 2.8 G      -18 ± 0.4 H     -29 ± 4.7 B     -47 ± 0.7 C    -47 ± 0.9 C
  pH 7.2)         (n=10)           (n=30)          (n=10)          (n=10)         (n=10)
  ∆𝐺𝑠𝑤𝑠 ,          12.8D        -27.7 ± 1.1 E  -30.4 ± 6.5 B   -6.3 ± 11.0 C   48 ± 15.3 F
  mJ/m2                             (n=3)           (n=3)           (n=3)          (n=3)
                                              44


Two pH values (4.2 and 7.2) were selected. pH 7.2 represented that of saliva of a
healthy person (6.2 to 7.6 typical range [61]). pH 4.2 was chosen to explore HAdV5
deposition under different conditions of virus-balm electrostatic interaction (Fig.5). In
what follows we present results of XDLVO modeling for two colloid/lip balm pairs
(SiO2/Burt’s Bees and HAdV5/Carmex). Results for the other colloid/lip balm pairs are
given in SM (Fig. 14 – 31).
Figure 6 shows XDLVO energy profiles for Burt’s Bees-HAdV5 interaction. For dry lip
balms, the XDLVO model indicates that at both pH values (4.2 and 7.2) the total energy
                                           𝑑𝐸𝑡𝑜𝑡
of interaction is attractive: 𝐸𝑡𝑜𝑡 ≤ 0 and       ≥ 0 (Fig. 6a, 6c). Figure 7 illustrates
                                            𝑑𝑟
XDLVO energy profiles for Burt’s Bees-SiO2 interaction. For dry balms (Fig. 7a, 7c), the
trends were the same as for HAdV5. In fact, for all five colloids and for both pH values,
the overall interaction of colloids with dry lip balm is always attractive (Fig. 14 – 31). The
main reason for the favorable interaction is the high hydrophobicity of dry lip balms (Fig.
4) and the resulting strong short-range hydrophobic attraction.
                                               45


Table 2. Values (in units of kT) of the primary maximum (𝑬𝒎𝒂𝒙 ) and secondary minimum
(𝑬𝒎𝒊𝒏 ) in the XDLVO energy profile for colloid interaction with hydrated lip balms.
 Lip balm      pH      HAdV5         HAdV40           P22          MS2           SiO2
 (hydrated)        𝐸𝑚𝑎𝑥    𝐸𝑚𝑖𝑛   𝐸𝑚𝑎𝑥    𝐸𝑚𝑖𝑛    𝐸𝑚𝑎𝑥    𝐸𝑚𝑖𝑛  𝐸𝑚𝑎    𝐸𝑚𝑖𝑛  𝐸𝑚𝑎𝑥   𝐸𝑚𝑖𝑛
 Burt’s Bees   4.2    -      -     10.9    -3.1   221.7   -1.1 381.4   -0.5  1099    -2.6
 Burt’s Bees   7.2    -      -       -       -    158.0   -1.3 378.4   -0.5  1191    -2.6
 Carmex        4.2  13.7   -3.9    29.8    -2.9   262.5   -1.2 445.3   -0.5  1448    -2.6
 Carmex        7.2   5.2   -3.1    68.6    -2.2   293.0   -1.0 451.0   -0.5  1337    -2.5
 ChapStick     4.2    -      -     45.7    -1.9   257.0   -0.8 403.5   -0.4  1192    -1.8
 ChapStick     7.2  13.4   -2.8    12.6    -2.3   139.5   -0.9 272.7   -0.4   922    -1.9
 Vaseline      4.2 337.7   -2.2   433.3    -1.7   477.0   -0.9 447.0   -0.4  1562    -2.2
 Vaseline      7.2 388.0   -2.0   488.0    -1.6   526.0   -0.8 493.0   -0.4  1673    -2.0
                                               46


     (a)                                                   (b)
     (c)                                                   (d)
Figure 6. XDLVO energies of the interaction of human adenovirus 5 with Carmex lip balm at pH 4.2 (a, b) and pH 7.2 (c,
d) in dry (a, c) and hydrated (b, d) states. DLVO and XDLVO energy profiles for HAdV5 interaction with the other three lip
balms (Burt’s Bees, ChapStick and Vaseline) are given in SM. DLVO total energy is provided for reference.
                                                           47


   (a)                                                           (b)
   (c)                                                           (d)
Figure 7. DLVO and XDLVO energies of the interaction of SiO2 particles with Burt’s Bees lip balm at pH 4.2 (a, b) and pH
7.2 (c, d) in dry (a, c) and hydrated (b, d) states. DLVO and XDLVO energy profiles for HAdV5 interaction with the other
three lip balms (Carmex, ChapStick and Vaseline) are given in SM. DLVO total energy is provided for reference.
                                                              48


For hydrated lip balms, XDLVO modeling gives a more nuanced prediction. Across the
20 colloid/lip balm combinations evaluated in this work, interaction energy profiles
covered a range 𝐸𝑚𝑎𝑥 (zero to thousands kT) and 𝐸𝑚𝑖𝑛 (zero to ~ 4 kT) values (Table 2;
Fig. 6, Fig. 7, A14-A21). Low 𝐸𝑚𝑎𝑥 (< several kT) or no primary energy barrier points to
the likely irreversible attachment into the primary minimum. If the primary energy barrier
(𝐸𝑚𝑎𝑥 ) is high, deposition into the secondary minimum, provided it exists, is possible.
Deposition into a secondary minimum is reversible and is stronger (or less reversible)
for deeper minima. In the case of SiO2, hydrating Burt’s Bees lip balm flipped its short-
range interaction with the colloids from strongly attractive to strongly repulsive (Fig. 7a
vs Fig. 7b, Fig. 7c vs Fig. 7d). At the same time, hydration did not have the same
dramatic effect on the lip balm’s interaction with HAdV5 (Fig. 6a vs Fig. 6b, Fig. 6c vs
Fig. 6d). The contrast in the predicted total energy of interaction with hydrated lip balm
(attractive interaction with HAdV5 virions versus strongly repulsive interaction with SiO2
colloids) underscores the importance of the physicochemical properties of the colloids.
2.4.5 Attachment of SiO2 colloids and HAdV5 virions to lip balm-coated surfaces:
Preliminary QCM-D study
To our knowledge, there have been no prior studies of particle adhesion to lip balms or
other similar lipophilic personal care products. Establishing the experimental
methodology for such measurements was one of the goals of this work. To study
deposition onto dry lip balms one would need to work with aerosols and address the
number of issues related to the transient processes of droplet evaporation and hydration
                                              49


of the surface upon droplet deposition. The preliminary tests performed in this study
were restricted to particle deposition from an aqueous solution onto hydrated lip balms.
Longer term QCM-D deposition data were obtained for two colloid/lip balm pairs:
SiO2/Burt’s Bees and HAdV5/Carmex.
Silica was selected to avoid any uncertainty related to any other suspended materials
possibly present in the feed stock. The conditions of the QCM-D tests with SiO2/Burt’s
Bees pair were not conducive to adhesion relative to other pairs (e.g. involving HAdV5
and 40) and could be viewed as a conservative estimate of the extent of particle
adhesion to lip balms.
The first test was performed using a 1.05 mg(SiO2)/ml suspension (Fig. 8a). QCM-D
frequency data indicated significant deposition (Fig. 8). The monotonous increase in the
dissipation signal was consistent with the deposition of colloids onto the lip balm.
                                             50


     (a)
     (b)
Figure 8. QCM-D measurements of the deposition of SiO2 colloids onto hydrated Burt’s
Bees lip balm under conditions when SiO2 loading is either a) constant (1.05 mg/ml) or
b) increases stepwise from 0.52 mg/ml to 1.05 mg/ml. In both tests, deposition occurs
from 150 mM NaCl electrolyte at pH 7.2. The mass values are calculated based on
Sauerbrey equation (eq. (3)) with 𝒏 = 3. The results for 𝒏 = 5 are shown in SM (Fig. 32).
The total mass of deposited SiO2 (65 µg) was ~ 0.7 % of the mass flown (~ 9.4 mg) over
the lip balm-coated sensor. In early stages of the experiment when the lip balm surface
was relatively SiO2-free, the deposition was determined by SiO2-lip balm interactions.
The surface loading of 74 µg(SiO2)/cm2 recorded ~ 1 h into QCM-D test (Fig. 8a) is
equivalent to ~ 3 monolayers of colloids. We speculate that this relatively large amount
                                            51


of SiO2 deposited despite repulsive SiO2-SiO2 interactions can be due to the partial
“burial” of SiO2 in the soft hydrated surface of the lip balm. At pH 7.2, SiO2 and Burt’s
Bees balm have surface charges of opposite signs (-36 mV and 10 mV, respectively;
Fig. 5b) so that electrostatic interactions between these surfaces are favorable. The
hydrophobic interactions are strongly repulsive because SiO2 and hydrated Burt’s Bees
are both hydrophilic (Table 1, Fig. 4). As a result, XDLVO predicts a very high (> 1000
kT) primary barrier and a shallow (~ 2.6 kT) secondary minimum (Table 3, Fig. 7d). We
conclude that SiO2 deposition occurred into the secondary minimum and should be
reversible.
In a companion test (Fig. 8b) on SiO2 deposition, the lip balm-coated sensor was
charged for 30 min with a two-times less concentrated SiO2 suspension (0.52
mg(SiO2)/ml) before reverting to the same feed as in test 1 (1.05 mg(SiO2)/ml). A much
weaker deposition was observed throughout the test. At the end of the first stage, the
deposit was a submonolayer with the average distance between deposited particles of ~
0.38 µm (~ 2 particle diameters). At the end of the long (160 min) second stage, the
mass of deposited particles was smaller than in the first test even though the total mass
of SiO2 introduced into the QCM-D chamber in test 2 was higher. The total mass of SiO2
deposited during test 2 (38 µg) was ~ 0.14% of the mass flown (27.5 mg) over the lip
balm-coated sensor. We conclude that the deposition history is important and that the
total mass load of colloids that a lip balm surface is exposed to cannot be a sole
predictor of the extent of deposition.
                                              52


QCM-D tests with HAdV5 were done using as-purchased highly purified stock with
HAdV5 concentration of 109 GC/mL and Carmex lip balm. Even when undiluted, the
mass concentration of HAdV5 stock (~ 0.76 µg/mL) was ~ 680 times smaller than that of
the silica suspension in QCM-D tests with SiO2. Indeed, a much smaller deposited mass
was measured for HAdV5, accompanied by a weaker dissipation signal (Fig. 9). The
total mass of deposited HAdV5 (~ 2.4 µg) was ~ 10.6 % of the mass flown (~ 22.8 µg)
over the lip balm-coated sensor. The % deposited value is significantly higher than that
for SiO2 colloids, which is consistent with the much smaller primary energy barrier, 𝐸𝑚𝑎𝑥 :
5.2 kT for HAdV5/Carmex versus 1337 kT for SiO2/Burt’s Bees (Table 2). The
secondary minimum, 𝐸𝑚𝑖𝑛 , was also deeper for HAdV5/Carmex (- 3.1 kT) than for
SiO2/Burt’s Bees (- 2.6 kT). We attribute the stronger (relative to the mass loading)
deposition of HAdV5 to a more likely association with the lip balms surface through the
secondary energy minimum and a possible irreversible attachment into the primary
minimum.
                                             53


Figure 9. QCM-D measurements of the deposition of human adenovirus 5 onto
hydrated Carmex lip balm from 150 mM NaCl electrolyte at pH 7.2. HAdV5
concentration in the QCM-D feed is ~109 GC/mL (~ 0.76 µg/mL). The mass values are
calculated based on Sauerbrey equation (eq. (3)) with 𝒏 = 3. The results for 𝒏 = 5 are
shown in SM (Fig. 33).
Higher mass concentrations in the feed are required for a higher QCM-D signal. To
more accurately assess the mass flux towards the sensor surface for a given mass
concentration in the feed, one would need to quantify the mass transfer of colloids to the
depositional plane. This requires solving the Graetz problem of diffusion-limited
transport to a surface from a crossflow [62, 63]. In the absence of such solution, the
best approach is to employ virus stocks with as high virus titer as possible while still of
high purity.
                                            54


2.4.6 Implications for virus control and public health protection
The results reported in this work indicate that dry lip balm can serve as a “hot spot” for
virus deposition. Given the intended application of lip balms, there is a clear risk to the
health of individuals using these products. Low humidity environments (e.g. typical for
long air travel) are of particular concern as they promote dehydration. Designing
materials to retain surface moisture is one possible approach to staving off adhesion of
colloids to lip balms and similar products. A multilayer design with a lipophilic core and a
hydrophilic outer layer is one possible strategy.
Likely contagion scenarios should be identified and studied. Possible routes include
ingestion of lip balm with associated viruses, ingestion of saliva laden with viruses
detached from lip balm, and breathing in viruses resuspended from the lip balm surface
into the flow of inhaled air. Given the importance of surface interactions, there is likely a
difference between non-enveloped viruses (e.g., adenoviruses, coxsackieviruses,
rotavirus) and enveloped viruses (influenza H1N1, human coronaviruses, herpesviruses,
hepatitis C) in their propensity to adhere to a lipophilic surface. Future work should
explore deposition from other relevant media (e.g. air, respiratory fluid) and onto other
surfaces (face and hand creams) as well as virus resuspension into saliva, common
drinks (e.g. low pH sodas, milk) and relevant fluid flows (e.g. breathed air).
                                              55


2.5 Conclusions
The study reports a protocol for preparing lip balm coatings to enable charge and
surface energy measurements as well as adhesion studies with these materials.
Surface charge and hydrophobicity were determined for four brands of lip balms. Also
measured were size, charge and hydrophobicity of human adenovirus 5. The measured
values were used in XDLVO modeling of adenovirus adhesion to lip balms. Adhesion of
four other colloids (HAdV40, MS2 and P22 bacteriophages and SiO2) spanning a range
of sizes, charges and surface energies was also evaluated.
The study tested the hypothesis that a drying-induced increase in lip balm
hydrophobicity enhances virus adhesion due to strong hydrophobic colloid-surface
interactions. Indeed, drying was shown to result in a dramatic decrease of surface
energy (𝛿(∆𝐺𝑠𝑤𝑠 ) ≥ 83.0 mJ/m2) of lip balms making their surfaces highly hydrophobic.
XDLVO modelling predicts that attachment to a dry lipstick (∆𝐺𝑠𝑤𝑠 < - 65 mJ/m2) is
highly favorable as a result of strong short-range hydrophobic attraction. Lip balms
hydrated in a solution with the ionic strength of human saliva are hydrophilic (∆𝐺𝑠𝑤𝑠 > 15
mJ/m2) and resist colloid attachment. Physicochemical properties of colloids are also
important. Adhesion occurs into shallow secondary minima for hydrophilic colloids such
as SiO2, MS2 and P22. Because of the hydrophobicity of adenoviruses, primary maxima
in XDLVO profiles are low or non-existent making irreversible deposition into primary
minima possible. Preliminary QCM-D tests with SiO2 colloids and human adenovirus 5
confirm deposition even on a hydrated lip balm.
                                             56


Prior work focused on virus adhesion to human skin [2-8], or, in two instances [6, 7],
employed freshly applied Vaseline as a adhesion barrier for MS2. The present work
extends these earlier investigations to study virus adhesion to lipophilic personal care
products. The proposed methodology can help direct the compositional design of lip
balms and similar materials and develop usage guidelines to minimize virus adhesion.
Future work should explore deposition from other relevant media (e.g. air, respiratory
fluid) and onto other surfaces (e. g. hand creams) as well as virus resuspension into
saliva, common drinks (e.g. low pH sodas, milk) and relevant fluid flows (e.g. breathed
air).
Acknowledgements
This material is based upon work supported in part by the National Science Foundation
Partnerships for International Research and Education program under Grant IIA-
1243433 and in part by the Center for European, Russian and Eurasian Studies at
Michigan State University. We thank Dr. Hang Shi (CTI and Associates, Inc.), Dr. Irene
Xagoraraki (Civil and Environmental Engineering, Michigan State University) and Dr.
Kristin Parent (Biochemistry and Molecular Biology, Michigan State University) for
useful discussions as well as to Dr. Wei Zhang (Department of Plant, Soil and Microbial
Sciences, Michigan State University) for providing access to Malvern Zetasizer Nano ZS.
We are also thankful to Mrs. Carol Flegler (Center for Advanced Microscopy, Michigan
State University) for recording SEM images and advising on appropriate sample
                                            57


preparation procedures.
                        58


APPENDIX
   59


                                        APPENDIX
A1. SEM imaging of lip balm coatings
The low vacuum (LV) mode enables examination of non-conducting samples or volatile
samples up to a vapor pressure of 270 Pa [64]. While the main components of lip balm
are petrolatum and petroleum, many other ingredients may be present (Table 4). The
vapor pressure of petrolatum is 1.3 Pa, while for petroleum the value is in the 30 to 60
Pa range [65]. These values fall within the safety range of the JEOL 6610 LV operated
in the LV mode. However, lip balm may contain components (e.g. volatile oils) with
vapor pressures above the upper limit of the LV range. To avoid potential damage to the
microscope, osmium tetroxide (OsO4) vapor fixation was employed to stabilize lip balm
samples. OsO4 reacts with the double bonds of lipids and oils and increases their vapor
pressure [66]. After 5 days of exposure to OsO4 vapor, lip balm samples turned slightly
darker, which indicated that the reaction did occur.
A2. QCM-D sensor cleaning procedure
QCM-D sensors were cleaned by following the procedure recommended by the
manufacturer. After each experiment, the sensors were soaked in a 5:1:1 mixture (RCA-
1 cleaning protocol [67]) of DI water, hydrogen peroxide (30 wt%) and ammonia (25
wt%) at 75 °C for 5 min, sonicated in DI water overnight, rinsed thoroughly with DI water,
dried with high-purity N2and treated in a UV/ozone chamber (UV/ozone Procleaner™ ,
Bioforce Nanosciences) for 60 min.
                                            60


A3. Additional information on Human Adenovirus 5 (ATCC® VR­1516™)
The Adenovirus Reference Material [68] contains purified Adenovirus Type 5 formulated
as a sterile liquid in 20 mM TRIS, 25 mM NaCl, 2.5% glycerol, pH 8.0 at room
temperature, and stored frozen at -70 °C. The U.S. Food and Drug Administration has
recommended using the Adenovirus Reference Material when characterizing adenoviral
gene therapy products [69]. As stated by the supplier, the viability “is warranted for 30
days from the date of shipment” [70]. In all experiments described in this work, the virus
was used within this time frame.
A4. Protocols of lip balm processing evaluated in the study
To produce a thin and homogenous layer of coating on sensors for QCM-D
measurement, four different preparation protocols A, B, C, D were designed and
evaluated. In protocol A (Fig. 1), a grain of lip balm was cut, put it in a vial, and melted
on a heater. After the sample was liquified, it was coated on the QCM-D sensors.
However, because of a very high viscosity of the melted sample it was difficult to
produce a homogenous layer with a proper thickness for QCM-D measurement.
Because of a high solidification point of the sample. it turned solid quickly upon contact
with the sensor surface. It was also found that heating the sensor prior to the contact
with the melted lip balm did not help achieve coatings of acceptable quality. In order to
prepare a lip balm sample with appropriate viscosity for spin-coating, a solvent was
                                              61


needed. Of the five solvents evaluated (ethyl acetate, isopropanol, methanol, pentane,
hexane, and petroleum distillate), petroleum distillate yielded the most homogenous lip
balm solution. After lip balm was melted and mixed with petroleum distillate, the mixture
was stirred for 24 h (Protocol C, Fig. 1). Protocol C successfully produced lip balm
solution with proper viscosity for spin-coating. However, the added material changed the
surface free energy of lip balm sample (Fig. 11). To solve this, the mixture with lip balm
and solvent was left in the fume hood for 5 days to remove the residual solvent. The
comparison of surface free energy of samples processed with protocol A and B (Fig. 11)
indicated that drying in the fume hood for 5 days can effectively remove the solvent. The
impact of filtration was also evaluated. Filtration with 0.45 μm filter after dissolution was
found to enhance the homogeneity of the sample-solvent mixture, which makes the
sample layer on QCM-D sensors more homogenous (Protocol D). However, it was
found that majority of the sample was rejected by the 0.45 μm filter, which indicated that
filtration removed most of the components from the sample. Therefore, protocol D was
not employed for the sample processing.
A5. 𝜻-potential measurements of HAdV5
In the study by Trilsky and Lenhoff [8], the following 20 mM buffers were used in
measurements of ζ-potential and aggregation of HAdV5 as a function of pH: acetate
buffer for pH 3.8, 4.5 and 5; MES for 5.5 < pH < 6.7; HEPES for pH 7.8 and for pH 7.5
and 8.5; ethanolamine for pH 8.5 and 10. In that study, no additional salt was employed
in order to keep the ionic strength (I) minimal because ζ peaks widen with increasing I.
                                              62


Also, at a lower I, the likelihood of electrolysis of the gold-plated electrodes of the folded
capillary cell is lower. The gold plating from the electrodes tends to contaminate the
sample at low pH and/or high. Trilsky and Lenhoff determined the ζ potential of HAdV5
using light scattering (Zetasizer Nano ZS particle analyzer, Malvern Instruments
ZEN3600) and folded capillary cells (Malvern DTS1060). The ζ potential was found from
the electrophoretic mobility using Henry’s equation [71].
In contrast, our study employed 0.1 mM KCL as the background electrolyte. The pH of
virus samples was adjusted by KOH and HCl. Firstly, 10 ml mixture of 0.1 mM KCl
solution and HAdV5 sample were added into two sterile beakers. Both beakers had 10
ml mixture. Then the pH of the solution in the first beaker was adjusted to 7, 6, 5, 4, 3,
and the pH in the second beaker was adjusted to 8, 9, 10 using KOH and HCl. The
mobility and ζ -potential of the virus sample was measured (Zetasizer Nano ZS)
immediately after each pH adjustment.
A6. XDLVO and DLVO modeling of colloid-lip balm interactions
The DLVO and XDLVO modeling was performed for each colloid/lip balm combination
Results for HAdV5/Burt’s Bees and SiO2/Burt’s Bees are given in Figures 6 and 7 in the
main manuscripts. Results for the other 18 combinations are given in Figures 13 – 30.
Table 7 can be used as a guide to locate the Figures.
                                               63


Table 3. Concentrations of various electrolytes in human saliva [9].
           𝐶𝑚𝑖𝑛 ,    𝐶𝑚𝑎𝑥 ,     𝐶𝑎𝑣𝑒 ,              𝐶𝑎𝑣𝑒 𝑍 2,
                                            𝑍
            mM        mM        mM                    mM
 Na+              2       21      11.5          1       11.5
 K+            10         36         23         1          23
 Ca2+         1.2        2.8           2        2            8
 Mg2+        0.08        0.5      0.29          2       1.16
 Cl-              5       40      22.5         -1       22.5
 HCO3-         25         25         25        -1          25
 PO43-        1.4         39      20.2         -3      181.8
                                           64


Table 4. Composition of the four lip balms evaluated in this work.
Notes:
A     Components marked as “main ingredients” by lip balm manufacturers.
B     Burt Bees includes Lanolin, tocopherol, rosemary, limonene.
      ChapStick includes jojoba esters, caprylic / capric triglyceride, tocopheryl acetate, shea butter, octyldodecanol,
      tocopherol, glyceryl stearate.
      Carmex includes lanolin, cetyl esters, theobroma cacao seed butter, paraffinum liquidum, enthol, salicylic acid,
      phenol, vanillin.
                                                            Components
    Lip balm                 Solvents                                                          Flavor or
                                                   OilsA    WaxA     Camphor      Vitamin E                 Other
                   PetrolatumA      PetroleumA                                                 fragrance
    Burt’s Bees                                      +        +                        +
                                                                                                              See notes B
    ChapStick            +                           +        +                                     +
    Carmex               +                                    +           +                         +
    Vaseline                             +                                                                   None
                                                               65


