We have examined the effect of growth rate on the evolution of two polymorphs of thin films of tetracene on SiO2 using synchrotron X-ray radiation and molecular beam techniques. Ex situ X-ray reflectivity shows that tetracene forms two phases on SiO2: a thin-film phase and a bulk phase. We have used in situ, real-time grazing incidence diffraction during growth to reveal the nature of growth concerning these two phases. We observe that there is initially growth of only the thin-film phase, up to a thickness of several monolayers. This is followed by the nucleation of the bulk phase, growth of both phases, and finally growth of only the bulk phase. We find that the deposited thickness when the bulk phase nucleates increases with increasing growth rate. Similarly, we find that the deposited thickness at which the thin-film phase saturates also increases with increasing growth rate. These apparent dependencies on growth rate are actually a consequence of the local coverage, which depends on growth rate, particularly for the former effect. At low growth rates, there is 3D growth resulting from the upward transport of tetracene at island edges, resulting in tall features where molecules escape the influence of the substrate and form into the bulk phase. Increasing the growth rate leads to growth that is more 2D and uniform in coverage, delaying the formation of the bulk phase.

The growth of thin films of organic molecules for application in organic electronics has been a topic of great interest as researchers search for effective combinations of materials and explore the complex behaviors that arise from the breadth of these combinations.1–3 Organic molecules can assemble into crystalline thin films, but they interact via relatively weak van der Waals interactions compared to covalent/ionic/metallic bonds in inorganic systems. In addition, due to their often complex, low-symmetry shapes, thin films of organic molecules can form a variety of polymorphs,4,5 a process that can be strongly influenced by the nature of the substrate.6 Even relatively simple molecules have been shown to exhibit multiple polymorphs, such as perylenetetracarboxylic dianhydride (PTCDA),7,8 pentacene,9–12 perfluoropentacene,13 and diindenoperylene (DIP).14,15 These different polymorphs naturally have different properties, and one particular structure may be more desirable than another.11,16 Therefore, it is of critical importance to understand how these polymorphs arise during thin film growth to obtain the desired properties.

Tetracene is another example of a molecule that exhibits multiple phases at room temperature in thin-film form.17–21 Gompf et al. and Milita et al. have observed the formation of two phases and have reported their relative amounts as determined by ex situ, post-deposition techniques.17,18 Both studies found that lower growth rates produced more bulk phase while higher growth rates produced more thin-film phase. Lacking in these works was a determination of the evolution of the two phases and why might the growth rate affect the relative proportion of the two crystalline structures. Here, we seek to provide a better understanding of how and why these phases arise during thin film growth. We will present the results from our investigation on the formation of the thin-film phase and the bulk phase of tetracene on SiO2, where we use in situ real-time grazing incidence diffraction (GID) at Cornell High Energy Synchrotron Source (CHESS) to monitor the evolution of these crystalline phases. We will show that this technique allows us to determine when each of the phases begins to grow, which will also lend insight into the dynamics of the growth of thin films of tetracene.

Thin films of tetracene were grown on clean SiO2 using a supersonic molecular beam source in a custom-designed UHV chamber fitted with Be windows, detailed elsewhere,22 in the G3 station at CHESS. Briefly, a supersonic beam of tetracene is generated by passing He carrier gas over a heated vessel containing tetracene (99.99%, Sigma-Aldrich) and then expanding this gas through a 150 μm diameter nozzle into a source chamber. This beam then passes through a trumpet-shaped skimmer, a differentially pumped ante-chamber, and then a beam-defining aperture before striking the sample, which is at a nominal Ts = 30 °C, in the main scattering chamber (base pressure ∼5 × 10−9 Torr). We vary the growth rate by controlling the temperature of the heated vessel containing the tetracene. Further details regarding the estimated incident kinetic energy (2.5-2.6 eV) and sample preparation are given elsewhere.23 Growth rate was estimated by ex situ atomic force microscopy (AFM) by calculating the volume over a given area (the scan size) to find thickness.23 AFM was also used to measure RMS roughness.

