desorption kinetics of methanol, ethanol, and water from graphene

Desorption Kinetics of Methanol, Ethanol, and Water from GrapheneR. Scott Smith,* Jesper Matthiesen,† and Bruce D. Kay*

Fundamental and Computational Sciences Directorate, Pacific Northwest National Laboratory, Richland, Washington 99352, UnitedStates

ABSTRACT: The desorption kinetics of methanol, ethanol, and water fromgraphene covered Pt(111) are investigated. The temperature programmed desorption(TPD) spectra for both methanol and ethanol have well-resolved first, second, third,and multilayer layer desorption peaks. The alignment of the leading edges isconsistent with zero-order desorption kinetics from all layers. In contrast, for water,the first and second layers are not resolved. At low water coverages (<1 monolayer(ML)) the initial desorption leading edges are aligned but then fall out of alignment athigher temperatures. For thicker water layers (10−100 ML), the desorption leadingedges are in alignment throughout the desorption of the film. The coverage dependence of the desorption behavoir suggests thatat low water coverages the nonalignment of the desorption leading edges is due to water dewetting from the graphene substrate.Kinetic simulations reveal that the experimental results are consistent with zero-order desorption. The simulations also show thatfractional order desorption kinetics would be readily apparent in the experimental TPD spectra.

I. INTRODUCTION

The desorption rates of astrophysically relevant molecules suchas methanol, ethanol, and water are needed to understand andquantify the evaporation behavior of interplanetary ices andcomets.1−8 Typically, the desorption kinetics are determined inthe laboratory at temperatures where they can be measured in areasonable amount of time (e.g., less than a day). Theseparameters are extrapolated to predict desorption rates at lowertemperatures, which are usually more relevant to astrophysicalprocesses. Dust grains, which are the core substrate forastrophysical ices, are believed to be composed of mostlycarbonaceous and silicaceous materials.5,8 As such, manylaboratory studies use carbon based surfaces such as amorphouscarbon or highly oriented pyrolytic graphite (HOPG) asanalogs for interstellar dust grains.5,8−11

In a series of papers, the desorption of methanol,12 ethanol,13

water,14 and ammonia15 from an HOPG substrate was studied.In that work, the multilayers of these species were observed todesorb with fractional-order desorption kinetics. For example,multilayer methanol, ethanol, water, and ammonia werereported to have desorption orders of 0.35, 0.08, 0.26, and0.25, respectively.6,8 These results are in contrast to the zero-order desorption kinetics expected for multilayer desorption.The desorption order will affect the overall desorption ratekinetics and thus is an important factor for modeling thedesorption behavior of astrophysical ices.2−6,8

In this paper we study the desorption kinetics of methanol,ethanol, and water deposited on a layer of graphene grown onPt(111). Molecular beams are used to deposit well-calibrateddoses onto the graphene substrate at 25 K and the desorptionkinetics are determined using temperature programmeddesorption (TPD). Both methanol and ethanol have well-resolved first, second, third, and multilayer layer desorptionpeaks. The alignment of the desorption leading edges isconsistent with zero-order desorption kinetics from all layers.

For water at low coverages (<1 monolayer (ML)) the initialdesorption leading edges are aligned but then fall out ofalignment at higher temperatures. In addition, the first andsecond layers are not resolved. For thicker water layers (10−100 ML), the desorption leading edges are in alignmentthroughout the desorption of the film. These observations areconsistent with water dewetting from the graphene substrate atlow water coverages and zero-order desorption at highercoverages. Kinetic simulations reveal that the experimentalresults are consistent with zero-order desorption. Thesimulations also show that fractional-order desorption kineticswould be readily apparent in the experimental TPD spectra.The zero-order desorption observed here is in contrast to thefractional-order desorption kinetics reported for desorption onHOPG (graphite). The conflicting results may be due todifferences in the substrate (HOPG versus graphene) or thefilm dosing technique (molecular beam versus backgrounddosing).

II. EXPERIMENTAL SECTION

All experiments were conducted utilizing an ultrahigh vacuumsystem (UHV) with a base pressure of <1 × 10−10 Torr, whichhas been previously described in more detail.16,17 The substratewas a graphene coated 1 cm diameter Pt(111) substrate spot-welded to tantalum leads. The substrate was cooled by a closedcycle helium cryostat to a base temperature of ∼25 K, whichcan be resistively heated to 1200 K. A K-type thermocouplespot-welded to the back of the Pt(111) substrate was used tomeasure temperature with a precision of better than ±0.01 K.