Table 5. Contact angles (measured), surface energy parameters (calculated), and free interfacial energy of surface-
surface interaction in water (calculated) for dry and hydrated lip balm samples and for human adenovirus 5. Lip balm
coatings were prepared according to Protocol A (see Fig. 1).
                                Contact angles, 𝜃 (°)                   Surface energy parameters, 𝛾 (mJ/m2)                ∆𝐺𝑠𝑤𝑠
             Surface
                             H2O      Glycerol      DID        𝛾 𝐿𝑊          𝛾+        𝛾−          𝛾 𝐴𝐵          𝛾 𝑡𝑜𝑡     (mJ/m2)
               Carmex      55.2±1.5   69.0±0.6    37.8±1.0   40.7±0.5        0      38.8±2.7        0          40.7±0.5   17.9±5.0
  hydrated
              ChapStick    60.0±1.5   72.3±0.6    49.5±0.7   34.6±0.4        0      35.4±2.7        0          34.6±0.4   15.3±5.1
             Burt’s Bees   58.0±0.6   72.5±0.9    40.8±0.9   39.2±0.5        0      38.2±1.6        0          39.2±0.5   17.7±3.1
              Vaseline     55.5±1.7   78.4±0.5    45.2±0.7   36.9±0.4        0      51.1±3.5        0          36.9±0.4   38.5±6.5
               Carmex      77.3±1.3   57.1±0.7    26.4±0.7   45.6±0.3    0.8±0.2    2.8±0.8      2.9±1.3       48.6±1.3   - 65.0±4.0
              ChapStick    86.9±1.4   68.7±0.4    42.4±0.8   38.4±0.4    0.5±0.1    1.4±0.6      1.6±0.7       40.0±0.8   - 72.2±4.5
  dry
             Burt’s Bees   96.7±1.1   77.0±0.6    40.8±1.7   39.2±0.9    0.1±0.1    0.1±0.1      0.2±0.0       39.4±0.9   - 94.7±4.2
              Vaseline     90.9±1.8   72.1±0.5    30.7±0.8   43.9±0.3    0.1±0.0    0.6±0.5      0.3±0.5       44.3±0.6   - 90.7±6.3
             HAdV5         72.3±0.4   70.5±1.2    28.5±1.0   36.1±1.9        0      14.9±1.2        0          36.1±1.9   - 27.7±1.1
                                                                        66


Table 6. Contact angles (measured), surface energy parameters (calculated), and free interfacial energy of surface-
surface interaction in water (calculated) for three types of lip balm (dry) with four different preparation protocols (A, B. C,
and D.) The protocols are described in Fig. 1. Values for protocol A duplicate corresponding values (protocol A) given for
dry lip balms in Table 5.
                 Protoc        Contact angles, 𝜃 (°)                        Surface energy parameters, 𝛾 (mJ/m2)                 ∆𝐺𝑠𝑤𝑠
 Surface
                                                                  𝐿𝑊            +          −           𝐴𝐵                𝑡𝑜𝑡    (mJ/m2)
                   ol       H2O      Glycerol      DID        𝛾             𝛾          𝛾           𝛾                 𝛾
                   A      77.3±1.3   57.1±0.7    26.4±0.7   45.6±0.3    0.8±0.2      2.8±0.8     2.9±1.3           48.6±1.3    - 65.0±4.0
    Carmex
                   B      70.8±1.0   59.1±0.5    32.6±0.8   54.1±1.3        0        7.7±1.0           0           54.1±1.3    -60.3±2.0
                   C      73.2±0.5   69.3±1.1    25.1±1.0   38.7±1.8        0       12.5±1.1           0           38.7±1.8    -35.4±1.3
                   D      70.5±0.6   67.6±1.4    25.1±0.5   39.4±2.3        0       14.5±1.5           0           39.4±2.3    -30.3±1.5
                   A      86.9±1.4   68.7±0.4    42.4±0.8   38.4±0.4    0.5±0.1      1.4±0.6     1.6±0.7           40.0±0.8    - 72.2±4.5
    ChapStick
                   B      73.1±0.9   57.9±3.2    32.6±0.8   58.2±2.8        0        5.0±0.9           0           58.2±2.8    - 74.2±3.1
                   C      70.9±1.7   65.1±1.7    26.1±0.7   39.2±0.9        0       11.8±1.5           0           44.0±1.6    - 40.3±1.7
                   D      66.6±2.3   60.2±1.7    29.4±1.3   44.5±0.6   0.05±0.1     13.5±2.9    1.7±29.7           46.1±29.7   - 34.5±7.5
                   A      96.7±1.1   77.0±0.6    40.8±1.7   39.2±0.9    0.1±0.1      0.1±0.1     0.2±0.0           39.4±0.9    - 94.7±4.2
    Burts Bees
                   B      95.7±0.9   78.6±1.0    41.6±1.7   40.4±1.9        0        0.4±0.3           0           40.4±1.9    -95.3±3.9
                   C      76.9±1.0   76.4±0.1    32.9±0.3   30.2±0.6        0       14.1±1.3           0           30.2±0.6    -27.4±1.0
                   D      84.8±2.5   106±0.9     35.9±2.8   3.8±0.8         0       36.9±5.9           0            3.8±0.8     5.8±2.9
                   A      90.9±0.2   72.1±3.9    30.7±0.8   43.9±0.3   0.05±0.2     0.57±0.6    0.32±0.6           44.3±0.7    -90.7±9.7
    Vaseline
                   B      84.5±0.8   56.9±2.3    29.1±0.8   44.6±0.3    1.6±0.5     0.22±0.2     1.2±0.7           45.8±0.8    -77.5±5.1
                   C      71.5±2.5   63.9±1.1    25.1±0.9   46.3±2.7        0       10.3±2.9           0           46.3±2.7    -46.4±3.7
                   D      73.1±2.4   67.6±2.8    25.1±0.5   41.4±5.0        0       11.2±3.7           ±           41.4±5.0    -40.5±4.6
                                                                       67


Table 7. Guide to Figures with results of XDLVO modeling. Two of the 20 Figures are
shown in the main manuscript while the other 18 are shown in the SM file.
                                               Lip balms
                      Carmex      ChapStick         Burt’s Bees   Vaseline
             HAdV5    Figure 6    Figure 15         Figure 14     Figure 16
             HAdV40   Figure 17   Figure 18         Figure 19     Figure 20
  Colloids
             P22      Figure 21   Figure 22         Figure 23     Figure 24
             MS2      Figure 25   Figure 26         Figure 27     Figure 28
             SiO2     Figure 29   Figure 30         Figure 7      Figure 31
                                              68


Figure 10. Particle size distribution in the stock HAdV5 and ST-ZL SiO2 suspensions.
Lines are added to guide the eye.
                                              69


                      (a)                                       (b)
                      (c)                                       (d)
Figure 11. Free energy of interaction in water for three types of lip balm as a function of
surface preparation method (see Fig. 1).
                                           70


                       (a)                                      (b)
                       (c)                                      (d)
Figure 12. Free energy, ∆𝐆𝐬𝐰𝐬 , for four lip balms as a function of drying time: a)
Carmex; b) Chapstick, c) Burt’s Bees, d) Vaseline. Lip balms coatings were prepared
using protocol B. In each graph, the solid line corresponds to the surface prepared
using protocol A (melted lip balm with no dissolution and drying steps involved), which
was used as the baseline: a) 65.03 mJ/m2 for Carmex, b) -72.18 mJ/m2 for Chapstick,
c) -95.62 mJ/m2 for Burt’s Bees. d) -90.71 mJ/m2 for Vaseline. Measured contact angles
used to calculate ∆𝐆𝐬𝐰𝐬 values are given Table 6.
                                             71


Figure 13. Surface charge of ST-ZL SiO2 colloids as a function of pH. Vertical dashed
lines indicate pH values (4.2 and 7.2) used in QCM-D measurements of SiO2
attachment to lip balm surfaces. Lines connecting experimental data points are added to
the guide the eye.
                                           72


     (a)
                                                                    (b)
     (c)                                                            (d)
Figure 14. XDLVO energy of interaction of human adenovirus 5 with Burt’s Bees lip balm at pH 4.2 (a, b) and pH 7.2 (c,
d) in dry (a, c) and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                              73


     (a)                                                              (b)
              (c)                                                     (d)
Figure 15. XDLVO energy of interaction of human adenovirus 5 with ChapStick lip balm at pH 4.2 (a, b) and pH 7.2 (c, d)
in dry (a, c) and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                              74


     (a)                                                        (b)
     (c)                                                        (d)
Figure 16. XDLVO energy of interaction of human adenovirus 5 with Vaseline lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in
dry (a, c) and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                              75


     (a)                                                            (b)
     (c)                                                            (d)
Figure 17. XDLVO energy of interaction of human adenovirus 40 with Carmex lip balm at pH 4.2 (a, b) and pH 7.2 (c, d)
in dry (a, c) and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                              76


     (a)                                                              (b)
     (c)                                                              (d)
Figure 18. XDLVO energy of interaction of human adenovirus 40 with Chapstick lip balm at pH 4.2 (a, b) and pH 7.2 (c, d)
in dry (a, c) and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                              77


     (a)                                                              (b)
     (c)                                                              (d)
Figure 19. XDLVO energy of interaction of human adenovirus 40 with Burt’s Bees lip balm at pH 4.2 (a, b) and pH 7.2 (c,
d) in dry (a, c) and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                              78


     (a)                                                         (b)
     (c)                                                         (d)
Figure 20. XDLVO energy of interaction of human adenovirus 40 with Vaseline lip balm at pH 4.2 (a, b) and pH 7.2 (c, d)
in dry (a, c) and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                              79


    (a)                                                           (b)
    (c)                                                           (d)
Figure 21. XDLVO energy of interaction of P22 with Carmex lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             80


    (a)                                                           (b)
    (c)                                                           (d)
Figure 22. XDLVO energy of interaction of P22 with Chapstick lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             81


    (a)                                                            (b)
    (c)                                                            (d)
Figure 23. XDLVO energy of interaction of P22 with Burt’s Bees lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             82


    (a)                                                            (b)
    (c)                                                            (d)
Figure 24. XDLVO energy of interaction of P22 with Vaseline lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             83


    (a)                                                            (b)
    (c)                                                            (d)
Figure 25. XDLVO energies of interaction of MS2 with Carmex lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             84


    (a)                                                          (b)
    (c)                                                          (d)
Figure 26. XDLVO energy of interaction of MS2 with Chapstick lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             85


    (a)                                                          (b)
    (c)                                                          (d)
Figure 27. XDLVO energies of interaction of MS2 with Burt’s Bees lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c)
and hydrated (b, d) states. DLVO total energy is provided for reference.
                                                            86


    (a)                                                         (b)
    (c)                                                         (d)
Figure 28. XDLVO energy of interaction of MS2 with Vaseline lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             87


    (a)                                                           (b)
    (c)                                                           (d)
Figure 29. XDLVO energies of interaction of SiO2 with Carmex lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             88


    (a)                                                           (b)
     (c)                                                          (d)
Figure 30. XDLVO energy of interaction of SiO2 with Chapstick lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             89


    (a)                                                          (b)
    (c)                                                          (d)
Figure 31. XDLVO energies of interaction of SiO2 with Vaseline lip balm at pH 4.2 (a, b) and pH 7.2 (c, d) in dry (a, c) and
hydrated (b, d) states. DLVO total energy is provided for reference.
                                                             90


     (a)
     (b)
Figure 32. Deposition of SiO2 colloids onto hydrated Burt’s Bees lip balm under
conditions when SiO2 loading is either a) constant (1.05 mg/ml) or b) increases stepwise
from 0.52 mg/ml to 1.05 mg/ml. In both tests deposition occurs from 150 mM NaCl
electrolyte at pH 7.2. The mass values are calculated based on Sauerbrey equation (eq.
(3)) with 𝒏 = 5. The results for 𝒏 = 3 are similar and shown in the manin manuscript (Fig.
8).
                                             91


Figure 33. QCM-D measurements of the deposition of human adenovirus 5 onto
hydrated Carmex lip balm from 150 mM NaCl electrolyte at pH 7.2. HAdV5
concentration in the QCM-D feed is ~109 GC/mL (~ 0.76 µg/mL). The mass values are
calculated based on Sauerbrey equation (eq. (3)) with 𝒏 = 5. The results for 𝒏 = 3 are
similar and shown in the manin manuscript (Fig. 9).
                                         92


REFERENCES
    93


                                       REFERENCES
[1] B. Stephens, P. Azimi, M.S. Thoemmes, M. Heidarinejad, J.G. Allen, J.A. Gilbert,
Microbial exchange via fomites and implications for human health, Curr. Poll. Rep.
(2019) 1-16.
[2] S.A. Ansari, S.A. Sattar, V.S. Springthorpe, G.A. Wells, W. Tostowaryk, Rotavirus
survival on human hands and transfer of infectious virus to animate and nonporous
inanimate surfaces, J. Clin. Microbiol. 26(8) (1988) 1513-1518.
[3] S.A. Ansari, V.S. Springthorpe, S.A. Sattar, S. Rivard, M. Rahman, Potential role of
hands in the spread of respiratory viral infections: Studies with human parainfluenza
virus 3 and rhinovirus 14, J. Clin. Microbiol. 29(10) (1991) 2115-2119.
[4] J.N. Mbithi, V.S. Springthorpe, J.R. Boulet, S.A. Sattar, Survival of hepatitis A virus
on human hands and its transfer on contact with animate and inanimate surfaces, J.
Clin. Microbiol. 30(4) (1992) 757-763.
[5] T.R. Julian, J.O. Leckie, A.B. Boehm, Virus transfer between fingerpads and fomites,
J. Appl. Microbiol. 109(6) (2010) 1868-1874.
[6] A.K. Pitol, H.N. Bischel, T. Kohn, T.R. Julian, Virus transfer at the skin–liquid
interface, Environ. Sci. Technol. 51(24) (2017) 14417-14425.
[7] A.K. Pitol, H.N. Bischel, A.B. Boehm, T. Kohn, T.R. Julian, Transfer of enteric viruses
adenovirus and coxsackievirus and bacteriophage MS2 from liquid to human skin, Appl.
Environ. Microbiol. 84(22) (2018) e01809-18.
[8] A.M. Wilson, K.A. Reynolds, M.P. Verhougstraete, R.A. Canales, Validation of a
stochastic discrete event model predicting virus concentration on nurse hands, Risk
Anal. 39 (2019) 1812-1824.
[9] R. Gfatter, P. Hackl, F. Braun, Effects of soap and detergents on skin surface pH,
stratum corneum hydration and fat content in infants., Dermatology 195 (1997) 258-262.
[10] S.E. Stewart, M.D. Parker, A. Amézquita, T.L. Pitt, Microbiological risk assessment
for personal care products, Int. J. Cosmet. Sci. 38(6) (2016) 634-645.
[11] L. Jimenez, Microorganisms in the environment and their relevance to
pharmaceutical processes, Microbial contamination control in the pharmaceutical
industry, CRC Press2004, pp. 19-32.
[12] E. Neza, M. Centini, Microbiologically contaminated and over-preserved cosmetic
products according Rapex 2008–2014, Cosmetics 3(1) (2016) 3.
                                              94


[13] M.O. Osungunna, B.B. Oluremi, A. Adetuyi, Bacteriological and antibiotic sensitivity
patterns of bacterial isolates from creams and lotions hawked in Sagamu, Ogun State,
Pak. J. Nutr. 9 (2010) 773-775.
[14] P.A. Geis, R.T. Hennesy, Preservative development, CRC Press, Florida, 2000, pp.
143-161.
[15] F.R. Reid, T.O. Wood, Pseudomonas corneal ulcer: the causative role of
contaminated eye cosmetics, Arch. Ophthalmol. 97(9) (1979) 1640-1641.
[16] M.D. Lundov, C. Zachariae, Recalls of microbiologically contaminated cosmetics in
EU from 2005 to May 2008, Int. J. Cosmetic Sci. 30(6) (2008) 471-474.
[17] J. Wang, A.B. Kay, J. Fletcher, M.K. Formica, T.E. McAlindon, Is lipstick associated
with the development of systemic lupus erythematosus (SLE)?, Clin. Rheumatol. 27
(2008) 1183-1187.
[18] The SCCS notes of guidance for the testing of cosmetic ingredients and their safety
evaluation, 9th revision SCCS/1564/15 Scientific Committee on Consumer Safety
Brussels, 2015.
[19] B. Yuan, M. Pham, T.H. Nguyen, Deposition kinetics of bacteriophage MS2 on a
silica surface coated with natural organic matter in a radial stagnation point flow cell,
Environ. Sci. Technol. 42(20) (2008) 7628-7633.
[20] M. Tong, Y. Shen, H. Yang, H. Kim, Deposition kinetics of MS2 bacteriophages on
clay mineral surfaces, Colloids Surfaces B 92 (2012) 340-347.
[21] H. Shi, I. Xagoraraki, K.N. Parent, M.L. Bruening, V.V. Tarabara, Elution is a critical
step for recovering human adenovirus 40 from tap water and surface water by cross-
flow ultrafiltration, Appl. Environ. Microbiol. 82(16) (2016) 4982-4993.
[22] K. Wong, B. Mukherjee, A.M. Kahler, R. Zepp, M. Molina, Influence of inorganic
ions on aggregation and adsorption behaviors of human adenovirus, Environ. Sci.
Technol. 46(20) (2012) 11145-11153.
[23] R. Lu, Q. Li, T.H. Nguyen, Random sequential adsorption of human adenovirus 2
onto polyvinylidene fluoride surface influenced by extracellular polymeric substances, J.
Colloid Interface Sci. 466 (2016) 120-127.
[24] D.K. Roper, S. Nakra, Adenovirus type 5 intrinsic adsorption rates measured by
surface plasmon resonance, Anal. Biochem. 348(1) (2006) 75-83.
[25] H.T.T. Dang, V.V. Tarabara, Virus deposition onto polyelectrolyte-coated surfaces:
A study with bacteriophage MS2, J. Colloid Interface Sci. 540 (2019) 155-166.
[26] ARMWG, Adenovirus Type 5 Reference Material (Product Information Sheet for
VR-1516), in: A.T.C.C. (ATCC) (Ed.) 2010.
                                               95


[27] B.G. Huyghe, X. Liu, S. Sutjipto, B.J. Sugarman, M.T. Horn, H.M. Shepard, C.J.
Scandella, P. Shabram, Purification of a type 5 recombinant adenovirus encoding
human p53 by column chromatography, Hum. Gene Ther. 6(11) (1995) 1403-1416.
[28] C. Boo, S. Lee, M. Elimelech, Z. Meng, S. Hong, Colloidal fouling in forward
osmosis: Role of reverse salt diffusion, J. Membr. Sci. 390 (2012) 277-284.
[29] S.R. Dasari, V.V. Goud, Comparative extraction of castor seed oil using polar and
non polar solvents, Int. J. Curr. Eng. Technol. 1 (2013) 121-123.
[30] H. Shi, V.V. Tarabara, Charge, size distribution and hydrophobicity of viruses:
Effect of propagation and purification methods, J. Virol. Methods 256 (2018) 123-132.
[31] C.J. van Oss, Interfacial forces in aqueous media, CRC Press2006.
[32] C.J. van Oss, R.F. Giese, The hydrophilicity and hydrophobicity of clay minerals,
Clay Clay Miner. 43 (1995) 474-477.
[33] E.V. Pasco, H. Shi, I. Xagoraraki, S.A. Hashsham, K.N. Parent, M.L. Bruening, V.V.
Tarabara, Polyelectrolyte multilayers as anti-adhesive membrane coatings for virus
concentration and recovery, J. Membr. Sci. 469 (2014) 140-150.
[34] M. Pham, E.A. Mintz, T.H. Nguyen, Deposition kinetics of bacteriophage MS2 to
natural organic matter: Role of divalent cations, J. Colloid Interface Sci. 338(1) (2009) 1-
9.
[35] M.C. Dixon, Quartz crystal microbalance with dissipation monitoring: enabling real-
time characterization of biological materials and their interactions, J. Biomol. Tech. 19
(2008).
[36] A. Armanious, M. Aeppli, R. Jacak, D. Refardt, T. Sigstam, T. Kohn, M. Sander,
Viruses at solid-water interfaces: a systematic assessment of interactions driving
adsorption, Environ. Sci. Technol. 50(2) (2015) 732-743.
[37] H.T.T. Dang, Virus attachment to surfaces: Assessing relative contributions of
electrostatic, van der Waals, and acid-base interactions. Ph.D. dissertation, Civil and
Environmental Engineering, Michigan State University, 2018.
[38] N.F. Steinmetz, E. Bock, R.P. Richter, J.P. Spatz, G.P. Lomonossoff, D.J. Evans,
Assembly of multilayer arrays of viral nanoparticles via biospecific recognition: a quartz
crystal microbalance with dissipation monitoring study, Biomacromol. 9(2) (2008) 456-
462.
[39] X. Huang, J. Xu, H.-F. Ji, G. Li, H. Chen, Quartz crystal microbalance based
biosensor for rapid and sensitive detection of maize chlorotic mottle virus, Anal. Method.
6(13) (2014) 4530-4536.
                                             96


[40] L. Gutierrez, S.E. Mylon, B. Nash, T.H. Nguyen, Deposition and aggregation
kinetics of rotavirus in divalent cation solutions, Environ. Sci. Technol. 44(12) (2010)
4552-4557.
[41] G. Tiliket, G. Ladam, Q.T. Nguyen, L. Lebrun, Polyethylenimine surface layer for
enhanced virus immobilization on cellulose, Appl. Surf. Sci. 370 (2016) 193-200.
[42] G. Ross, Essentials of human physiology, Yearbook Medical Publishers Chicago,
1978.
[43] G.W. Gary, J.C. Hierholzer, R.E. Black, Characteristics of noncultivable
adenoviruses associated with diarrhea in infants: a new subgroup of human
adenoviruses, J. Clin. Microbiol. 10(1) (1979) 96-103.
[44] G. Winberg, G. Wadell, Structural polypeptides of adenovirus type 16 incomplete
particles, J. Virol. 22(2) (1977) 389-401.
[45] G. Sauerbrey, Verwendung von Schwingquarzen zur Wägung dünner Schichten
und zur Mikrowägung, Zeitschrift für physik 155(2) (1959) 206-222.
[46] B. Derjaguin, L. Landau, Theory of the stability of strongly charged lyophobic sols
and of the adhesion of strongly charged particles in solutions of electrolytes, Acta
Physico Chimica URSS 14 (1941).
[47] E.J.W.O. Verwey, J. Th. G., Theory of the stability of lyophobic colloids, Elsevier,
Amsterdam, 1948.
[48] C. Dika, M.H. Ly-Chatain, G. Francius, J.F.L. Duval, C. Gantzer, Non-DLVO
adhesion of F-specific RNA bacteriophages to abiotic surfaces: importance of surface
roughness, hydrophobic and electrostatic interactions, Colloids Surfaces A 435 (2013)
178-187.
[49] C.J. van Oss, M.K. Chaudhury, R.J. Good, Interfacial Lifshitz-van der Waals and
polar interactions in macroscopic systems, Chem. Rev. 88 (1988) 927-941.
[50] Y. Chen, L. Shen, R. Li, X. Xu, H. Hong, H. Lin, J. Chen, Quantification of interfacial
energies associated with membrane fouling in a membrane bioreactor by using BP and
GRNN artificial neural networks, J. Colloid Interface Sci. 565 (2020) 1-10.
[51] J. Teng, L. Shen, Y. Xu, Y. Chen, X.-L. Wu, Y. He, J. Chen, H. Lin, Effects of
molecular weight distribution of soluble microbial products (SMPs) on membrane fouling
in a membrane bioreactor (MBR): Novel mechanistic insights, Chemosphere 248 (2020).
[52] Y. Xing, A. Ellis, M. Magnuson, W.F. Harper, Adsorption of bacteriophage MS2 to
colloids: Kinetics and particle interactions, Colloids Surfaces A 585 (2020) 124099.
[53] J.A. Brant, A.E. Childress, Assessing short-range membrane–colloid interactions
using surface energetics, J. Membr. Sci. 203(1-2) (2002) 257-273.
                                               97


[54] A.R. Fernandes, M.F. Dario, C.A.S. de Oliveira Pinto, T.M. Kaneko, A.R. Baby,
M.V.R. Velasco, Stability evaluation of organic lip balm, Braz. J. Pharmaceut. Sci. 49
(2013) 293-299.
[55] I. Foubert, I.V. Cauteren, K. Dewettinck, A. Huyghebaert, Differential scanning
calorimetry as a means of predicting chocolate fat-blooming, in: N. Widlak, R. Hartel, S.
Narine (Eds.), Crystallization and solidification properties of lipids, AOCS Press2001, pp.
79-86.
[56] S. Pan, N. Germann, Mechanical response of industrial benchmark lipsticks under
large-scale deformations, Acta Mechanica 231 (2020) 3031-3042.
[57] A.W. Adamson, A.P. Gast, Physical Chemistry of Surfaces, Wiley -
Interscience1997.
[58] E.I. Trilisky, A.M. Lenhoff, Sorption processes in ion-exchange chromatography of
viruses, J. Chromatogr. A 1142 (2007) 2-12.
[59] J. Kim, M.J. Park, M. Park, H.K. Shon, S.-H. Kim, J.H. Kim, Influence of colloidal
fouling on pressure retarded osmosis, Desalination 389 (2016) 207-214.
[60] A. Zdziennicka, K. Szymczyk, B. Jańczuk, Correlation between surface free energy
of quartz and its wettability by aqueous solutions of nonionic, anionic and cationic
surfactants, J. Colloid Interface Sci. 340(2) (2009) 243-248.
[61] S. Baliga, S. Muglikar, R. Kale, Salivary pH: A diagnostic biomarker, J. Indian Soc.
Periodont. 17(4) (2013) 461.
[62] L. Graetz, Ueber die Wärmeleitungsfähigkeit von Flüssigkeiten, Annalen der Physik
254(79) (1882).
[63] W.M. Deen, Analysis of Transport Phenomena, Oxford University Press, New York,
Oxford, 1998.
[64] B.W. Robinson, E.H. Nickel, A useful new technique for mineralogy: the
backscattered-electron/low vacuum mode of SEM operation, Am. Mineral. 64(11-12)
(1979) 1322-1328.
[65] R.C. Wilhoit, B.J. Zwolinski, Handbook of vapor pressures and heats of
vaporization of hydrocarbons and related compounds, Thermodynamics Research
Center, College Station, TX1971.
[66] V.B. Wigglesworth, The use of osmium in the fixation and staining of tissues, Proc.
R. Soc. London. B - Biol. Sci. 147(927) (1957) 185-199.
[67] W. Kern, The evolution of silicon wafer cleaning technology, J. Electrochem. Soc.
137(6) (1990) 1887-1892.
                                              98