During deposition, thin-film growth was monitored in situ with real-time GID. The X-ray beam was incident on the substrate at an angle of 0.13°, and the energy of the X-ray beam was 10.06 keV. The intensity of the scattered X-ray beam was measured continuously using a PILATUS 100K area detector (DECTRIS, Ltd.) with a dwell time of 1 s. In this manner, we collect scattering data throughout the duration of growth, as compared to approaches using ex situ analysis, which can only characterize the thin film post-deposition. Fitting of in-plane diffraction peaks was performed using an approximate 2D Voigt function (the product of a Gaussian and a Lorentzian function). Post-deposition X-ray reflectivity (XRR) and GID were measured ex situ in the G2 station at CHESS. Ex situ GID data were indexed using previously developed methods.24,25

In Figure 1, we display the ex situ XRR of a thin film of tetracene (∼160 ML) grown on SiO2 at a nominal rate of 0.47 ML s−1. Here, we observe two sets of peaks, indicating two distinct phases of tetracene. One phase displays peaks at qz = 0.487, 0.971, and 1.456 Å−1, giving an average dz-spacing of 12.93 ± 0.016 Å. The other phase displays peaks at qz = 0.516, 1.031, and 1.547 Å−1, giving an average dz-spacing of 12.19 ± 0.0025 Å. The latter matches well with the bulk-phase structure of tetracene reported by Campbell and co-workers (dz-spacing of 12.26 Å)26 and by Holmes and co-workers (dz-spacing of 12.10 Å).27 The former, with the larger dz-spacing is a thin-film phase, which was reported in a previous study of thin films of tetracene on SiO2 by Gompf and co-workers.17 Pentacene, a molecule very similar to tetracene, has also been observed to form both a thin-film phase and a bulk phase for a variety of growth conditions.10,12,28,29

FIG. 1.

Ex situ XRR of a ∼160 ML thin film of tetracene grown on SiO2 at a nominal rate of 0.47 ML s−1. The vertical lines (from left to right) indicate the expected positions from dz = 12.93 Å (a thin-film phase) and 12.19 Å (the bulk phase). Inset on the bottom left is the chemical structure of tetracene.

FIG. 1.

Ex situ XRR of a ∼160 ML thin film of tetracene grown on SiO2 at a nominal rate of 0.47 ML s−1. The vertical lines (from left to right) indicate the expected positions from dz = 12.93 Å (a thin-film phase) and 12.19 Å (the bulk phase). Inset on the bottom left is the chemical structure of tetracene.

Close modal

We display in Figure 2ex situ GID from the same thin film examined by XRR in Figure 1. We also show the results from indexing the two phases. The lattice parameters for the bulk phase, indicated by the red symbols and text, were refined from the parameters given by Holmes and co-workers.27 The lattice parameters for the thin-film phase, indicated by the green symbols and text, were iterated upon starting with a shift in parameters analogous to that seen for the bulk phase and thin-film phase of pentacene.29 The lattice parameters are summarized in Table I. Given that tetracene and pentacene possess very similar crystal structures, we unsurprisingly observe a similarity between these data and GID for thin films of pentacene.29–31 These data, collected ex situ, tell us nothing about time evolution of the two phases. Namely, it is not yet clear if these two phases begin forming at the same time, or does one phase grows after the other, and is this transition one that is smooth or abrupt? We can use in situ, real-time GID to interrogate the evolution of these two phases during thin-film growth.

FIG. 2.

Ex situ GID of a ∼160 ML thin film of tetracene grown on SiO2 at a nominal rate of 0.47 ML s−1, same as that shown in Figure 1.

FIG. 2.

Ex situ GID of a ∼160 ML thin film of tetracene grown on SiO2 at a nominal rate of 0.47 ML s−1, same as that shown in Figure 1.

Close modal
TABLE I.

Summary of the lattice parameters of the thin-film phase and bulk phase of both tetracene and pentacene.

Phasea (nm)b (nm)c (nm)α (deg)β (deg)γ (deg)
Tetracene       
Thin-film (this work) 0.592 0.76 1.32 79.8 86.4 89.6 
Bulk (this work) 0.604 0.792 1.32 76.6 72 86 
Bulka 0.606 0.784 1.301 77.1 72.1 85.8 
Pentacene       
Thin-filmb 0.592 0.754 1.563 81.5 87.2 89.9 
Bulkc 0.606 0.79 1.501 81.6 77.2 85.8 
Phasea (nm)b (nm)c (nm)α (deg)β (deg)γ (deg)
Tetracene       
Thin-film (this work) 0.592 0.76 1.32 79.8 86.4 89.6 
Bulk (this work) 0.604 0.792 1.32 76.6 72 86 
Bulka 0.606 0.784 1.301 77.1 72.1 85.8 
Pentacene       
Thin-filmb 0.592 0.754 1.563 81.5 87.2 89.9 
Bulkc 0.606 0.79 1.501 81.6 77.2 85.8 
a

Holmes et al. (Ref. 27).

b

Nabok et al. (Ref. 29).

c

Campbell et al. (Ref. 26).

