Surface hydrophobization of pulp fibers in paper sheets via gas phase reactions

.


Introduction
Hydrophobization is an important route to adjust the properties and the processability of cellulosic materials [1][2][3][4][5][6][7].In such processes, the cellulosic materials are subjected either to pure reagents (i.e. in liquid, plasma or vapor state), and solutions of hydrophobic precursors [8][9][10][11][12].The precursors usually consist of at least one reactive moiety and an alkyl/aryl substituent.The reactive groups on the precursors decompose to form either a physical coating or a chemical bond during the coating process, rendering the cellulose materials hydrophobic [11,[13][14][15][16]. Gas phase reactions have several advantages over solution-based reactions as they typically require low amounts of reagents and lead to homogenous coatings [17].The most common precursors to modify cellulose via the gas phase are acid chlorides and anhydrides [8,12,14,[18][19][20][21], which form ester bonds on the surface of cellulosic materials.The reaction kinetics and the resulting degree of substitution strongly depend on the reaction conditions as well as the properties of the cellulose materials (i.e.accessibility, crystallinity and presence of impurities) [1].However, acid chlorides release HCl during ester bond formation which partially depolymerizes the cellulosic materials, thereby potentially affecting e.g.mechanical and structural integrity [1].Anhydrides (either pure or mixed ones) also create the corresponding acid as a byproduct which may deteriorate the performance of the cellulosic materials as well.In other areas than cellulose, the most employed gas phase modifiers are silanes in the form of alkoxy, hydroxy or aminosilanes [22].These reactive groups can covalently attach to OH rich surfaces such as silicon wafers, glass and cellulose [23][24][25][26].One of the most employed precursors deposited via the gas phase on silicon surfaces in an industrial context is hexamethyldisilazane, HMDS [27].The two trimethylsilyl groups in HMDS can be both attached to the substrate.The main byproduct, NH 3 , catalyzes the silylation reaction, without potentially harming the cellulose structure [28].A potential disadvantage of the silyl groups on the surface is their stability, particularly at low pH values where desilylation readily occurs [29][30][31][32][33].This instability is maybe one reason why the use of HMDS for the modification of cellulose substrates via the gas phase has not been extensively explored so far [17] since coating applications typically require chemical robustness.
We are interested in the interaction of liquids with cellulosic substrates such as papers and cellulose nanomaterials [29,31,32].When a liquid comes into contact with e.g. an untreated paper, it is strongly wetting the fiber surfaces and quickly absorbs.Liquidcellulose interactions are important in many other applications as well.Tissues are designed to take up a certain amount of water, while in oil absorbing materials the surface free energy is adjusted to efficiently absorb oils from water/oil mixtures [34].Hydrophobic coatings on cellulose fibers result in water-repellent behavior of textiles, which ideally also feature water vapor transmission [35].Two main factors are influencing the penetration behavior of liquids into porous cellulosic materials, namely the chemistry of the cellulose surface as well as the pore structure.The standard process for hydrophobization of paper is sizing where hydrophobic sizing agents are applied either to the surface of the paper (surface sizing) [36] or attached directly to the fibers during the papermaking process [37,38] (internal sizing).Common internal sizing agents are ASA (Alkenyl Succinic Anhydride), AKD (Alkyl ketene dimer) or rosin acids, all creating a bulk hydrophobization of the paper.All internal sizing processes are carried out in the water phase.As a consequence, the wetting and re-drying of the paper leads to severe changes in the pore structure [39] as well as the mechanical properties of the paper when these hydrophobization processes are applied to existing papers [40].The design of paper based substrates, which allows to tune hydrophobicity without compromising the pore structure and mechanical properties would therefore significantly contribute to designing porous sheets with specific liquid-surface interactions, e.g. for printable electronics [41], microfluidic devices [42], and lab-on-a-chip applications [43].The interaction of inkjet droplets with hydrophobically tuned porous substrates [44] is particularly relevant since all of these applications are often realized with inkjet printing, at least on the pilot scale [45].
Here, we address this challenge by screening three compounds for their potential to coat unsized raw paper via the gas phase.We explore how these reagents affect paper properties, i.e. wettability, mechanical strength and printability.We chose three models for this study, namely, an acid chloride (palmitoyl chloride, PCl), a mixed anhydride (TFAA/ Ac 2 O) and HMDS.The goal is to realize a system that changes hydrophobicity without changing the pore structure and the mechanical properties.
The used paper type for all experiments was an unsized base paper which is converted to different business paper grades, produced by Mondi AG.The paper is made from bleached eucalyptus pulp, it contains 21.25% calcium carbonate filler particles.Neither sizing agents nor calendaring (surface smoothing) have been applied to these papers.The most important paper properties are listed in Table 1.
2.2.Gas phase hydrophobization using HMDS and TFAA/Ac 2 O A5 paper sheets were placed in a desiccator (volume = 10 l), which was equipped with a rubber seal, an intermediate base and a glass bowl containing the hydrophobization agent (3-5 ml).The paper sheets were fixed with a yarn on both sides and were mounted in the desiccator.Afterwards, the desiccator was evacuated to a pressure of 400 mbar.For reaction of HMDS, the desiccator was placed in a laboratory furnace at 40°temperatures to promote the evaporation of HMDS.The anhydrides (mixture of TFAA and Ac 2 O in a ratio 1:2) did not require any elevated temperatures for the gas phase reaction and were kept at room temperature.The reaction time was 24 h for HMDS, and 5 min for the TFAA/ Ac 2 O.
The vapor pressure of HMDS at this temperature is ca 4000 Pa at 40 °C and TFAA is ca 54,000 Pa at 25 °C according to the literature [46].

