Influence of Hydrogen Bonding on the Surface Diffusion of Molecular

Jun 26, 2017 - (31) Because they are related by substitution of a single functional group that alters the degree of hydrogen bonding, these three mole...
1 downloads 13 Views 2MB Size
Subscriber access provided by University of South Florida

Article

Influence of Hydrogen Bonding on the Surface Diffusion of Molecular Glasses: Comparison of Three Triazines Yinshan Chen, Men Zhu, Audrey Laventure, Olivier Lebel, Mark D. Ediger, and Lian Yu J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.7b05333 • Publication Date (Web): 26 Jun 2017 Downloaded from http://pubs.acs.org on July 5, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Influence of Hydrogen Bonding on the Surface Diffusion of Molecular Glasses: Comparison of Three Triazines Yinshan Chen†, Men Zhu‡, Audrey Laventure§, Olivier Lebel§, M. D. Ediger‡, Lian Yu*†‡ † School of Pharmacy, University of Wisconsin-Madison, Madison, WI, 53705, USA ‡ Department of Chemistry, University of Wisconsin-Madison, Madison, WI 53706, USA § Department of Chemistry and Chemical Engineering, Royal Military College of Canada, Kingston, ON, K7K 7B4, Canada

ABSTRACT Surface grating decay measurements have been performed on three closely related molecular glasses to study the effect of intermolecular hydrogen bonds on surface diffusion. The three molecules are derivatives of bis(3,5-dimethyl-phenylamino)-1,3,5-triazine and differ only in the functional group R at the 2-position, with R being -C2H5, -OCH3, and -NHCH3, and referred to as “Et”, “OMe”, and “NHMe”, respectively. Of the three molecules, NHMe forms more extensive intermolecular hydrogen bonds than Et and OMe and was found to have slower surface diffusion. For Et and OMe, surface diffusion is so fast that it replaces viscous flow as the mechanism of surface grating decay as temperature is lowered. In contrast, no such transition was observed for NHMe under the same conditions, indicating significantly slower surface diffusion. This result is consistent with the previous finding that extensive intermolecular hydrogen bonds slow down surface diffusion in molecular glasses and is attributed to the persistence of hydrogen bonds even in the surface environment. This result is also consistent with the lower stability of the vapor deposited glass of NHMe relative to Et and OMe and supports the view that surface mobility controls the stability of vapor-deposited glasses. 1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

INTRODUCTION Glasses are amorphous solids prepared by cooling liquids, condensing vapors, and evaporating solutions

while

avoiding

crystallization.

Glasses

are

out-of-equilibrium

materials,

thermodynamically driven to crystallize and to age toward the equilibrium liquid state. This nature of glasses makes their physical stability an important issue in developing amorphous materials, since crystallization and aging lead to changes in materials properties. It has recently been recognized that surface mobility plays an important role in the physical stability of glasses. Because surface molecules are more mobile than bulk molecules, crystallization can be substantially faster on the surface of a glass than in its interior.1,2 Because surface molecules can equilibrate faster than bulk molecules, it is possible to use vapor deposition to prepare highdensity, high-kinetic stability glasses that have properties expected for liquid-cooled glasses that have been aged for millions of years.3,4 These results indicate that surface mobility has a central role in understanding and controlling the physical stability of glasses. There has been recent progress in measuring the surface diffusion coefficients Ds of amorphous solids. Surface diffusion refers to the lateral translation of molecules or atoms. Recent studies have measured the Ds of amorphous silicon,5,6 molecular glasses,7,8,9,10,11,12 and metallic glasses. 13 In addition, surface diffusion in glass-forming liquids has been studied by computer simulations 14,15,16,17 and theories. 18,19 A remarkable finding of these studies is that surface diffusion can be vastly faster than bulk diffusion, by up to 8 orders of magnitude at the laboratory glass transition temperature Tg, consistent with the observations of fast surface crystallization and formation of stable glasses by vapor deposition.

