Droplet slipperiness despite surface heterogeneity at molecular scale

Friction determines whether liquid droplets slide off a solid surface or stick to it. Surface heterogeneity is generally acknowledged as the major cause of increased contact angle hysteresis and contact line friction of droplets. Here we challenge this long-standing premise for chemical heterogeneity at the molecular length scale. By tuning the coverage of self-assembled monolayers (SAMs), water contact angles change gradually from about 10° to 110° yet contact angle hysteresis and contact line friction are low for the low-coverage hydrophilic SAMs as well as high-coverage hydrophobic SAMs. Their slipperiness is not expected based on the substantial chemical heterogeneity of the SAMs featuring uncoated areas of the substrate well beyond the size of a water molecule as probed by metal reactants. According to molecular dynamics simulations, the low friction of both low- and high-coverage SAMs originates from the mobility of interfacial water molecules. These findings reveal a yet unknown and counterintuitive mechanism for slipperiness, opening new avenues for enhancing the mobility of droplets.

Trichlorosilane molecules can form either smooth self-assembled monolayers via self-limiting adsorption on the substrate surface (as illustrated in the main text Fig. 1) or rougher films where silane molecules react with each other in a process termed polymerization, as has been presented earlier by Fadeev et al. 1 Water has a complex role in trichlorosilane chemistry because it is a critical reactant in the hydrolysis of trichlorosilanes and it is also a reaction product of the condensation and polymerization reactions.The selflimiting growth type requires a dry reaction environment, whereas the polymerization growth type occurs when the environment contains water.When the reaction environment is not sufficiently dry, the overall growth can be a mixture of these two growth types.In this supplementary note we discuss the aspects that support the smooth monolayer type of growth over the growth type involving polymerization.
In this work, the growth of OTS films is performed using an atomic layer deposition reactor capable of providing excellent control on minimizing the water content in the reactor atmosphere by having the reactor in vacuum, purging it with nitrogen gas and having it at temperature of 60 °C (reactor is always kept at least at 60 °C to avoid accumulation of water in the reactor walls prior to the OTS growth process).The reactor has a base pressure of ca.7 Pa, which is mostly due to leakages of normal air that has ca.1% water content.The OTS pulse pressure is 50 ± 10 Pa, which is much higher than the partial pressure of residual water in the reactor atmosphere.Therefore, OTS molecules in vapor phase are unlikely to hydrolyze and polymerize with other OTS molecules in the reactor atmosphere.
The OH-group rich SiO 2 substrate surface is hygroscopic and may have a thin film of surface-bound water despite the otherwise dry reaction environment.Therefore, the hydrolysis of the OTS molecules likely occurs only at or in the vicinity of the substrate surface, and due to the proximity of the surface OH groups, the hydrolyzed OTS molecules are also likely to bond to the surface.Once a first Si-O-Si bond between an OTS molecule and a surface is formed, it is likely that the remaining two Si-OH moieties of the hydrolyzed OTS molecule will also bond to the substrate surface or possibly to a neighboring OTS molecule already adsorbed to the surface if it still has a remaining Si-OH moiety.Therefore, OTS should not substantially polymerize vertically with other OTS molecules while bonding to the substrate is the more probable scenario.
Experimental evidence also suggests growth of smooth self-assembled monolayers without substantial vertical polymerization.Firstly, the formed OTS films maintain the smoothness of the underlaying silicon wafer substrate according to atomic force microscopy measurements (see Figure 2d-g in the main text).Vertical polymerization would lead to increased RMS roughness, which we do not observe at any OTS SAM coverage regime.Secondly, the coverage increase follows a standard monolayer growth model (Supplementary Note 4) that assumes growth rate being proportional to the fraction of unoccupied surface area.This indicates that OTS molecules do not grow on top of other OTS molecules that have bound earlier to the surface, i.e., the growth is clearly self-limiting.Lastly, ellipsometry shows that the growth of OTS film also saturates to thickness that is nearly equal to the length of a single OTS molecule.For the above reasons, the OTS film can be assumed to form as a smooth monolayer with negligible amount of vertical polymerization.