Special Issue: A. W. Castleman, Jr. Festschrift

Received: January 29, 2014Revised: March 20, 2014Published: March 21, 2014

Article

pubs.acs.org/JPCA

© 2014 American Chemical Society 8242 dx.doi.org/10.1021/jp501038z | J. Phys. Chem. A 2014, 118, 8242−8250

pubs.acs.org/JPCA

The absolute temperature was calibrated to an accuracy ofbetter than ±2 K utilizing the desorption of Ar, Kr, and H2Omultilayers. The Pt(111) substrate was cleaned by Ne+ ionsputtering at 1.5 kV followed by O2 exposure and subsequentUHV annealing at 1000 K. Graphene was then grown byheating the Pt(111) substrate to 1100 K and exposing it to amolecular beam of decane.18,19 This procedure produces a well-ordered, single layer of carbon with the structure of graphite.The details of the characterization of the film by LEED aredescribed in detail elsewhere.19 A well-ordered graphene filmthat covers the entire surface (no exposed Pt) has a Kr TPDthat has a well-defined monolayer desorption peak with a zero-order line shape. A Kr TPD check was performed regularly toconfirm the integrity of the graphene film. The integrity of thegraphene film was not affected by repeated dose and desorptioncycles.A quasi-effusive molecular beam at normal incidence was

utilized to deposit the films by expanding 1.0 Torr of methanol,1.0 Torr of ethanol, or 2.0 Torr of water through a 1 mmdiameter orifice. The beam was collimated by three stages ofdifferential pumping before impinging on the substrate at a rateof 0.44 ML/s (methanol), 0.5 ML/s (ethanol), or 0.87 ML/s(water) with a beam diameter slightly larger than the 1 cmdiameter of the Pt(111) substrate. For methanol and ethanol, 1ML was defined using the time required to saturate the secondlayer. The saturation of the second layer was used because itsareal density is more likely to be representative of the bulk thanthe first layer because of structural ordering that can occur onthe substrate. The approximate number of molecules/cm2 wasobtained by converting the densities of amorphous methanol(0.984 g/cm3)20 and ethanol (0.958 g/cm3)21 to the number ofmolecules per cm3 and taking the 2/3 root. Using thisprocedure, 1 ML for methanol corresponds to ∼7.0 × 1014

molecules/cm2 and for ethanol 1 ML is ∼5.4 × 1014 molecules/cm2. For water, 1 ML is defined as the monolayer saturationcoverage on the Pt(111) substrate and corresponds to ∼1.1 ×1015 molecules/cm2.22

TPD spectra were obtained using a linear heating rate andutilizing an Extrel quadrapole mass spectrometer in a line-of-sight configuration. Methanol and ethanol desorption weremonitored at m/z = 31 and for water m/z = 18. The intensityof the mass spectrometer signal was converted into an absolutedesorption rate in ML/s utilizing the respective ML definitionsdescribed above.

III. RESULTS AND DISCUSSIONA. Methanol Desorption Kinetics from Graphene.

Figure 1 displays TPD spectra for methanol films deposited at25 K on graphene and heated at 0.6 K/s. The top panel, Figure1a, contains desorption spectra with coverages from 0.1 to 0.8ML which is the saturation coverage of the first layer. Theleading edge for all of the desorption spectra are aligned, whichis a signature of zero-order desorption kinetics. Zero-orderdesorption for submonolayer coverages has been observedbefore and is attributed to the formation of a two-dimensionalequilibrium between individual absorbates (gas phase) andislands (condensed phase).23 The two-dimensional, two-phasecoexistence thus uniquely determines the chemical potential ofthe system and its temperature dependent vapor pressure(desorption rate). If the desorption rate is slow relative to thekinetic processes (e.g., surface diffusion) required to maintain atwo-dimensional surface equilibrium, then zero-order kineticsshould persist for the entire temperature and coverage range of

the desorption process. Therefore, the desorption rate is onlydefined by the temperature and not the coverage. This resultsin the alignment of the desorption leading edges (i.e., zero-order desorption kinetics).The bottom panel, Figure 1b, displays methanol desorption

spectra from both the first layer (blue) and the second layer(0.9−1.7 ML, red). The TPD leading edges for desorptionfrom the second layer are aligned and give rise to a peak that iswell-resolved from the first layer peak. Following desorption ofthe second layer, continued desorption aligns with the first layerdesorption curves. This shows that methanol desorptionproceeds by a layer-by-layer desorption mechanism.Figure 2 displays methanol TPD spectra from 0.1 to 5.0 ML

plotted on a log scale for clarity. The spectra show thealignment of the leading edges for desorption from the first(0.1−0.8 ML, blue lines), second (0.9−1.8 ML, red lines), andthird (1.8−2.5 ML, black lines) layers. Above the third layer,the desorption for coverages of 3.0−5.0 ML all align on thesame curve and we designate this as multilayer desorption. Theresults clearly show that the desorption process is layer-by-layerand that the desorption kinetics are zero-order. Figure 3 is anArrhenius plot of the TPD spectra from Figure 2 for thesaturation coverages of the first (blue), second (red), third(black), and fifth (green) layers. The leading edges for all thespectra are well-fit by an Arrhenius function (dashed lines), and

Figure 1. TPD spectra for methanol films deposited on a graphenecovered Pt(111) substrate at 25 K and heated at 0.6 K/s. (a) TPDspectra (blue) for methanol desorption from the first layer. The firstlayer coverages are 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, and 0.8 ML. Formethanol, 1 ML is defined as 7.0 × 1014 molecules/cm2 (see text fordetails). (b) Both first (blue) and second layer (red) methanol TPDspectra from graphene. The second layer coverages are 0.9, 1.0, 1.1,1.2, 1.3, 1.4, 1.5, 1.6, and 1.7 ML.