2 ACS Paragon Plus Environment

Page 3 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Another intriguing result from recent work is that surface diffusion in molecular glasses slows down significantly in the presence of extensive intermolecular hydrogen bonds.20 Prevalent in organic materials, hydrogen bonds are stronger and more directional than van der Waals forces. For the molecular glasses whose surface diffusion

has

been

measured

to

date

(Scheme 1), there is a significant variation in the extent of hydrogen bonding, ranging from none (OTP and TNB) to limited (IMC and NIF) to extensive (polyalcohols). In the same order, the rate of surface diffusion is observed to systematically decrease, by at least 4 orders of magnitude when compared at Tg.20 This large change of surface diffusion rate is in contrast to the relatively constant bulk diffusion rate at Tg, with Dv ~ 10−20 m2/s. 21,22,23,24 This effect of hydrogen bonding on surface mobility has been attributed to the “robustness” of hydrogen bonds; that is, the number of hydrogen

Scheme 1. Molecular structures of the glasses whose surface diffusion has been measured. In the box are the systems of this study (Et, OMe, NHMe). IMC: indomethacin; NIF: nifedipine; OTP: ortho-terphenyl; TNB: tris-naphthyl benzene.

bonds per molecule is largely unchanged on going from the bulk to the surface environment.20,25,26,27,28 This presumably preserves a bulk-like kinetic barrier for diffusion in the surface environment. The effect of hydrogen bonding on surface diffusion has also been suggested by computer simulations. Consistent the comments

3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 27

above, surface diffusion is significantly slower in water than in the Lennard-Jones liquid parametrized for Ar when compared at the same bulk diffusivity.16,29 Although the existing results are informative on the effect of hydrogen bonding on surface diffusion, it is desirable to study the problem at a higher resolution. This is the goal of the present study, where we measure the surface diffusion in three closely related glass-forming triazines that differ in the extent of hydrogen bonding (Scheme 1, molecules in the box). These molecules were originally synthesized to study the effect of molecular structure on glass-forming ability.30 The 3 molecules share the same bis(3,5-dimethyl-phenylamino)-1,3,5-triazine core but different functional groups (Et, OMe, and NHMe) at the 2-position of the triazine ring. In these molecular glasses, hydrogen bonding occurs through the NH group and the NHMe derivative can form more intermolecular hydrogen bonds per molecule than Et and OMe, by approximately a factor of 1.6. 31 Because they are related by substitution of a single atom that alters the degree of hydrogen bonding, these three molecules provide a sensitive test of the effect of hydrogen bonding on surface diffusion. As yet another motivation for this study, we note that vapor deposition of the three triazines produces glasses of significantly different stability – measured by their density and resistance to thermal transformation, with the NHMe glass being less stable than the Et and OMe glasses.31 Given that surface mobility enables the formation of stable glasses by vapor deposition, this result suggests a lower surface mobility in NHMe than in Et and OMe, a conclusion to be tested in the present study. We report that of the 3 triazine glasses, NHMe has much slower surface diffusion than Et and OMe, by at least one order of magnitude. This result is consistent with and adds precision to the previous conclusion that extensive intermolecular hydrogen bonding slows down surface

4 ACS Paragon Plus Environment

Page 5 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

diffusion in molecular glasses. Our result is also consistent with the lower stability of the vapordeposited glasses of NHMe31 and supports the view that surface mobility is a primary factor enabling the formation of stable glasses by vapor deposition.