In Figure 3, we present a subset of data that was collected by the area detector during growth of the same thin film examined in Figures 1 and 2. While data are recorded effectively every second, we display data for five times during growth: 20, 40, 60, 80, and 200 s. At 20 s of thin film growth, it is clear that the only phase that exists at this point is the thin-film phase. At 40 s of thin film growth, the (021) peak of the bulk phase becomes faintly visible. Beyond 40 s of thin film growth, peaks belonging to the bulk phase become more apparent. At 200 s of thin film growth, the intensity of the (021) peak of the bulk phase is similar in magnitude to the (021) peak of the thin-film phase.

FIG. 3.

In situ real-time GID during the growth of the same thin film as in Figures 1 and 2. Scattering data are shown for 20 s, 40 s, 60 s, 80 s, and 200 s from the start of growth.

FIG. 3.

In situ real-time GID during the growth of the same thin film as in Figures 1 and 2. Scattering data are shown for 20 s, 40 s, 60 s, 80 s, and 200 s from the start of growth.

Close modal

To track the formation of both phases, we have used approximate 2D Voigt functions to fit simultaneously the (021)TF and (021)B Bragg peaks for every frame of data. We can extract the intensity, position, and the full-width-at-half-maximum (FWHM) of the peaks. Due to low signal-to-noise at early times, we use a threshold of three times the standard error, 3σ, of the counts above zero to determine the appearance of a diffraction peak. The intensity of either peak before this time is considered to be zero. The integrated intensities of the (021)TF and (021)B peaks, which are directly proportional to the volume of these phases, for the same experiment shown in Figure 3 are shown in Figure 4(a). These data reveal several important facts concerning the evolution of the two crystalline phases. First, at short times, only the thin-film phase is observed, where the integrated intensity of the diffraction peak from the thin-film phase begins to rise immediately, but that of the bulk phase does not begin rising for some time after growth has begun. Second, once the bulk phase has appeared, we observe that there is growth of both phases for a significant period of time. We also note that there is clear acceleration in the rate of growth of the bulk phase and possibly also, to a lesser extent, with the thin film phase. Third, at longer times the bulk phase eventually dominates the growth, as the growth of the thin-film phase decelerates, and the thickness/amount of the thin-film phase plateaus.

FIG. 4.

(a) Integrated intensity of the (021)TF and (021)B peaks from Figure 3 displayed as a function of time. Every 4th data point is shown, and the data have been normalized to the maximum intensity observed for the bulk phase. (b) Numerical derivatives of the data shown in (a) as a function of time.

FIG. 4.

(a) Integrated intensity of the (021)TF and (021)B peaks from Figure 3 displayed as a function of time. Every 4th data point is shown, and the data have been normalized to the maximum intensity observed for the bulk phase. (b) Numerical derivatives of the data shown in (a) as a function of time.

Close modal

This behavior is made somewhat clearer in Figure 4(b), where we have used numerical differentiation of the data given in Figure 4(a) to estimate the rates of growth of the two phases. We see that these data reveal four stages of growth: (i) initially, we only have growth of the thin film phase; followed by (ii) growth of both phases, where there is strong acceleration in the rate of growth of the bulk phase; followed by (iii) deceleration of the rate of growth of the thin film phase; followed by (iv) only growth of the bulk phase.

Is this behavior observed for other growth rates? In Figure 5(a), we display the integrated intensity of the (021)B Bragg peak as a function of time for several growth rates. For all cases, there is time delay in the formation of the bulk phase, and the diffraction feature does not rise as soon as growth begins. Furthermore, the slope of the intensity increases until there is an approximately constant slope, consistent with the results we showed in Figure 4(b). Thus, for all cases, there is a delay in nucleation of the bulk phase, followed by a period where the growth accelerates and eventually becomes constant. In Figure 5(b) we plot the integrated intensity vs. the total thickness based on the estimated growth rates from AFM. As may be seen, these curves do not overlap. This can be interpreted as a consequence of the amounts of the two phases that are present as a function of the growth rate. Namely, these data indicate that the fraction of the thin film that is in the bulk phase at a particular total thickness decreases as the rate of growth increases.

FIG. 5.

(a) Integrated intensity of the (021)B peaks at various rates of growth displayed as a function of time. Every 4th data point is shown. (b) The same data from (a) displayed as a function of the total thickness.

FIG. 5.

(a) Integrated intensity of the (021)B peaks at various rates of growth displayed as a function of time. Every 4th data point is shown. (b) The same data from (a) displayed as a function of the total thickness.