The Journal of Physical Chemistry A Article

dx.doi.org/10.1021/jp501038z | J. Phys. Chem. A 2014, 118, 8242−82508243

the fit parameters are given in Table 1. Note that due tocompensation effects, equally good fits can be obtained for arange of prefactor and activation energy combinations. Toestimate the error for an individual parameter, one parameterwas held fixed and the other was varied. For this and allsubsequent Arrhenius fits, we estimate an error of a factor of 2for ν and ±5% for Ea.The straight lines in the Arrhenius plot indicate zero-order

kinetics and that there is no change in phase (crystallization) ofthe methanol film during desorption. In previous work usingboth infrared and gas permeation techniques,24−28 we haveconfirmed pure methanol films heated at less than 1 K/s,crystallize at temperatures less than 120 K. This is consistentwith the multilayer desorption being from the crystalline phaseand a straight line in the Arrhenius plot (i.e., the lack of a phasetransition).

B. Ethanol Desorption Kinetics from Graphene. Figure4 displays TPD spectra for ethanol films deposited on grapheneand heated at a rate of 0.6 K/s. The TPD spectra for desorptionfrom the first layer (0.1−0.8 ML) are shown in Figure 4a andall align on a common leading edge. Similarly, Figure 4b showsthat the leading edges are completely aligned for desorptionfrom the second layer (0.9−1.8 ML, red) and that, followingthe desorption of the second layer, the leading edges fordesorption of the first layer are also aligned. These observationsare consistent with zero-order desorption from both layers.Figure 5 displays ethanol TPD spectra for initial coverages

from 0.1 to 5.0 ML plotted on a log scale. The spectra areconsistent with layer-by-layer desorption from the first (blue),second (red), and third layers (black). For coverages greaterthan 3 ML, the spectra (4.0−5.0 ML, green) are initially alignedwith desorption from the third layer, but at higher temperaturesthe rate decreases below that of the third layer. This is moreapparent in Figure 6, which displays an Arrhenius plot of theTPD for desorption from the saturated first (blue), second(red), third (black), and fifth (green) layers. The curves fordesorption from the first and second layers are straight,indicating no phase changes during desorption. In contrast,both the curves for desorption from the third and fifth layerschange slope during desorption. The dashed lines are Arrheniusfits to the desorption curves, and the fit parameters are given inTable 1. For the Arrhenius fit to the third layer desorption, thefit was confined to desorption prior to the change in slope andfor the fifth layer the fit was confined to desorption after thechange in the initial slope. A lower desorption rate (dashedgreen line) than expected from an extrapolation of an Arrheniusfit to the initial desorption rate (dashed black line) is consistentwith crystallization of an initially amorphous film. Vapordeposition of ethanol (methanol, water, and many others) atlow temperature creates films that are initially amor-phous.24,25,29,30 When heated, the amorphous films willcrystallize to the lower free energy crystalline phase, whichresults in a lower desorption rate.

Figure 2. Log plot of TPD spectra for methanol films on a graphenecovered Pt(111) substrate. The first (blue) and second (red) layerTPD spectra are the same as in Figure 1. The third layer (black)coverages are 1.8, 1.9, 2.0, and 2.5 ML and the multilayer coverages(green) are 3.0, 4.0, 4.5, and 5.0 ML.

Figure 3. Arrhenius plot of methanol TPD spectra from a graphenecovered Pt(111) substrate. TPD spectra for desorption from the first(0.8 ML, blue), second (1.7 ML, red), third (2.5 ML, black), and fifth(5 ML, green) layers are displayed. The corresponding dashed linesare Arrhenius fits and the parameters are given in Table 1.

Table 1. Arrhenius Fit Parameters

Arrhenius parametersa

absorbate figure layerheating rateβ (K/s) ν (ML/s)

Ea(kJ/mol)

methanol 3 first 0.6 3.2 × 10+15 45.43 second 0.6 1.1 × 10+16 44.43 third 0.6 9.8 × 10+15 43.63 multilayer 0.6 3.1 × 10+16 44.6

ethanol 6 first 0.6 4.5 × 10+15 52.06 second 0.6 6.7 × 10+16 50.66 third 0.6 7.5 × 10+16 48.66 fifth 0.6 5.7 × 10+15 45.97 first 0.1 1.5 × 10+16 53.67 second 0.1 5.6 × 10+16 50.47 third 0.1 1.1 × 10+17 49.17 multilayer 0.1 2.2 × 10+17 50.8

water 10 multilayeramorphous

0.6 1.5 × 10+15 46.4

10 multilayercrystalline

0.6 2.4 × 10+15 47.7

aDue to compensation effects, equally good fits can be obtained for arange of any individual Arrhenius fit parameter. To estimate theindividual parameter error, one parameter was held fixed and the otherwas varied. Typical errors were a factor of 2 for ν and ±5% for Ea.