MATERIALS AND METHODS Et, OMe and NHMe were synthesized in the Lebel group according to published procedures.30 Differential Scanning Calorimetry (DSC) was performed with a TA Instruments Q2000 unit. The glass transition temperature (Tg) was measured during a heating run at 10 K/min after the sample was cooled at 10 K/min to 40 K below Tg. The Tg values (onset temperatures) are: 314 K (Et), 330 K (OMe), and 360 K (NHMe). To print a surface grating, a master was placed on a supercooled liquid at ~ Tg + 45 K, and removed below its Tg. A Linkam THMS 600E stage was used to control temperature, which was purged with dry nitrogen and housed inside a glove bag, also purged with nitrogen. The masters were purchased from Rainbow Symphony (1000 and 2000 nm), separated from data storage discs (DVD for 740 nm and Blu-ray for 330 nm), or replicated from glass gratings through a Norland UV-curing optical adhesive (420, 3300 and 8200 nm). The masters were gold-coated to minimize the transfer of contaminants. The thickness of each triazine glass film was 50-100 µm, much larger than the wavelength of the surface gratings, ensuring that the evolution of the top surface was unaffected by the film substrate. The smoothing of surface gratings over time was monitored by atomic force microscopy (AFM, Bruker Veeco Mutiple Mode IV) or laser diffraction. Both types of experiments were performed in flowing nitrogen at a constant temperature maintained with the Linkam stage and data collection began after temperature equilibration. For AFM, the sample was scanned in the tapping mode at the ambient temperature in flowing nitrogen. The height profile was Fourier 5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 27

transformed to obtain the amplitude of a sinusoidal grating or in the case of a sawtooth profile, the amplitude of the first harmonic. For laser diffraction, a HeNe laser (λ = 632.8 nm, Uniphase Corp.) was passed through the sample and the first-order diffraction in transmission was recorded with a silicon amplified detector (Thorlabs) interfacing with a National Instruments LabVIEW program. The grating amplitude was verified to be proportional to the square root of the diffraction intensity. The sample surface temperature was checked against the melting points of fine crystals sprinkled onto the sample and a small correction (up to 3 K) was made if necessary. This temperature difference between the top surface of the sample and the metal block of the hot stage resulted from the lower ambient temperature and the poor thermal conductivity of the organic glass. Viscosity was measured with a TA Advanced Rheometric Expansion Systems (ARES). This is a parallel-plate rheometer with a plate diameter of 8 mm. Each sample was mounted at Tg + 90 K and held at Tg + 70 K for 10 min to ensure complete melting and remove bubbles. The gap between the plates was adjusted so that the sample occupied the entire space between the plates (typically between 1.0 and 1.2 mm). After temperature equilibration, viscosity was measured over a range of frequencies.

6 ACS Paragon Plus Environment

Page 7 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

RESULTS Figure 1 shows the typical surface grating decay curves for the triazines. I/I0 is the normalized first-order diffraction intensity. The decay at each temperature is fitted with an exponential function or a stretched exponential function, I/I0 = exp[-(KIt)β], with β close to 1. The grating-amplitude decay constant K is given by KI/2 because of the relation I ∝ h2. Figure 2a shows the grating decay constant K vs. temperature for the three systems with 1000 nm wavelength grating. For each system, the measured K covers 5 to 6 decades.

Figure 1. Typical decay kinetics for the surface gratings of Et, OMe, and NHMe at the wavelength of 1000 nm. I/I0 is normalized diffraction intensity. The curves are exponential or stretchedexponential fits.

According to Mullins,32 K for a sinusoidal grating of wavelength λ is given by K = Fq+Aq2+(A’+C)q3+Bq4

(1)

where q = 2π/λ F = γ/2η A = p0γΩ2/(2πm)1/2(kT)3/2 D = A’ + C = ρ0DGγΩ2/(kT) + DvγΩ/(kT) B = DsγΩ2ߥ/(kT) 7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 27