[68] B. Hutchins, Development of a reference material for characterizing adenovirus
vectors, Bioprocess. J. 1(1) (2002) 25-28.
[69] S.L. Simek, FDA Perspectives on the use of Adenoviral Reference Material, 2002.
[70] ARMWG, Adenovirus Type 5 Reference Material (Product Information Sheet for
VR-1516), in: A.T.C.C. (ATCC) (Ed.) 2010.
[71] E.I. Trilisky, A.M. Lenhoff, Sorption processes in ion-exchange chromatography of
viruses, J. Chromatogr. A 1142 (2007) 2-12.
                                              99


                                      CHAPTER THREE
Virus adhesion to archetypal fomites: A study with human adenovirus and human
                                 respiratory syncytial virus
3.1 Abstract
Adhesion of two viruses – one enveloped (human respiratory syncytial virus, HRSV)
and one non-enveloped (human adenovirus 5, HAdV5) – to four fomites (silica, nylon,
stainless steel, polypropylene) was quantified and interpreted based on
physicochemical properties of viruses and fomites. The selected fomites are tentatively
identified as “archetypes” representing groups of materials distinctly different in
mechanisms of their interfacial interactions. The surfaces are typified on the basis of
their surface energy components including the dispersive (Lifshitz-van der Waals)
component and two polar (electron donor and electron acceptor) components. Virus-
fomite interactions are predicted using the extended Derjaguin-Landau-Verwey-
Overbeek (XDLVO) theory and are experimentally assessed in tests with quartz crystal
microbalance with dissipation (QCM-D). Polar interactions (manifested as hydrophobic
attraction for all virus-fomite pairs but HAdV5/silica) governed virus attachment to
fomites from a solution of high ionic strength typical for a respiratory fluid, while
dispersive interactions played a relatively minor role. For both HAdV5 and HRSV, the
areal mass density of deposited viruses correlated with the free energy of virus-fomite
interfacial interaction in water, ∆𝐺𝑣𝑤𝑓 . The dependence of virus-fomite attachment
probability on ∆𝐺𝑣𝑤𝑓 collapsed into one trend for both HAdV5 and HRSV pointing to the
                                              100


possibility of using ∆𝐺𝑣𝑤𝑓 as a predictor of virus adhesion. Fomite rinsing with DI water
resulted in a partial virus removal attributable to longer range repulsive electrostatic
interactions. The proposed methodology can guide screening and selection of materials
that discourage virus adhesion. The information on the efficiency of virus attachment to
materials as a function of their surface energy components can help design anti-
adhesive surfaces, develop surface cleaning solutions and protocols, and inform
transport and fate models for viruses in indoor environments.
3.2 Introduction
Fomites are any inanimate surfaces that, when contaminated with pathogenic
microorganisms, can serve as a means of transferring the pathogens to a new human
host. There is growing evidence that fomites play an important role in the spread of
viruses [1-3]. Higher probability of transfer is associated with fomites in indoor
environments such as classrooms, hospitals, nursing homes, cruise ships, and
restaurants and especially with surfaces that are frequently touched (e.g. doorknobs,
refrigerator handles, dishcloths, faucets) or facilitate virus transmission through other
uses (e.g. airducts, hospital linen) [3-11]. Personal protection equipment such as rubber
gloves, N95 particulate respirators, surgical masks, gowns may also act as fomites for
viral cross-infection [12]. Indoor airflow may enhance transport of virus and virus-laden
particles, which could be sucked into the ventilation system through return vents [13].
                                              101


The COVID-19 pandemic highlighted the importance of understanding the relative
importance of a various pathways for pathogen transmission. While it appears that
airborne transport is the dominant mechanism for the transmission of SARS-CoV-2,
indirect transmission via fomites does contribute to the spread of this virus [14-18].
Indeed, viable SARS-CoV-2 virus has been found on many surfaces and objects
contaminated by respiratory secretions or droplets expelled by infected individuals [19-
26]. Even if the probability of transfer on a single touch is low, high persistence of
viruses on a surface translates into a higher number of touches and a higher overall
infection risk [26]. A number of studies have shown that respiratory pathogens are
capable of surviving on fomites for extended periods of time - from hours to months [2,
27, 28]. While the knowledge of SARS-CoV-2 transmission mechanisms is still
incomplete, fomites have been firmly established as a transfer route for many other viral
pathogens of significant concern to human health including influenza virus [29, 30] and
norovirus [31, 32]. There is a large and growing knowledge base on the persistence of
viruses at various surface [33]; at the same time, physicochemical bases of virus
attachment to and removal from surfaces under different conditions remain relatively
unexplored [1, 33]. Given the operational definition of risk as a product of availability
and infectivity, adhesion and removal studies are needed to quantify the contribution of
surface-meditated transmission to the availability term – that is the likelihood that a
particular fomite surface has an attached virus that can be detached and transmitted to
a human host.
                                              102


In their recent review, Castaño et al. [1] described how separate virus-fomite
interactions are accounted by the classical DLVO and XDLVO theories. There have
been many studies that applied XDLVO modeling to describe virus adhesion to various
surfaces including membrane filters [34], iron oxide particles [35], personal care
products [36], foods and food-contact surfaces [37], polyelectrolyte multilayers [38] and
sand [39]. Experimental techniques used to study virus attachment to surfaces range
from simple direct contact tests and traditional adsorption studies to record kinetics and
isotherms of adsorption [37] to more complex methods such as those employing
sensors based on surface plasmon resonance (SPR) [40] and quartz crystal
microbalance [38, 41]. SPR is a powerful surface sensing technique especially suitable
for studying interfacial kinetics and affinity characterization. SPR is sensitive to the
vertical position of individual viruses and, coupled with microscopy, enables high-
throughput imaging of single viruses [42]. Liu et al. recently employed plasmonic
imaging technology to study the interfacial dynamics of SARS-CoV-2 pseudovirus
(SARS-CoV-2 surrogate) adsorption on self-assembled monolayers with amino and
carboxyl terminal groups from solutions including artificial saliva, artificial lung fluid and
surface water [40]. SPR-based techniques are limited by the requirement of having a
noble metal substrate to excite the plasmon resonance. Free of this constraint, QCM-D
method has been used to explore virus adhesion to various surfaces while providing
(through the dissipation data) additional information on the rheology of the adsorbed
layers. Studied surfaces include silica [41], natural organic matter [43, 44],
polyelectrolyte multilayers [38], household paints [45] and self-assembled monolayers
[46]. Yet, to our knowledge neither XDLVO nor QCM-D prior studies had focused on
                                               103


fomites with the specific goal of exploring virus adhesion to common indoor surfaces.
The dearth of systematic knowledge in this area is likely due to the sheer diversity of
fomites in terms of their chemical makeup, morphology and physicochemical properties
of their surface. Viruses too, differ significantly in their size, charge, hydrophobicity and
morphology. The broad range of possible deposition, attachment and resuspension
scenarios adds to this complexity. Once viruses are attached, their survival on fomites
depends on virus type, strain and inoculation titer; as a broad example, enteric viruses
(which are mostly non-enveloped) are known to maintain their infectivity for longer that
respiratory viruses (mostly enveloped). While a reductionist approach with its
incremental accumulation of data based on studies of specific virus-fomite pairs is
certainly possible, it would be desirable to identify a limited range of representative
fomites, viruses and deposition conditions that can typify practically relevant scenarios
and help reach generalizable conclusions.
The practical goal of the present work is to develop an approach for reducing the
complexity of the broad range of surface chemistries and morphologies presented by
fomites to a relatively small subset of “archetype” surfaces. An “archetype” is
operationally defined as representing a set of surface properties that are characteristic
of a distinct group of materials and define their adhesive behavior with respect to a virus
of concern. The practical value of the approach is in facilitating both the selection of
surfaces with desired virus adhesion characteristics and the design of surface cleaning
solutions and protocols. With the premise that fomites can be grouped based on the
values of their surface energy components, we select four specific materials - silica,
                                               104


stainless steel, nylon, and polypropylene - and rationalize the selection using the van
Oss theory, a three-component model for surface energy [47]. SiO2 represents high
surface energy metal oxides with a dominant electron donor component, nylon 6,6
typifies monopolar polymer with strong dispersive interactions, stainless steel is a pure
metal coated by an oxide-rich passivation film with commensurate electron donor and
electron acceptor components, while polypropylene represents low surface energy
apolar polymers with weak dispersive interactions only.
The study employs two pathogenic microorganisms - human respiratory syncytial virus
(HRSV) and human adenovirus 5 (HAdV5) - as representatives of enveloped and
nonenveloped viruses, respectively. HRSV and HAdV5 are selected based on the high
relevancy of both viruses for public health, demonstrated importance of fomites for their
transmission, as well as for practical reasons (both are available commercially in high
purity and titer). HAdV, a large nonenveloped virus [48, 49], is highly resistant to both
monochloramine and UV irradiation [50] and can survive on fomites for many days [2].
HAdV is primarily spread by the fecal-oral and respiratory routes through person-to-
person contact and fomite-mediated transfer [51, 52]. Adenoviruses can cause a range
of clinical diseases, including respiratory, gastrointestinal, and conjunctival illness.
HRSV, which features a distribution of sizes and morphologies (spherical or filamentous)
[53], can cause severe disease, especially in children, the elderly and
immunocompromised adults [54]. HRSV is primarily spread by large droplets and via
fomites, and can survive on nonporous surfaces, skin, and gloves for many hours [55,
56]. The makeup of the solution from which deposition occurs affects the likelihood of
                                             105


virus attachment. A large number of illnesses are associated with more than 150
different types of pathogenic respiratory viruses transmittable from bodily secretions to
surfaces. Indeed, viruses deposit onto fomites either as a result of direct contact with an
infected human or from respiratory droplets produced by such person. Given that both
HRSV and HAdV5 are present in the human respiratory tract, the present study focuses
on the latter transmission route and explores virus deposition from a high ionic strength
electrolyte (150 mM NaCl) typical for respiratory fluid (9 g·L-1, physiological
concentration [57]). While the composition of respiratory fluid is complex and varies with
individual’s health status [58, 59], some aggregate characteristics such as pH and ionic
strength can be captured in modeling studies.
The study explores the hypothesis that virus attachment to fomites in a high ionic
strength solution is governed by hydrophobic interactions. To test the hypothesis, we
use quartz crystal microbalance with dissipation (QCM-D) and QCM-D sensors with
specialty coatings to experimentally determine attachment efficiency for HRSV and
HAdV5 depositing onto four “archetypal” fomites that span a range of hydrophobicities.
The experimental measurements are complemented by modeling based on the
extended Derjaguin-Landau-Verwey Overbeek (XDLVO) theory. Experimental and
modeling results are interpreted in terms of polar, dispersive, and electrostatic
interactions between viruses and fomites.
                                            106


3.3 Materials and Methods
3.3.1 Reagents, fomites, viruses
All chemical reagents were of high purity (>99%). NaCl, KCl, glycerol, sodium dodecyl
sulfate (SDS), ethylene glycol (EG) and diiodomethane (DID) were purchased from
Sigma Aldrich. QCM-D sensors coated with materials representing fomites – silica,
nylon 6,6 (hereafter “nylon”), stainless steel (SS), and polypropylene (PP) – were
purchased from Nanoscience. Deconex 11 (Fisher Scientific) was used as the cleaning
solution for PP and nylon sensors (see Supplementary Material (SM)). Silica and SS
sensors were cleaned with 2% SDS solution and Hellmanex II (Hellma GmbH & Co.
KG), respectively.
Human adenovirus 5 (HAdV5) was purchased from American Type Culture Collection
(ATCC® VR-1516TM) [60]. The product is an aqueous suspension of HAdV5 in 20 mM
TRIS, 25 mM NaCl, 2.5% glycerol (pH 8.0) [61] purified by single column
chromatography [62]. HAdV5 is a non-enveloped dsDNA virus with an icosahedral
nucleocapsid. Human respiratory syncytial virus (HRSV; strain: Long) was also
purchased from ATCC (VR-26PQTM) [63]. The product is an aqueous suspension of
HRSV in 50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA prepared by concentrating
another HRSV stock (ATCC VR-26TM) via sucrose cushion centrifugation [64]. HRSV is
an ssRNA virus with a helical nucleocapsid surrounded by matrix protein and an
envelope. Spherical and filamentous HRSV virions have been identified [53, 65, 66].
                                            107


3.3.2 Virion characterization: Hydrodynamic size, 𝜻-potential, concentration
Electrophoretic mobility and hydrodynamic diameter of HRSV virions were measured by
laser doppler micro-electrophoresis (Zetasizer Nano ZS, Malvern) and dynamic light
scattering (DLS), respectively. The 1 mM KCl diluent used in these measurements was
pre-filtered through 0.22 μm syringe filter. The charge and size of HAdV5 (ATCC® VR-
1516TM) were determined previously [36]. These measurement techniques capture
averaged values of virus charge and size. While the electrical charge has a certain
distribution over the virus surface, streaming potential is an aggregate estimate of the
surface potential averaged over the surface. Similarly, DLS estimates particle’s diffusion
coefficient, which is then converted to particle size assuming that the particle is
spherical. Yet viruses (including HAdV5 [36] and HRSV [53, 65, 67]) have complex
morphology. The adopted approach where viruses are treated as spherical colloids with
a chemically homogenous surface are justified as both XDLVO modeling (section 2.4)
and interpretation of experimental data on virus adhesion obtained by QCM-D (sections
2.5 and 3.1) rely on these assumptions. As a complement to DLS measurements, both
HRSV and HAdV5 were imaged using transmission electron microscopy (TEM, JEM-
1400 Flash, Jeol, Nieuw-Vennep). TEM sample preparation and imaging procedures
are described in SM, section S1. Virus concentration was measured by fluorometry
(Qubit fluorometer, Invitrogen) with Qiagen DNA and RNA mini kits used to extract
dsDNA from HAdV5 and ssRNA from HRSV. The ssRNA High Sensitivity and dsDNA
                                             108


High Sensitivity modes were utilized for HRSV and HAdV5 genome quantification,
respectively.
3.3.3 Quantifying surface energy of viruses and fomites
3.3.3.1 Approach
Hydrophobicity of a solid (𝑠) can be quantified in terms of the free energy of its
interfacial interaction with an identical material when immersed in water (𝑤), ∆𝐺𝑠𝑤𝑠 [47].
We applied this approach to evaluate hydrophobicity of viruses (𝑣) and fomites (𝑓) by
computing ∆𝐺𝑣𝑤𝑣 and ∆𝐺𝑓𝑤𝑓 , respectively. The calculation relies on the knowledge of
the solid’s surface energy in terms of its three components: two Lewis acid-base
(electron acceptor, 𝛾𝑠+ , and electron donor, 𝛾𝑠− ) components and the Lifshitz-van der
Waals component, 𝛾𝑠𝐿𝑊 . The Lewis acid-base components are also described as polar
while the Lifshitz-van der Waals component is often referred to as dispersive or apolar.
The components can be determined by measuring contact angles (𝜃) of three probe
liquids (𝑙) with known 𝛾𝑙𝐿𝑊 , 𝛾𝑙+ and 𝛾𝑙− on the surface of the solid and substituting these
values into the Young-Dupré equation [68, 69]
(1 + 𝑐𝑜𝑠𝜃)𝛾𝑙𝑇𝑂𝑇 = 2 (√𝛾𝑠𝐿𝑊 𝛾𝑙𝐿𝑊 + √𝛾𝑠+ 𝛾𝑙− + √𝛾𝑠− 𝛾𝑙+ ),                                   (1)
where 𝛾𝑙𝑇𝑂𝑇 is the total surface energy of the probe liquid:
𝛾 𝑡𝑜𝑡 = 𝛾 𝐿𝑊 + 𝛾 𝐴𝐵 = 𝛾 𝐿𝑊 + 2√𝛾 − 𝛾 + .                                                   (2)
The free energy of solid-solid interfacial interaction in water is given by
                                               109


                                2
∆𝐺𝑠𝑤𝑠 =   −2 (√𝛾𝑠𝐿𝑊   −    𝐿𝑊
                        √𝛾𝑤 )     − 4 (√𝛾𝑠+ 𝛾𝑠− + √𝛾𝑤+ 𝛾𝑤− − √𝛾𝑠+ 𝛾𝑤− − √𝛾𝑠− 𝛾𝑤+ )      (3)
where 𝛾𝑤𝐿𝑊 , 𝛾𝑤+ and 𝛾𝑤− are surface energy components of water. A positive value of
∆𝐺𝑠𝑤𝑠 indicates a hydrophilic surface, while negative ∆𝐺𝑠𝑤𝑠 corresponds to a
hydrophobic surface. The absolute value of ∆𝐺𝑠𝑤𝑠 indicates the degree of hydrophilicity
(or hydrophobicity, when ∆𝐺𝑠𝑤𝑠 < 0) of the surface. As in the case with size and charge
measurements, the determination of virus and fomite surface energies based on contact
angles of probe liquids treats these surfaces as chemically homogeneous. Thus, the
presence of hydrophobic or hydrophilic “patches” on a surface is not accounted for and
energy values describe interaction of “equivalent” chemically homogenous surfaces.
3.3.3.2 Experiments
Contact angles of three probe liquids - DI water, glycerol, and DID - on fomite surfaces
and HRSV lawns were measured using the sessile drop method (goniometer /
tensiometer model 250, ramé-hart). Additional contact angle measurements with EG as
the fourth probe liquid were done for the polypropylene surface. Virus lawns were
formed by filtering virus stock suspensions through a polyethersulfone ultrafiltration
membrane (50 kDa, Pall Life Sciences) to form a multilayer (> 5 monolayers) cake, or
lawn, of virions [70]. Prior to recording contact angle values, the lawn was allowed to dry
in air until the contact angle of water on the virus lawn stabilized. A separate set of
measurements was performed to study the effect of pH and ionic strength on water
contact angles on fomite surfaces. All measurements were performed in the air at the
ambient temperature of 22 °C and the relative humidity of 47 %. The droplet volume
                                                110


was 6 μl. Contact angle values were calculated by DROPimage Advanced software
based on recorded droplet shapes.
3.3.4 Modeling virus-fomite interactions. Extended Derjaguin-Landau-Verwey-
Overbeek theory
The interactions between colloidal and surfaces can be predicted by XDLVO theory.
                                                             𝑋𝐷𝐿𝑉𝑂
The theory describes the total energy of interaction 𝑈𝑣𝑤𝑓           between a spherical particle
(representing a virus (𝑣)) and a flat surface (representing a fomite (𝑓)) in water (𝑤) as a
                                                         𝐿𝑊
sum (see eq. (7)) of Lifshitz-van der Waals (LW), 𝑈𝑣𝑤𝑓        , electrostatic double layer (EL),
  𝐸𝐿                                   𝐴𝐵
𝑈𝑣𝑤𝑓 , and Lewis acid-base (AB), 𝑈𝑣𝑤𝑓     , energies expressed as follows:
  𝐿𝑊
               𝐴𝑎          𝑎                                                                   (4)
𝑈𝑣𝑤𝑓 (𝑑) = −       = 2𝜋𝑑02 ∆𝐺𝑑𝐿𝑊
               6𝑑          𝑑       0
  𝐸𝐿
                                  1 + 𝑒 −𝜅𝑑
𝑈𝑣𝑤𝑓  (𝑑) = 𝜋𝜀𝑟 𝜀0 𝑎 [2𝜓𝑣 𝜓𝑓 𝑙𝑛 (           ) + (𝜓𝑣2 + 𝜓𝑓2 ) 𝑙𝑛(1 − 𝑒 −2𝜅𝑑 )]                  (5)
                                  1 − 𝑒 −𝜅𝑑
  𝐴𝐵
                              𝑑0 − 𝑑
𝑈𝑣𝑤𝑓 (𝑑) = 2𝜋𝑎𝜆∆𝐺𝑑𝐴𝐵    𝑒𝑥𝑝 (        )                                                         (6)
                      0
                                 𝜆
  𝑋𝐷𝐿𝑉𝑂     𝐿𝑊       𝐸𝐿      𝐴𝐵
𝑈𝑣𝑤𝑓    = 𝑈𝑣𝑤𝑓  + 𝑈𝑣𝑤𝑓   + 𝑈𝑣𝑤𝑓                                                                (7)
where 𝑎 is the virus radius, 𝑑 is the virus-fomite minimal interfacial separation distance,
𝐴 = −12𝜋𝑑02 ∆𝐺𝑑𝐿𝑊0
                    is the Hamaker constant, 𝜀𝑟 is the relative dielectric permittivity of
water (for water at 25 °C, 𝜀𝑟 ≈ 78.3), 𝜀0 is the dielectric permittivity of vacuum (𝜀0 =
8.854×1012 C·V-1·m-1), 𝜓𝑣 and 𝜓𝑠 are surface potentials of the virus and fomite,
respectively, 𝜅 is the inverse Debye screening length, λ is the characteristic delay length
of AB interactions in water (λ = 0.6 nm), and 𝑑0 is the minimum separation distance (𝑑0
                                               111


= 0.158 nm) due to Born repulsion. Surface potentials 𝜓𝑣 and 𝜓𝑓 are commonly
approximated by 𝜁-potentials (𝜁𝑣 and 𝜁𝑓 ). Values of ∆𝐺𝑑𝐿𝑊 0
                                                             and 𝐺𝑑𝐴𝐵
                                                                    0
                                                                      in eq. (4) and eq. (6)
are given by:
                                                                                          (8)
∆𝐺𝑑𝐿𝑊
    0
       = 2 (√𝛾𝑣𝐿𝑊 − √𝛾𝑤𝐿𝑊 ) (√𝛾𝑤𝐿𝑊 − √𝛾𝑓𝐿𝑊 )
                                                                                          (9)
∆𝐺𝑑𝐴𝐵
    0
       = 2 √𝛾𝑤+ (√𝛾𝑓− + √𝛾𝑣− − √𝛾𝑤− ) + 2√𝛾𝑤− (√𝛾𝑓+ + √𝛾𝑣+ − √𝛾𝑤+ )
                                −2√𝛾𝑓+ 𝛾𝑣− − 2√𝛾𝑓− 𝛾𝑣+
The free energy of virus-fomite interfacial interaction in water, ∆𝐺𝑣𝑤𝑓 , is
∆𝐺𝑣𝑤𝑓 = ∆𝐺𝑑𝐿𝑊 0
                + ∆𝐺𝑑𝐴𝐵
                      0
                                                                                        (10)
Note that eq. (3) can be obtained from eqs. (8 - 10) by substituting subscripts 𝑣 and 𝑓
for 𝑠. Lewis acid-base interactions are also referred to as electron donor/electron
acceptor or polar interactions. Hydrophobic attraction and hydrophilic repulsion (i. e.
hydration pressure) are two types of polar interactions [47].
Virion size, surface charge and surface energy components as well as surface charge
and surface energy components of each of the four fomites were used as inputs to the
XDLVO model. In its description of the energy of sphere-plate interaction energy, the
model assumes that both surfaces are smooth. The effect of surface roughness can be
taken into account by adding the sphere-asperity term to the calculation of the total
interaction energy [71, 72]. Applied to the calculation of the energy of interaction
between household paint coatings and human adenovirus 40 (a virus similar to HAdV5),
this approach showed that the presence of ~ 27 nm asperities on paint-coated QCM-D
                                             112


sensors had only a minor effect (~ 10% change in the total energy) [38]. Based on the
manufacturer’s data, the root mean square surface roughness of QCM-D sensors is < 1
nm for stainless steel and silica and ~ 3.5 nm for nylon and polypropylene [73] – well
below the 27 nm value for which XDLVO predictions were only weakly affected.
Most data were obtained in experiments as described in sections 2.2 and 2.3. Size and
charge of HAdV5 were determined in our earlier study and used in the present work [36].
Values of the 𝜁-potential of the four fomites as a function of pH were adopted from
literature (Zemljic et al. [74] for PP, Hedberg et al. [75] for SS, Zhang et al. [76] for nylon,
and Wang et al. [36] for silica).
                                             113