To test that crystallization was occurring in the thickerethanol layers, experiments were conducted at a heating rate of0.1 K/s. The lower heating rate was used to change the time the

films spent at lower temperatures and thus change the relativedesorption and crystallization rates. Figure 7 is an Arrhenius

plot of TPD spectra for desorption from the first (blue), second(red), third (black), fifth (green), and fiftieth (light blue) layersof ethanol heated at 0.1 K/s. As was observed for the 0.6 K/sexperiments (Figure 6), the first and second layers for the 0.1K/s experiments (Figure 7) are straight indicating no phasechanges during desorption. In contrast to the data in Figure 6,at this slower heating rate desorption from the third layer(black) curve is relatively straight until desorption from thethird layer is complete. This likely means that desorption fromthe third layer was complete prior to the onset of crystallization.The desorption curve for the fifth layer (green) has a couple

of changes in slope until it eventually aligns with the second

Figure 4. TPD spectra for ethanol films deposited on a graphenecovered Pt(111) substrate at 25 K and heated at 0.6 K/s. (a) TPDspectra (blue) for desorption from the first layer. The first layercoverages are 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, and 0.8 ML. For ethanol, 1ML is defined as 5.4 × 1014 molecules/cm2 (see text for details). (b)Both first (blue) and second layer (red) ethanol TPD spectra fromgraphene. The second layer coverages are 0.9, 1.0, 1.1, 1.2, 1.3, 1.4, 1.5,1.6, 1.7, and 1.8 ML.

Figure 5. Log plot of TPD spectra for ethanol films on a graphenecovered Pt(111) substrate. The first (blue) and second (red) layerTPD spectra are the same as in Figure 4. The third layer (black)coverages are 1.9, 2.0, 2.5, and 3.0 ML and the multilayer coverages(green) are 4.0, 4.5, 5.0 ML.

Figure 6. Arrhenius plot of ethanol TPD spectra from a graphenecovered Pt(111) substrate for films deposited at 25 K and heated at 0.6K/s. TPD spectra for desorption from the first (0.8 ML, blue), second(1.8 ML, red), third (2.8 ML, black), and fifth (5.0 ML, green) layersare displayed. The corresponding dashed lines are Arrhenius fits andthe parameters are given in Table 1.

Figure 7. Arrhenius plot of ethanol TPD spectra from a graphenecovered Pt(111) substrate for films deposited at 25 K and heated at 0.1K/s. TPD spectra for desorption from the first (0.8 ML, blue), second(1.8 ML, red), third (2.8 ML, black), fifth (5.0 ML, green), and fiftieth(50 ML, light blue) layers are displayed. The corresponding dashedlines are Arrhenius fits and the parameters are given in Table 1.



and first layer desorption curves. Also note that the desorptionrate at a given temperature from the fifth layer is less than thatof the third layer, indicating that it has a lower free energy. Thelower free energy for the thicker layer is an indicator of a morestable structure or crystalline phase. The desorption spectrumfor a 50 ML thick film is displayed in Figure 7 and isrepresentative of the TPD spectra for thicker films (10−50ML). The TPD spectra from films with ethanol layerthicknesses greater than 10 ML (not shown) have leadingedges that are all aligned on a straight-line curve. The lowerdesorption rate and the straight line desorption curve meansthat the thicker ethanol films have likely crystallized prior to theonset of significant desorption. This is consistent with ourprevious work.24−26 The earlier crystallization of thicker films(>10 ML) may be due to an increase in the probability offinding a nucleation center or the possibility of forming a morestable extended structure. The crystallization kinetics for bothethanol and methanol are complicated with both havingmultiple crystalline phases21,31−33 and an in-depth study ofthe crystallization kinetics is beyond the scope of the presentstudy. The main point is that above 10 ML, the desorptionkinetics for multilayer ethanol films are zero-order, which isconfirmed by the straight line for the 50 ML film in Figure 7and the alignment of the leading edges for the TPD spectra forfilms from 10 to 50 ML (not shown). The dashed lines areArrhenius fits to desorption from the first, second, third, andfiftieth layers and the parameters are given in Table 1.C. Water Desorption Kinetics from Graphene. Figure 8

displays TPD spectra for water films on graphene that weredeposited at 25 K and heated at a rate of 0.6 K/s. Figure 8ashows TPD spectra for films with coverages from 0.2 to 1.0 ML(first layer, blue). The leading edges for the desorption spectraare initially aligned but at higher temperatures and coveragesthe TPD fall out of alignment. The desorption spectra from thesecond layer (Figure 8b, red) are also initially aligned but in thiscase, with the exception of the 1.2 ML spectrum, remainaligned throughout desorption. The second layer desorptionbehavior, however, cannot be represented by a singleexponential. Instead, there is an apparent “bump” in thedesorption rate between 145 and 150 K. Also note that, unlikethe observations for methanol and ethanol, the second layerdesorption peak is at higher temperature than the first layer andno first layer desorption peak can be resolved when desorbingmore than 1 ML. The “dip” in the desorption spectra and theshift to higher temperature suggest a transition to a more stablestructure. These results are consistent with previous workwhere it was shown that submonolayer water films on graphenedewet to form a metastable bilayer water structure.19 Thehigher desorption rates observed for submonolayer coverages(<1 ML, blue lines) in Figure 8b are due to desorption fromfilms that initially wet the graphene. At higher temperaturesthese films undergo a dewetting transition to the more stablebilayer structure. Similarly, for coverages above 1 ML (redlines), the “dip” in the desorption rate is due to the formationof the metastable bilayer. A similar 2D to 3D dewettingtransition has been observed on graphite.34