Gas phase hydrophobization using PCl
Palmitoyl chloride (2 ml) was added into a 10 ml round flask equipped with a stirring bar.Paper stripes were mounted above the flask and fixed by a Teflon disk with holes.The flask was evaporated (10 mbar) and heated to 180 °C for a period of 2 h.To reduce condensation of the reagent in the reaction chamber the whole chamber was insulated with aluminum foil.

Wettability testing
Contact angle measurements were carried out with a Fibro DAT 1100 instrument with a drop volume of 4 μl.The used liquid was distilled H 2 O.During the experiment the calculated contact angle is plotted against the time, where the change in contact angle over time is related to wetting and penetration processes.To calculate the contact angle a tangent is fitted to the outline of the interface of the drop and the paper surface.The initial contact angle, which was used to compare the hydrophobization methods to each other, was measured after 200 ms, because before the drop shape was instable due to application of the drop on the paper.For the measurement, a paper strip with a width of 15 mm is used and 10 drops were applied on the strip during the experiment.The measurement time for one drop was set to 1 min.The contact angle measurements were carried out with all samples modified with different hydrophobization agents.
To determine the contact angle hysteresis and furthermore the receding and advancing contact angle, a Krüss DSA 100 device was used.A 5 μl water drop was placed on the modified paper sheet via a syringe and the volume was increased with 1 μl/s until the contact angle did not change anymore to obtain the advancing contact angle.After reaching the advancing contact angle 0.3 μl/s were removed till the contact angle did not change anymore.At this point the receding contact angle was recorded.The contact angle hysteresis, advancing and receding contact angle were determined from the HMDS-modified sample.
The surface energy of the HMDS-modified paper sample was experimentally analyzed with contact angle measurements with three different test liquids using a Dataphysics OCA 200 device.The overall surface tensions (ST) and their polar and dispersive components are shown in Table 2.
On each sample five drops of each test liquid (V = 2 μl) were applied, and the contact angle was measured.The surface energy and their polar and dispersive component were calculated via the OWRK method [47,48].

.Penetration dynamics analyzer (PDA)ultrasonic liquid penetration
For all modified substrates the Emtec Penetration Analyzer 2.0 was used, and the measurement frequency was set to 2 MHz.The paper sheets were cut into rectangles of 7 cm × 5 cm and were fixed on the sample holder with a two-sided adhesive tape.The sample holder was placed in the measurement cell which was filled with distilled H 2 O. On one side of the cell, there is an ultrasonic emitter while an ultrasonic receiver is fixed on the other end.The paper sample is quickly immersed in the liquid and ultrasonic waves are sent through the liquid and the sample from the emitter to the receiver.The ultrasonic intensity measured at the receiver is recorded over time.The ultrasonic waves get scattered, absorbed or reflected by the sample during liquid penetration and thus the decline of signal intensity over time provides a measure for the liquid penetration into the paper [49].