3.3.5 QCM-D studies of HAdV5 and HRSV attachment to fomites
3.3.5.1 Approach: Quantifying virus-fomite attachment efficiency
To accurately assess the mass flux of viruses towards the QCM-D sensor surface one
needs to solve the Graetz problem of diffusion-limited transport to a flat plate from a
crossflow [77, 78]. Given the complexity of the QCM-D chamber geometry and the need
to add surface reaction (describing the finite probability of virus attachment to the
sensor surface), the solution would need to be numerical. Instead, we propose a simple
model that describes the rate of virus deposition (ng/min) onto QCM-D sensor surface
as
𝑑𝑚𝑣
     = 𝛼𝑣𝑓 𝑗𝑚 𝐴𝑠                                                                      (11)
 𝑑𝑡
where 𝛼𝑣𝑓 (unitless) is the virus-fomite attachment efficiency, 𝐴𝑠 (cm2) is surface area of
the sensor and 𝑗𝑚 (ng·cm-2·min-1) is the mass flux of viruses towards the sensor. We
define mass transfer coefficient, 𝑘 (m·s-1), as follows:
𝑗𝑚 = 𝑘(𝐶𝑏 − 𝐶𝑠 )                                                                      (12)
where 𝐶𝑏 and 𝐶𝑏 are virus concentrations in the bulk of the flow and at the sensor
surface, respectively. For early stages of deposition 𝐶𝑠 ≪ 𝐶𝑏 , the sensor is mostly virus-
free and the rate of virus deposition is approximately constant:
𝑑𝑚𝑣
     ≈ 𝛼𝑣𝑓 𝑘𝐶𝑏 𝐴𝑠                                                                     (13)
  𝑑𝑡
The linearity of the deposition rate dependence on concentration was confirmed in tests
with different 𝐶𝑏 (see SM, Fig. 12). Because the rate of mass transfer across the viscos
                                             114


sublayer at the QCM-D sensor surface depends on virus diffusivity, mass transfer
coefficient, 𝑘, needs to be determined for each virus separately.
3.3.5.2 QCM-D experiments
The QCM-D E4 system (Biolin Scientific) was used to quantify the deposition of viruses
onto sensor surfaces. Prior to measurement, QCM-D sensors were cleaned (see SM,
section S4) and then mounted into the flow chamber to determine their resonance
frequency in air. QCM-D tests were carried out at 25 °C in a continuous flow mode (0.15
ml/min) using a digital peristaltic pump (IPC, four channels, ISMATEC). The sensors
were first equilibrated with DI water (pH 5.8), then with 0.22 µm filtered NaCl electrolyte
(pH 5.8) and only after the vibration frequency stabilized were challenged with a virus
suspension. Procedures performed to avoid air bubble formation on the surface of
hydrophobic PP sensors are described in SM, section S2. QCM frequency and
dissipation were recorded every 1 min. The frequency shifts were fitted into the
Sauerbrey equation [79] to compute the change in areal mass density, ∆𝑚 (ngcm-2):
         𝐶∆𝑓
∆𝑚 = −                                                                                  (14)
           𝑛
where 𝐶 is the mass sensitivity constant (𝐶 = 17.7 ngHz-1cm-2), 𝑛 is the overtone
number and ∆𝑓 is the frequency shift (Hz). All virus suspensions used in QCM-D tests
were prepared by diluting the virus stock in a background electrolyte pre-filtered through
a 0.22 μm filter. Background electrolytes were either 150 mM NaCl (pH 5.8) or 1 mM
NaCl (pH 4.1 in tests with HAdV5 and pH 3.8 in tests with HRSV). The 150 mM NaCl
solution was used in most tests with all 8 fomite-virus pairs. The 1 mM NaCl was used
                                            115


in additional tests with HAdV5/PP and HRSV/PP pairs to simulate conditions of 𝛼𝑣𝑓 = 1
(see SM, section S6). Prior to use in QCM-D experiments, virus suspensions were
vortexed for 15 s to ensure a uniform dispersion of virions.
After ~ 60 min of QCM-D measurement, the sample chamber and tubing were flushed
with the background electrolyte followed by DI water (pH 5.8). Each sensor was used in
5 to 7 different QCM-D experiments and cleaned after each test. To ascertain that
cleaning did not affect the hydrophobicity of the sensor coatings, contact angles of three
probe liquids were re-measured after multiple cycles of QCM-D tests and cleaning (see
SM; Tables 4 – 6).
                                           116


3.4 Results and Discussion
3.4.1 Virus concentration, size, charge and hydrophobicity
The concentrations of HAdV5 and HRSV in stock suspensions were verified by
fluorometry (Qubit, Invitrogen) and found to be somewhat different from the values
given by ATCC for these specific lots. The measured concentrations were 3.45×1011
GC/ml for HAdV5 (lower than ATCC-reported 5.8×1011 GC/ml [61]) and 2.65×1010
GC/ml for HRSV (higher than ATCC-reported 1.7×1010 GC/ml [64]). Based on the
measured values of the hydrodynamic size (see below) and the approximate virion
density, the corresponding mass concentrations of HAdV5 and HRSV in the feed
suspensions were estimated to be 28 ng/mL and 12 ng/mL.
The hydrodynamic diameter, 𝑑ℎ , of HAdV5 was ~ 102 nm as determined by DLS in our
previous study [36]. This size is derived using Stokes-Einstein equation (eq. (15)) based
on the measured value of HAdV5 diffusivity, 𝐷 = 4.27×10-12 (m2·s-1).
      𝑘𝐵 𝑇
𝐷=                                                                                  (15)
     3𝜋𝜇𝑑ℎ
The difference between HAdV5 diameter values estimated from TEM images (~ 90 nm,
Figures 10G and 10H) and measured by DLS is likely due to the presence of fibers on
the HAdV5 surface [39, 48], which slow down diffusion.
                                            117


Figure 34. -potential as a function of pH (A, D), size distribution (B, E), and TEM images (C, F, G) of HRSV (A-C) and
HAdV5 (D-G). In A, B, D and E, lines are added to guide the eye. Charge and size values for HAdV5 (D, E) are adopted
from our earlier report [36]. Additional TEM images are shown in SM, Figure 42.
                                                               118


Table 8. Contact angle of probe liquids, surface energy parameters (𝜸𝑳𝑾 , 𝜸+ , 𝜸− , 𝜸𝑨𝑩 , 𝜸𝒕𝒐𝒕 ) and the free energy of
interfacial interaction in water (∆𝑮𝒔𝒘𝒔 ) of four fomites (clean, before use) and two viruses. Error estimates were obtained
by propagating experimental errors in measured contact angles through the calculation of surface energy components
(eqs (1) and (2)) and ∆𝑮𝒔𝒘𝒔 (eq. (3)). Additional contact angle measurements were performed for the four fomites using
150 mM NaCl solution at pH 5.8; the contact angle data and calculated surface energy values are given in SM (Table 11).
               Parameter                                       Fomites                                      Viruses
                                      Silica       Nylon        Stainless steel   Polypropylene    HAdV5 B             HRSV C
                           H2O A    12.1 ± 0.9   50.1 ± 1.4       60.7 ± 1.9       100.5 ± 1.5    72.3 ± 0.4          73.5 ± 2.6
    Contact           Glycerol      14.7 ± 1.5   50.5 ± 1.0       31.4 ± 0.8       89.4 ± 1.6     70.5 ± 1.2          63.6 ± 1.3
     angle                  DID     31.4 ± 2.6   12.4 ± 0.9       50.2 ± 1.5       58.6 ± 2.9     28.5 ± 1.0          45.2 ± 2.0
                             EG        n/a          n/a               n/a          77.5 ± 1.8         n/a                n/a
                            𝛾 𝐿𝑊    43.6 ± 1.1   49.6 ± 0.2       34.2 ± 0.8       29.4 ± 1.7     36.1 ± 1.9          36.9 ± 1.1
                             𝛾+     2.2 ± 0.2    0.02 ± 0.02       6.8 ± 0.5            0              0              0.4 ± 0.2
    Surface energy           𝛾−     44.5 ± 0.7   26.9 ± 2.1        7.4 ± 1.5        0.5 ± 0.2     14.9 ± 1.2          9.0 ± 2.6
                            𝛾 𝐴𝐵
       (mJ·m-2)
                                    19.7 ± 0.6    1.4 ± 0.4       14.2 ± 1.6            0              0              3.8 ± 1.1
                            𝛾 𝑡𝑜𝑡   63.3 ± 1.2   51.0 ± 0.4       48.4 ± 1.8       29.4 ± 1.7     36.1 ± 1.9          40.7 ± 1.5
                           ∆𝐺𝑠𝑤𝑠    15.7 ± 1.1   - 8.5 ± 3.9      - 25.5 ± 2.9     - 88.9 ± 2.5   - 27.7 ± 1.1         - 40.2 ±
                                                                                                                         7.8
                                                                    119


Notes: A DI water, pH 5.8; B Wang et al. [36]; C Contact angle measurements were
performed on top of a multilayer lawn of viruses assembled, by filtration, on the surface
of a hydrophilic (∆𝐺𝑠𝑤𝑠 = 18.9 mJ·m-2) polyethersulfone ultrafiltration membrane.
In addition, drying-induced shrinkage of virions during negative staining may have
decreased the diameters observed in TEM images [80]. A similar discrepancy between
DLS- and TEM-derived sizes was also reported for HAdV40 [81]. For the HRSV
suspension, the volume-based particle size distribution obtained by DLS (Fig. 10B; note
the logarithmic size scale) was bimodal with a larger peak at ~ 30 nm and a smaller
peak at ~ 132 nm, indicating varied particle size or morphology of HRSV [53, 65, 67].
Given that volume-based distributions are sensitive to the presence of larger particles,
the small intensity of the 132 nm peak points to their low abundance. Still, because
deconvolution of DLS scattered light frequency data assumes monodisperse scatterers,
multimodal distributions should be treated with caution as quantitatively inaccurate. In
such case, direct visualization by TEM is especially valuable. TEM imaging showed the
preponderance of smaller and spherical HSRV virions (Fig. 10C) although larger and
irregularly shaped virions were also occasionally observed. The latter observation is
consistent with earlier reports of HRSV polymorphism [53, 65, 67].
Based on the measured values of HRSV electrophoretic mobility as a function of pH
(Fig. 41), the isoelectric point (pI) of HRSV was estimated to be 3.91 ± 0.07.
Electrophoretic mobilities were converted to 𝜁-potentials (Fig. 10A) using Ohshima
equation [82] (see SM, section A5, Table 10). The Ohshima approach was used
                                             120


because for HRSV in 1 mM KCl, 𝜅𝑎 ≈ 6.4, making neither Smoluchowski expression
(valid when 𝜅𝑎 ≫ 1) nor Hückel expression (valid when 𝜅𝑎 ≪ 1) applicable. At pH 5.8,
which has been reported to fall within the pH range typical for human respiratory fluid
[83], the 𝜁-potential of HRSV is - 33.9 ± 2.6 mV. The electrophoretic mobility of HAdV5
was measured earlier [36]; the pI of HAdV5 is 4.6 ± 0.03 while 𝜁-potential at pH 5.8 is -
18.2 ± 0.1 mV (Fig. 10D).
Hydrophobicity of adenoviruses is responsible for their low recoveries from water [84-86]
presumably in relation to virus loss to surfaces during sample handling (e. g. storage,
transfer, sample concentration) [81]. The high negative values of ∆𝐺𝑠𝑤𝑠 reported for
HAdV40 (- 30.4 mJ·m-2 [81]) and HAdV5 (- 27.7 mJ·m-2 [36]) confirm their
hydrophobicity. The propensity of these viruses to attach to surfaces also implies a
higher likelihood of fomite-mediated transfer. In the present work, HRSV was
determined to be even more hydrophobic (than HAdV5 and HAdV40) with ∆𝐺𝑠𝑤𝑠 of -
40.2 mJ·m-2. In terms of surface energy components, the hydrophobicity of HAdV40,
which is a monopolar virus (𝛾 + = 0; Table 3), was due to the small value of the only non-
zero component of virus-water polar interactions, √𝛾𝑠− 𝛾𝑤+ , when compared to the polar
interaction between water molecules, √𝛾𝑤+ 𝛾𝑤− (see eq. (3)). The latter term represents
the hydrogen bonding energy of the cohesion of water. To our knowledge the present
study is the first report of HRSV surface energy and, more generally, of HRSV
hydrophobicity.
                                            121


3.4.2 Fomite hydrophobicity
In the order from most hydrophilic to most hydrophobic, the four fomites ranked as
follows: silica > nylon > SS > PP, with ∆𝐺𝑓𝑤𝑓 values of 15.7, - 8.5, -25.5, and - 88.9
mJ·m-2, respectively (Table 3). These estimates were computed based on contact
angles measured using DI water at pH 5.8 as one of the probe liquids. For PP, √𝛾𝑠+ took
on a small negative value; this is a common observation for monopolar or near-
monopolar materials where an experimental error prevents solving eq. (1) to determine
surface tension components of the solid [87]. In this study, we assumed that 𝛾𝑠+ for PP
is zero and computed 𝛾𝑙− using a graphical approach described by McCafferty [88] (see
SM, section S8).
Surfaces in contact with aqueous solutions are most hydrophobic at pH near the pI of
the surface. Indeed, isoelectric points can be determined by contact angle titration [89].
Cuddy et al. [90] performed such measurements for several common QCM-D sensors
(Al2O3, Au, SiO2, Ag, Ti). Because pH of the deposition and cleaning solutions can vary,
it is important to evaluate hydrophobicity of fomites at different pH. The pH dependence
of ∆𝐺𝑓𝑤𝑓 for each fomite is shown in Fig. 11. The trends closely followed those of water
contact angles (Fig. 44). For all four fomites the dependence of ∆𝐺𝑓𝑤𝑓 on pH was a
curve with a minimum. The dependence was strongest for stainless steel, which was
considerably hydrophobic (∆𝐺𝑓𝑤𝑓 = - 24.2 mJ·m-2) near its pI but became hydrophilic
(∆𝐺𝑓𝑤𝑓 > 0) at pH ≲ 2.1 and pH ≳ 8.4. Hydrophobicity of nylon, PP, silica and SS
peaked at pH of ~ 2.5, 2.6, 3.9 and 4.0 respectively. The estimate for silica is in good
                                             122


agreement with the pI value determined for silica-coated QCM-D sensors by Cuddy et al.
[90].
Virga et al. [91] showed that the contact angle of an aqueous electrolyte is a stronger
function of pH for higher ionic strength electrolytes. For the solutions employed in this
work (DI water, 1 mM NaCl, 150 mM NaCl), the effects of the ionic strength of the
contact angle (Fig. 44) and, by extension, hydrophobicity (Fig. 11) were not statistically
significant. It was also assumed that the dependence of the surface tension of the
aqueous solution on its ionic strength and pH was insignificant. Indeed, for the
electrolyte used in the present work (150 mM NaCl) the effect was reported to be small:
~ 2 mJ·m-2 increase over the surface tension of DI water [92].
3.4.3 Quantifying virus-fomite interactions. Four archetypal fomites
Table 4 presents values of ∆𝐺𝑣𝑤𝑓 and its dispersive and polar constituents (∆𝐺𝑑𝐿𝑊      0
                                                                                         and
∆𝐺𝑑𝐴𝐵
    0
      , see eqs (8-10)) for all eight virus-fomite pairs. The net interfacial interaction is
attractive (∆𝐺𝑣𝑤𝑓 < 0) for all pairs except HAdV5/silica. A detailed analysis of ∆𝐺𝑣𝑤𝑓 in
terms of surface energy components (i. e. relative contributions of various terms in eqs.
(8) and (9)) can identify surface properties responsible for the strength, or weakness, of
the overall interaction. Potentially, such analysis can help with the selection of surfaces
resisting virus adhesion as well as the optimal makeup of surface cleaning solutions.
                                              123


Figure 35. Free energy of interfacial interaction in water (∆𝑮𝒇𝒘𝒇 ) of four fomites as a
function of pH. Depending on the type of aqueous solution used as a probe liquid,
values are shown using either empty symbols (150 mM NaCl at pH 5.8; Table 10) or
gray symbols (DI water at pH 5.8; Table 3) or black symbols (1 mM NaCl water at pH
3.8 and 4.1; Table 14).
                                            124


Table 9. Free energy of virus-fomite interfacial interaction in water, ∆𝑮𝒗𝒘𝒇 , and its
dispersive (∆𝑮𝑳𝑾                  𝑨𝑩
               𝒅𝟎 ) and polar (∆𝑮𝒅𝟎 ) components, for eight different virus-fomite pairs.
Standard deviations were obtained by propagating experimental errors in measured
contact angles (Table 3) through the calculation of surface energy parameters (eqs. (1)
and (2)) and ∆𝑮𝒗𝒘𝒇 (eqs. (8), (9), and (10)).
                                                   Interaction energy (mJ·m-2)
        Virus             Fomite
                                             ∆𝐺𝑑𝐿𝑊
                                                 0
                                                                ∆𝐺𝑑𝐴𝐵
                                                                    0
                                                                                  ∆𝐺𝑣𝑤𝑓
                     Silica               - 5.5 ± 0.4        - 0.4 ± 3.2       - 5.8 ± 3.2
                     Nylon                - 6.7 ± 0.4       - 19.0 ± 1.8      - 25.7 ± 1.9
 Human
 respiratory
                     Stainless steel      - 3.3 ± 0.3       - 30.6 ± 3.3      - 33.9 ± 3.2
 syncytial virus
                     Polypropylene        - 2.1 ± 0.5       - 59.1 ± 4.6      - 61.3 ± 4.6
                     Silica               - 5.2 ± 0.6         7.9 ± 1.1         2.7 ± 1.3
                     Nylon                - 6.4 ± 0.8       - 10.4 ± 0.0      - 16.7 ± 0.8
 Human
 adenovirus 5        Stainless steel      - 3.2 ± 0.4       - 29.4 ± 0.8      - 32.5 ± 0.9
                     Polypropylene        - 2.0 ± 0.5       - 56.0 ± 0.0      - 58.0 ± 0.5
A common feature for all four fomites and the two viruses studied in this work is that
their 𝛾 𝐿𝑊 component is higher than that of water (𝛾𝑤𝐿𝑊 = 21.8 mJ·m-2); as a result, for
none of the virus-fomite pairs the conditions (𝛾𝑓𝐿𝑊 < 𝛾𝑤𝐿𝑊 < 𝛾𝑣𝐿𝑊 or 𝛾𝑣𝐿𝑊 < 𝛾𝑤𝐿𝑊 < 𝛾𝑓𝐿𝑊 )
necessary for dispersive interaction to be repulsive (∆𝐺𝑑𝐿𝑊 0
                                                               > 0; eq. (8)) were fulfilled.
Thus, for both HAdV5 and HRSV their dispersive interactions with each of the four
fomites were attractive. Notably, some common materials such as PTFE (𝛾𝑤𝐿𝑊 = 18.6
                                             125


mJ·m-2; see Table 2.3 in the book by Kinloch [93]) would have repulsive dispersive
interactions with HAdV5 and HRSV in water.
In several other respects, the four selected fomites differed. In what follows, the
differences are described and presented as the basis for considering the four fomites as
typifying distinct groups of materials (i. e. as “archetypes”).
1. Silica is characterized by a large electron donor component of its surface tension
    such that 𝛾𝑓− ≫ 𝛾𝑓+ . Silica’s polar interactions with HAdV5 are repulsive due to
    hydration of both silica and virus surfaces (large 𝛾𝑓− 𝛾𝑤+ and 𝛾𝑣− 𝛾𝑤+ terms in eq. (9)).
    Repulsive polar interactions (∆𝐺𝑑𝐴𝐵  0
                                            > 0) are referred to as hydration pressure. For the
    HAdV5/silica pair, hydration pressure overcomes water cohesion given by 𝛾𝑤− 𝛾𝑤+ so
    that even in the presence of attractive dispersive forces (∆𝐺𝑑𝐿𝑊   0
                                                                          < 0), the overall
    interaction is a mild repulsion (∆𝐺𝑣𝑤𝑓 > 0).
2. Nylon is less polar than silica. A near-monopolar surface (𝛾𝑓+ ≅ 0), nylon also has a
    smaller 𝛾𝑓− . As a result, the electron-donor attraction between nylon and viruses is
    weaker than water cohesion leading to ∆𝐺𝑑𝐴𝐵     0
                                                       < 0. Attractive polar interactions (∆𝐺𝑑𝐴𝐵
                                                                                               0
    < 0) are referred to as hydrophobic attraction. In case of nylon, the attraction is
    further enhanced due to the high (the highest among the four fomites) dispersive
    component of nylon’s surface tension, 𝛾𝑓𝐿𝑊 .
3. Stainless steel surface has a substantial electron acceptor component such that
    𝛾𝑓+ ≅ 𝛾𝑓− . This gives stronger polar interactions with viruses (i. e. higher 𝛾𝑓+ 𝛾𝑣− and
                                                126


     𝛾𝑓− 𝛾𝑣+ ). Notably, these terms are independent of the properties of the continuous
     phase.
4. Polypropylene is an apolar material (𝛾𝑓+ ≅ 0; 𝛾𝑓− ≅ 0). Because the dispersive
     component of PP’s surface tension, 𝛾𝑓𝐿𝑊 , is close to that of water (29.4 vs 21.8
     mJ·m-2; Table 3), dispersive interactions of PP with any virus are also weak (eq. (8)).
     For viruses such that 𝛾𝑣𝐿𝑊 > 𝛾𝑤𝐿𝑊 (which is the case for both HRSV and HAdV5), the
     dispersive interactions are attractive. As a result, hydrophobic attraction is the
     dominant mechanism of virus interaction with PP surface.
Many common materials are monopolar or nearly so with 𝛾𝑓+ very close to zero [94]. For
such materials, substituting eq. (8) and eq. (9) into eq. (10), and posing ∆𝐺𝑣𝑤𝑓 = 0, gives
the following relationship between 𝛾𝑓− and 𝛾𝑓𝐿𝑊 :
                                                                                          2
          √𝛾𝑣𝐿𝑊 𝛾𝑤𝐿𝑊 + √𝛾𝑓𝐿𝑊 𝛾𝑤𝐿𝑊 − √𝛾𝑣𝐿𝑊 𝛾𝑓𝐿𝑊 − 𝛾𝑤𝐿𝑊 − 2√𝛾𝑤− 𝛾𝑤+ + √𝛾𝑤− 𝛾𝑣+ + √𝛾𝑤+ 𝛾𝑣−
𝛾𝑓−  =
                                         √𝛾𝑣+ − √𝛾𝑤+                                           (16)
        [                                                                               ]
In the 2D space with 𝛾𝑓− and 𝛾𝑓𝐿𝑊 as coordinates, eq. (16) corresponds to the boundary
separating all monopolar materials into those that have an overall attractive interaction
(∆𝐺𝑣𝑤𝑓 < 0) and those having an overall repulsive interaction (∆𝐺𝑣𝑤𝑓 > 0) with a given
colloid. Figure 12 illustrates such boundary for the case when the colloid is HRSV2. Both
the monopolar PP and the near-monopolar nylon (Table 3) interact with HRSV favorably.
2
  In his book “Interfacial Forces in Aqueous Media”, van Oss pointed out that electron donor-only
monopolar (𝛾𝑠+ = 0) surfaces with 𝛾𝑠𝐿𝑊 = 40 mJ·m-2 (“a typical value for most biological and many
other organic materials”) are hydrophilic when their 𝛾𝑠− is above 28.3 mJ·m-2. In effect, this
comment referred to a specific point on the ∆𝐺𝑠𝑤𝑠 = 0 curve on a graph similar to Fig. 3 but
drawn for ∆𝐺𝑠𝑤𝑠 .
                                                 127


Note that ∆𝐺𝑣𝑤𝑓 does not account for electrostatic interactions, which, under appropriate
conditions (see section 3.5), can create a substantial primary barrier and prevent
adhesion. Such graph constructed for a given virus with known surface tension
components would allow screening of various candidate materials and selecting ones
with desirable adhesive properties vis-à-vis the virus.
Figure 36. Free energy, ∆𝑮𝒗𝒘𝒇 , of interfacial interaction of HRSV with monopolar
fomites (𝜸𝒇+ = 0) in DI water. The solid line corresponds to ∆𝑮𝒗𝒘𝒇 = 0 (i. e. solution of eq.
(16)).
                                             128