Figure 9 displays water TPD spectra for coverages from 0.2to 5.0 ML plotted on a log scale. The leading edge desorptionrate decreases with increasing coverage for desorption from thefirst (blue) and second (red) layers. Above 2 ML, the TPDspectra for desorption from the third (black) and higher layers(green) are very nearly aligned. These spectra also have anapparent inflection (“bump”) in the desorption rate between

140 and 145 K. This “bump” is due to the crystallization of theinitially amorphous water ice.29,30

Figure 8. TPD spectra for water films deposited on a graphenecovered Pt(111) substrate at 25 K and heated at 0.6 K/s. (a) TPDspectra (blue) for water desorption from the first layer. First layercoverages are 0.2, 0.4, 0.6, 0.8, and 1.0 ML. For water, 1 ML is definedas 1.1 × 1015 molecules/cm2 (see text for details). (b) Both first (blue)and second layer (red) water TPD spectra from graphene. The secondlayer coverages are 1.2,1.4, 1.6, 1.8, and 2.0 ML.

Figure 9. Log plot of TPD spectra for water films on a graphenecovered Pt(111) substrate. The first (blue) and second (red) layerTPD spectra are the same as in Figure 8. The third layer (black)coverages are 2.2, 2.4, 2.6, 2.8, and 3.0 ML and the multilayercoverages (green) are 3.2, 3.4, 3.6, 3.8, 4.0, 4.2, 4.4, 4.6, 4.8, and 5.0ML.



An Arrhenius plot of the 1, 3, and 5 ML TPD spectra fromFigure 9 and the TPD spectra from thicker films (10 to 50 ML)is displayed in Figure 10. The desorption spectra from thicker

films (50, 70, and 100 ML, green lines) are aligned on a straightline at low temperature and then transition to a lower rate andremain aligned on a second straight line. The two desorptionregimes, above and below the transition, are fit to an Arrheniusfunction (dashed lines) and the parameters are given in Table1. The desorption rate parameters are similar to thosepreviously obtained for amorphous (low temperature) andcrystalline (high temperature) water.35 The desorption spectrafor coverages less than 5 ML (blue lines) have leading edgedesorption rates that are well above those of the thicker films.For those films the higher desorption rates are likely due todesorption from a water film prior to dewetting and/or fromthe metastable bilayer. For intermediate thickness films, >5 ML(red lines), the leading edge desorption rates are more closelyaligned with the thicker film rates but show some curvatureafter crystallization. Thus, for relatively thin films (<25 ML)determination of the desorption kinetics is complicated bywater dewetting from the graphene substrate. For thicker films(>25 ML) the desorption kinetics are zero-order as issupported by the straight line desorption curves (before andafter crystallization) shown in Figure 10.D. Simulation of Fractional Order Desorption Kinetics.

The experimental TPD spectra presented in sections IIIA−Csuggest that the desorption kinetics for multilayers of methanol,ethanol, and water are zero-order. Zero-order desorption is alsoobserved for monolayer coverages of methanol and ethanol ongraphene. In recent studies,6,8 fractional desorption orders formethanol, ethanol, and water multilayers on an HOPG(graphite) substrate were reported (0.35,12 0.08,13 and 0.26,14

respectively). In this section we use kinetic simulations to testthe sensitivity of the TPD spectra to the desorption order.The desorption rate is described by the Polanyi−Wigner

equation,

θ θ ν θ θ− = θ−

tT T

dd

( , ) ( , )e E k T n( )/a B

(1)

where θ is coverage, T is temperature, Ea is the desorptionactivation energy, kB is the Boltzmann constant, ν is theprefactor, and n is the desorption order. In these simulations weassume that the prefactor, ν, and the desorption activationenergy, Ea, are coverage and temperature independent (i.e., theyare constant throughout desorption). Equation 1 was integratedusing a constant heating rate, β = dT/dt, to generate thesimulated TPD spectra.Figure 11 displays TPD spectra simulated using the

desorption parameters for multilayer methanol (ν = 3.1 ×

1016 and Ea = 44.57 kJ/mol, Table 1) and a heating rate of 0.6K/s. A desorption order of n = 0.25 was chosen so as to be inthe range of the fractional desorption orders reported formultilayer water, methanol, and ammonia.6,8 The simulations ofmultilayer (2−5.0 ML) desorption in Figure 11a clearly showthat the leading desorption edges are not aligned. An Arrheniusplot of the simulated spectra is displayed in Figure 11b and alsoshows the nonalignment of the leading edges, in addition to aslight curvature in the desorption curves. These are bothsignatures of non-zero-order desorption kinetics. Thus, adesorption order of 0.25 would be readily observable in ourexperimental system.To test the ultimate sensitivity of TPD experiments to the

desorption order, multilayer simulations were conducted with avalue of n = 0.05 and the results are displayed in Figure 12. In

Figure 10. Arrhenius plot of water TPD spectra from a graphenecovered Pt(111) substrate for films deposited at 25 K and heated at 0.6K/s. TPD spectra are for coverages of 1, 3, and 5 ML (blue); 10 and25 ML (red); and 50, 70, and 100 ML (green). The dashed lines arefits to desorption before (amorphous) and after (crystalline) the phasetransition. The Arrhenius fit parameters are given in Table 1.