IR-spectroscopy
Infrared spectra were recorded on an ALPHA FT-IR spectrometer from Bruker.The measurements were carried out with an attenuated total reflectance (ATR) attachment using a diamond ATR element, with 48 scans at a resolution of 4 cm −1 and with a scan range between 400 and 4000 cm −1 .The recorded data was analyzed using OPUS 4.0 software.

X-ray photoelectron spectroscopy
Surface chemistry was evaluated with XPS, using a Kratos AXIS Ultra photoelectron spectrometer with monochromated Al Kα irradiation at low power (100 W) and under neutralization.Prior to the experiment samples were pre-evacuated overnight, in order to stabilize experimental conditions.Samples were measured together with the lab-defined of in-situ reference of 100% cellulose (Whatman).Both wide energy resolution scans, with 1 eV step and 80 eV CAE, and high-resolution scans of C1s and O1s regions with 0.1 eV step and 20 eV CAE were recorded on 2-3 locations.Nominal analysis spot size is 400 × 800 μm.All binding energies were charge-corrected using the cellulose C1s main component of C\ \O at 286.7 eV, after fitting the high-resolution carbon data into four Gaussian components with equal half widths, using CasaXPS software.
HMDS coated paper samples were also rinsed with excess THF prior to measurement.
Silicon content (extracted from Si2p) were 0.17 ± % (blank), 0.93 ± 0.05% (HMDS) and 0.87 ± 0.13 (HMDS + rinsed).This means an effective increase of 0.76 at.% silicon in the papers upon HMDS treatment.The blank was measured at three spots while for the HMDS and HMDS/rinsed sheets three spots on two different sheets were explored.
The DS was calculated by considering the change in molecular composition when silicon is present in the molecule.The molecular formula C (6+3x) O 5 Si x depicts the steady increase of silyl groups as detected by XPS.Based on this formula a plot can be made, where on the y-axis the Si-atomic % (according to the formula above) is plotted against the DS.The plot is not completely linear as the total amount of atoms increases with increasing silicon content, while the oxygen content remains constant.The plot in the region between a DS Si of 0.05 and 0.25 is depicted in Fig. 1 .
The regression line can be used to calculate the DS of the papers that have been treated with HMDS (Si at.% = 0.76), leading to a value 0f 0.087.

Tensile testing
The tensile tests were carried out according to ISO 1924-2:2008.Three stripes of each sample were tested with this method.The tensile test was carried out for the HMDS-modified and the TFAA/Ac2Omodified paper sheets.

Roughness measurements
The roughness of the paper samples was determined with a Bendtsen roughness tester, according to ISO 8791-2.Three individual spots on one paper sheet were measured.The optical properties were analyzed with a Datacolor Elrepho spectrophotometer.Brightness was measured according to DIN ISO 11476:2016 and opacity according to ISO 2471:2008.

Mercury porosimetry
Mercury porosimetry was measured with an Autopore IV 9500 device [50].This method is based on the penetration of a non-wetting liquid (mercury) at a high pressure.Here the pore size is determined from the mercury volume forced into the pore structure at increasing pressures.

Printing tests
The printing tests were carried out with a commercially available Epson XP-342 printer equipped with an ink tank system, connected to the ink cartridges via tubes.A commercial water-based pigment ink from Canon Production Printing was used.
In the printer settings, high printing quality was chosen and furthermore the paper settings were adjusted to "Epson Matte", to ensure a high ink deposition while printing.
To analyze the smearing of the ink, a rectangular finger tester was used, equipped with a two-sided adhesive tape.On this tape, a counter-paper was fixed and 10 s after printing the tester was pulled over the wet ink with a defined pressure.Afterwards, the color density of the counter-paper was measured.The color density was measured in a square arrangement in 9 different positions (3 × 3 arrangement).Each sample was tested three times.
Additionally, color density measurements of printed cyan, yellow, black and magenta areas were carried out using a Gretag Macbeth D19 c densitometer.Color density was measured in 9 different positions in a square 3 × 3 arrangement within the 100% intensity print area.
Finally, printed lines and areas were imaged using an Alicona Infinite Focus Microscope (IFM) with 50× magnification to compare feathering effects of the ink on the different paper surfaces.