Supplementary Note 2: Detailed analysis of OTS SAM growth process
Covalent bonding of OTS molecules to SiO 2 surfaces releases HCl as by-product.At the same time, the partial pressure of OTS vapor decreases in the reactor slowing down the growth rate.Furthermore, over time air leaks into the ALD reactor gradually bringing in moisture and other airborne impurities.For these reasons, the reactor was replenished periodically for the OTS SAM depositions of 20 min and longer by reconnecting the reactor to vacuum pump and removing excess OTS and reaction byproduct HCl and applying a new dose of OTS.As the growth of OTS SAM is fast in the beginning, the first exposure step was kept short, maximum 15 min.The next 1-4 steps were slightly longer, up to 27.5 min, and the last steps after that were each 60 min.Table 1 shows list of exposures used for each prepared SAM surface.The SAM growth was monitored in-situ with operando ellipsometry, i.e., monitoring the SAM growth occurring in the ALD reactor chamber in real time.Fig. 1a-e shows the SAM thickness development of each prepared SAM surface.In each case, the growth starts by a rapid increase of thickness (marked with first * in the main text Fig. 2a).This can be accounted for rapid physisorption of OTS molecules onto the substrate surface.The physisorbed OTS molecules start soon covalently bonding to the substrate OH groups, releasing HCl by-product.This mass loss causes a step decrease in the observed film thickness.After roughly 1-2 minute of growth, observed thickness starts again to increase as rate of HCl removal decreases below OTS adsorption rate.
Once first exposure is finished, the ALD reactor is reconnected to vacuum pump for 5 s, which removes most of produced HCl and also excess OTS from the reactor.This effectively removes physisorbed OTS from the sample surface, which causes a rapid yet small decrease of observed film thickness, and the remaining thickness represents covalently adsorbed SAM.For those runs with multiple exposures, a new OTS dose is introduced into the ALD reactor chamber (marked with second * in the main text Fig. 2a), which causes a similar peak in the observed thickness as with the first exposure.Again, more molecules get physisorbed onto the substrate surface, and the molecules start binding covalently to the substrate over time before the next reactor purge that removes all the physisorbed molecules.Each following OTS exposure cycle behaves similarly, however the amount of physisorption gets smaller as OTS coverage increases, which appears as reduction of thickness decrease right after the dosing.
Fig. 1f shows the decrease of observed thickness after the last purge for each SAM growth.This thickness decrease is due to the removal of physisorbed molecules like discussed above.The decrease gets smaller towards the longer runs, and after 24 h deposition no removal is observed, which coincides with the slowing of the growth.The introduction of first replenishment event (second OTS dose) increases the thickness decrease after the deposition.Therefore, it is likely that replenishment cycles increase the growth rate, affecting the total growth slightly (i.e., 20 min growth would have had lower final thickness had it been performed with one OTS dose only).After the thickness decrease, the remaining OTS is covalently bonded, as introduction of ca. 100 Pa and 1 s water vapor pulse into the reactor does not cause observable changes in the SAM thickness, see Fig. 2. Had there been substantial amounts of unhydrolyzed physisorbed OTS on the surface, the hydrolysis would have likely caused an observable change in ellipsometric SAM thickness.Ellipsometry can be used to obtain SAM thickness and refractive index by fitting an optical model describing the surface structure to the measured  and  values.In the case of OTS SAM deposited on Si wafer with native oxide layer, the optical model composes of three layers.The bottom layer represents the crystalline Si substrate, and it is modeled with tabulated values for refractive index.The middle layer represents native oxide of the silicon wafer.Its thickness is set to 1.5 nm (see below for details) and its refractive index is modeled with Cauchy dispersion equation ( ) , where is the wavelength of light and , , and are constants that are fitted from data recorded just before the start of SAM growth.The top layer represents the OTS SAM.To estimate the average SAM thickness with the model, refractive index was fixed to 1.45 2 , and to estimate the refractive index, thickness was fixed to 1.0 nm, which corresponds to OTS SAM maximum height obtained in the MD simulations in this work.As the SAM thickness is in the order of 0.1 -1.0 nm, ellipsometry is not sensitive enough so that both thickness and refractive index of the SAM could be fitted simultaneously.