Figure 11. Simulated TPD spectra using the Arrhenius parameters formultilayer methanol desorption (ν = 3.1 × 1016 ML0.75/s and Ea = 44.6kJ/mol, Table 1), a heating rate of 0.6 K/s, and a desorption order of n= 0.25. (a) Simulated TPD spectra with coverages of 2.0, 2.5, 3.0, 3.5,4.0, 4.5, and 5.0 ML. (b) Arrhenius plot of the TPD spectra in (a).



this case the leading edges are more closely aligned (Figure12a) but there is a slight, yet noticeable deviation fromalignment near the peak of desorption. The Arrhenius plot(Figure 12b) shows that the curves are very nearly aligned on astraight line with perhaps a small amount of curvature near thedesorption peak. Determination of a desorption order of n =0.05 would be difficult but may be possible with very carefulexperiments. On the basis of these simulations, distinguishingbetween zero-order and a fractional desorption order of lessthan n = 0.05 would be extremely difficult using desorptiontechniques.

IV. DISCUSSION AND CONCLUSIONSThe experimental results in section IIIA clearly show thatmethanol desorption from a graphene covered Pt(111)substrate occurs with zero-order desorption kinetics. This isthe case for both monolayer and multilayer coverages. Evidencefor zero-order desorption is given in Figures 1 and 2 where themethanol TPD spectra show the alignment of the desorptionleading edges for clearly resolved monolayer, second layer, thirdlayer, and multilayer desorption features. Further evidence isgiven in the Arrhenius plot in Figure 3 where the desorptionrate from each layer falls onto a straight line. As can be seen

from eq 1, this is expected behavior for coverage independentkinetics. An Arrhenius plot of the rate (not scaled by thecoverage) for a non-zero-order process would show curvature.The results for ethanol desorption from graphene covered

Pt(111) are similar to those of methanol. The alignment of thedesorption leading edges for the TPD spectra in Figure 4 andFigure 5 are consistent with zero-order desorption kinetics. TheArrhenius plots in Figures 6 and 7 show resolved monolayer,second layer, third layer, and multilayer desorption features.Desorption of films with coverages greater than 3 ML iscomplicated by the crystallization of the initially amorphousfilm. In this case the desorption kinetics from the crystallinemultilayer were determined in Figure 7 using a much thickerfilm (50 ML) and a slower heating rate (0.1 K/s). These resultsare consistent with the zero-order desoroption kinetics for boththe monolayer and multilayer coverages of ethanol ongraphene.The desorption of the water monolayer on graphene is

complicated by the dewetting that occurs on the graphenesubstrate during the desorption process.19,36 However, the TPDspectra for thicker water films are aligned (Figures 9 and 10)and this is indicative of zero-order desorption. As is well-knownfor water films deposited at low temperatures, the irreversibleamorphous to crystalline transition manifests itself as a “bump”in the TPD spectra. For thicker water films (>10 ML), thedesorption rate before and after the phase transition are well-described by an Arrhenius equation fit. The desorption rate ofthinner water films (<10 ML) is initially greater than anextrapolation of the amorphous desorption rate and this may bedue to desorption from the higher free energy metastablebilayer state or a pre-dewetted film.The desorption kinetics for methanol, ethanol, and water

from graphite have been previously reported.5,6,8,12−14 In thosestudies fractional desorption orders were observed for thedesorption of multilayer methanol (0.26), ethanol (0.08), andwater (0.35) and also for monolayer methanol (1.23).However, another laboratory has reported that the desorptionkinetics for methanol, ethanol, and water from a graphitesubstrate are zero-order.36 Our results are consistent with zero-desorption for methanol, ethanol, and water multilayers ongraphene. The simulations presented in Figures 11 and 12confirm that desorption experiments with the detectionsensitivity of our instrument should be able to distinguishdesorption orders as low as n = 0.05 and we see no evidence forfractional desorption kinetics.The differences between the zero-order kinetics observed

here and the fractional order kinetics reported elsewhere maybe due to several factors. The graphene and graphite substratesmay be slightly different. In our case the graphene layer film is asingle layer on the Pt(111) metal whereas the graphitesubstrate is created by cleaving a bulk sample and consists ofmultiple carbon layers. Although the outer carbon surface maybe similar in both cases, the interaction energy may be affectedby the subsurface layer (Pt for graphene versus another carbonlayer for graphite) and thus affect the desorption kinetics.Another factor could be the dosing method. In our case, themolecular beam dosing allows for uniform flux and precisecontrol of the coverage. On the other hand, the use ofbackground dosing makes it difficult to calibrate the exactcoverage. Finally, the detection sensitivity in the line-of-sightconfiguration of our instrument easily allows for the detectionof desorption rates as low as 10−4 ML/s and this can be usefulin determining the desorption order. All of these factors may

Figure 12. Simulated TPD spectra using the Arrhenius parameters formultilayer methanol desorption (ν = 3.1 × 1016 ML0.95/s and Ea = 44.6kJ/mol, Table 1), a heating rate of 0.6 K/s, and a desorption order of n= 0.05. (a) Simulated TPD spectra with coverages of 2.0, 2.5, 3.0, 3.5,4.0, 4.5, and 5.0 ML. (b) Arrhenius plot of the TPD spectra in (a).



account for the differences between the graphene and graphitein the literature results.