Close modal

We can gain additional insight into the evolution of the two phases if we consider similar results for the thin-film phase. As before with the (021)B peak, we can also fit the (021)TF Bragg peak to an approximate 2D Voigt function. In Figure 6(a), we display the integrated intensity of the (021)TF peak as a function of time. This diffraction feature begins to grow essentially immediately after the substrate is exposed to the molecular beam of tetracene. At long times, we can again see that the intensity of this feature plateaus, indicating that the growth of the thin-film phase eventually stops. We also observe that the intensity reached at the plateau (proportional to the amount of the thin-film phase) appears to scale with the rate of growth.

FIG. 6.

(a) Integrated intensity of the (021)TF peaks at various rates of growth displayed as a function of time. Every 4th data point is shown. We also display, for this same (021)TF peak, (b) the out-of-plane FWHM, (c) |q021|, and (d) the out-of-plane peak position at various rates of growth as a function of time. Every 4th data point is shown.

FIG. 6.

(a) Integrated intensity of the (021)TF peaks at various rates of growth displayed as a function of time. Every 4th data point is shown. We also display, for this same (021)TF peak, (b) the out-of-plane FWHM, (c) |q021|, and (d) the out-of-plane peak position at various rates of growth as a function of time. Every 4th data point is shown.

Close modal

Similar to the analysis presented above for the bulk phase, in Figure 7(a) we plot the integrated intensity for the thin-film phase vs. the total thickness based on the estimated growth rates from AFM. Here we see that there is considerable overlap between these sets of data, reflecting the dominance of the thin-film phase at smaller thickness. Deviations become apparent at larger thicknesses where the amount of the thin-film phase plateaus, and this plateau becomes larger at higher growth rates.

FIG. 7.

(a) Integrated intensity of the (021)TF peaks at various rates of growth displayed as a function of the total coverage. Every 4th data point is shown. We also display, for this same (021)TF peak, (b) the out-of-plane FWHM, (c) |q021|, and (d) the out-of-plane peak position at various rates of growth as a function of the total coverage. Every 4th data point is shown.

FIG. 7.

(a) Integrated intensity of the (021)TF peaks at various rates of growth displayed as a function of the total coverage. Every 4th data point is shown. We also display, for this same (021)TF peak, (b) the out-of-plane FWHM, (c) |q021|, and (d) the out-of-plane peak position at various rates of growth as a function of the total coverage. Every 4th data point is shown.

Close modal

Additional information concerning the nature of the thin film that is grown can be found from an analysis of the width of the Bragg peak for the thin-film phase. We display the out-of-plane full-width-at-half-maximum (FWHM) as a function of time in Figure 6(b) and as a function of the total coverage in Figure 7(b). We can expect that the out-of-plane FWHM will be larger for thinner films and smaller for thicker films because of the crystallites being a finite size, as described by the Scherrer equation. As expected, the out-of-plane FWHM is initially larger before decaying to a smaller value as the thin film becomes thicker. This width becomes nearly equivalent for all growth rates (∼0.03 Å−1) at large thicknesses, and this value is greater than the expected smearing of the diffraction spot due to the size of the sample and the beam (Δqz ∼ 0.002 Å−1). This observation implies that the out-of-plane size of the crystallites may be similar for all cases. We note in passing that analysis of the in-plane width of the peaks (not shown here) gave results where the deviations were below the threshold for smearing of the diffraction spots due to the size of the sample and the beam (Δqxy ∼ 0.02 Å−1). Thus, we will not comment further on those results.

Analysis of the position of the in-plane diffraction peak for the thin-film phase is of interest, however. We display the magnitude of the vector defining the position of the (021)TF peak, |q021|, as a function of time in Figure 6(c) and as a function of the total coverage in Figure 7(c). Here, there is a modest change in |q021| from ∼1.65 Å−1 to ∼1.66 Å−1, representing a shift of less than 1%. As a function of total coverage, the results are quite similar for the different rates of growth. We display the out-of-plane position (qz) for this same (021)TF peak as a function of time in Figure 6(d) and as a function of the total coverage in Figure 7(d). Here we observe an unexpected, yet significant, change in the position of this peak. For all cases, the peak appears first at q021,⊥ ∼ 0.1 Å−1 and then increases and plateaus at a value of q021,⊥ ∼ 0.17 Å−1. As a function of total coverage, the results are quite similar for the different rates of growth. The relatively large changes in this quantity are due to much more subtle changes in the lattice parameters, including the unit cell angles. We note that the entire peak shifts instead of the formation of multiple peaks, suggesting that every portion of the thin film that is thin-film phase is changing in concert. Kowarik and co-workers observed a similar phenomenon during the growth of DIP on SiO2.14,15