3.4.4 Virus adhesion to fomites: QCM-D measurements
3.4.4.1 Model of mass transfer in QCM-D chamber. Assumptions
Virus transport to the QCM-D sensor surface can be viewed a s a two-step process –
long range transport from the bulk of the flow to the surface followed by a collision event,
which may or may not result in an attachment. The attachment efficiency, 𝛼𝑣𝑓 , is defined
as the probability that a collision results in attachment. The QCM-D part of the present
work is designed to determined 𝛼𝑣𝑓 values so that they can be related to the energy of
virus-fomite interactions. In the simple mass transfer model given by eq. (13), the long-
range transport of a virus and short-range virus-fomite interactions are described by 𝑘
and 𝛼𝑣𝑓 , respectively. The linearity of eq. (13) was tested and confirmed (𝑅 2 > 0.99) in a
subset of QCM-D tests with HAdV5 and SS (see SM, Fig. 43). Further, it was assumed
that deposited virions formed a laterally homogeneous film. In reality, the deposited
layer is laterally heterogeneous, consisting of discrete virions with a solvation shell
contributing to the QCM-D signal to different extents at different coverages. The use of
the Sauerbrey equation is justified, however, in view of the near-overlapping time
dependencies of ∆𝑓𝑛 ⁄𝑛 for different harmonics (see section 3.4.3) and a relatively weak
dissipation signal ∆𝐷𝑛 ⁄(∆𝑓𝑛 ⁄𝑛) << 4×10-7 Hz-1) [41, 95].
                                               129


3.4.4.2 Quantifying mass transfer in QCM-D chamber: Mass transfer coefficients
for HAdV5 and HRSV and size of depositing virions
Based on 𝑑𝑚𝑣 ⁄𝑑𝑡 values measured in tests with PP under conditions of attractive
electrostatic interactions (i. e. assuming 𝛼𝑣𝑓 = 1; Fig. 49a, Fig. 49c), eq. (13) predicted
the mass transfer coefficient, 𝑘, for HRSV and HAdV5 to be 5.97×10-9 (m/s) and
1.87×10-9 (m/s), respectively. The value of 𝑘 for HAdV5 together with HAdV5 diffusion
coefficient measured by DLS (section 3.1) can be used to estimate the effective
thickness of the mass transfer boundary layer in the QCM-D chamber:
     𝐷
𝛿≈                                                                                     (17)
     𝑘
The thickness of boundary layer is defined by the hydrodynamics of the flow in the
chamber and should be the same for both viruses. Given the very narrow size
distribution for the HAdV5 stock (Fig. 10E) and, therefore, a more accurate 𝐷 estimate
for this virus, eq. (17) was applied to HAdV5 to predict 𝛿 ≈ 2.3 mm. This value of 𝛿
together with the value of 𝑘 for HRSV can be used to predict the effective hydrodynamic
diameter of HRSV based on Stokes-Einstein equation. Combining eq. (15) and eq. (17)
gives:
        𝑘𝐵 𝑇
𝑑ℎ =                                                                                   (18)
      3𝜋𝜇𝑘𝛿
Eq. (18) predicts HRSV hydrodynamic diameter of 32 nm, which is an almost exact
match to the higher peak in the size distribution for this virus (Fig. 10B). We conclude
that the smaller size fraction of HRSV is the morphological subset of HRSV virions that
predominantly deposit on the QCM-D sensor surface. While the polydispersity of HRSV
                                             130


(Fig. 10B) confounds a direct application of size data in mass transfer calculations for
this virus, the above analysis shows that QCM-D signal is due to the deposition of
virions from the smaller size fraction of the HRSV population.
3.4.4.3 Virus deposition onto fomites. Virus-fomite attachment efficiency, 𝜶𝒗𝒇
Figure 13 shows a QCM-D dataset recorded in experiments on virus deposition from
150 mM NaCl solution. Representative QCM-D results (both frequency and dissipation
signals) from tests with all virus-fomites pairs are given in SM (Figures 50 – 57). During
the virus deposition stage, the dissipation signal increased monotonously, with a
relatively constant deposition rate, 𝑑𝑚𝑣 ⁄𝑑𝑡, indicating early stages of the deposition
process, far from the jamming limit. Indeed, based on the areal mass density and virion
size, the surface coverage was below 4 % for HRSV and below 1 % for HAdV.
Three additional observations in QCM-D tests require a commentary. First, changes in
the QCM-D signal in response to changes of the background solution (from DI to 150
mM and back) were not immediate. The transient period is due to a finite retention time
within the QCM-D chamber. The lower bound on the retention time in the tubing and the
QCM-D chamber is ~ 1.5 min; the estimate considers QCM-D chamber as an ideal
completely mixed flow reactor so that the actual retention time should be higher.
                                             131


Figure 37. Example QCM-D data set: deposition of HRSV on stainless steel.
Representative QCM-D data for all virus-fomite pairs and deposition conditions are
given in SM (Figures 50 – 57). All calculations based on the QCM-D data were
performed using the signal for the 5th overtone (𝒏 = 5 in eq. (14)). The reason for
choosing the 5th harmonic was that the signals for the 3rd and 1st harmonics were
unstable, likely due to their high sensitivity to mounting stress caused by the O-rings
holding the sensor within the QCM-D chamber [96].
Second, averaged over all 24 QCM-D experiments (8 virus-fomite pairs, triplicate tests),
the frequency shift due to the change of the solution from DI water to 150 mM NaCl was
20.9 ± 4.0 Hz. This was smaller than the shift predicted by the Kanazawa-Gordon
equation (64.6 Hz; see eq. (S1)) based on the density and viscosity of the two solutions
at 20 °C [97, 98]. The discrepancy may be due to a limited sensitivity of the sensors or a
variation in the density of the quartz sensor. Third, the high reproducibility of 𝑑𝑚𝑣 ⁄𝑑𝑡
values for each condition indicated that the limited reuse of QCM-D sensors (for
cleaning protocols see SM, section S4) had little impact on virus adhesion. This is
                                              132


consistent with results of t-testing, which showed that sensor surface energy, ∆𝐺𝑓𝑤𝑓 ,
remained stable (𝑝 > 0.1) after repeated usage-cleaning cycles (Tables 12 and 13 vs
Table 11).
Figure 14 shows values of the attachment efficiency computed based on 𝑑𝑚𝑣 ⁄𝑑𝑡 slopes
in QCM-D tests with various virus-fomite pairs. For both HAdV5 and HRSV, more
favorable interfacial interaction (quantitatively expressed in terms of ∆𝐺𝑣𝑤𝑓 )
corresponded to a higher probability of attachment. Remarkably, normalization by 𝐶𝑏 (a
step in the computation of 𝛼𝑣𝑓 ), made the 𝛼𝑣𝑓 vs ∆𝐺𝑣𝑤𝑓 dependencies for both viruses
collapse into one trend, pointing to the possibility of using of ∆𝐺𝑣𝑤𝑓 as a predictor of
virus adhesion. Whether this result holds true for other viruses and fomites warrants
further testing.
As described in section 2.5.1, the determination of 𝛼𝑣𝑓 relies on quantifying the mass
transfer coefficient, 𝑘, which is based on tests performed under conditions of highly
favorable virus-fomite interaction with 𝛼𝑣𝑓 = 1. To model this scenario, QCM-D tests
were performed with polypropylene as the most hydrophobic of the four fomites (to
                                      𝐴𝐵
maximize hydrophobic attraction 𝑈𝑣𝑤𝑓     ), and under conditions when the electrostatic
                               𝐸𝐿
interactions are favorable (𝑈𝑣𝑤𝑓   < 0) and strongest. The latter conditions were achieved
by a) adjusting pH to be within the pH range bracketed by the pI values of
polypropylene and the virus in question (see SM, section S6) and b) switching to low
ionic strength electrolyte (1 mM NaCl) to minimize screening of attractive electrostatic
interactions.
                                              133


Figure 38. Virus-fomite attachment efficiency as a function of the free energy of virus-
fomite interaction in water. Empty symbols correspond to tests with 150 mM NaCl
electrolyte as the deposition solution. Filled symbols correspond to tests with virus
deposition onto polypropylene from 1 mM NaCl at pH 4.1 for HAdV5 and at pH 3.8 for
HRSV where virus-fomite electrostatic interaction is favorable and at its maximum
absolute value. (See SM, section S6 for the algorithm used to select these pH values.).
Figure 15 gives examples of two QCM-D datasets that correspond to the least
hydrophobic fomite-virus pair (silica and HAdV5 with ∆𝐺𝑓𝑤𝑓 of 15.7 and - 27.7 mJ·m-2,
respectively) and the most hydrophobic one (HRSV and PP with ∆𝐺𝑓𝑤𝑓 of - 40.2 and -
88.9 mJ·m-2, respectively). As mentioned earlier, one simplification behind the
presented approach is that the deposited layer is assumed to be laterally homogeneous.
We note that this assumption provides a conservative estimate on the difference
between hydrophilic and hydrophobic fomites in terms of their adhesiveness. The
contribution of the solvent to the QCM-D signal should be higher for attached particles
                                             134


with a larger shell of associated solvent, especially at lower coverages [95], such as
those (< 4%) that occurred in our QCM-D tests. Thus, we expect that due to the higher
contribution of the solvation shell, the values of 𝛼𝑣𝑓 are overpredicted for more
hydrophilic virion-fomite pairs; therefore, the contrast between PP and silica should be
more pronounced than what is apparent from Fig. 14.
                                              135


 (a)
 (b)
Figure 39. Mass and dissipation QCM-D data for the deposition of a) HRSV on
polypropylene and b) HAdV5 on silica. Insets illustrate areal mass density values
computed based on 5th, 7th, 9th, and 11th harmonics of the QCM-D signal.
                                           136


3.4.5 XDLVO predictions
Virus properties required as inputs to XDLVO model included virus size as well as
surface energy and 𝜁-potential of viruses and fomites. The 𝜁-potential of the four fomites
as a function of pH was obtained from literature [36, 74, 75]. There was a significant
variation in the reported values of 𝜁-potential of stainless steel [75, 99, 100] likely due to
different degrees of passivation of the surface. While pure metals have very high
surface energy, fast oxidation in air passivates the surface and minimizes its energy
through the formation of a metal oxide bilayer. The passive film on the stainless steel
surface consists of an inner sub-layer with segregated chromium oxide and an outer
sublayer enriched in iron oxyhydroxide [101-103]. For XDLVO simulations, we adopted
surface charge data reported by Hedberg et al. [75].
For both viruses, the XDLVO model predicted attractive interactions with nylon, SS and
PP at all virus-fomite separation distances (Figures 16a, 45a, 45c, 46a, 47a, 47c, 48a)
due to strong hydrophobic attraction. The XDLVO energy of interaction between silica
and each of the two viruses featured a primary maximum and a secondary minimum.
For HAdV5/silica (Fig. 16c), the corresponding energy values were 200 kT and - 3.0 kT;
the reason for the very high primary energy barrier was the strong repulsive Lewis acid-
base interaction between HAdV5 and silica (“hydration pressure”) as indicated by the
positive value of ∆𝐺𝑑𝐴𝐵
                      0
                        (~7.9 mJ·m-2; Table 4). For HRSV/silica (Fig. 48c), the primary
maximum was much lower (21 kT) while the secondary minimum was shallower (– 2.3
kT). The repulsive electrostatic interaction between HRSV and silica results in slightly
                                             137


repulsive total interaction even though LW and AB interactions are both attractive (∆𝐺𝑑𝐿𝑊     0
= -5.5 J·m-2, ∆𝐺𝑑𝐴𝐵
                  0
                     ≈ - 0.4 J·m-2; Table 4). During DI water rinse, there was always an
energy barrier for all virus-fomite pairs except for the condition of 𝛼𝑣𝑓 = 1. In XDLVO
simulations describing the DI rinse, the ionic strength of the solution was assumed to be
10-5 M.
In XDLVO energy profiles corresponding to the conditions of DI water rinse (Fig. 16b,
Fig 16d), the energy barriers were observed in all the cases at longer distance (> 5 nm)
except for HAdV5/silica, where barrier was located at the distance of ~ 0.4 nm. The
appearance of energy barriers is due to the increased importance of repulsive
electrostatic interactions. In high ionic strength solution, the electrostatic interaction is
                                                                               𝐸𝐿
limited due to the compression of the electric double layer, making the 𝑈𝑣𝑤𝑓       decrease
much faster with distance. With the lowering of the ionic strength, this restriction
                        𝐸𝐿
disappears, making 𝑈𝑣𝑤𝑓     higher at longer distance. This is consistent with QCM-D
results showing a much more effective removal of viruses when switching from 150 mM
rinse to DI water rinse.
                                              138


                     (a)                                       (b)
                     (c)                                       (d)
Figure 40. XDLVO total energy of interactions during virus deposition (a, c) and
cleaning (b, d) stages of QCM-D experiments. The results are for HRSV interaction with
polypropylene (a, b) and HAdV5 interaction with silica (c, d).
                                         139


While outside of the scope of the present work, testing non-polar liquids as cleaning
agents for virus removal is also of interest.
The XDLVO modeling complements interfacial energy analysis (Table 4) by providing
insights into the relative importance of electrostatic interactions as a function of ionic
strength. Accounting for classical DLVO (LW, EL) as well as polar (AB) interactions, the
model predicts that at lower ionic strengths a solution with an appropriate pH deters
adhesion and can help resuspend attached viruses. Separation distances that
correspond to primary maxima (barriers for irreversible adhesion) and secondary
minima (loci for possible reversible adhesion) correspond to a fraction of a virus
diameter. Thus, for solution chemistries that discourage adhesion, a very minor
disturbance in a virus’ position at the surface may lead to detachment. While almost no
removal of viruses was observed by flowing 150 mM NaCl solution after virus deposition,
removal of 28.8% on average was possible after switching to DI water as the rinsing
solution. These findings are also consistent with the results reported by Liu et al who
employed SPR imaging to study adsorption of SARS-CoV-2 pseudovirus onto model
surfaces [40]. Under conditions when the virus and the surface carried the charge of the
same sign, adsorption was stronger at a higher ionic strength where the repulsive
electrostatic interactions are screened out.
In general, anti-adhesion surfaces should have high positive value of the interfacial free
energy of its interaction with a virus, ∆𝐺𝑣𝑤𝑓 (see Fig. 12). Consistent with XDLVO
                                             140


modelling results, QCM-D data show that for both HAdV and HRSV, the efficiency of
their attachment to fomites is correlated with ∆𝐺𝑣𝑤𝑓 . For a virus with known surface
tension parameters, this knowledge can be used to select surfaces with as high ∆𝐺𝑣𝑤𝑓
as practically possible to deter virus adhesion. The proposed approach can also guide
the selection of surface cleaning solutions and protocols. When the virus and surface
carry electrical charges of the same sign, a cleaning solution with low ionic strength will
allow for electrostatic repulsion. Whereas if the virus and surface are oppositely charged,
cleaning solution with a high ionic strength can screed out attractive electrostatic
interactions. Both strategies should promote virus detachment.
3.5 Conclusions
The present work considers interactions of two human enteric viruses – one enveloped
(HRSV) and one non-enveloped (HAdV5) – with four fomites (silica, nylon, stainless
steel, polypropylene). The selected fomites are tentatively identified as “archetypes”
representing surfaces that are distinctly different in mechanisms of their interfacial
interactions. The surfaces are typified on the basis of their surface energy components:
− Archetype 1 (e. g. silica) is characterized by a large electron donor component such
    that 𝛾𝑓− ≫ 𝛾𝑓+ . For these surfaces, hydration pressure can overcome water cohesion
    to result in an overall repulsion of viruses.
− Archetype 2 (e. g. nylon) is less polar than archetype 1 and has high dispersive
    component, 𝛾𝑓𝐿𝑊 . A near-monopolar surface (𝛾𝑓+ ≅ 0), archetype 2 also has a smaller
                                              141


    𝛾𝑓− . As a result, the electron-donor attraction between archetype 2 and viruses is
    weaker than water cohesion leading to hydrophobic attraction that is further
    enhanced due to dispersive forces.
− Archetype 3 (e. g. stainless steel) has a substantial electron acceptor component
    such that 𝛾𝑓+ ≅ 𝛾𝑓− . This gives stronger polar interactions with viruses (i. e. higher
    𝛾𝑓+ 𝛾𝑣− and 𝛾𝑓− 𝛾𝑣+ ).
− Archetype 4 (e. g. polypropylene) is an apolar material (𝛾𝑓+ ≅ 0; 𝛾𝑓− ≅ 0) with the
    dispersive component, 𝛾𝑓𝐿𝑊 , close to that of water. Hydrophobic attraction is the
    dominant mechanism of virus interaction with such surfaces.
Areal mass density on the fomite surface is found to correlate with the free energy of
virus-fomite interfacial interaction in water, ∆𝐺𝑣𝑤𝑓 . For both HAdV5 and HRSV, more
negative ∆𝐺𝑣𝑤𝑓 values correspond to higher virus-fomite attachment efficiencies.
Moreover, 𝛼𝑣𝑓 vs ∆𝐺𝑣𝑤𝑓 dependencies for HAdV5 and HRSV collapse into one trend
pointing to the possibility of using ∆𝐺𝑣𝑤𝑓 as a predictor of virus adhesion. Virus
deposition from 150 mM NaCl electrolyte is defined by polar interactions. Under
conditions of low ionic strength, however, electrostatic forces emerge as a dominant
interaction at a longer range.
Through the analysis of relative contributions of separate surface tension components
to the energy of interfacial interaction, the study describes a possible approach to the
selection of surfaces with desired adhesion properties. The methodology helps fill the
                                               142


knowledge gap on virus adhesion to fomites – an important component of models on
virus transport and fate in built environments. The study can help guide screening and
selection of materials that discourage virus adhesion, design of anti-adhesive surfaces,
as well as development of surface cleaning solutions and protocols.
Acknowledgements
This material is based upon work supported in part by the U.S. National Science
Foundation under Grant OISE-1952438 and in part by the Research Council of Norway
under project #310074. We are grateful to Dr. Wei Zhang (Department of Plant, Soil and
Microbial Sciences, Michigan State University) for making available the Malvern
Zetasizer Nano-ZS instrument in his laboratory as well as to Dr. Irene Xagoraraki and
Brijen Miyani (Department of Civil and Environmental Engineering, Michigan State
University) for providing training on the procedures for RNA and DNA extraction.
                                             143


Nomenclature
𝛼𝑣𝑓    virus-fomite attachment efficiency
𝛾𝑙𝐿𝑊   Lifshitz-van der Waals component of the surface energy of probe liquid
𝛾𝑠𝐿𝑊   Lifshitz-van der Waals component of the surface energy of solid
𝛾𝑤𝐿𝑊   Lifshitz-van der Waals component of the surface energy of water
𝛾𝑙+    electron acceptor component of the surface energy of probe liquid
𝛾𝑠+    electron acceptor component of the surface energy of solid
𝛾𝑤+    electron acceptor component of the surface energy of water
𝛾𝑙−    electron donor component of the surface energy of probe liquid
𝛾𝑠−    electron donor component of the surface energy of solid
𝛾𝑤−    electron donor component of the surface energy of water
𝛿      effective diffusion distance
𝜀𝑟     relative dielectric permittivity of water
𝜀0     dielectric permittivity of vacuum
𝜁      𝜁-potential
𝜃      contact angle
𝜅      inverse Debye screening length (also referred to as Hϋckel parameter)
λ      (=0.6 nm) the decay length for Lewis acid-base interaction in water
𝜇      dynamic viscosity
𝜓𝑣     surface potential of a virus
𝜓𝑓     surface potential of a fomite
                                            144


𝐴     Hamaker constant
𝐴𝑠    surface area of the QCM sensor
𝑎     virus radius
𝐶     QCM sensor mass sensitivity constant
𝐶𝑏    virus concentration in the bulk of the flow
𝐶𝑠    virus concentration at the sensor surface
𝐷     diffusion coefficient
𝑑     virus-fomite minimal separation distance
𝑑ℎ    hydrodynamic diameter
𝑑0    minimum separation distance due to Born repulsion
𝑘     mass transfer coefficient
𝑘𝐵    Boltzmann’s constant
𝐸𝑚𝑎𝑥  primary maximum in the total XDLVO energy of interaction
𝐸𝑚𝑖𝑛  secondary minimum in the total XDLVO energy of interaction
∆𝑓    QCM vibration frequency shift
∆𝐺𝑓𝑤𝑓 free energy of interfacial interaction of two identical fomites in water
∆𝐺𝑠𝑤𝑠 free energy of interfacial interaction of two identical solids in water
∆𝐺𝑣𝑤𝑣 free energy of interfacial interaction of two identical viruses in water
∆𝐺𝑣𝑤𝑓 free energy of virus-fomite interfacial interaction in water
𝑗𝑚    mass flux of virus towards the QCM sensor
                                          145


𝑚       areal mass density
𝑛       QCM vibration overtone number
  𝑋𝐷𝐿𝑉𝑂 total XDLVO energy of interaction between a sphere and a plate
𝑈𝑠𝑙𝑣
  𝐿𝑊
𝑈𝑣𝑤𝑓    energy of Lifshitz-van der Waals interaction between a sphere and a plate
  𝐸𝐿
𝑈𝑣𝑤𝑓    energy of electrostatic interaction between a sphere and a plate
  𝐴𝐵
𝑈𝑣𝑤𝑓    energy of Lewis acid-base interaction between a sphere and a plate
𝑇       absolute temperature
                                           146


APPENDIX
  147


                                          APPENDIX
A1. Transmission electron microscopy of HAdV5 and HRSV
In correlation with using dynamic light scattering as a method to measure the particle
size, transmission electron microscopy can also be used to determine particle size,
shape, and chemical makeup. TEM provides the primary particle size (exact particle
diameter) in comparison to the hydrodynamic diameter, obtained using DLS, which
tends to be larger. If agglomeration of the particles is present, larger particle size
distributions or multi-model particle size distributions will most likely occur using DLS.
TEM is the preferred method for determining the primary particle size of dilute colloidal
dispersions [104]. In order to avoid exposure of lab personnel and the public by
accidentally released viruses during handling of HAdV5 and HRSV, efficient and
feasible safety precautions should be addressed. Virus sample inactivation was
performed to guarantee biosafety. The inactivation procedures must not only inactivate
the virus, but also preserve their morphology. In diagnostic negative staining EM,
Formaldehyde (FA) at a concentration of 2% to 4% is commonly used because it
stabilizes biological structures [105]. In our study, 37% FA was added in the virus
suspension to reach a final concentration of 3.7% of FA. The mixture was vortexed for
15 s and then incubated in 25 °C for 12 h as suggested by Moeller et al. [105]. To
prepare specimen for TEM imaging, one drop of purified virus stock was applied to a
carbon-coated formvar grid. The virus sample was incubated on the grid for 5 min and
the washed away by drops of water. Uranyl acetate (1%) was then used to stain the grid,
and excess stain was removed with filter paper. The grid was air-dried prior to TEM
                                             148


imaging. Images were recorded using JEM-1400 Flash (Jeol, Nieuw-Vennep, Tokyo).
High accelerating voltages give higher resolution, but less contrast, and can result in
greater specimen damage. For these reasons, studies of biological samples tend to
employ low accelerating voltages (60 kV to 100 kV), while studies of inorganic materials,
which often require higher resolution, usually employ an accelerating voltage of 200 kV.
In this study, 100 kV accelerating voltage was utilized.
A2. QCM-D experiments: Avoiding air bubbles in tests with hydrophobic sensors
When an aqueous solution is brought into contact with a dry hydrophobic surface of a
QCM-D sensor, air bubbles tend to form on the surface confounding QCM-D
measurements [106-108]. Accordingly, in tests with hydrophobic PP sensors, we
adopted a different protocol to establish a baseline All sample solutions (except virus
suspensions) were first sonicated for at least 30 min in the air, and then degassed in a
vacuum desiccator. Virus suspensions were degassed in a vacuum desiccator but not
sonicated. An Erlenmeyer flask was filled with solution so there was no head space.
The sample solution was kept in the vacuum desiccator for at least 15 s and was
immediately sealed with parafilm after being taken out of the vacuum desiccator to
avoid re-aeration. The QCM-D tubing was inserted into the solution through a parafilm
cover. Further, before each QCM-D test, the PP sensors were flowed first with ethanol
for at least 2 h at 0.1 ml/min [106] and then with the DI water for at least 60 min while
QCM frequency was recorded. No significant frequency changes during this time
indicated absence of air bubbles. Finally, the baseline was established with DI water at
                                            149