The exact thickness of the native oxide is difficult to estimate with sub-Ångström accuracy.Therefore, we tested how sensitive our analysis for SAM thickness is for the set native oxide thickness.Fig. 3a shows the obtained SAM thickness for three different native oxide thickness values used in the analysis.The SAM thickness is within the error margins for each case, meaning the analysis is not sensitive for the set value of the native oxide thickness (in range from 1 nm to 2 nm).
The recorded SAM thickness and refractive index are dependent on the assumed value of the other parameters.For example, Fig. 3b shows how thickness of a 4 h SAM is different depending on what refractive index value is used in fitting of the optical model to the ellipsometer data.This leads to a systematic error if there is a difference between the assumed and real refractive indices, meaning that the absolute obtained ellipsometer values may have an inaccuracy of ca.10% but the relative difference between the samples is still accurate.
The obtained refractive index can be further transformed into SAM coverage, i.e., areal density of OTS molecules 1,2 .In short, mass uptake per unit area of the film can be calculated using Equation 1( ) where M and are molecular mass and molar refractivity of the SAM, respectively, d is SAM thickness and n is refractive index of the SAM.Areal density of molecules adsorbed on the surface is obtained by dividing both sides of Equation 1 with mass of a single molecule, yielding Equation 2 In the case of OTS, = 189, = 46.7 (values obtained using principle explained by Cuypers et al. 3 ), = 3.14 • 10 −25 kg and = 1.0 nm as assumed during determination of refractive index from the ellipsometer data.Again, it must be noted that calculation of molecule areal density is sensitive to the assumption of SAM thickness.If thickness is assumed to be 0.9 nm or 1.1 nm instead of 1.0 nm, the coverage shifts ± 10%, see Fig. 3c.The substrate stability inside the ALD reactor is essential when growing SAMs for several days.We verified the surface stability by keeping a substrate inside an idle ALD reactor over 72 h time period.First, a Si substrate was pre-treated similarly as described in the Methods section, and then inserted into the ALD reactor pre-heated to 60 °C.Next, the reactor was pumped down to vacuum and a 20 sccm N 2 flow was added for 72 h.The sample surface was monitored with operando ellipsometry during that time.The sample was modelled with two-layer model: the first layer is the crystalline Si substrate (fixed values for refractive index) and the second layer is the native oxide that is modelled with Cauchy dispersion.In the beginning of the experiment, the thickness of the native oxide was set to 1.50 nm, and Cauchy parameters A, B, and C were fitted to obtain the refractive index of the native oxide.Then the Cauchy parameters were fixed to the fitted values, and thickness of the native oxide was fitted during the in-situ monitoring of the sample.
The results are shown in Fig. 4 below.The detected changes are very small and random.This indicates that the native oxide thickness remains stable inside the ALD reactor, and it is valid to assume a constant native oxide thickness during SAM growth monitoring.We note also that the fit of native oxide thickness is sensitive for any changes at its surface, in particular if the amount of surface bound water would change.Therefore, the amount of surface bound water is also very stable, which is important for the SAM growth stability.
After the 72 h waiting period, another, freshly pre-treated Si substrate was inserted into the ALD reactor.Next, a normal OTS SAM deposition was performed for the two substrates inside the ALD, after which advancing and receding contact angles of both samples were measured.The results are in Table 2, and they show that SAM grew very similarly on both samples.This indicates that surface OH group density has also remained very constant during the 72 h wait period inside the ALD reactor, which is important for the late stage SAM growth and quantification of OH groups with the metal reactants that occurred after the SAM growth.