Results and discussion
A disadvantage of PCl for paper hydrophobisation is its low vapor pressure.The reaction required a pressure of 10 mbar at a temperature of 180 °C to achieve homogenous surface modification.After exposing the papers for 120 min to PCl vapors, the static water contact angle was 117 ± 3°indicating a successful hydrophobization of the paper (supporting information, Fig. S1).The IR spectra of the paper before and after the modification, however, showed hardly any differences.Only a band with very weak intensity at 1730 cm −1 indicated the presence of an ester bond (supporting information, Fig. S2).A major disadvantage of this coating procedure is that it requires harsh reaction conditions which have caused an undesired deterioration of mechanical paper properties.This approach is therefore not suitable for the homogeneous coating of large substrates.In contrast to that, the other two employed compounds (HMDS, TFAA/Ac 2 O) exhibited much higher vapor pressures.Therefore, the coating procedures were carried out in desiccators at a pressure of 400 mbar and moderate temperatures (HMDS: 40 °C, TFAA/Ac 2 O: 25°).We studied the kinetics of the process as a function of the water contact angle (acquired after 200 ms settling time of the droplets on the surface) for the different compounds (Fig. 1).We expected that prolonged exposure times led to a more effective hydrophobization of the paper sheets.For the TFAA/Ac 2 O system, we observed an increase in hydrophobicity from 76°± 3 (5 min) to 98°± 3 (15 min, Fig. 1a).However, longer exposure times led to a reduction of the static contact angle to 86°± 5°at 30 min, and 80°± 4°a t 60 min).In addition, the samples turned yellowish and started to develop a different haptics, which became more pronounced with increasing exposure time.At the same time, roughness increased dramatically and mechanical properties deteriorated (supporting information).Probably the filler (CaCO 3 ) degraded due to the formation of acetic and trifluoracetic acid.In contrast, the samples exposed for 5 min also showed an increase in roughness but the haptics and also the tensile strength did not feature any difference to the unsized paper (supporting information, Fig. S3).The IR spectra of the samples showed an intense band at 1673 cm −1 which is characteristic for C_O vibrations (supporting information, Fig. S2).The strong intensity of this band indicated that a rather high DS was reached already within 5 min reaction time.At similar exposure times as for the TFAA/Ac 2 O system, hardly any lasting effect of the HMDS on the paper surface properties was observed.Although the initial contact angles (10 ms after deposition) were high (91°), the hydrophobization effect disappeared already after a short time.After 200 ms, the contact angles were only slighter higher than those of the unsized raw paper.Longer exposure times to HMDS (24 h), however, led to completely hydrophobized papers with static water contact angles of 106 ± 3°which do not change within 1 s (Fig. 1b).
In addition to the contact angle measurements over time, the contact angle hysteresis (advancing θ adv and receding θ rec contact angle) was investigated.The determination of θ adv and θ rec was challenging due to high roughness of the paper sheets.At a certain drop size (approx.6 μl) the contact angle became smaller again, because the shape of the drop changed due to the high paper roughness.θ adv (131 ± 1°) was determined from the mean of the three highest measured contact angles and θ red (93 ± 1°) was determined from the three points right before the drop shape changed.Further, we explored the stability of the HMDS modified papers during storage in a defined environment (22 °C, 50% relative humidity).For storage up to 7 days, hardly any difference in wettability was observed, After this period, the static water contact angles decreased to some extent.After three weeks, the materials still exhibited clearly hydrophobic character with static contact angles larger than 90°(Fig.2).
IR spectroscopy should provide evidence for the formation of a silyl ether which feature typical bands at 840 (ν Si-C ), 1254 (ν SiOC ) and 1050 (δ SiOC ) cm −1 .However, the IR spectra were identical for the coated and the nontreated papers, indicating a rather low degree of substitution with silyl groups.In order to elucidate the amount of silicon in the papers, we performed XPS.
The XPS spectra (Fig. 3) showed the presence of additional contributions in the C1s (287 eV), O1s (533 eV) and Si2p (101 eV) region stemming from trimethylsilyl groups, accounting for 0.93 at.% in total.As the papers contained traces of Si already before the treatment (0.17 at.%), the effective increase by the HMDS treatment was 0.76 at.%.The amount of silicon in the papers can be used to determine an average degree of substitution as only trimethylsilyl groups can be formed.The DS TMS (actually it is the DS Si,surface on the XPS accessible area of the outermost layers of the paper sheets) was 0.087 which means that only ca every 40th AGU is modified with a silyl group which is extremely low for a hydrophobic coating.An interesting question is whether the silyl groups are covalently attached or whether there is a significant portion of groups that is physically adsorbed on the cellulose.For this purpose, we rinsed the modified paper sheets with THF, which is an excellent solvent for HMDS.The overall silicon content slightly decreased from 0.93 ± 0.05 at.% to 0.87 ± 0.13 at.%, which however is within the statistical error.It can be concluded that the silicon introduced by HMDS is not removed by the THF treatment, which is a strong indication that there are covalently bound silyl groups present on the cellulose fibers.However, one may argue that this is an average value for a limited area of a paper sheet as measurement spots are typically in the size of tens to hundreds of microns in XPS.Another option is to determine the coating efficiency via an indirect method.As XPS and wettability are capable to exclusively detect surface effects, a penetration profile  S1. would give conclusions about the coating of the inner structure of the paper sheets.