0.15 ml/min flow rate. The rest of the procedure was the same as with hydrophilic (SiO 2,)
and less hydrophobic (nylon-6,6, SS) sensors.
A3. QCM-D experiments: Accounting for the effect of liquid viscosity and density
The effect of fluid’s viscosity and density on QCM-D sensor vibration frequency when
switching from DI water to the electrolyte solutions (150 mM NaCl) was accounted for
using eq. (S1), which is commonly attributed to Kanazawa and Gordon [109, 110]. (See
Revyakine et al. [111] for references to earlier reports where the equation appears). The
relationship expresses the change in the oscillation frequency of a quartz crystal in
contact with a fluid as a function of materials parameters of the fluid and the quartz:
             𝜂𝐿 𝜌𝐿
∆𝑓 = −√𝑓03                                                                           (S1)
            𝜋𝜇𝑄 𝜌𝑄
where 𝜇𝑄 = 2.974×1011 g∙cm-1·s-2 (2.974×1010 Pa) and 𝜌𝑄 = 2.648 g∙cm-3 are the shear
modulus and density of quartz, respectively; 𝑓0 = 25 MHz is the nominal vibration
frequency of the dry quartz crystal at the 5th harmonic, 𝜂𝐿 and 𝜌𝐿 are the absolute
viscosity and density of the liquid respectively [109, 110]. For pure water at 25 °C, 𝜂𝐿
and 𝜌𝐿 are 0.89×10-3 Pa·s and 0.997 g∙cm-3, respectively. For 150 mM NaCl solution at
25 °C, 𝜂𝐿 and 𝜌𝐿 are 0.9×10-3 Pa·s [112] and 1.003 g·cm-3 [113], respectively.
                                             150


A4. QCM-D experiments: Cleaning protocols
Protocol for cleaning QCM-D tubing
After the test with viruses, the fluid path (including both tubing and QCM-D chamber)
was first disinfected with bleach using the following steps:
   1) Priming the fluid path with deionized water.
   2) Flushing the fluid path with a 0.625% NaOH solution. The solution was prepared
        by diluting commercial bleach (12.5% NaOH) with DI water.
   3) Pausing the priming sequence.
   4) After 30 min of leaving the fluid path filled with the bleach solution and no
        pumping, priming the fluid path using DI water to remove the remaining bleach
        solution from the tubing and into the waste container.
After disinfection, the fluid path was cleaned to remove any virions possibly remaining
on the inner surfaces of the tubing and the QCM-D chamber. The procedure included
the following steps:
   1) Flushing the fluid path with a 2% SDS solution
   2) Pausing the priming sequence.
   3) After 30 min of leaving the fluid path filled with the SDS solution and no pumping,
        resume flushing and prime the fluid path using deionized water to remove the
        remaining SDS solution from the tubing and into the waste container.
                                               151


Protocols for cleaning stainless steel and silica sensors
The cleaning protocol for stainless steel sensor included the following seven steps:
   1. Immersing in 1% Hellmanex II for 12 h at room temperature.
   2. Rinsing with DI water.
   3. Drying with nitrogen gas.
   4. Sonicating in 99% ethanol for 10 min.
   5. Rinsing with DI water.
   6. Drying with nitrogen gas.
   7. Treating with UV and O3 (Bioforce UV-Ozone Cleaner) for 60 min.
The cleaning protocol for SiO2 sensor (QSX 303) included the following five steps:
   1. Treating with Bioforce UV-Ozone Cleaner for 10 min.
   2. Sonicating in 2% SDS aqueous solution for 60 min.
   3. Rinsing with DI water.
   4. Drying with nitrogen gas.
   5. Treating with UV and O3 (Bioforce UV Ozone Cleaner) for 60 min.
Protocols for cleaning polypropylene and nylon sensors
The manufacturer does not provide the standard cleaning protocols for polypropylene
and nylon-6,6 sensors. The cleaning procedure for these two types of sensors was
designed based in part on that for polystyrene and included the following five steps:
   1. Immersing in 1% Deconex 11 solution for 60 min at room temperature.
   2. Rinsing with DI water and keep immersed in DI water for 2 h.
                                            152


    3. Rinsing with 99% ethanol.
    4. Drying with nitrogen gas.
A5. Determination of ζ–potential based on measured electrophoretic mobility
Experimentally determined values of HRSV electrophoretic mobility as a function of pH
(Fig. 41) were converted to 𝜁–potential values, using an expression derived by Ohshima
[114]:
                      2 𝜀𝑟 𝜀0                 1
                  𝜇=          𝜁 1+                         3                        (S2)
                      3 𝜂                       2.5
                                   2 [1 +                ]
                               [          𝜅𝑎[1 + 2𝑒 −𝜅𝑎 ] ]
where 𝜅 (nm-1) is the Hückel parameter (𝜅 −1 (nm) is the Debye screening length), 𝜂 is
the viscosity of the electrolyte where viruses are suspended, and 𝑎 is the radius of the
virus. Equation (S2) provides an accurate (< 1% error) estimate of ζ–potential for any
value of 𝜅𝑎. For HRSV in 1 mM KCl, 𝜅𝑎 ≈ 6.4, making neither Smoluchowski expression
(assumes 𝜅𝑎 ≫ 1) nor Hückel expression (assumes 𝜅𝑎 ≪ 1) for the relationship
between 𝜁 and 𝜇 applicable. The Ohshima equation was employed in our earlier study
to determine 𝜁-potential of HAdV5 [115].
                                             153


Figure 41. Electrophoretic mobility of HRSV virions as a function of pH. Table 10 lists
mobility and zeta potentials as functions of pH in a tabular format.
A6. Algorithm for selecting the pH value where electrostatic interactions between
a virus and a fomite surface are most attractive
                                  𝐸𝐿
To determine the pH when 𝐸𝑣𝑤𝑓         is most attractive, a few pH values were first selected
                                                                                    𝐸𝐿
between isoelectric point (pI) of virus and PP to ensure the attraction of 𝑈𝑣𝑤𝑓        . For
example, the pls of HAdV5 and PP were 4.6 and 3.8 respectively, so 3.9, 4.0, 4.1, 4.2,
                                             𝐸𝐿
4.3, 4.4, 4.5 were selected. Then the 𝐸𝑣𝑤𝑓        at these pH were calculated and compared.
                         𝐸𝐿
The pH at which the 𝑈𝑣𝑤𝑓      reaches the highest value was selected for these QCM-D
                       1+𝑒 −𝜅𝑑
tests. In eq. (5), 𝑙𝑛 (1−𝑒 −𝜅𝑑 ) is positive for all distances, (𝜓𝑣2 + 𝜓𝑓2 ) 𝑙𝑛(1 − 𝑒 −2𝜅𝑑 ) is
                                            𝐸𝐿
negative for all distances, Thus, for 𝑈𝑣𝑤𝑓      (𝑑) to be highly negative (virus-fomite
attraction), the 𝜓𝑣 𝜓𝑓 should be highly negative.
                                                 154


 𝐸𝐿
                                1 + 𝑒 −𝜅𝑑
𝑈𝑣𝑤𝑓 (𝑑) = 𝜋𝜀𝑟 𝜀0 𝑎 [2𝜓𝑣 𝜓𝑓 𝑙𝑛 (          ) + (𝜓𝑣2 + 𝜓𝑓2 ) 𝑙𝑛(1 − 𝑒 −2𝜅𝑑 )]          (5)
                                1 − 𝑒 −𝜅𝑑
               𝐸𝐿
Based on the 𝑈𝑣𝑤𝑓   calculated, 4.1 and 3.85 was selected as the experimental pH for
HAdV5-PP and HRSV-PP respectively.
                                             155


                     (a)                                   (b)
                     (c)                                   (d)
Figure 42. TEM images of (a, b) HAdV5 virions and (c, d) HRSV virions.
                                       156


A7. Verification of linearity of the concentration dependence of virus mass flux to
the sensor
                                                                         𝒅𝒎
Figure 43. Rate of HAdV5 deposition on stainless steel QCM-D sensors ( 𝒅𝒕𝒗) as a
function of HAdV5 concentration in the feed suspension (𝑪𝒃 ). The intermediate value of
𝑪𝒃 and deposition conditions are the same as in all other QCM-D tests. Solution is 150
                                                                          𝒅𝒎
mM NaCl. The observed linearity validates the assumption behind eq. (3): 𝒅𝒕𝒗 ≈
𝜶𝒗𝒇 𝒌𝑪𝒃 𝑨.
A8. Computing surface tension components for polypropylene
Eq. (1) was recast as
(1 + 𝑐𝑜𝑠𝜃)𝛾𝑙𝑇𝑂𝑇 − 2√𝛾𝑠𝐿𝑊 𝛾𝑙𝐿𝑊            𝛾𝑙+                                (S3)
                              = √𝛾𝑠+ + √     √𝛾 −
            2√𝛾𝑙−                        𝛾𝑙− 𝑠
                                            157


and 𝛾𝑠− was determined from the slope of a linear dependence of the left-hand side of
eq. (S3) on √𝛾𝑙+ ⁄𝛾𝑙− with the intercept (√𝛾𝑠+ ) set to zero. To improve the reliability of the
prediction, additional contact angle measurements with EG, a probe liquid with a high
√𝛾𝑙+ ⁄𝛾𝑙− ratio, were performed.
                                             158


                     a)                                              b)
                     c)                                              d)
Figure 44. Contact angle of four fomites: a) silica, b) nylon, c) stainless steel, d)
polypropylene, as a function of pH. Depending on the type of aqueous solution used,
values are shown using either empty symbols (150 mM NaCl at pH 5.8) or gray symbols
(DI water at pH 5.8) or black symbols (1 mM NaCl at pH 3.8 and 4.1).
                                           159


                     (a)                                        (b)
                     (c)                                        (d)
Figure 45. XDLVO total energy of interactions during virus deposition (a, c) and
cleaning (b, d) stages of QCM-D experiments. The results are for HRSV interaction with
polypropylene (a, b) and stainless steel (c, d).
                                            160


                      (a)                                       (b)
                      (c)                                       (d)
Figure 46. XDLVO total energy of interactions during virus deposition (a, c) and
cleaning (b, d) stages of QCM-D experiments. The results are for HRSV interaction with
nylon (a, b) and silica (c, d).
                                         161


                     (a)                                        (b)
                     (c)                                        (d)
Figure 47. XDLVO total energy of interactions during virus deposition (a, c) and
cleaning (b, d) stages of QCM-D experiments. The results are for HAdV5 interaction
with polypropylene (a, b) and stainless steel (c, d).
                                           162


                      (a)                                       (b)
                      (c)                                       (d)
Figure 48. XDLVO total energy of interactions during virus deposition (a, c) and
cleaning (b, d) stages of QCM-D experiments. The results are for HAdV5 interaction
with nylon (a, b) and silica (c, d).
                                         163


                     (a)                                        (b)
                     (c)                                        (d)
Figure 49. XDLVO total energy of interactions during virus deposition (a, c) and
cleaning (b, d) stages of QCM-D experiments. The results are for HRSV (a, b) and
HAdV5 (c, d) interaction with polypropylene Conditions correspond to 𝜶𝒗𝒇 = 1.
                                          164


Figure 50. Example QCM-D data set: deposition of HRSV on polypropylene. The data
are based on the 5th harmonic of the QCM-D signal.
                                        165


Figure 51. Example QCM-D data set: deposition of HRSV on stainless steel. The data
are based on the 5th harmonic of the QCM-D signal.
                                        166


Figure 52. Example QCM-D data set: deposition of HRSV on nylon. The data are based
on the 5th harmonic of the QCM-D signal.
                                         167


Figure 53. Example QCM-D data set: deposition of HRSV on silica. The data are based
on the 5th harmonic of the QCM-D signal.
                                         168


Figure 54. Example QCM-D data set: deposition of HAdV5 on polypropylene. The data
are based on the 5th harmonic of the QCM-D signal
                                        169


Figure 55. Example QCM-D data set: deposition of HAdV5 on stainless steel. The data
               are based on the 5th harmonic of the QCM-D signal.
                                       170


Figure 56. Example QCM-D data set: deposition of HadV5 on nylon. The data are
based on the 5th harmonic of the QCM-D signal
                                        171


Figure 57. Example QCM-D data set: deposition of HAdV5 on silica. The data are
               based on the 5th harmonic of the QCM-D signal.
                                    172


Table 10. 𝜻-potential values computed using Ohshima’s expression (eq. (S1)) based on
measured electrophoretic mobilities of HRSV as a function of pH.
     pH                   Mobility (μm/s)/(V/cm)                 𝜁-potential (mV)
 3.22 ± 0.02                  0.613 ± 0.079                          10.0 ± 1.3
 3.82 ± 0.07                  0.119 ± 0.059                          1.9 ± 1.0
 4.51 ± 0.05                  -0.793 ± 0.092                        -12.9 ± 1.5
 4.74 ± 0.04                  -1.527 ± 0.079                        -24.8 ± 1.3
 5.74 ± 0.11                  -2.054 ± 0.143                        -33.4 ± 2.3
 6.65 ± 0.11                  -2.546 ± 0.107                        -41.4 ± 1.7
 7.71 ± 0.08                  -2.852 ± 0.086                        -46.4 ± 1.4
 9.00 ± 0.10                  -2.985 ± 0.142                        -48.6 ± 2.3
                                           173


Table 11. Contact angle of probe liquids, surface energy parameters (𝜸𝑳𝑾 , 𝜸+ , 𝜸− , 𝜸𝑨𝑩 , 𝜸𝒕𝒐𝒕 ) and the free energy of
interfacial interaction in water (∆𝑮𝒔𝒘𝒔 ) of four fomites (clean, before use). Surface energies are calculated based on
contact angles of 150 mM NaCl instead of DI water. Corresponding surface energy values calculated based on contact
angles of DI water are given in Table 3. Error estimates were obtained by propagating experimental errors in measured
contact angles through the calculation of 𝜸 (eq. (1)) and ∆𝑮𝒔𝒘𝒔 (eq. (3)).
Notes: A pH 5.8; B The data are the same as those reported in Table 3.
                                                                                         Fomites
                                        Parameter
                                                              Silica        Nylon              Stainless steel   Polypropylene
                                        150 mM NaCl A       11.3 ± 1.0    47.8 ± 0.7             59.0 ± 0.8       99.0 ± 0.7
 Surface energy (mJ/m2) Contact angle
                                             Glycerol B     14.7 ± 1.5    50.5 ± 1.0             31.4 ± 0.8       89.4 ± 1.6
                                                    DID B   31.4 ± 2.6    12.4 ± 0.9             50.2 ± 1.5       58.6 ± 2.9
                                                    EG B       n/a            n/a                    n/a          77.5 ± 1.8
                                                            43.6 ± 1.1    49.6 ± 0.2             34.2 ± 0.8        29.4± 1.7
                                                            2.2 ± 0.2    0.003 ± 0.008             6.5 ± 0.4           0
                                                            44.9 ± 0.7    30.1 ± 1.2               8.8 ± 0.7       0.7 ± 0.2
                                                            19.7 ± 0.6    0.59 ± 0.15            15.1 ± 0.8            0
                                                            63.3 ± 1.2    50.2 ± 0.22            49.3 ± 1.1       29.4 ± 1.7
                                                            16.1 ± 1.1    - 2.6 ± 2.3           - 23.6 ± 1.4      - 85.9 ± 2.8
                                                                                         174


Table 12. Contact angle of probe liquids, surface energy parameters (𝜸𝑳𝑾 , 𝜸+ , 𝜸− , 𝜸𝑨𝑩 , 𝜸𝒕𝒐𝒕 ) and the free energy of
interfacial interaction in water (∆𝑮𝒔𝒘𝒔 ) of four fomites (after use and subsequent cleaning; see section S4 for the
description of cleaning protocols). Error estimates were obtained by propagating experimental errors in measured contact
angles through the calculation of 𝜸 (eq. (1)) and ∆𝑮𝒔𝒘𝒔 (eq. (3)).
Notes: A pH 5.8; B The data are the same as those reported in Table 3.
                                                                                      Fomites
                                        Parameter
                                                            Silica       Nylon              Stainless steel   Polypropylene
                                                H2O A     14.9 ± 1.6   51.6 ± 2.1             63.1 ± 1.2       102.4 ± 2.3
 Surface energy (mJ/m2) Contact angle
                                              Glycerol    24.4 ± 2.7   47.5 ± 2.3             37.2 ± 1.3       91.0 ± 0.6
                                                    DID   37.6 ± 2.0   15.7 ± 1.9             45.3 ± 1.6       59.7 ± 1.2
                                                    EG       n/a           n/a                    n/a          78.1 ± 1.5
                                                          40.8 ± 1.0   48.9 ± 0.5             36.9 ± 0.8       28.8 ± 0.7
                                                          1.8 ± 0.3    0.14 ± 0.13              4.9 ± 0.5           0
                                                          48.0 ± 1.9   25.5 ± 2.1               6.9 ± 1.1      0.38 ± 0.14
                                                          18.5 ± 1.1    3.7 ± 0.7             11.7 ± 1.1            0
                                                          59.3 ± 1.5   52.7 ± 0.9             48.6 ± 1.4       28.8 ± 0.7
                                                          22.0 ± 2.3   - 10.9 ± 3.8          - 31.3 ± 2.5      - 90.5 ± 2.3
                                                                                      175


Table 13. Contact angle of probe liquids, surface energy parameters (𝜸𝑳𝑾 , 𝜸+ , 𝜸− , 𝜸𝑨𝑩 , 𝜸𝒕𝒐𝒕 ) and the free energy of
interfacial interaction in water (∆𝑮𝒔𝒘𝒔 ) of four fomites (after use and subsequent cleaning; see section S4 for the
description of cleaning protocols). Surface energies are calculated based on contact angles of 150 mM NaCl instead of DI
water. Error estimates were obtained by propagating experimental errors in measured contact angles through the
calculation of 𝜸 (eq. (1)) and ∆𝑮𝒔𝒘𝒔 (eq. (3)).
Notes: A pH 5.8; B The data are the same as those reported in Table 11.
                                                                                       Fomites
                                        Parameter
                                                              Silica       Nylon             Stainless steel   Polypropylene
                                        150 mM NaCl A       12.7 ± 1.1   47.5 ± 0.7            62.0 ± 1.6       101.7 ± 1.9
 Surface energy (mJ/m2) Contact angle
                                             Glycerol B     24.4 ± 2.7   47.5 ± 2.3            37.2 ± 1.3       91.0 ± 0.6
                                                    DID B   37.6 ± 2.0   15.7 ± 1.9            45.3 ± 1.6       59.7 ± 1.2
                                                    EG B       n/a          n/a                    n/a          78.1 ± 1.5
                                                            40.8 ± 1.0   48.9 ± 0.5            36.9 ± 0.8       28.8 ± 0.7
                                                            1.7 ± 0.3    0.09 ± 0.10             4.7 ± 0.5           0
                                                            49.3 ± 1.7   28.2 ± 2.0              7.9 ± 1.4      0.47 ± 0.16
                                                            18.4 ± 1.1    3.2 ± 0.7            12.2 ± 1.3            0
                                                            59.2 ± 1.4   52.1 ± 0.8            49.1 ± 1.6       28.8 ± 0.7
                                                            23.5 ± 2.1   - 5.9 ± 3.6          - 29.8 ± 3.1      - 89.1 ± 2.3
                                                                                       176


Table 14. Contact angle of probe liquids, surface energy parameters (𝜸𝑳𝑾 , 𝜸+ , 𝜸− , 𝜸𝑨𝑩 , 𝜸𝒕𝒐𝒕 ) and the free energy of
interfacial interaction in water (∆𝑮𝒔𝒘𝒔 ) of polypropylene at pH 4.1 and 3.8.
                         Contact angle                                       Surface energy (mJ/m2)
    pH
            1 mM NaCl Glycerol         DID              𝛾 𝐿𝑊        𝛾+        𝛾−        𝛾 𝐴𝐵        𝛾 𝑡𝑜𝑡         ∆𝐺𝑠𝑤𝑠
    4.1     102.8 ± 1.4 91.0 ± 0.6     59.7 ± 1.2    29.4 ± 1.7   0±0     0.23 ± 0.09  0±0      29.4 ± 1.7      -93.4 ± 1.9
    3.8     102.3 ± 1.9 91.0 ± 0.6     59.7 ± 1.2    29.4 ± 1.7   0±0     0.28 ± 0.11  0±0      29.4 ± 1.7      -92.5 ± 2.2
                                                               177


REFERENCES
    178


                                       REFERENCES
[1] N. Castaño, S. Cordts, M.K. Jalil, K. Zhang, S. Koppaka, A. Bick, R. Paul, S.K. Tang,
Fomite transmission, physicochemical origin of virus–surface interactions, and
disinfection strategies for enveloped viruses with applications to SARS-CoV-2, ACS
Omega (2021). https://doi.org/10.1021/acsomega.0c06335
[2] S.A. Boone, C.P. Gerba, Significance of fomites in the spread of respiratory and
enteric viral disease, Appl. Environ. Microbiol. 73(6) (2007) 1687-1696.
https://doi.org/10.1128/aem.02051-06
[3] B. Stephens, P. Azimi, M.S. Thoemmes, M. Heidarinejad, J.G. Allen, J.A. Gilbert,
Microbial exchange via fomites and implications for human health, Curr. Poll. Rep.
(2019) 1-16. https://doi.org/10.1007/s40726-019-00123-6
[4] F. Pancic, D.C. Carpenter, P.E. Came, Role of infectious secretions in the
transmission of rhinovirus. , J. Clin. Microbiol. 12 (1980) 567-571.
https://dx.doi.org/10.1128/JCM.12.4.567-571.1980
[5] A.A. Chughtai, S. Stelzer-Braid, W. Rawlinson, G. Pontivivo, Q. Wang, Y. Pan, D.
Zhang, Y. Zhang, L. Li, C.R. MacIntyre, Contamination by respiratory viruses on outer
surface of medical masks used by hospital healthcare workers, BMC Infect. Dis. 19(1)
(2019). https://doi.org/10.1186/s12879-019-4109-x
[6] M.Z. Hassan, K. Sturm-Ramirez, M.Z. Rahman, K. Hossain, M.A. Aleem, M.U.
Bhuiyan, M.M. Islam, M. Rahman, E.S. Gurley, Contamination of hospital surfaces with
respiratory pathogens in Bangladesh, PLOS One 14(10) (2019) e0224065.
https://doi.org/10.1371/journal.pone.0224065
[7] E.K. Kurgat, J.D. Sexton, F. Garavito, A. Reynolds, R.D. Contreras, C.P. Gerba, R.A.
Leslie, S.L. Edmonds-Wilson, K.A. Reynolds, Impact of a hygiene intervention on virus
spread in an office building, Int. J. Hyg. Environ. Health 222(3) (2019) 479-485.
https://doi.org/10.1016/j.ijheh.2019.01.001
[8] J.M.L. Brotherton, V.C. Delpech, G.L. Gilbert, S. Hatzi, P.D. Paraskevopoulos, J.M.
McAnulty, A large outbreak of influenza A and B on a cruise ship causing widespread
morbidity, Epidemiol. Infect. 130(2) (2003) 263-271.
https://doi.org/10.1017/s0950268802008166
[9] S. Xiao, Y. Li, T.-W. Wong, D.S.C. Hui, Role of fomites in SARS transmission during
the largest hospital outbreak in Hong Kong, PLOS One 12(7) (2017).
https://doi.org/10.1371/journal.pone.0181558
[10] S.F. Dowell, J.M. Simmerman, D.D. Erdman, J.-S.J. Wu, A. Chaovavanich, M.
Javadi, J.-Y. Yang, L.J. Anderson, S. Tong, M.S. Ho, Severe acute respiratory
                                             179