We have examined in situ and in real time the evolution of two polymorphs, a thin-film phase and a bulk phase, during the thin-film growth of tetracene on SiO2. We have also examined the effect of the rate of growth on the evolution of these two phases. GID revealed that these two phases of tetracene do not begin growing simultaneously and that there is not an abrupt transition between the growth of the two phases. Instead, there are four regimes of growth: (i) growth initially of only the thin-film phase; (ii) a regime where the bulk phase nucleates, and there is a rapid rate of acceleration in the rate of growth of that phase, while the thin-film phase continues to grow; (iii) a regime where the growth of the thin-film phase decelerates; and (iv) a regime where the thin film phase stops growing and only the bulk phase continues to grow. We note here that a significant difference between the two phases is their dz-spacing, where a transition from the thin-film to the bulk phase will involve tilting of the molecules of tetracene further from the surface normal, decreasing the dz-spacing. Such a transition from a larger dz-spacing to a smaller one has also previously been reported for pentacene and DIP.15,32 Given the fact that we observe growth of both phases for a significant period of time, there must be portions of the thin film which contain regions of both phases. This creates a diffuse interface between the phases, where initially the thin film is comprised of only the thin-film phase, while eventually only the bulk phase is grown. We will discuss how this transition from growth of the thin-film phase to growth of the bulk phase occurs and how it depends on the rate of growth.

When does the bulk phase begin to grow? A previous study on pentacene reported that growth must exceed a critical thickness of at least 100 nm of pentacene on SiO2 (at room temperature) before the bulk phase begins to grow,9 and another argued that surface energy was the driving force for this transition.33 Watanabe et al. observed the rise of the bulk phase of pentacene after ∼6.5 ML of growth of the thin-film phase using in situ GID.34 In another study, Mayer et al. used in situ X-ray reflectivity and argued that both the thin-film and bulk phase of pentacene nucleate at nearly the same thin-film thickness (∼1 ML).12 However, these measurements were of poor time resolution (each scan took 65-90 s) and conclusions were drawn from fitting a model to data that did not begin until the thin film of pentacene was already thicker than 60 nm (∼40 ML).

Our results lack the uncertainties of these previous studies due to high signal-to-noise ratio data, short acquisition times (1 s), and the use of in situ, real time techniques. The results we have presented here clearly show that, for tetracene, the bulk phase does not nucleate in the first monolayer. This is very likely also true for pentacene, as it is a very similar molecule, and the observations by Bouchoms et al. concerning the critical thickness for the transition from bulk to thin-film phase corroborate this.9 Is there also a critical thickness for tetracene at which a similar transition from the thin-film phase to the bulk phase occurs? Moreover, is this concept of a critical thickness even valid for the case of tetracene grown on SiO2?

Using the intensities of the bulk phase shown in Figure 5 (vide supra), we can determine the time at which diffraction from the bulk phase becomes apparent and use that time and the mean growth rate to determine an onset thickness. More specifically, we use the time at which the counts were above zero by 3σ, as described earlier. We display the results from this analysis in Figure 8(a). The onset thickness shows a strong dependence on growth rate, varying from ∼5 ML at 0.056 ML s−1 to ∼17 ML at 0.45 ML s−1. We recall that the intensity of the plateau for the thin-film phase also displayed a dependence on growth rate [cf. Figures 6(a) and 7(a)]. To quantify this effect, we have fit straight lines to the linear portion of the rising intensity of the thin-film phase, and another to the plateaued region, to find the intersection of these two lines and determine a thickness. In Figure 8(a), we display these (total) thicknesses as a function of the rate of growth. We see that these thicknesses increase with growth rate from ∼16 ML at 0.056 ML s−1 to ∼92 ML at 0.45 ML s−1. These results unambiguously indicate that the amount of the thin-film phase present in the thin films is greater for higher growth rates. This agrees with a report from Gompf and co-workers, where they observe that increasing growth rate from 0.2 nm s−1 (0.15 ML s−1) to 1.7 nm s−1 (1.3 ML s−1) results in a greater amount of the thin-film phase.17 Similarly, Milita and co-workers have also previously studied the formation of two phases of tetracene on SiO2 and reported an increase in the amount of the phase with higher dz-spacing (note: their value was 13.2 Å) as the growth rate was increased from 0.025 nm s−1 (0.019 ML s−1) to 0.2 nm s−1 (0.15 ML s−1).18 

FIG. 8.

(a) Thicknesses for the onset of the bulk phase (right ordinate, closed circles) and the thickness where the (021)TF intensity plateaus (left ordinate, open squares) as a function of the rate of growth. (b) Thicknesses for the onset of the bulk phase (right ordinate, closed circles) and the RMS roughness of ∼19 ML thick films of tetracene on SiO2 (left ordinate, open diamonds) as a function of the rate of growth.