■ AUTHOR INFORMATIONCorresponding Authors*R. S. Smith: tel, (509) 371-6156; e-mial, [email protected].*B. D. Kay: tel, (509) 371-6143; e-mail, [email protected] Address†J. Matthiesen: Nano-Science Center, Department of Chem-istry, University of Copenhagen, Denmark.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the U.S. Department of Energy(DOE), Office of Basic Energy Sciences, Division of ChemicalSciences, Geosciences, and Biosciences. The research wasperformed using EMSL, a national scientific user facilitysponsored by DOE’s Office of Biological and EnvironmentalResearch and located at Pacific Northwest National Laboratory,which is operated by Battelle operated for the DOE underContract No. DE-AC05-76RL01830.

■ REFERENCES(1) Sandford, S. A.; Allamandola, L. J. Condensation and Vapor-ization Studies of CH3OH and NH3 Ices - Major Implications forAstrochemistry. Astrophys. J. 1993, 417, 815−825.(2) Fraser, H. J.; Collings, M. P.; McCoustra, M. R. S.; Williams, D.A. Thermal Desorption of Water Ice in the Interstellar Medium. Mon.Not. R. Astron. Soc. 2001, 327, 1165−1172.(3) Collings, M. P.; Anderson, M. A.; Chen, R.; Dever, J. W.; Viti, S.;Williams, D. A.; McCoustra, M. R. S. A Laboratory Survey of theThermal Desorption of Astrophysically Relevant Molecules. Mon. Not.R. Astron. Soc. 2004, 354, 1133−1140.(4) Viti, S.; Collings, M. P.; Dever, J. W.; McCoustra, M. R. S.;Williams, D. A. Evaporation of Ices near Massive Stars: Models Basedon Laboratory Temperature Programmed Desorption Data. Mon. Not.R. Astron. Soc. 2004, 354, 1141−1145.(5) Brown, W. A.; Bolina, A. S. Fundamental Data on the Desorptionof Pure Interstellar Ices. Mon. Not. R. Astron. Soc. 2007, 374, 1006−1014.(6) Williams, D. A.; Brown, W. A.; Price, S. D.; Rawlings, J. M. C.;Viti, S. Molecules, Ices and Astronomy. Astron. Geophys. 2007, 48, 25−34.(7) Thrower, J. D.; Burke, D. J.; Collings, M. P.; Dawes, A.; Holtom,P. D.; Jamme, F.; Kendall, P.; Brown, W. A.; Clark, I. P.; Fraser, H. J.;McCoustra, M. R. S.; Mason, N. J.; Parker, A. W. Desorption of HotMolecules from Photon Irradiated Interstellar Ices. Astrophys. J. 2008,673, 1233−1239.(8) Burke, D. J.; Brown, W. A. Ice in Space: Surface ScienceInvestigations of the Thermal Desorption of Model Interstellar Ices onDust Grain Analogue Surfaces. Phys. Chem. Chem. Phys. 2010, 12,5947−5969.(9) Katz, N.; Furman, I.; Biham, O.; Pirronello, V.; Vidali, G.Molecular Hydrogen Formation on Astrophysically Relevant Surfaces.Astrophys. J. 1999, 522, 305−312.(10) Pirronello, V.; Liu, C.; Roser, J. E.; Vidali, G. Measurements ofMolecular Hydrogen Formation on Carbonaceous Grains. Astron.Astrophys. 1999, 344, 681−686.(11) Perry, J. S. A.; Price, S. D. Detection of Rovibrationally ExcitedH2 Formed through the Heterogeneous Recombination of H Atomson a Cold HOPG Surface. Astrophys. Space Sci. 2003, 285, 769−776.(12) Bolina, A. S.; Wolff, A. J.; Brown, W. A. Reflection AbsorptionInfrared Spectroscopy and Temperature Programmed Desorption