A powerful tool to study penetration dynamics is to exploit scattering and reflection of ultrasonic waves.Two main parameters can be extracted from these measurements, namely the wetting time, t w , and the penetration speed.In PDA measurements, the wetting time is defined by the time it takes to achieve a 100% intensity of the ultrasonic signal, i.e. more hydrophobic materials feature t w in the seconds regime while hydrophilic ones are wetted within a few milliseconds.Inhomogeneities on the paper surface may unravel as wetting will proceed via dislocations or non-uniformly coated areas on the surface.The slope of the decay of the intensity correlates with the penetration speed [49].As expected for a hydrophilic substrate, the unsized paper showed a very short t w (0.03 s), which is already on the edge of the resolution of the instrument (Fig. 4).
Also the TFAA/Ac 2 O coated papers showed a similar wetting time, despite their hydrophobicity, indicating an inhomogeneous coating of the paper surface.Both, the unsized paper and the TFAA/Ac 2 O modified ones exhibited penetration speeds of 537 ± 18 and 442 ± 15 s −1 .The PCl and HMDS coated papers exhibited much longer wetting time of 0.72 and 2.29 s and much smaller penetration speeds of 1.54 ± 0.04 and 0.42 ± 0.01 s −1 .
Further, an experiment to elaborate whether the HMDS can diffuse in the gas phase also through the paper substrate was performed.Prior to the exposure to vaporous HMDS, we covered one half of a paper sheet with aluminum foil, and fixed it with adhesive tape to prevent diffusion of HMDS underneath the aluminum foil.After removal of the aluminum foil, the wettability of the covered part of the paper sheet (104°± 1°) was identical to the one without aluminum protection (105°± 1°).Although the experiment is simple and may have limitations, it is a further indication that HMDS is able to penetrate the whole paper network rather than modifying just the surface of the sheets.As a consequence, also the pore structure should be hardly affected by the coating procedure.Fig. 5 depicts the incremental specific intrusion of mercury which is plotted against the pore diameters of the paper network.The Hg-porosity measurements revealed a porosity of 40.9 for the unsized paper and 44.1 for the HMDS coated papers.This is, given the accuracy of the method, a very good agreement.
Further experimental evidence that paper integrity has not been compromised by the HMDS coating procedure was obtained by mechanical testing of the paper sheets.Tensile strength of the papers was determined 4069 ± 13 and 4037 ± 58 N•m −1 for the unsized and the HMDS coated papers, respectively (supporting information, Fig. S3).For instance, the TFAA/Ac 2 O modified paper sheets showed a decrease in tensile strength to 3586 ± 230 N•m −1 .
The optical properties of the papers (opacity, brightness) were not affected by the coatings (supporting information, Fig. S4).
Moreover, it is of interest if the hydrophobized papers are comparable to industrially sized paper.The HMDS coated paper behaves similarly to paper sized industrially with alkyl ketene dimer (AKD) in terms of contact angle and penetration dynamics (supporting information, Fig. S6).We correlated the gas phase modifications to the printing behavior of the modified papers.We used a commercial printer that operates with a water-based pigment ink with a surface tension of 32 mN•m −1 .Yellow, black, cyan and magenta colored rectangles were printed on the sheets and color density, penetration and smearing of the ink was analyzed.
Fig. 6 compares photographs of the different printouts of the TFAA/ Ac 2 O and the HMDS modified papers with the unsized sheets.Already by visual inspection hardly any difference between the HMDS and the unsized papers can be observed.This was confirmed by analysis of the mean color densities which did not reveal any differences (0.94 ± 0.06 vs 0.93 ± 0.05) for the unsized and HMDS coated paper, respectively (supporting information, Fig. S5).The TFAA/Ac 2 O modified papers, however, showed a significant reduction in color density (0.68 ± 0.05).The lower color density of the TFAA/Ac 2 O modified sheets originated from an incomplete coverage of the fibers, probably caused by the inhomogeneous surface modification as observed in the PDA measurements.The micrographs clearly showed that there were parts on the paper where there was no ink, which thus leads to lower color densities (Fig. 7).The unsized papers showed feathering to a minor extent, i.e. the ink unevenly spread on the surface concomitant with reduced penetration of ink into the paper.In contrast, the HMDS coated papers did not feature feathering, probably due to a better match of the surface free energy of the coated papers (48.9 mN m −1 ) with the surface tension of the ink (32 mN m −1 ) compared to the unsized sheets (63.0 mN m −1 ).
For the smearing test, we placed another paper sheet on the printouts and employed a rectangular finger tester, which was moved afterwards.Smearing directly impacts the color density, i.e. more smearing leads to higher color densities.Fig. S5 depicts the color density after smearing of the different papers.The unsized and HMDS coated sheets both exhibited similarly low color density (i.e.low smearing tendency).The TFAA/Ac 2 O coated papers, by contrast, exhibited higher smearing tendency.The formation of calcium trifluoroacetate from CaCO 3 during the gas phase treatment could fixate the ink pigments on the paper surface similarly as in CaCl2 priming [51].