FIG. 8.

(a) Thicknesses for the onset of the bulk phase (right ordinate, closed circles) and the thickness where the (021)TF intensity plateaus (left ordinate, open squares) as a function of the rate of growth. (b) Thicknesses for the onset of the bulk phase (right ordinate, closed circles) and the RMS roughness of ∼19 ML thick films of tetracene on SiO2 (left ordinate, open diamonds) as a function of the rate of growth.

Close modal

One more factor to consider in interpreting our results concerns the morphology/topography of the deposited thin films. As we have shown in previous work, the topography of tetracene exhibits an unusual dependence on the rate of growth—3D islanded growth is favored at low growth rates, while 2D layer-by-layer and smoother growth occurs at high growth rates.22 In Figure 8(b) we plot the RMS roughness for thin films of tetracene grown on SiO2 under the same conditions we consider here. These films were all 19 ± 2 ML in thickness; placing them near the upper limit of values we found for the thickness of the onset of the bulk phase. Interestingly, we see that for the lowest growth rate considered here, the RMS roughness approaches the nominal thickness, which can only occur if the film is discontinuous, and possesses regions that are much thicker than what one would observe from, say, stochastic, random deposition.

Do any diffraction features from the thin-film phase provide clues as to the onset of nucleation of the bulk phase? The most relevant data are provided in Figure 7. Of the three sets of results concerning peak positions and FWHMs, the set most in line with the observed appearance of the bulk phase (at ∼5-17 ML s) would be the results for the position of the out-of-plane component for the (021)TF peak. We note that the initial increase in q021,⊥ corresponds to a decrease in the dz-spacing. We determine the thickness at which the out-of-plane position of the (021)TF peak begins to plateau by fitting a line to the early time rising data, fitting a line to the long-time plateau, and finding the intersection. Interestingly, this thickness exhibits no apparent trend with growth rate and has an average of ∼8 ± 2 ML. Thus, it is unclear if these data indicate an evolution of the thin-film phase unit cell towards one that is more like the bulk phase and might initiate nucleation.

We have addressed when the bulk phase begins to grow and now examine why it begins to grow. First, we discard the notion that the bulk phase will form on clean SiO2, at least for the conditions we consider here. The thin-film phase forms because of the interface between the tetracene thin film and SiO2 and the interfacial energy associated with it. This interfacial interaction is sufficient to overcome the energetic cost of forming the thin-film phase over the bulk phase. Why would a bulk phase form on some regions of SiO2, but not others? The most logical explanation for the appearance of the bulk phase is that the influence of the substrate has dissipated sufficiently. The presence of the thin-film phase is the entity that of course changes the effect of the substrate. The two likely factors concerning the thin film phase that might affect nucleation of the bulk phase are its thickness and topography.

Let us assume that a critical thickness does exist for the nucleation of the bulk phase, perhaps similar to the value (∼8 ML) suggested by the analysis of the data in Figure 7(d). This, by itself, would not predict a dependence on growth rate, which we clearly observe here. We must recognize, however, that what likely is important is the local thickness and not the mean thickness. As shown in Figure 8(b) we find for the lowest growth rates that we can expect that the RMS roughness can be significant, i.e., on the order of or even exceeding that of the mean thickness. Thus, in this scenario, the critical thickness is reached at the smallest mean coverage at the lowest rate of growth, while it requires a larger mean coverage at the higher rates of growth where the thin films are smoother. This is a compelling argument, but it may not tell the complete story. Surface roughness itself may also play a role in nucleation. It is known for many thin film/substrate combinations that nucleation is often facilitated at step edges, single-layer, or multiple-layer steps. In this case, the thin films deposited at lower rates of growth might also be expected to provide a higher density of steps sites (due to greater roughness) than those grown at higher rates.

To summarize our observations, we consider the schematic diagram in Figure 9. Here we display the four stages of growth, indicating how the two phases may evolve as a function of time and/or thickness. First, there is an initial period, Stage I, where only the thin film phase is formed. Even for relatively high rates of thin-film growth there is significant roughness, as indicated. Second, as we have argued above, at some point the local thickness exceeds the critical thickness and the bulk phase nucleates. If this phase grows in a 3D islanded mode, we expect its rate to accelerate as we have observed for Stage II of growth. In Stage III, the bulk phase continues to grow, while growth of the thin-film phase begins to decelerate. Finally, in Stage IV, growth of the thin-film phase has ceased, the bulk phase covers the surface, and only the bulk phase is subsequently grown.

FIG. 9.

Schematic representation of the evolution of the two phases of tetracene, the thin-film phase (reddish tones) and the bulk phase (blueish tones).