Investigations of the Interaction of Methanol with a Graphite Surface.J. Chem. Phys. 2005, 122, 044713.(13) Burke, D. J.; Wolff, A. J.; Edridge, J. L.; Brown, W. A. TheAdsorption and Desorption of Ethanol Ices from a Model GrainSurface. J. Chem. Phys. 2008, 128, 104702.(14) Bolina, A. S.; Wolff, A. J.; Brown, W. A. Reflection AbsorptionInfrared Spectroscopy and Temperature-Programmed DesorptionStudies of the Adsorption and Desorption of Amorphous andCrystalline Water on a Graphite Surface. J. Phys. Chem. B 2005, 109,16836−16845.(15) Bolina, A. S.; Brown, W. A. Studies of Physisorbed AmmoniaOverlayers Adsorbed on Graphite. Surf. Sci. 2005, 598, 45−56.(16) Smith, R. S.; Zubkov, T.; Kay, B. D. The Effect of the IncidentCollision Energy on the Phase and Crystallization Kinetics of VaporDeposited Water Films. J. Chem. Phys. 2006, 124, 114710.(17) Zubkov, T.; Smith, R. S.; Engstrom, T. R.; Kay, B. D.Adsorption, Desorption, and Diffusion of Nitrogen in a ModelNanoporous Material. I. Surface Limited Desorption Kinetics inAmorphous Solid Water. J. Chem. Phys. 2007, 127, 184707.(18) Tait, S. L.; Dohnalek, Z.; Campbell, C. T.; Kay, B. D. N-Alkaneson Pt(111) and on C(0001)/Pt(111): Chain Length Dependence ofKinetic Desorption Parameters. J. Chem. Phys. 2006, 125, 234308.(19) Kimmel, G. A.; Matthiesen, J.; Baer, M.; Mundy, C. J.; Petrik, N.G.; Smith, R. S.; Dohnalek, Z.; Kay, B. D. No Confinement Needed:Observation of a Metastable Hydrophobic Wetting Two-Layer Ice onGraphene. J. Am. Chem. Soc. 2009, 131, 12838−12844.(20) Steytler, D. C.; Dore, J. C.; Montague, D. C. Neutron-Diffraction Studies of Amorphous Methyl-Alcohol. J. Non-Cryst. Solids1985, 74, 303−312.(21) Srinivasan, A.; Bermejo, F. J.; deAndres, A.; Dawidowski, J.;Zuniga, J.; Criado, A. Evidence for a Supercooled Plastic-Crystal Phasein Solid Ethanol. Phys. Rev. B 1996, 53, 8172−8175.(22) Kimmel, G. A.; Stevenson, K. P.; Dohnalek, Z.; Smith, R. S.;Kay, B. D. Control of Amorphous Solid Water Morphology UsingMolecular Beams. I. Experimental Results. J. Chem. Phys. 2001, 114,5284−5294.(23) Daschbach, J. L.; Peden, B. M.; Smith, R. S.; Kay, B. D.Adsorption, Desorption, and Clustering of H2O on Pt(111). J. Chem.Phys. 2004, 120, 1516−1523.(24) Ayotte, P.; Smith, R. S.; Teeter, G.; Dohnalek, Z.; Kimmel, G.A.; Kay, B. D. A Beaker without Walls: Formation of DeeplySupercooled Binary Liquid Solutions of Alcohols from NanoscaleAmorphous Solid Films. Phys. Rev. Lett. 2002, 88, 245505.(25) Smith, R. S.; Ayotte, P.; Kay, B. D. Formation of SupercooledLiquid Solutions from Nanoscale Amorphous Solid Films of Methanoland Ethanol. J. Chem. Phys. 2007, 127, 244705.(26) Matthiesen, J.; Smith, R. S.; Kay, B. D. Mixing It Up: MeasuringDiffusion in Supercooled Liquid Solutions of Methanol and Ethanol atTemperatures near the Glass Transition. J. Phys. Chem. Lett. 2011, 2,557−561.(27) Smith, R. S.; Matthiesen, J.; Kay, B. D. Breaking through theGlass Ceiling: The Correlation between the Self-Diffusivity in andKrypton Permeation through Deeply Supercooled Liquid NanoscaleMethanol Films. J. Chem. Phys. 2010, 132, 124502.(28) Matthiesen, J.; Smith, R. S.; Kay, B. D. Using Rare GasPermeation to Probe Methanol Diffusion near the Glass TransitionTemperature. Phys. Rev. Lett. 2009, 103, 245902.(29) Smith, R. S.; Matthiesen, J.; Knox, J.; Kay, B. D. CrystallizationKinetics and Excess Free Energy of H2O and D2O Nanoscale Films ofAmorphous Solid Water. J. Phys. Chem. A 2011, 115, 5908−5917.(30) Smith, R. S.; Petrik, N. G.; Kimmel, G. A.; Kay, B. D. Thermaland Nonthermal Physiochemical Processes in Nanoscale Films ofAmorphous Solid Water. Acc. Chem. Res. 2012, 45, 33−42.(31) Torrie, B. H.; Binbrek, O. S.; Strauss, M.; Swainson, I. P. PhaseTransitions in Solid Methanol. J. Solid State Chem. 2002, 166, 415−420.(32) Galvez, O.; Mate, B.; Martin-Llorente, B.; Herrero, V. J.;Escribano, R. Phases of Solid Methanol. J. Phys. Chem. A 2009, 113,3321−3329.



mailto:[email protected]




(33) Ramos, M. A.; Shmyt’ko, I. M.; Arnautova, E. A.; Jimenez-Rioboo, R. J.; Rodriguez-Mora, V.; Vieira, S.; Capitan, M. J. On thePhase Diagram of Polymorphic Ethanol: Thermodynamic andStructural Studies. J. Non-Cryst. Solids 2006, 352, 4769−4775.(34) Chakarov, D. V.; Osterlund, L.; Kasemo, B. Water-Adsorptionand Coadsorption with Potassium on Graphite(0001). Langmuir 1995,11, 1201−1214.(35) Speedy, R. J.; Debenedetti, P. G.; Smith, R. S.; Huang, C.; Kay,B. D. The Evaporation Rate, Free Energy, and Entropy of AmorphousWater at 150 K. J. Chem. Phys. 1996, 105, 240−244.(36) Ulbricht, H.; Zacharia, R.; Cindir, N.; Hertel, T. ThermalDesorption of Gases and Solvents from Graphite and CarbonNanotube Surfaces. Carbon 2006, 44, 2931−2942.



desorption kinetics of methanol, ethanol, and water from graphene

Documents