Supplementary Note 6. Fourier transform infrared spectroscopy (FTIR) of SAMs
Level of SAM crystallinity can be determined with FTIR as the peak position of the asymmetric CH 2 stretching band depends on the alkyl tail configuration 5,6 .For alkyl tails in all-trans configuration the peak occurs near 2917 cm -1 , and the more cis conformers the tail has the more the peak is blue shifted, i.e., wavenumber increases.Fig. 5 shows FTIR spectra obtained for five SAMs with different growth times.The asymmetric CH 2 peak occurs redshifts from 2929 cm -1 for 2 min growth time to 2926.5 cm -1 for 168 h growth time.This indicates a small straightening of alkyl tails during the SAM growth, but no crystallization within the explored growth time range.

Supplementary Note 7: Alkyl chain tilt angle and OTS molecule height of simulated SAM surfaces
We extended the MD simulations of the generated SAMs (Supplementary Note 18) by further 42 ns.The last 40 ns were used for analyzing the average tilt angle of the alkyl chains of the OTS molecules.(See Video 1 how alkyl tails wiggle due to thermal motion.)For that, we calculated the angle as depicted on Fig. 6.The vector following the alkyl chain goes from the silane oxygen bonded to surface to the last carbon of the alkyl chain (the methyl group).
At lower coverages, nearly all molecules are lying down with peak maximum between 75° and 80°.As the coverage increases from 0.77 molecules nm -2 to 1.10 molecules nm -2 , an increasing fraction of molecules start to have tilt angles around 50°, because some molecules cannot find space to lie along the surface.This behavior is intensified for medium coverages, where we see a transition from most of the chains having tilt angles between 45° to 50°.At 2.4 molecules nm -2 coverage a significant share of molecules tends to be perpendicular to the surface, with tilt angles of less than 30°.For higher coverages, the tilt angles start moving to lower values below 20°.This trend is in accordance with earlier reports of tilt angle for other silane chains with different alkyl chain lengths 7 .
We measure the length of the chain as the distance between the silane oxygen bonded to surface and last carbon of the alkyl chain (in an all-trans configuration), obtaining a value of 0.975 nm.We multiplied this value by the cosine of the mean angle for each coverage density (Fig. 6), obtaining an estimate for the mean molecule height in our simulated systems, as shown in the main text Fig. 2k.Comparing the latter to the ellipsometric thickness presented in the main text Fig. 2b we see that the mean molecule height overlaps the experimental thickness data from ellipsometry.

Supplementary Note 8: Labelling OH vacancies with metal reactants
Let us first consider how the metal reactant compounds bind to the SAM surfaces during the vapor phase deposition step under the dry vacuum condition.The metal reactant compounds react with a surface OH group via reaction , where is the central metal atom, is ligand, is the coordination number and marks surface group [8][9][10][11] .For diethylzinc (DEZ) = 2 and the reaction can occur for one ligand or for both of the ligands, if there are two adjacent OH-groups where a DEZ molecule can bind to (Fig. 7a).For titanium tetrachloride (TiCl 4 ), tetrakis(dimethylamido)hafnium (TDMAHf), and titanium tetraisopropoxide (TTIP) = 4 and the reaction can occur maximum three times due to molecule tetrahedral geometry, as one of the ligands is always pointing out of the surface (Fig. 7b).Once the surfaces are removed from the dry vacuum conditions, the remaining ligand(s) of the molecules react with moisture of the air where is the number of remaining ligands.Therefore, the deposition of the metal reactant compounds effectively adds only metal atoms and OH groups to the SAM surface.Lastly, the density of OH groups on the surface is high, almost 10 groups nm -2 based on analysis done in the main text section "Quantification of OH vacancies of SAM surfaces".Therefore, practically each vacancy of the SAM large enough for metal reactant adsorption contains at least one OH group allowing covalent bonding of the reactant to that vacancy.
The possible physisorption of metal reactants to the SAM surfaces needs to be considered, too.Based on earlier observations made by Hong et al. 12 , adsorption of TDMAHf does not occur on top of a dense octadecyltrichlorosilane (ODTS) layer.OTS is chemically similar to ODTS, so it is safe to assume that TDMAHf does not physisorb on OTS.Adsorption of other metal reactants as function of OTS coverage resemble that of TDMAHf adsorption, implying that those reactants would not have physisorption either on top of the alkyl tails.
Physisorption of metal reactants could in principle occur also to the vacancies of the SAM.Possibility for this was checked by depositing DEZ on top of uncoated SiO 2 surface similarly as described in the Methods section.After deposition, the Zn/Si elemental ratio was determined with XPS at different temperatures starting from 20 °C and ending to 300 °C.Results are shown in Fig. 8, and no substantial desorption of Zn from the surface can be observed.The good temperature stability of the Zn/Si ratio indicates that DEZ is mainly covalently bonded to the surface.We note that the metal reactant deposition was also itself performed at 150 °C to minimize the possibility for physisorption of the reactant molecules.Lastly, the trend of metal/Si ratio is very similar for all metal reactants as function of SAM coverage, as is visible in the main text Fig. 3c.Had there been substantial physisorption of these reactants, it would be likely visible in the original XPS results, as all of the reactants have different ligands and thus likely different physisorption properties.geometry always causes one ligand to point away from the surface, thus preventing it from bonding to surface OH groups.b-l) Water droplet on SAM surfaces with coverages 0.08 -3.87 molecules nm -2 .The color bar represents mass density in g cm -3 of the oxygen from the water molecules.For SAM up to coverage of 2.92 molecules nm -2 , the densities calculated averaging the oxygen atom positions of the last 5 ns of the simulations.For SAM coverages of 3.44 molecules nm -2 and 3.87 molecules nm -2 , the average was done in the last 0.25 ns of the simulations due to the high mobility of the droplets in those cases.For plain SiO 2 , the average was done in the last 15 ns of the simulations.