FIG. 9.

Schematic representation of the evolution of the two phases of tetracene, the thin-film phase (reddish tones) and the bulk phase (blueish tones).

Close modal

We have examined the effect of growth rate on the evolution of two polymorphs of thin films of tetracene that are grown on SiO2 at room temperature. From ex situ X-ray reflectivity, we find that tetracene forms two phases on SiO2—a thin-film phase with dz = 12.93 Å and the bulk phase with dz = 12.19 Å. We use in situ, real-time grazing incidence diffraction to determine directly the evolution of the two phases with time and thin film thickness. Initially, up to several monolayers of deposition, only the thin-film phase is grown. We find no evidence for the formation of the bulk phase in the monolayer regime, rather, there is a significant delay in the onset of growth of the bulk phase. At a growth rate of 0.056 ML s−1, the bulk phase begins growing at ∼5 ML, while the thin-film phase saturates after the growth of ∼16 ML (total) of deposited thin film. As growth rate is increased, both of these thicknesses marking changes in the stage of growth increase. At a growth rate of 0.45 ML s−1, the bulk phase begins growing at ∼17 ML, while the growth of the thin-film phase persists until ∼88 ML total of deposited thin film. We propose that the delay, in terms of a thin film thickness, in the formation of the bulk phase at the higher growth rates is associated with the formation of smoother thin films at these rates. At lower growth rates, due to the effects of upward transport of tetracene at island edges, the thin films are rougher, and a smaller total coverage is able to produce thin films where the critical thickness is exceeded, enabling the growth of the bulk phase. The bulk phase is more likely to form on tall features, of which there are many at low growth rates where there is 3D growth. Increasing the growth rate kinetically traps molecules from moving upwards and promotes more sustained growth of the thin-film phase. From this work, it is clear that the influence of the substrate on growth may dictate what the initial growing phase of organic thin film is, but the kinetics of thin-film reorganization can play a significant role in determining when and what phase ultimately becomes the dominant contributor to growth.

We would like to thank Dr. Arthur R. Woll for invaluable technical contributions, as well as James Dong and Jade M. Noble for technical assistance. We also thank Detlef-M. Smilgies for his assistance on indexing the GID data. This work was supported in part by Cornell University’s David R. Atkinson Center for a Sustainable Future (ACSF). This work made use of the Nanobiotechnology Center shared research facilities at Cornell and is based upon research conducted at the Cornell High Energy Synchrotron Source (CHESS), which is supported by the National Science Foundation under NSF Award No. DMR-1332208.