syndrome coronavirus on hospital surfaces, Clin. Infect. Dis. 39(5) (2004) 652-657.
https://doi.org/10.1086/422652
[11] A.J. Prussin II, J.A. Belser, W. Bischoff, S.T. Kelley, K. Lin, W.G. Lindsley, J.P.
Nshimyimana, M. Schuit, Z. Wu, K. Bibby, L.C. Marr, Viruses in the Built Environment
(VIBE) meeting report, Microbiome 8 (2020) 1. https://doi.org/10.1186/s40168-019-
0777-4
[12] H. Sakaguchi, K. Wada, J. Kajioka, M. Watanabe, R. Nakano, T. Hirose, H. Ohta, Y.
Aizawa, Maintenance of influenza virus infectivity on the surfaces of personal protective
equipment and clothing used in healthcare settings, Environmental Health and
Preventive Medicine 15(6) (2010) 344-349. https://doi.org/10.1007/s12199-010-0149-y
[13] F. Cui, X. Geng, O. Zervaki, D.D. Dionysiou, J. Katz, S.-J. Haig, M. Boufadel,
Transport and fate of virus-laden particles in a supermarket: Recommendations for risk
reduction of COVID-19 spreading, J. Environ. Eng. 147(4) (2021) 04021007.
https://doi.org/10.1061/(ASCE)EE.1943-7870.0001870
[14] A.K. Pitol, T.R. Julian, Community transmission of SARS-CoV-2 by fomites: Risks
and risk reduction strategies, Environ. Sci. Technol. Lett. 8 (2021) 263−269.
https://dx.doi.org/10.1021/acs.estlett.0c00966
[15] A. Meiksin, Dynamics of COVID-19 transmission including indirect transmission
mechanisms: A mathematical analysis, Epidemiol. Infect. 148 (2020) e257.
https://doi.org/10.1017/s0950268820002563
[16] L. Owen, K. Laird, The role of textiles as fomites in the healthcare environment: a
review of the infection control risk, PeerJ 8 (2020) e9790.
https://doi.org/10.7717/peerj.9790
[17] A. Brlek, Š. Vidovič, S. Vuzem, K. Turk, Z. Simonović, Possible indirect
transmission of COVID-19 at a squash court, Slovenia, March 2020: Case report,
Epidemiol. Infect. 148 (2020) e120. https://doi.org/10.1017/s0950268820001326
[18] Z. Zhang, L. Zhang, Y. Wang, COVID‐19 indirect contact transmission through the
oral mucosa must not be ignored, J. Oral Pathology Med. 49(5) (2020) 450-451.
https://doi.org/10.1111/jop.13019
[19] T. Yamagishi, M. Ohnishi, N. Matsunaga, K. Kakimoto, H. Kamiya, K. Okamoto, M.
Suzuki, Y. Gu, M. Sakaguchi, T. Tajima, S. Takaya, N. Ohmagari, M. Takeda, S.
Matsuyama, K. Shirato, N. Nao, H. Hasegawa, T. Kageyam, I. Takayama, S. Saito, K.
Wada, R. Fujita, H. Saito, K. Okinaka, M. Griffith, A.E. Parry, B. Barnetson, J. Leonard,
Environmental sampling for severe acute respiratory syndrome coronavirus 2 during a
COVID-19 outbreak on the Diamond Princess cruise ship, J. Infect. Dis. 222(7) (2020)
1098-1102. https://doi.org/10.1093/infdis/jiaa437
[20] S.W.X. Ong, Y.K. Tan, P.Y. Chia, T.H. Lee, O.T. Ng, M.S.Y. Wong, K. Marimuthu,
Air, surface environmental, and personal protective equipment contamination by severe
                                             180


acute respiratory syndrome coronavirus 2 (SARS-CoV-2) from a symptomatic patient,
JAMA 323(16) (2020) 1610-1612. https://doi.org/10.1001/jama.2020.3227
[21] B. Pastorino, F. Touret, M. Gilles, X. de Lamballerie, R. Charrel, Prolonged
infectivity of SARS-CoV-2 in fomites, Emerg. Infect. Dis. 26(9) (2020) 2256-2257.
https://dx.doi.org/10.3201/eid2609.201788
[22] Z.-D. Guo, Z.-Y. Wang, S.-F. Zhang, X. Li, L. Li, C. Li, Y. Cui, R.-B. Fu, Y.-Z. Dong,
X.-Y. Chi, Aerosol and surface distribution of severe acute respiratory syndrome
coronavirus 2 in hospital wards, Wuhan, China, 2020, Emerg. Infect. Dis. 26(7) (2020)
1583-1591. https://doi.org/10.3201/eid2607.200885
[23] P.Y. Chia, K.K. Coleman, Y.K. Tan, S.W.X. Ong, M. Gum, S.K. Lau, X.F. Lim, A.S.
Lim, S. Sutjipto, P.H. Lee, Detection of air and surface contamination by SARS-CoV-2
in hospital rooms of infected patients, Nature Comm. 11(1) (2020) 1-7.
https://doi.org/10.1038/s41467-020-16670-2
[24] N. Van Doremalen, T. Bushmaker, D.H. Morris, M.G. Holbrook, A. Gamble, B.N.
Williamson, A. Tamin, J.L. Harcourt, N.J. Thornburg, S.I. Gerber, Aerosol and surface
stability of SARS-CoV-2 as compared with SARS-CoV-1, New Engl. J. Med. 382(16)
(2020) 1564-1567. https://doi.org/10.1056/NEJMc2004973
[25] Transmission of SARS-CoV-2: implications for infection prevention precautions:
Scientific Brief, 09 July 2020, World Health Organization, 2020.
[26] A.P. Harvey, E.R. Fuhrmeister, M. Cantrell, A.K. Pitol, J.M. Swarthout, J.E. Powers,
M.L. Nadimpalli, T.R. Julian, A.J. Pickering, Longitudinal monitoring of SARS-CoV-2
RNA on high-touch surfaces in a community setting, Environ. Sci. Technol. Lett. 8(2)
(2021) 168-175. https://doi.org/10.1021/acs.estlett.0c00875
[27] F.X. Abad, C. Villena, S. Guix, S. Caballero, R.M. Pintó, A. Bosch, Potential role of
fomites in the vehicular transmission of human astroviruses, Appl. Environ. Microbiol.
67(9) (2001) 3904-3907. http://dx.doi.org/10.1128/AEM.67.9.3904-3907.2001
[28] A. Kramer, I. Schwebke, G. Kampf, How long do nosocomial pathogens persist on
inanimate surfaces? A systematic review, BMC Infect. Dis. 6(1) (2006) 130.
https://doi.org/10.1186/1471-2334-6-130
[29] A. Chabrelie, J. Mitchell, J. Rose, D. Charbonneau, Y. Ishida, Evaluation of the
influenza risk reduction from antimicrobial spray application on porous surfaces, Risk
Anal. 38(7) (2018) 1502-1517. https://doi.org/10.1111/risa.12952
[30] B. Killingley, J. Nguyen‐Van‐Tam, Routes of influenza transmission, Influenza
Other Resp. 7 (2013) 42-51. https://doi.org/10.1111/irv.12080
[31] R.A. Canales, K.A. Reynolds, A.M. Wilson, S.L. Fankem, M.H. Weir, J.B. Rose, S.
Abd-Elmaksoud, C.P. Gerba, Modeling the role of fomites in a norovirus outbreak, J.
                                             181


Occup. Environ. Hyg. 16(1) (2019) 16-26.
https://doi.org/10.1080/15459624.2018.1531131
[32] E.L. Jones, A. Kramer, M. Gaither, C.P. Gerba, Role of fomite contamination during
an outbreak of norovirus on houseboats, Int. J. Environ. Health Res. 17(2) (2007) 123-
131. https://doi.org/10.1080/09603120701219394
[33] M.O. Aydogdu, E. Altun, E. Chung, G. Ren, S. Homer-Vanniasinkam, B. Chen, M.
Edirisinghe, Surface interactions and viability of coronaviruses, J. R. Soc. Interface 18
(2021). https://doi.org/10.1098/rsif.2020.0798
[34] B. Michen, F. Meder, A. Rust, J. Fritsch, C. Aneziris, T. Graule, Virus removal in
ceramic depth filters based on diatomaceous earth, Environ. Sci. Technol. 46(2) (2012)
1170-1177. https://doi.org/10.1021/es2030565
[35] J.-A. Park, S.-B. Kim, DLVO and XDLVO calculations for bacteriophage MS2
adhesion to iron oxide particles, J. Contam. Hydrol. 181 (2015) 131-140.
https://doi.org/10.1016/j.jconhyd.2015.01.005
[36] X. Wang, R. Şengür-Taşdemir, İ. Koyuncu, V.V. Tarabara, Lip balm drying
promotes virus attachment: Characterization of lip balm coatings and XDLVO modeling,
J. Colloid Interface Sci. 581 (2021) 884-894. https://doi.org/10.1016/j.jcis.2020.07.143
[37] I. Samandoulgou, I. Fliss, J. Jean, Adhesion of norovirus to surfaces: Contribution
of thermodynamic and molecular properties using virus-like particles, Food Environ.
Virol. (2021). https://doi.org/10.1007/s12560-021-09471-3
[38] H.T.T. Dang, V.V. Tarabara, Virus deposition onto polyelectrolyte-coated surfaces:
A study with bacteriophage MS2, J. Colloid Interface Sci. 540 (2019) 155-166.
https://doi.org/10.1016/j.jcis.2018.12.107
[39] K. Wong, B. Mukherjee, A.M. Kahler, R. Zepp, M. Molina, Influence of inorganic
ions on aggregation and adsorption behaviors of human adenovirus, Environ. Sci.
Technol. 46(20) (2012) 11145-11153. https://doi.org/10.1021/es3028764
[40] Y.-N. Liu, Z.-T. Lv, S.-Y. Yang, X.-W. Liu, Optical tracking of the interfacial
dynamics of single SARS-CoV‐2 pseudoviruses, Environ. Sci. Technol. 55 (2021)
4115−4122. https://dx.doi.org/10.1021/acs.est.0c06962
[41] A. Armanious, M. Aeppli, R. Jacak, D. Refardt, T. Sigstam, T. Kohn, M. Sander,
Viruses at solid-water interfaces: a systematic assessment of interactions driving
adsorption, Environ. Sci. Technol. 50(2) (2015) 732-743.
https://doi.org/10.1021/acs.est.5b04644
[42] X.L. Zhou, Y. Yang, S. Wang, X.W. Liu, Surface plasmon resonance microscopy:
From single-molecule sensing to single-cell imaging, Angew. Chem., Int. Ed. 59(5)
(2020) 1776−1785. https://doi.org/10.1002/anie.201908806
                                            182


[43] B. Yuan, M. Pham, T.H. Nguyen, Deposition kinetics of bacteriophage MS2 on a
silica surface coated with natural organic matter in a radial stagnation point flow cell,
Environ. Sci. Technol. 42(20) (2008) 7628-7633. https://doi.org/10.1021/es801003s
[44] M. Pham, E.A. Mintz, T.H. Nguyen, Deposition kinetics of bacteriophage MS2 to
natural organic matter: Role of divalent cations, J. Colloid Interface Sci. 338(1) (2009) 1-
9. https://doi.org/10.1016/j.jcis.2009.06.025
[45] H.T.T. Dang, V.V. Tarabara, Attachment of human adenovirus onto household
paints, Colloids Surf. B 204 (2021). https://doi.org/10.1016/j.colsurfb.2021.111812
[46] X. Huang, J. Xu, H.-F. Ji, G. Li, H. Chen, Quartz crystal microbalance based
biosensor for rapid and sensitive detection of maize chlorotic mottle virus, Anal.
Methods 6(13) (2014) 4530-4536. https://doi.org/10.1039/C4AY00292J
[47] C.J. Van Oss, Interfacial forces in aqueous media, CRC Press, 2006.
[48] G.R. Nemerow, P.L. Stewart, V.S. Reddy, Structure of human adenovirus, Curr.
Opin. Virol. 2(2) (2012) 115-121. https://doi.org/10.1016/j.coviro.2011.12.008
[49] C. San Martín, Latest insights on adenovirus structure and assembly, Viruses 4(5)
(2012) 847-877. https://doi.org/10.3390/v4050847
[50] K. Sirikanchana, J.L. Shisler, B.J. Marinas, Effect of exposure to UV-C irradiation
and monochloramine on adenovirus serotype 2 early protein expression and DNA
replication, Appl. Environ. Microbiol. 74(12) (2008) 3774-3782.
https://doi.org/10.1128/aem.02049-07
[51] S. Jing, J. Zhang, M. Cao, M. Liu, Y. Yan, S. Zhao, N. Cao, J. Ou, K. Ma, X. Cai,
Household transmission of human adenovirus type 55 in case of fatal acute respiratory
disease, Emerg. Infect. Dis. 25(9) (2019) 1756. https://doi.org/10.3201/eid2509.181937
[52] G. La Rosa, M. Fratini, S.D. Libera, M. Iaconelli, M. Muscillo, Viral infections
acquired indoors through airborne, droplet or contact transmission, Ann. Ist. Super.
Sanità 49 (2013) 124-132. https://doi.org/10.4415/ANN_13_02_03
[53] L. Liljeroos, M.A. Krzyzaniak, A. Helenius, S.J. Butcher, Architecture of respiratory
syncytial virus revealed by electron cryotomography, PNAS 110(27) (2013) 11133-
11138. https://doi.org/10.1073/pnas.1309070110
[54] A.R. Falsey, E.E. Walsh, Respiratory syncytial virus infection in adults, Clin.
Microbiol. Rev. 13(3) (2000) 371-384. https://doi.org/10.1128/CMR.13.3.371
[55] C.B. Hall, R.G. Douglas Jr, Modes of transmission of respiratory syncytial virus, J.
Pediatr. 99(1) (1981) 100-103. https://doi.org/10.1016/S0022-3476(81)80969-9
                                             183


[56] C.B. Hall, R.G. Douglas Jr, J.M. Geiman, Possible transmission by fomites of
respiratory syncytial virus, J. Infect. Dis. 141(1) (1980) 98-102.
https://doi.org/10.1093/infdis/141.1.98
[57] E.P. Vejerano, L.C. Marr, Physico-chemical characteristics of evaporating
respiratory fluid droplets, J. R. Soc. Interface 15(139) (2018) 20170939.
https://doi.org/10.1098/rsif.2017.0939
[58] B.K. Walsh, D.J. Mackey, T. Pajewski, Y. Yu, B.M. Gaston, J.F. Hunt, Exhaled-
breath condensate pH can be safely and continuously monitored in mechanically
ventilated patients, Respir. Care 51(10) (2006) 1125-1131.
[59] R.M. Effros, K.W. 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 ) 663-9. https://doi.org/10.1164/ajrccm.165.5.2101018
[60] ATCC Product Sheet. Human adenovirus 5 (ATCC® VR1516™).
[61] ATCC Certificate of Analysis. Adenovirus type 5 Reference Material.
[62] ATCC Product Information Sheet for VR-1516™.
[63] ATCC Product Sheet. Human respiratory syncytial virus, High titer (ATCC® VR-
26PQ™).
[64] ATCC Certificate of Analysis. Human respiratory syncytial virus, High titer (Strain:
Long).
[65] Z. Ke, R. Dillard, T. Chirkova, F. Leon, C. Stobart, C. Hampton, J. Strauss, D.
Rajan, C. Rostad, J. Taylor, H. Yi, R. Shah, M. Jin, T. Hartert, R. Peebles, B. Graham,
M. Moore, L. Anderson, E. Wright, The morphology and assembly of respiratory
syncytial virus revealed by cryo-electron tomography, Viruses 10(8) (2018) 446.
https://doi.org/10.3390/v10080446
[66] C. Griffiths, S.J. Drews, D.J. Marchant, Respiratory syncytial virus: infection,
detection, and new options for prevention and treatment, Clin. Microbiol. Rev. 30(1)
(2017) 277-319. https://doi.org/10.1128/cmr.00010-16
[67] G. Kiss, J.M. Holl, G.M. Williams, E. Alonas, D. Vanover, A.W. Lifland, M. Gudheti,
R.C. Guerrero-Ferreira, V. Nair, H. Yi, B.S. Graham, P.J. Santangelo, E.R. Wright,
Structural analysis of respiratory syncytial virus reveals the position of M2-1 between
the matrix protein and the ribonucleoprotein complex, J. Virol. 88(13) (2014) 7602-7617.
https://doi.org/10.1128/jvi.00256-14
[68] C.J. van Oss, M.K. Chaudhury, R.J. Good, Interfacial Lifshitz-van der Waals and
polar interactions in macroscopic systems, Chem. Rev. 88(6) (1988) 927-941.
https://doi.org/10.1021/cr00088a006
                                              184


[69] M.E. Schrader, Young-Dupre revisited, Langmuir 11 (1995) 3585-3589.
https://doi.org/10.1021/la00009a049
[70] H. Shi, V.V. Tarabara, Charge, size distribution and hydrophobicity of viruses:
Effect of propagation and purification methods, J. Virol. Methods 256 (2018) 123-132.
https://doi.org/10.1016/j.jviromet.2018.02.008
[71] E.M.V. Hoek, G.K. Agarwal, Extended DLVO interactions between spherical
particles and rough surfaces, J. Colloid Interface Sci. 298 (2006) 50-58.
https://doi.org/10.1016/j.jcis.2005.12.031
[72] X. Huang, S. Bhattacharjee, E.M.V. Hoek, Is surface roughness a "Scapegoat" or a
primary factor when defining particle - substrate interactions?, Langmuir 26 (2010)
2528-2537. https://doi.org/10.1021/la9028113
[73] Nanoscience Instruments (personal communication).
[74] L.F. Zemljič, O. Plohl, A. Vesel, T. Luxbacher, S. Potrč, Physicochemical
characterization of packaging foils coated by chitosan and polyphenols colloidal
formulations, Int. J. Mol. Sci. 21(2) (2020) 495. https://doi.org/10.3390/ijms21020495
[75] Y. Hedberg, X. Wang, J. Hedberg, M. Lundin, E. Blomberg, I.O. Wallinder, Surface-
protein interactions on different stainless steel grades: effects of protein adsorption,
surface changes and metal release, J. Mater. Sci.: Mater. Med. 24(4) (2013) 1015-1033.
https://doi.org/10.1007/s10856-013-4859-8
[76] X. Zhang, R. Bai, Adsorption behavior of humic acid onto polypyrrole-coated nylon
6, 6 granules, J. Mater. Chem. 12(9) (2002) 2733-2739.
https://doi.org/10.1039/B201364A
[77] L. Graetz, Ueber die Wärmeleitungsfähigkeit von Flüssigkeiten, Annalen der Physik
254 (79) (1882) 79-94. https://doi.org/10.1002/andp.18822540106
[78] W.M. Deen, Analysis of Transport Phenomena, Oxford University Press, 1998.
[79] G. Sauerbrey, Verwendung von Schwingquarzen zur Wägung dünner Schichten
und zur Mikrowägung, Zeitschrift für Physik 155(2) (1959) 206-222.
https://doi.org/10.1007/BF01337937
[80] J.-L. Putaux, A. Buléon, R. Borsali, H. Chanzy, Ultrastructural aspects of
phytoglycogen from cryo-transmission electron microscopy and quasi-elastic light
scattering data, Int. J. Biol. Macromol. 26(2) (1999) 145-150.
https://doi.org/10.1016/S0141-8130(99)00076-8
[81] H. Shi, I. Xagoraraki, K.N. Parent, M.L. Bruening, V.V. Tarabara, Elution is a critical
step for recovering human adenovirus 40 from tap water and surface water by cross-
flow ultrafiltration, Appl. Environ. Microbiol. 82(16) (2016) 4982.
https://doi.org/10.1128/AEM.00870-16
                                              185


[82] H. Ohshima, A simple expression for Henry's function for the retardation effect in
electrophoresis of spherical colloidal particles, J. Colloid Interface Sci. 168(1) (1994)
269-271. https://doi.org/10.1006/jcis.1994.1419
[83] H. Fischer, J.H. Widdicombe, Mechanisms of acid and base secretion by the airway
epithelium, J. Membr. Biol. 211(3) (2006) 139-150. https://doi.org/10.1007/s00232-006-
0861-0
[84] N. Albinana-Gimenez, P. Clemente-Casares, B. Calgua, J.M. Huguet, S. Courtois,
R. Girones, Comparison of methods for concentrating human adenoviruses,
polyomavirus JC and noroviruses in source waters and drinking water using quantitative
PCR, J. Virol. Methods 158(1) (2009) 104-109.
http://dx.doi.org/10.1016/j.jviromet.2009.02.004
[85] C. Gibbons, R. Rodriguez, L. Tallon, M. Sobsey, Evaluation of positively charged
alumina nanofibre cartridge filters for the primary concentration of noroviruses,
adenoviruses and male-specific coliphages from seawater, J. Appl. Microbiol. 109(2)
(2010) 635-641. https://doi.org/10.1111/j.1365-2672.2010.04691.x
[86] H. Shi, E.V. Pasco, V.V. Tarabara, Membrane-based methods of virus
concentration from water: a review of process parameters and their effects on virus
recovery, Environ. Sci. Water Res. Technol. 3 (2017) 778-792.
https://doi.org/10.1039/C7EW00016B
[87] G. Hwang, C.-H. Lee, I.-S. Ahn, B.J. Mhin, Determination of reliable Lewis acid-
base surface tension components of a solid in LW-AB approach, J. Ind. Eng. Chem. 17
(2011) 125-129. https://doi.org/10.1016/j.jiec.2010.12.009
[88] E. McCafferty, Acid-base effects in polymer adhesion at metal surfaces, J. Adhes.
Sci. Technol. 16(3) (2002) 239-255. https://doi.org/10.1163/156856102317295478
[89] L.-K. Chau, M.D. Porter, Surface isoelectric point of evaporated silver films:
Determination by contact angle titration, J. Colloid Interface Sci. 145(1) (1991) 283−286.
https://doi.org/10.1016/0021-9797(91)90121-N
[90] M.F. Cuddy, A.R. Poda, L.N. Brantley, Determination of isoelectric points and the
role of pH for common quartz crystal microbalance sensors, ACS Appl. Mater. Interf. 5
(2013) 3514−3518. https://doi.org/10.1021/am400909g
[91] E. Virga, E. Spruijt, W.M. de Vos, P.M. Biesheuvel, Wettability of amphoteric
surfaces: The effect of pH and ionic strength on surface ionization and wetting,
Langmuir 34 (2018) 15174−15180. https://doi.org/10.1021/acs.langmuir.8b02875
[92] P.K. Weissenborn, R.J. Pugh, Surface tension of aqueous solutions of electrolytes:
Relationship with ion hydration, oxygen solubility, and bubble coalescence, J. Colloid
Interface Sci. 184 (1996) 550-563. https://doi.org/10.1006/jcis.1996.0651
[93] A.J. Kinloch, Adhesion and Adhesives: Science and Technology, Springer, 1987.
                                             186


[94] C.J. van Oss, M.K. Chaudhury, R.J. Good, Monopolar surfaces, Adv. Colloid
Interface Sci. 28 (1987) 35-64. https://doi-org.proxy1.cl.msu.edu/10.1016/0001-
8686(87)80008-8
[95] I. Reviakine, D. Johannsmann, R.P. Richter, Hearing what you cannot see and
visualizing what you hear: Interpreting quartz crystal microbalance data from solvated
interfaces, Anal. Chem. 83(23) (2011) 8838-8848. https://doi.org/10.1021/ac201778h
[96] M. Rodal, How are QCM results influenced by the lateral distribution of the
measured layer? Surface Science Blog, Biolin Scientific, 2019.
https://www.biolinscientific.com/blog/how-are-qcm-results-influenced-by-the-lateral-
distribution-of-the-measured-layer. (Accessed April 9 2021).
[97] K.K. Kanazawa, J.G. Gordon, Frequency of a quartz microbalance in contact with
liquid, Anal. Chem. 57(8) (1985) 1770-1771. https://doi.org/10.1021/ac00285a062
[98] K.K. Kanazawa, J.G. Gordon, The oscillation frequency of a quartz resonator in
contact with liquid, Anal. Chim. Acta 175 (1985) 99-105. https://doi.org/10.1016/S0003-
2670(00)82721-X
[99] L. Boulangé-Petermann, A. Doren, B. Baroux, M.-N. Bellon-Fontaine, Zeta potential
measurements on passive metals, J. Colloid Interface Sci. 171(1) (1995) 179-186.
https://doi.org/10.1006/jcis.1995.1165
[100] S. Wu, S. Altenried, A. Zogg, F. Zuber, K. Maniura-Weber, Q. Ren, Role of the
surface nanoscale roughness of stainless steel on bacterial adhesion and microcolony
formation, ACS Omega 3(6) (2018) 6456-6464.
https://doi.org/10.1021/acsomega.8b00769
[101] G. Okamoto, Passive film of 18-8 stainless steel structure and its function,
Corrosion Sci. 13 (1973) 471-489. https://doi.org/10.1016/0010-938X(73)90031-0
[102] N. Sato, M. Cohen, The kinetics of anodic oxidation of iron in neutral solution: II
The initial stages, J. Electrochem. Soc. 111 (1964) 512-522.
https://doi.org/10.1149/1.2426171
[103] L. Boulange-Petermann, C. Jullien, P.E. Dubois , T. Benezech, C. Faille, Influence
of surface chemistry on the hygienic status of industrial stainless steel, Biofouling 20(1)
(2004) 25-33. https://doi.org/10.1080/08927010410001655542
[104] G.C. Laine, Characterization of and controlling morphology of ultra-thin
nanocomposites, University of Kentucky, 2013.
[105] L. Moeller, L. Schuenadel, A. Nitsche, I. Schwebke, M. Hanisch, M. Laue,
Evaluation of virus inactivation by formaldehyde to enhance biosafety of diagnostic
electron microscopy, Viruses 7(2) (2015) 666-679. https://doi.org/10.3390/v7020666
                                            187