Conclusion
We have demonstrated that a commonly used compound for gas phase coatings in semi-conductor industry, HMDS, can be also applied to manufacture hydrophobic papers with static water contact angles above 100°.We showed that HMDS is able to diffuse into the paper, thereby homogenously modifying the individual fibers throughout the whole paper substrate.The degree of hydrophobicity can be tuned to some extent via the reaction time, or as we showed, by simply storing the samples for several weeks.As the DS TMS is low, the porosity of the sheets and other relevant paper technological parameters were not affected by the gas phase hydrophobization which is a difference to most of the other reported surface functionalization techniques.The stability of the silyl layer is sufficiently large to study interactions of liquids with such paper substrates.Ink deposition during printing revealed major differences between HMDS and TFAA/Ac 2 O coated papers in color density and smearing properties.There is a direct correlation between the surface tension of the ink and the surface free energy of the coated papers.This provides the opportunity to systematically study the influence of surface free energy on liquid-substrate interactions.In conclusion, HMDS hydrophobization provides a powerful tool to tune the hydrophobicity of porous cellulosic sheets without affecting the pore structure or the mechanical properties of the substrate.

Fig. 1 .
Fig. 1.Evolution of water contact angles [°] as a function of settling time of unsized paper sheets in dependence of exposure time to TFAA/Ac 2 O (a) and HMDS (b).

Fig. 2 .
Fig. 2. Static water contact angles after different storage times.

Fig. 3 .
Fig. 3. XPS spectra of the paper sheets hydrophobized with HMDS (a) and unsized paper (b).The elemental composition is listed in TableS1.

Fig. 4 .
Fig.4.Penetration dynamic analysis of the modified and unsized raw paper.

Table 1
Properties of the used papers.