Here, γ is the surface energy, η the viscosity,

p0

the

equilibrium

vapor

pressure, Ω the molecular volume, m the molecular mass, ρ0 the equilibrium vapor density, DG the diffusion coefficient of evaporated atmosphere,

molecules Dv

the

in

the

bulk

inert

diffusion

coefficient, Ds the surface diffusion coefficient, and ߥ the number of molecules per unit area of surface. For our systems, the following estimates are made: γ = 0.05 N/m, Ω ≈ 4×10-28 m3. The terms in eq. (1) correspond to surface smoothing by viscous

flow

(F),

Figure 2. (a) Decay constant K vs. T at λ = 1000 nm for Et, OMe, and NHMe. The curves are the calculated Fq term after a vertical shift on the log scale by -1.25 for Et, -0.55 for OMe, and -0.2 for NHMe (b) K vs. grating wavelength λ at a fixed −1 temperature (shown). K ∝ λ indicates decay by −4 viscous flow; K ∝ λ indicates decay by surface diffusion.

evaporation-

condensation (A and A’), bulk diffusion (C), and surface diffusion (B). We find that the viscous-flow term accounts very well for the observed decay constant at high temperatures for all 3 systems and in the case of NHMe, the entire range of temperature studied. The solid curves in Figure 2a are the calculated K for the viscous-flow term: Fq = (γ/2η)∙(2π/λ). For this calculation, we used experimentally measured viscosities reported in Figure 3. The detailed measurement results of viscosity as a function of frequency can be found in Figure S1 in the Supporting Information. Notice that the calculated Fq term matches well the observed decay constant K for NHMe at all temperatures studied and for Et and OMe at higher temperatures.

8 ACS Paragon Plus Environment

Page 9 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

This agreement indicates that viscous flow is the mechanism of decay at the indicated temperatures. We note that to make the calculated Fq coincide with the observed K in Figure 2a, a small vertical shift C is applied to Fq on the log scale, with C = -1.25 for Et, -0.55 for OMe, and -0.2 for NHMe. This shift is likely a result of combined experimental errors in K, viscosity, and surface tension. The above assignment that viscous flow is responsible

for

high-temperature

decays

is

confirmed by the wavelength dependence of K. According to eq. (1), if viscous flow is the mechanism of decay, K is proportional to λ-1. As Figure 2b shows, this relation is indeed observed

Figure 3. Viscosities of Et, OMe and NHMe vs. temperature. The curves are the Vogel– Fulcher–Tammann (VTF) fits of the data.

at temperatures at which viscous flow is assigned the mechanism of decay.

9 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

To

further

evaluate

our

Page 10 of 27

mechanistic

assignment, we find that the evaporationcondensation

terms

make

negligible

contributions to the observed decay constant K. To calculate these terms, we used: DG = 0.1 cm2/s for organic molecules diffusing in ambient atmosphere; p0 = 1.84 × 10-9 Pa for Et, 4.56 × 10-9 Pa for OMe, and 1.24 × 10-9 Pa for NHMe

Figure 4. Surface diffusion coefficients (Ds) of Et, OMe and the upper bound of NHMe plotted against T.

(predicted values at 298 K by ACD)33. We find that the A and A’ terms are at least 100 times smaller than the observed K. Similarly, we evaluated the bulk diffusion term (C) in eq. (1) under the assumption that the triazines’ Dv values are similar to those of other organic glasses at the same Tg–scaled temperature.21,22,23 We find that the C term is at least 1000 times smaller than the observed K. We now return to the low-temperature decay of Et and OMe (Figure 2a), where the observed K is much larger than the Fq term (viscous-flow mechanism). As evaporation-condensation and volume diffusion have been ruled out as possible mechanisms, the only mechanism responsible for the fast decay observed is surface diffusion. We further verified this conclusion through the wavelength dependence of K. If surface diffusion is the mechanism of decay, K is proportional to λ-4, and this relation is indeed observed for Et and OMe at low temperatures (see the data presented in Figure 2b). Figure 4 shows the surface diffusion constants Ds of Et and OMe calculated from K = Bq4 = DsγΩ2ν/(kT). For these two systems, we obtain Ds ~ 10-14 m2/s at Tg. For NHMe, since no decay by surface diffusion was observed, only an upper bound could be calculated for Ds from Bq4