[106] N.I. Rabiah, Y. Sato, A. Kannan, W. Kress, F. Straube, G.G. Fuller, Understanding
the adsorption and potential tear film stability properties of recombinant human lubricin
and bovine submaxillary mucins in an in vitro tear film model, Colloid Surface B 195
(2020) 111257. https://doi.org/10.1016/j.colsurfb.2020.111257
[107] H. Seo, M. Yoo, S. Jeon, Influence of nanobubbles on the adsorption of
nanoparticles, Langmuir 23(4) (2007) 1623-1625. https://doi.org/10.1021/la062763r
[108] X.H. Zhang, Quartz crystal microbalance study of the interfacial nanobubbles,
Phys. Chem. Chem. Phys. 10(45) (2008) 6842-6848. https://doi.org/10.1039/b810587a
[109] K.K. Kanazawa, J.G. Gordon, The oscillation frequency of a quartz resonator in
contact with liquid, Anal. Chim. Acta 175 (1985) 99-105. https://doi.org/10.1016/S0003-
2670(00)82721-X
[110] K.K. Kanazawa, J.G. Gordon, Frequency of a quartz microbalance in contact with
liquid, Anal. Chem. 57(8) (1985) 1770-1771. https://doi.org/10.1021/ac00285a062
[111] I. Reviakine, D. Johannsmann, R.P. Richter, Hearing what you cannot see and
visualizing what you hear: Interpreting quartz crystal microbalance data from solvated
interfaces, Anal. Chem. 83(23) (2011) 8838-8848. https://doi.org/10.1021/ac201778h
[112] H. Ozbek, Viscosity of aqueous sodium chloride solutions from 0 to 150 C,
American Chemical Society 29th Southeast Regional Meeting, Tampa, FL, 1977.
[113] G.M. Iliescu, C. Vasile, Caracteristici termofizice ale produselor alimentare,
Editura Tehnică, 1982.
[114] H. Ohshima, A simple expression for Henry's function for the retardation effect in
electrophoresis of spherical colloidal particles, J. Colloid Interface Sci. 168(1) (1994)
269-271. https://doi.org/10.1006/jcis.1994.1419
[115] H. Shi, I. Xagoraraki, K.N. Parent, M.L. Bruening, V.V. Tarabara, Elution is a
critical step for recovering human adenovirus 40 from tap water and surface water by
cross-flow ultrafiltration, Appl. Environ. Microbiol. 82(16) (2016) 4982.
https://doi.org/10.1128/AEM.00870-16
                                              188


                                    CHAPTER FOUR
  PVDF membrane ageing due to chemical cleaning: Understanding the evolution
                            of membrane-foulant interactions
4.1 Abstract
This study investigates the surface chemistry evolution of polyvinylidene fluoride (PVDF)
under the effect of intermittent Humic acid (HA) fouling and sodium hypochlorite (NaClO)
cleaning. The surface physicochemical evolutions of aged membranes are
characterized based on surface energy calculation. The membrane ageing tests show
that NaClO cleaning have opposite effect on pristine PVDF membrane and pristine
PVDF/PVP membrane. Specifically, NaClO cleaning enhances the hydrophilicity of
PVDF membrane but hydrophobicity of PVDF/PVP membrane. Static fouling-cleaning
tests show that the surface energy of the membrane floats between a lower value after
each cleaning stage and a higher value after each fouling stage. More experiments will
be performed, including FTIR and XPS analysis, to help interpret the surface energy
change from molecular level.
                                           189


4.2 Introduction
Membrane processes are being increasingly used in drinking water treatment due to
their effectiveness to remove many contaminants in raw water as well as their low cost.
Polyvinylidene fluoride (PVDF) is widely used as the base polymer for many
membranes due to its high hydrophobicity, chemical resistance, and excellent
mechanical strength. However, pure PVDF membranes are highly hydrophobic, making
them more susceptible to fouling in water treatment [1]. In this sense, hydrophilic
additives which have hydroxyl, amine, or carboxylic acid groups are often blended into
PVDF, which enhance the filtration performance of these membranes [2, 3]. Yet these
polymeric additives may leach out from the membrane due to their relatively small sizes
and their affinity to the water.
Indeed, membrane fouling is one of the critical challenges in the successful application
of membrane processes. Fouling arises from interactions between a membrane and
various components present in the raw water [4]. Natural organic matter (NOM), such
as humic acid, is ubiquitous in surface water – a common source of raw water for
drinking water treatment plants. NOM is one of the major membrane foulants as
documented in research literature and reports from professional practice community [4-
7]. Humic acid fouling is a complex phenomenon which involves both reversible and
irreversible processes [4].
                                           190


Chemical cleaning is necessary to remove irreversible pollution. There are several
categories of cleaning chemicals including surfactants, oxidants, acids, and chelating
agents. Among them, sodium hypochlorite (NaClO) is one of the most commonly used
cleaning reagents because of its low price and effectiveness against organic foulants [8].
However, chemical agents can cause damage to the membrane [9]. Numerous studies
have shown that changes in the characteristics of aged membranes were due to the
oxidation or decomposition of the hydrophilic additives in blended PVDF membranes [1].
The evolution of hydrophilicity of aged PVDF membranes was extensively studied [8, 10]
However, there are no unified conclusions obtained on the evolution of the intrinsic
characteristics of PVDF membranes when cleaned with NaClO solutions [10].
In this study, humic acid (Sigma-Aldrich) was chosen as the model organic foulant. This
humic acid has been thoroughly characterized and used in numerous membrane fouling
studies [4, 11, 12]. Humic acid (HA) is representative of natural organic matter, a
common component in surface water where it is present in both dissolved and
suspended fractions. HA is a complex and heterogeneous combination of many
macromolecules. The fouling potential of HA during membrane filtration of surface
waters varies greatly depending on the solution chemistry and membrane properties
[13].
The study explores the hypothesis that adlayer of residual foulants offers increasing
protection of the membrane polymer against bleach as it develops on the membrane
surface.
                                            191


4.3 Experimental
4.3.1 Reagents and materials
The humic acid sodium salt (HA), polyvinylpyrrolidone (PVP) powder, NaClO (10%),
Dimethylformamide (DMFA), glycerol, ethylene glycol (EG) and diiodomethane (DID)
were purchased from Sigma Aldrich. Polyvinylidene fluoride (PVDF) powder (Solef
1015, Solvay) and DMFA was used in PVDF membrane preparation (See
Supplementary Materials (SM), section S1). QCM-D sensors coated with PVDF were
purchased from Nanoscience.
4.3.2 Chemical cleaning study of membranes
The cleaning agent used in the study is 1% NaClO (pH 12). Pristine membranes were
kept in NaClO solution or DI water for 5, 10, 20, 30, 50, 60, 80, 100 hours. The contact
angles of four probe liquids (water, glycerol, DID, EG) on the PDVF membrane were
measured at the end of each cleaning period. The surface energy and surface tension
components were calculated and plotted as a function of CT values.
                                            192


4.3.3 Static fouling-cleaning test of membranes
Prior to static fouling-cleaning tests, 100 mg/ml HA solution was prepared and
maintained at a pH of 7. The ionic strength (IS) of HA was kept at 15 mM. The following
two steps were continuously performed for several cycles of fouling (HA for 24 hours)
and cleaning (1% NaClO for 5 hours). The solution was strred during the whole process
to keep its homogeneity.
4.3.4 Quantifying surface energy of PVDF membrane
The hydrophobicity of membrane was quantified
(1 + 𝑐𝑜𝑠𝜃)𝛾𝑙𝑇𝑂𝑇 = 2 (√𝛾𝑠𝐿𝑊 𝛾𝑙𝐿𝑊 + √𝛾𝑠+ 𝛾𝑙− + √𝛾𝑠− 𝛾𝑙+ ),                              (1)
where 𝛾𝑙𝑇𝑂𝑇 is the total surface energy of the probe liquid:
𝛾 𝑡𝑜𝑡 = 𝛾 𝐿𝑊 + 𝛾 𝐴𝐵 = 𝛾 𝐿𝑊 + 2√𝛾 − 𝛾 + .                                              (2)
The free energy of solid-solid interfacial interaction in water is given by
                               2
∆𝐺𝑠𝑤𝑠 =   −2 (√𝛾𝑠𝐿𝑊   −    𝐿𝑊
                        √𝛾𝑤 )    − 4 (√𝛾𝑠+ 𝛾𝑠− + √𝛾𝑤+ 𝛾𝑤− − √𝛾𝑠+ 𝛾𝑤− − √𝛾𝑠− 𝛾𝑤+ )     (3)
 where 𝛾𝑤𝐿𝑊 , 𝛾𝑤+ and 𝛾𝑤− are surface energy components of water. A positive value of
∆𝐺𝑠𝑤𝑠 indicates a hydrophilic surface, while negative ∆𝐺𝑠𝑤𝑠 corresponds to a
hydrophobic surface. The absolute value of ∆𝐺𝑠𝑤𝑠 indicates the degree of hydrophilicity
(or hydrophobicity, when ∆𝐺𝑠𝑤𝑠 < 0) of the surface.
                                               193


Contact angles of four probe liquids (DI water, glycerol, DID, and EG) were measured
using the sessile drop method (goniometer/tensiometer model 250, ramé-hart). The
droplet volume was 6 μl. All measurements were performed in the air at the room
temperature of 22 °C and the relative humidity of 46 %. Contact angle values were
calculated by DROPPimage Advanced software based on droplet shapes recorded by
camera.
4.3.5 Quantification of HA-membrane interactions
The free energy of virus-PVDF interfacial interaction in water ∆𝐺𝑣𝑤𝑚 , is calculated as
follows:
                                                                                        (8)
∆𝐺𝑑𝐿𝑊0
       = 2 (√𝛾𝑣𝐿𝑊 − √𝛾𝑤𝐿𝑊 ) (√𝛾𝑤𝐿𝑊 − √𝛾𝑚𝐿𝑊
                                           )
∆𝐺𝑑𝐴𝐵              − + 𝛾 − − 𝛾 − ) + 2 𝛾 − (√𝛾 + + √𝛾 + − √𝛾 + )
       = 2 √𝛾𝑤+ (√𝛾𝑚     √ 𝑣 √ 𝑤       √ 𝑤                                              (9)
     0                                          𝑚      𝑣       𝑤
                                      + −        −𝛾 +
                                −2√𝛾𝑚  𝛾𝑣 − 2√𝛾𝑚   𝑣
∆𝐺𝑣𝑤𝑚 = ∆𝐺𝑑𝐿𝑊 0
                 + ∆𝐺𝑑𝐴𝐵
                       0
                                                                                      (10)
                                           194


4.4 Results
4.4.1 Impact of NaClO cleaning on membrane characteristics
Figure 58. Free energy of interfacial interaction in water (∆𝑮𝒔𝒘𝒔 ) of membrane as a
function of cleaning time (cleaning agent: 1% NaClO).
From Fig. 58, one can observe that DI water has no influence on the surface energy of
PVDF and PVDF/PVP membrane. Fig. 58 also shows that pristine PVDF membrane is
hydrophobic (∆𝐺𝑠𝑤𝑠 = -29.3 mJ/m2), and PVP causes an increase in hydrophilicity of as-
                                            195


cast membrane (∆𝐺𝑠𝑤𝑠 = 2.5 mJ/m2 for PVDF/PVP membrane). Fig. 58 shows that
NaClO cleaning increases the hydrophilicity of pristine PVDF membrane. This is
consistent with one previous study, in which PVDF membrane is also shown to become
more hydrophilic with NaClO cleaning [14]. One possible explanation of this
hydrophilicity increase is the formation of carbonyl groups in PVDF polymeric structure.
For PVDF/PVP membrane, Fig. 58 shows that extended ageing in NaClO makes the
PVDF/PVP membrane more hydrophobic, presumably because the increase in
hydrophobicity caused by PVP leaching overcame the increase in hydrophilicity due to
the formation of carbonyl group. This result is consistent with the study conducted by
Ren [10].
                                             196


4.4.2 Static fouling and cleaning cycles
Figure 59. Free energy of interfacial interactions (∆𝑮𝒎𝒘𝒎 ) of PVDF (red) and
PVDF/PVP (blue) membrane as a function of fouling and cleaning cycles. The solid and
empty symbols refer to replicated tests.
Fig. 59 shows that surface energy of the membrane floats between a lower value after
each cleaning stage and a higher value after each fouling stage. The asymptotic value
corresponds to the chemically irreversible layer of foulants that is conditioned by
consecutive exposures to foulants and is comprised by the adsorbed foulant fraction
that is hard to oxidize further. The hydrophilicity of PVDF and PVDF/PVP membrane are
enhanced after each fouling may be attributed to the hydrophilic nature of humic acid.
PVDF/PVP membrane became more hydrophobic after three cycles of fouling-cleaning
compared to the pristine membrane, as presented in Fig. 59. This hydrophobicity
                                            197


increase may be due to the leaching of PVP from PVDF/PVP membrane. After the third
cycles of fouling-cleaning, PVDF/PVP membrane is still less hydrophobic than pristine
PVDF membrane, suggesting the existence of residual PVP in the membrane. However,
for the rest of the fouling-cleaning cycles, PVDF and PVDF/PVP membrane have similar
trends of surface energy change, as shown in Fig. 59, which may indicate that after
three cycles most PVP was leached out from PVDF/PVP membrane so PVDF structure
change prevails over PVP leaching as a factor responsible for the overall change in
membrane surface energy.
4.5 Discussion
In this study, reagent grade NaClO (10%, Sigma Aldrich) is used as the cleaning agent.
Some wastewater treatment facilities such as Traverse City Regional Wastewater
Treatment Plant, citric acid and NaClO are used for maintenance cleaning. Therefore,
more research is needed to study the effect of citric acid and NaClO cleaning on the
surface chemistry of PVDF membrane. Also, the fouling solution used in this study is
humic acid sodium salt (Sigma Aldrich), which is a standard humic acid. In general, the
raw water, which treated by municipal water treatment plant, usually contains various
components apart from humic acid. These components will have potential influence on
the kinetics of membrane fouling and cleaning. Future study is needed to study the
surface chemistry evolution of PVDF when raw water is involved.
                                             198


Furthermore, both PVDF and PVDF/PVP membranes used in this study were made in
the lab. In WWTPs, however, commercial membranes are more commonly used in the
filtration systems. Commercial membranes are usually modified with coatings or
blended with additives to enhance the membrane performance and mechanical strength.
Typically, the chemical components of these coatings or additives are not known since
they belong to the intellectual properties of manufacturers.
Lastly, the concentration of NaClO selected in this study (1%) is much higher than that
in Traverse city WWTPs (0.036%). The purpose of choosing this concentration (C) is to
accelerate the ageing effect, so that reasonable cleaning time (T) can be selected to
reach the same CT value as Traverse city WWTPs.
4.6 Conclusions
From the data collected so far, one can tentatively conclude that a firmly attached humic
acid layer may be formed on membrane surface after a few cycles of fouling and
cleaning. This humic acid layer can protect the membrane from being further aged by
NaClO. This study is still in progress. More experiment including FTIR, XPS analysis will
be performed to investigate the change of functional groups on the surface of
membrane. FTIR and XPS analysis will help understand the evolution of this humic acid
layer from molecular level. This study can help understand the adsorption kinetics of
humic acid on PVDF membrane as well as the NaClO cleaning effect on PVDF
membrane fouled with humic acid.
                                             199


APPENDIX
  200


                                         APPENDIX
S1. PVDF and PVDF/PVP membrane preparation
The PVDF and PVDF/PVP membranes were prepared using phase inversion method
by immersion precipitation. The preparation procedure is as follows:
Table 15. The amount of each ingredient to make 50 g 15% solution of PVDF and
PVDF/PVP in DMFA.
                                    PVDF               DMFA               PVP
    PVDF membrane                   7.5 g              45 ml                0
 PVDF/PVP membrane                  7.5 g             43.4 ml             1.5 g
    1. Preparation of PVDF and PVDF/PVP solution: certain amount of PVDF, PVP and
       DMFA (density 𝜌 = 0.9445 𝑔/𝑐𝑚3 ) were added in a beaker. The polymer solution
       was stirred and heated until complete dissolution. The solution must not boil,
       otherwise it will take additional time to remove bubbles.
    2. Degassing of the polymer solution: The polymer solution was degassed using a
       vacuum desiccator (Wheaton) and a vacuum pump (model UN726 FTP, KNF
       NEUBERGER).
    3. Casting of membranes: The membrane was casted using an automatic film
       applicator (Elcometer 4340) with a molding knife (Elcometer 3570) on a glass
                                             201


      plate. Homogenous polymer solution was poured into the molding knife. A
      uniform layer of polymer on the surface of the glass plate was made.
   4. Immersion precipitation: Immersing the glass with a layer of polymer solution into
      precipitating bath (stage of coagulation) for 5 min at 20 ℃. In our study, DI water
      was used as the coagulant.
The membranes were stored in the DI water at room temperature prior to use.
                                          202


REFERENCES
    203


                                          REFERENCES
[1] S.Z. Abdullah, P.R. Bérubé, Assessing the effects of sodium hypochlorite exposure
on the characteristics of PVDF based membranes, Water research 47(14) (2013) 5392-
5399.
[2] D. Musale, S. Kulkarni, Fouling reduction in poly (acrylonitrile-co-acrylamide)
ultrafiltration membranes, Journal of membrane science 111(1) (1996) 49-56.
[3] N. Hilal, O.O. Ogunbiyi, N.J. Miles, R. Nigmatullin, Methods employed for control of
fouling in MF and UF membranes: a comprehensive review, Separation Science and
Technology 40(10) (2005) 1957-2005.
[4] W. Yuan, A.L. Zydney, Humic acid fouling during ultrafiltration, Environmental
Science & Technology 34(23) (2000) 5043-5050.
[5] I.H. Suffet, M. Patrick, Aquatic humic substances: influence on fate and treatment of
pollutants, 1988.
[6] V. Lahoussine-Turcaud, M.R. Wiesner, J.-Y. Bottero, Fouling in tangential-flow
ultrafiltration: the effect of colloid size and coagulation pretreatment, Journal of
Membrane Science 52(2) (1990) 173-190.
[7] Y. Kaiya, Y. Itoh, K. Fujita, S. Takizawa, Study on fouling materials in the membrane
treatment process for potable water, Desalination 106(1-3) (1996) 71-77.
[8] Q. Wang, H. Zeng, Z. Wu, J. Cao, Impact of sodium hypochlorite cleaning on the
surface properties and performance of PVDF membranes, Applied Surface Science 428
(2018) 289-295.
[9] X. Gan, T. Lin, F. Jiang, X. Zhang, Impacts on characteristics and effluent safety of
PVDF ultrafiltration membranes aged by different chemical cleaning types, Journal of
Membrane Science 640 (2021) 119770.
[10] L. Ren, S. Yu, H. Yang, L. Li, L. Cai, Q. Xia, Z. Shi, G. Liu, Chemical cleaning
reagent of sodium hypochlorite eroding polyvinylidene fluoride ultrafiltration membranes:
Aging pathway, performance decay and molecular mechanism, Journal of Membrane
Science 625 (2021) 119141.
[11] A.I. Schäfer, L.D. Nghiem, N. Oschmann, Bisphenol A retention in the direct
ultrafiltration of greywater, Journal of Membrane Science 283(1-2) (2006) 233-243.
[12] C.Y. Tang, Y.-N. Kwon, J.O. Leckie, Fouling of reverse osmosis and nanofiltration
membranes by humic acid—effects of solution composition and hydrodynamic
conditions, Journal of Membrane Science 290(1-2) (2007) 86-94.
                                                204


[13] L.D. Nghiem, D. Vogel, S. Khan, Characterising humic acid fouling of nanofiltration
membranes using bisphenol A as a molecular indicator, Water research 42(15) (2008)
4049-4058.
[14] Y. Zhang, J. Wang, F. Gao, Y. Chen, H. Zhang, A comparison study: The different
impacts of sodium hypochlorite on PVDF and PSF ultrafiltration (UF) membranes,
Water research 109 (2017) 227-236.
                                         205


                                       CHAPTER FIVE
                                Conclusions and future work
Conclusions on individual projects are provided at the end of corresponding Chapters.
This Chapter lists several overarching conclusions drawn from the entirety of the work.
Overall, this dissertation considers the interactions of various viruses to various fomites
(personal care products, silica, nylon, stainless steel, polypropylene) as well as the
interactions of natural organic matter foulants to PVDF membrane. The methodology
proposed in this study helps fill the knowledge gap on virus adhesion to fomites – an
important component of models on virus transport and fate in the built environment. The
proposed approach can help guide screening and selection of materials that discourage
virus adhesion, design of anti-adhesive surfaces, as well as development of surface
cleaning solutions and protocols.
Virus adhesion can be affected by various variables including the characteristics of virus
(particle size, zeta potential, hydrophobicity, etc.), characteristics of fomites (surface
charge, hydrophobicity, surface roughness, etc.), and environmental conditions (ionic
strength, pH, temperature, etc.). This dissertation considers three types of interactions
between virus and fomites: van der Waals interaction, electrostatic interaction, and acid-
base (or hydrophobic) interaction. However, other forces (apart from the three
mentioned above) can play a role. For instance, steric repulsions and specific
interactions with divalent cations were shown as important for controlling rotavirus
                                             206


deposition [1]. Steric effects caused by irregular topographies of the virus (e.g., protein
loops that protrude from the capsid into solution) and the deposition surface may affect
adhesion [2-4].
This dissertation considers virus adhesion to surfaces with static properties. Yet,
surface characteristics can change over time due to environmental influences or
intrinsic evolution. For example, in water treatment plants, the surface chemistry of
membrane filters evolve due to fouling and cleaning. As another example, fomite
surfaces undergo changes through repeated use: for instance, doorknob surface can
change its roughness and attain a coat of oxides and/or foulants over an extended
period of usage, which can affect their interaction with pathogens of interest. More
research is needed to understand such changes and they affect virus adhesion to
surfaces.
One important direct for future research is on the design of surfaces with tailored
properties to control adhesion. Due to their size and unique functionalities,
nanomaterials are of particular interest in this regard. As a part of my future work, I
intend to use Janus particle to develop novel coatings to prevent bacterial and protein
fouling as well as the interactions of lipids and proteins in biomembranes. Janus
particles are particles containing at least two different chemical compounds, and this
chemical asymmetry gives Janus particles two or more different properties [5].
Biofouling is of great concern in numerous applications ranging from biosensors to
biomedical implants and devices, and from food packaging to industrial and marine
                                              207


equipment [6]. Janus nanoparticles have strong adsorption force, orientation, and
surface activity on the interface, which may reduce the interfacial tension of foulants and
materials [7].
Another research area that is of great interest for the design of surfaces with controlled
adhesion properties is to use machine learning to predict material characteristics based
on their molecular structure. Design of new membrane materials has been guided
mostly by experience and intuition. Synthesis of new polymeric materials and testing of
permeability is time-consuming and expensive. Machine learning (ML) models can be
selected and trained on the experimental database to correlate polymer chemical
structure and their surface properties. The information obtained from the modeling can
guide the design of new polymer compositions with target characteristics (resistance to
virus adhesion, antifouling, etc.).
                                            208


REFERENCES
    209


                                       REFERENCES
[1] L. Gutierrez, S.E. Mylon, B. Nash, T.H. Nguyen, Deposition and aggregation kinetics
of rotavirus in divalent cation solutions, Environmental science & technology 44(12)
(2010) 4552-4557.
[2] B. Yuan, M. Pham, T.H. Nguyen, Deposition kinetics of bacteriophage MS2 on a
silica surface coated with natural organic matter in a radial stagnation point flow cell,
Environmental science & technology 42(20) (2008) 7628-7633.
[3] S.L. Penrod, T.M. Olson, S.B. Grant, Deposition kinetics of two viruses in packed
beds of quartz granular media, Langmuir 12(23) (1996) 5576-5587.
[4] J.A. Redman, S.B. Grant, T.M. Olson, M.E. Hardy, M.K. Estes, Filtration of
recombinant Norwalk virus particles and bacteriophage MS2 in quartz sand: importance
of electrostatic interactions, Environmental science & technology 31(12) (1997) 3378-
3383.
[5] X. Li, L. Chen, D. Cui, W. Jiang, L. Han, N. Niu, Preparation and application of Janus
nanoparticles: Recent development and prospects, Coordination Chemistry Reviews
454 (2022) 214318.
[6] I. Banerjee, R.C. Pangule, R.S. Kane, Antifouling Coatings: Recent Developments in
the Design of Surfaces That Prevent Fouling by Proteins, Bacteria, and Marine
Organisms, Advanced Materials 23(6) (2011) 690-718.
https://doi.org/https://doi.org/10.1002/adma.201001215.
[7] L. Yang, J. Xu, J. Wang, F. Lang, B. Liu, Z. Yang, Responsive Single-Chain/Colloid
Composite Janus Nanoparticle, Macromolecules 53(6) (2020) 2264-2270.
https://doi.org/10.1021/acs.macromol.0c00109.
                                             210