1.
G.
Witte
and
C.
Wöll
,
J. Mater. Res.
19
,
1889
(
2004
).
3.
C. D.
Dimitrakopoulos
and
P. R. L.
Malenfant
,
Adv. Mater.
14
,
99
(
2002
).
4.
Y.
Diao
,
K. M.
Lenn
,
W.
Lee
,
M. A.
Blood-Forsythe
,
J.
Xu
,
Y.
Mao
,
Y.
Kim
,
J. A.
Reinspach
,
S.
Park
,
A.
Aspuru-Guzik
,
G.
Xue
,
P.
Clancy
,
Z.
Bao
, and
S. C. B.
Mannsfeld
,
J. Am. Chem. Soc.
136
,
17046
(
2014
).
5.
M.
Mas-Torrent
and
C.
Rovira
,
Chem. Rev.
111
,
4833
(
2011
).
6.
A. O. F.
Jones
,
B.
Chattopadhyay
,
Y. H.
Geerts
, and
R.
Resel
,
Adv. Funct. Mater.
26
,
2233
(
2016
).
7.
B.
Krause
,
A. C.
Dürr
,
K.
Ritley
,
F.
Schreiber
,
H.
Dosch
, and
D.
Smilgies
,
Phys. Rev. B
66
,
235404
(
2002
).
8.
S. R.
Forrest
,
Chem. Rev.
97
,
1793
(
1997
).
9.
I. P. M.
Bouchoms
,
W. A.
Schoonveld
,
J.
Vrijmoeth
, and
T. M.
Klapwijk
,
Synth. Met.
104
,
175
(
1999
).
10.
D.
Knipp
,
R. A.
Street
,
A. R.
Völkel
, and
J.
Ho
,
J. Appl. Phys.
93
,
347
(
2003
).
11.
C.
Ambrosch-Draxl
,
D.
Nabok
,
P.
Puschnig
, and
C.
Meisenbichler
,
New J. Phys.
11
,
125010
(
2009
).
12.
A.
Mayer
,
A.
Kazimirov
, and
G.
Malliaras
,
Phys. Rev. Lett.
97
,
105503
(
2006
).
13.
A.
Hinderhofer
,
T.
Hosokai
,
C.
Frank
,
J.
Novák
,
A.
Gerlach
, and
F.
Schreiber
,
J. Phys. Chem. C
115
,
16155
(
2011
).
14.
S.
Kowarik
,
A.
Gerlach
,
S.
Sellner
,
L.
Cavalcanti
,
O.
Konovalov
, and
F.
Schreiber
,
Appl. Phys. A
95
,
233
(
2009
).
15.
S.
Kowarik
,
A.
Gerlach
,
S.
Sellner
,
F.
Schreiber
,
L.
Cavalcanti
, and
O.
Konovalov
,
Phys. Rev. Lett.
96
,
125504
(
2006
).
16.
H.-L.
Cheng
,
Y.-S.
Mai
,
W.-Y.
Chou
,
L.-R.
Chang
, and
X.-W.
Liang
,
Adv. Funct. Mater.
17
,
3639
(
2007
).
17.
B.
Gompf
,
D.
Faltermeier
,
C.
Redling
,
M.
Dressel
, and
J.
Pflaum
,
Eur. Phys. J. E
27
,
421
(
2008
).
18.
S.
Milita
,
M.
Servidori
,
F.
Cicoira
,
C.
Santato
, and
A.
Pifferi
,
Nucl. Instrum. Methods Phys. Res., Sect. B
246
,
101
(
2006
).
19.
S.
Milita
,
C.
Santato
, and
F.
Cicoira
,
Appl. Surf. Sci.
252
,
8022
(
2006
).
20.
J.
Wünsche
,
G.
Tarabella
,
S.
Bertolazzi
,
M.
Bocoum
,
N.
Coppedè
,
L.
Barba
,
G.
Arrighetti
,
L.
Lutterotti
,
S.
Iannotta
,
F.
Cicoira
, and
C.
Santato
,
J. Mater. Chem. C
1
,
967
(
2013
).
21.
N.
Moriguchi
,
T.
Nishikawa
,
T.
Anezaki
,
A.
Unno
,
M.
Tachibana
, and
K.
Kojima
,
Phys. B
376–377
,
276
(
2006
).
22.
T. W.
Schroeder
, “
Thin film deposition employing supersonic molecular beams: Nucleation and growth of silicon, silicon germanium and pentacene
,” Ph.D. dissertation (
Cornell University
,
2004
).
23.
R. K.
Nahm
and
J. R.
Engstrom
,
J. Phys. Chem. C
120
,
7183
(
2016
).
24.
D.-M.
Smilgies
and
D. R.
Blasini
,
J. Appl. Crystallogr.
40
,
716
(
2007
).
25.
P.
Busch
,
M.
Rauscher
,
D.-M.
Smilgies
,
D.
Posselt
, and
C. M.
Papadakis
,
J. Appl. Crystallogr.
39
,
433
(
2006
).
26.
R. B.
Campbell
,
J. M.
Robertson
, and
J.
Trotter
,
Acta Crystallogr.
15
,
289
(
1962
).
27.
D.
Holmes
,
S.
Kumaraswamy
,
A. J.
Matzger
, and
K. P. C.
Vollhardt
,
Chem. - Eur. J.
5
,
3399
(
1999
).
28.
C. D.
Dimitrakopoulos
,
A. R.
Brown
, and
A.
Pomp
,
J. Appl. Phys.
80
,
2501
(
1996
).
29.
D.
Nabok
,
P.
Puschnig
,
C.
Ambrosch-Draxl
,
O.
Werzer
,
R.
Resel
, and
D.-M.
Smilgies
,
Phys. Rev. B
76
,
235322
(
2007
).
30.
H.
Yoshida
,
K.
Inaba
, and
N.
Sato
,
Appl. Phys. Lett.
90
,
181930
(
2007
).
31.
H.
Yoshida
and
N.
Sato
,
Appl. Phys. Lett.
89
,
101919
(
2006
).
32.
S.
Kowarik
,
A.
Gerlach
,
W.
Leitenberger
,
J.
Hu
,
G.
Witte
,
C.
Wöll
,
U.
Pietsch
, and
F.
Schreiber
,
Thin Solid Films
515
,
5606
(
2007
).
33.
L. F.
Drummy
and
D. C.
Martin
,
Adv. Mater.
17
,
903
(
2005
).
34.
T.
Watanabe
,
T.
Hosokai
,
T.
Koganezawa
, and
N.
Yoshimoto
,
Mol. Cryst. Liq. Cryst.
566
,
18
(
2012
).