Determination of contact angles
We determined the droplet's contact angle for each MD system.Due to the droplet's lateral movement in some of the systems, we reprocessed the trajectories to center the droplet on the same position in all frames and calculated the number density of water molecules for each configuration.The static contact angles of simulated droplets were obtained by fitting a circular arc into the edge of the droplet density profile and setting the baseline to the SAM average height, see Fig. 13a.Results are shown in Fig. 13b.The static contact angles agree with experimentally obtained ACA for coverages below 2.0 molecules nm -2 .Above that coverage, simulated droplets have higher static contact angle than the experimentally obtained ACA.This is due to the relatively high surface roughness in comparison to droplet size in the MD simulations.Contact angle dependency on the simulated droplet size was checked with droplets ranging in size from 1718 to 4274 water molecules without notable size sensitivity observed (see Fig. 13c).

SAM coverage
Supplementary Note 12: Prediction of SAM contact angle with Cassie's law The contact angle of a chemically heterogenous, smooth surface can be predicted with Cassie's law ∑ where f i is the fraction of surface component i and  i is its contact angle.For a case where SAM is grown on well cleaned, hydroxyl rich SiO 2 surface, Cassie's law can be written as where represents the normalized surface coverage of SAM, contact angle of homogenous OTS SAM surface and contact angle of homogeneous SiO 2 surface.Here we used = 110° and = 0°.For surface coverage , two different estimators were tested.The first one is the SAM coverage obtained from ellipsometry normalized by its maximum value.The second one is the covered area of SAM alkyl chains (see Fig. 14a) that is calculated from areal molecule density and average alkyl tilt angle where and are the radius and height of the SAM octyl tail, respectively, and  is the average molecule tilt angle as depicted in Fig. 6. = 0.25 nm and = 1.0 nm were used in the calculations.As tilted molecules may partially lie on top of each other, the covered area can be larger than the underlying area itself, and to be used as a coverage estimation in Cassie's law it requires normalization with its maximum value.Fig. 14b shows the difference in calculated by the two methods.
Fig. 14c shows the measured ACA and RCA of the SAM surfaces and the contact angle for each surface predicted with Cassie's law using the two above mentioned coverage estimators.As is seen from the graph, the areal density of SAM alone fails to predict the OTS SAM contact angle, especially with coverage of 0.5 molecules nm -2 and above.Considering the average molecule tilt, the prediction becomes better.The high tilt of molecules at low coverages increases the surface contact angle at early stage of deposition and lowering of the tilt in the later stage of growth explains the slower growth of contact angles.A note must be made that the simple normalization of the covered area is far too naive method as it ignores the varying fraction of molecules lying on top of each other on different SAM coverages, and thus Cassie angles calculated using this coverage estimator cannot be expected to be fully accurate.

Fig. 1 |
Fig. 1 | Monitoring OTS SAM growth with ellipsometer.a-e) SAM ellipsometric thickness growth of all depositions with total time ranging from 30 s to 168 h.f) Magnitude of thickness decrease at the end of each deposition.The inset shows how thickness decrease (i) is calculated by taking the difference of average thickness (ii) before and (iii) after the final purge (iv) final purge.Inset data from 2 min deposition.Different colors show how many replenishment (= evacuation + dose) steps there have been in each deposition process, and error bars represent pooled standard deviation of averaged thickness regions before and after the final purge.

Fig. 2 |Supplementary Note 3 :
Fig.2| Effect of water vapor on the OTS SAM thickness after final purge.The green shaded region represents the time when reactor contains OTS dose.The blue shaded region represents the time when reactor is being purged with N 2 .The red line represents the moment when water vapor pulse was introduced into the reactor.

Fig. 3 |
Fig. 3 | Ellipsometer accuracy.a) SAM final ellipsometric thickness with different values used for native oxide thickness.The 1.0 nm and 2.0 nm data series are shifted in x direction for better visibility.Error bars represent standard deviation of average SAM ellipsometric thickness after the final purge due to signal noise.b) Ellipsometric thickness recorded during 4 h OTS SAM growth with three different refractive index values used in the model fitting.c) SAM coverage of each sample for three different thickness values used in model fitting and coverage calculation.

Fig. 4 |
Fig. 4 | Stability of native oxide.Comparison of thickness change of native oxide and SAM over 72 h period.Inset shows a zoom-in to the measured thickness variations of the native oxide.The inset data is smoothed with running average to better visualize the long-term changes in the detected thickness.Table 2 | Advancing and receding contact angles of OTS SAM grown on UV-O 3 activated surfaces with 72 h agingand no aging prior to SAM growth.Contact angle error margin is ±0.7° for each measurement and is based on standard deviation of mean angles detected from three different surface locations.

Fig. 5 |
Fig. 5 | Fourier transform infrared spectroscopy of prepared SAM surfaces.The red dashed lines show peak locations of 168 h OTS SAM graph as guide for eye for comparison between spectra.The curves are background subtracted and leveled spectra with baseline shift for 30 min -168 h SAM spectra.

Fig. 6 |Supplementary Video 1 (
Fig. 6 | Angular distributions for all coverage densities simulated with MD.The inset shows how the tilt angle was calculated.Supplementary Video 1 (Supplementary_Video_1.mov) | Thermal motion of OTS SAM at room temperature.The video represents SAM surface with 0.77 molecules nm -2 coverage over a time period of 10 ns.

Fig. 7 |
Fig. 7 | Metal reactants bonding to surface OH groups.Metal reactant a) with two ligands can bond to one or two adjacent surface OH groups and b) with four ligands can bind to one, two, or three adjacent OH groups.Molecule

Fig. 8 |
Fig. 8 | Elemental ratio of Zn to Si resolved by XPS as function of substrate temperature.The error margins related to Zn/Si ratio are difficult to quantify (as they originate from multiple sources and are related to temperature measurement location, data fit accuracy etc.) and are thus not drawn in the figure.

Fig. 9 |Fig. 12 |
Fig. 9 | Elemental concentration of a) carbon (C), b) oxygen (O), c) silicon (Si), d) metal label (Zn/Ti/Hf) deposited on the SAM surface with varying SAM coverage.Elemental concentrations are measured from three different locations from each sample surface, which is represented by the three separate bars in each group.

Fig. 13 |
Fig. 13 | Comparison of experimental ACA and RCA to static CA of simulated droplet on OTS SAM surfaces.a) Schematics showing how CA were determined from simulated droplets.The white lines represent the circular arc fitted to the droplet number density profile and the SAM average height above the SiO 2 surface.b) Experimentally determined advancing and receding contact angles compared to contact angles calculated based on the MD simulations.c) Contact angles obtained from molecular dynamics simulations for varying water droplet sizes.Errors of the data points vary between 1.4° and 7.9° depending on system, yet not presented due to lack of statistics making error estimation imprecise.

Fig. 14 |
Fig. 14 | Prediction of OTS SAM surface contact angle.a) Illustration of accessible OH groups i) without molecule tilt and ii) with average molecule tilt considered.b) Normalized covered area of SAM molecules as function of normalized SAM coverage.The solid grey line represents 1:1 ratio.c) Experimentally observed contact angles and Cassie's law prediction for the contact angles with the two coverage estimators.

Fig. 16 |
Fig. 16 | Mass density profiles of water molecule oxygen atoms along the SAM-SiO 2 surface normal.The silica substrate edge is located at 0 nm, and it is defined by the z-coordinate position of topmost Si atom layer of the substrate.The vertical dashed black line indicates the location of the SAM edge for the coverage of 3.87 molecules nm -2 , which is the maximum coverage studied in this work.

Fig. 17 |
Fig. 17 | Residence correlation P() of water molecule oxygen atoms in the first hydration layer for each of the SAM coverages.The correlation decay allows determining the residence time.Plain SiO 2 and SAM coverages 0.08 -2.92 molecules nm -2 were analyzed at the SiO 2 interface and SAM coverages 3.44 -3.87 molecules nm -2 were analyzed above the SAM interface.Time (ps) 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0

Table 1 |
Prepared OTS SAM samples.Count and length of each replenishment steps are shown in the columns of the table.