Effect of Chain Length on Swelling Transitions of Brodie Graphite Oxide in Liquid 1‐Alcohols

Swelling is the most fundamental property of graphite oxides (GO). Here, a structural study of Brodie graphite oxide (BGO) swelling in a set of long chain 1‐alcohols (named C11 to C22 according to the number of carbons) performed using synchrotron radiation X‐ray diffraction at elevated temperatures is reported. Even the longest of tested alcohols (C22) is found to intercalate BGO with enormous expansion of the interlayer distance from ≈6Å up to ≈63Å, the highest expansion of GO lattice ever reported. Swelling transitions from low temperature α‐phase to high temperature β‐phase are found for BGO in all alcohols in the C11–C22 set. The transitions correspond to decrease of inter‐layer distance correlating with the length of alcohol molecules, and change in their orientation from perpendicular to GO planes to layered parallel to GO (Type II transitions). These transitions are very different compared to BGO swelling transitions (Type I) found in smaller alcohols and related to insertion/de‐insertion of additional layer of alcohol parallel to GO. Analysis of general trends in the whole set of 1‐alcohols (C1 to C22) shows that the 1‐alcohol chain length defines the type of swelling transition with Type I found for alcohols with C<10 and Type II for C>10.


Introduction
Graphite oxides (GO) are materials prepared by strong oxidation of graphite, most commonly using procedures developed by Hummers, [1] Brodie, [2] and Staudenmaier. [3]12a] Swelling also enables permeation of solvents across multilayered graphene oxide membranes prepared most often using deposition of aqueous dispersions. [9,14]The value of interlayer distance is swollen films, membranes or papers [6a,15] defines important properties of these materials for applications.15e,16] The match between size of ions and interlayer distance is also important for migration of ions in supercapacitor electrodes. [17]However, swelling of multi-layered graphene oxide laminates is not the same as in precursor graphite oxides and depends on air-ageing, [18] sample thickness, [19] oxidation degree, [20] flake size, deposition method and possibly several other parameters [21] typically not present in pristine graphite oxides.Therefore, studies of swelling performed using graphite oxides are of the most fundamental value.
Precise control of inter-layer distance is also of great importance for further work with chemical modification of GO.One example of chemical modification is "pillared GO structures" (PGO:s).PGO:s with high surface area could be useful for many possible applications, e.g.gas storage, [22] materials for supercapacitors, [23] and membranes. [24]Swelling provides a possibility to expand GO lattice and to insert pillaring molecules otherwise too large for direct insertion into GO interlayers.
Swelling of GO in several polar solvents has been studied in recent decade as a function of temperature and pressure revealing unusual anomalies.6d,26] Remarkably, GO swelling in several solvents has been found rather different for materials prepared using Brodie (BGO) and Hummers (HGO) methods. [5]27b] Therefore, BGO is more suitable as a model system for a fundamental understanding GO swelling and the structure of swollen GO states. [11]13a,18a,30] Recent studies revealed that 1-alcohol molecules in the set C1 to C9 are intercalated into BGO forming structures with the solvent molecule layers oriented parallel to GO planes.The number of solvent layers intercalated into BGO structure at ambient conditions increases from one (for methanol, ethanol and propanol) to three in butanol (C4), pentanol (C5), hexanol (C6), heptanol (C7) and finally to four in octanol (C8) and nonanol (C9).30a,b] Cooling the solvates in excess of solvent has been found to result in sharp step-like increase of inter-layer distance due to swelling transitions.30b] The phase transitions related to changes in the number of intercalated solvent layers (parallel to the graphene oxide plane orientation) will be named in the following as a Type I.The phase transitions were found absent in the BGO systems with pentanol and hexanol, [30a] while for BGO with butanol it was observed in some experiments and not found in other for unclear reasons. [11]t is interesting to note that progressively longer alcohols are less and less hydrophilic and already starting from hexanol are not miscible with water. [31]It is well known that GO does not swell in non-polar solvents.Therefore, it is surprising that 1alcohols with rather long hydrophobic chain can be intercalated into GO structure.Intercalation of 1-alcohols longer than nonanol (C9) into GO was reported already in early studies from 1960s and later confirmed in 1980s (but only as a part of PhD thesis published in German never published in per-reviewed literature). [32]Remarkably, GO swelling in longer 1-alcohols was reported to occur with increase of inter-layer distances almost linearly as a function chain length, up to ≈50Å for C18. [33]No studies were performed in modern times for alcohols longer than 1-nonanol, with the exception of our very recent report on BGO-Hexadecanol (C16) system. [34]urprisingly, the structure of BGO formed in an excess of C16 melt appeared to consist of two layers of alcohol molecules with orientation perpendicular to GO planes and inter-layer distance of ≈48.8Å (-phase). [34]The structure of the -phase is then very different compared to the structures reported earlier for BGO systems with C1 to C9 alcohols.Our temperature dependent study of BGO swelling in excess of C16 alcohol revealed a reversible phase transition with a decrease of interlayer distance to ≈30.9Å above 338K (-phase).Notably the change in the inter-layer distance at the point of transition significantly exceeds the thickness of one 1-alcohol layer (≈4Å), thus suggesting completely different nature compared to Type I transitions described above.The phase transition was found to occur with a change in the orientation of the intercalated molecules frfom perpendicular to the GO planes in the -phase to parallel in the -phase.This kind of phase transition will be named Type II in following discussions.
The cooling of the -phase in absence of excess C16 melt was found to result in formation of third phase (-phase) suggestively corresponding to the same composition as -phase but with change in orientation of alcohol molecules from parallel to perpendicular relative to GO planes. [34]27a] Part of the solvent intercalated in BGO interlayers at low temperature is released into the liquid phase above the point of incongruent melting.
The Type I transitions have been found for BGO in a set of alcohols C1 to C9 (methanol to nonanol) while the Type II transition so far was reported only for BGO-C16 system.It is anticipated that swelling of BGO in at least some of other long alcohols just above their melting points is likely to result in the formation of -phase.In this case, the swelling transitions of the Type II will be expected to occur at elevated temperatures.However, no studies of temperature dependent GO swelling in alcohols with C>9 were so far performed except for swelling in C16.The change in type of the phase transition (and change of structure for the low temperature phase) from Type I to Type II could be expected to occur for BGO in alcohols with length between C10 to C15.However, experimental data for these systems were not yet available.It is not clear also if swelling can be found in very long 1-alcohols.Longer alcohols are more hydrophobic and it might happen that swelling of GO completely cease starting from certain chain length.
In this study, we investigate the temperature dependent swelling of BGO in a set of long chain alcohols, from undecanol (C11) to behenyl alcohol (C22) using synchrotron X-ray diffraction and Differential Scanning Calorimetry (DSC).It is found that BGO immersed in an excess of liquid alcohol shows expansion of interlayer distance, from ≈6 Å (in pure BGO) up to ≈20-60 Å depending on the length of alcohol molecule chain.The value of inter-layer distance observed in BGO intercalated with C22 (≈63Å) is so far largest ever reported in the published literature on GO swelling and intercalation.The swelling phase transitions of Type II were found for all alcohols in the set C11 to C22.Very general trends in swelling of BGO in the whole set of 1-alcohols (C1 to C22) were revealed using analysis of new and earlier reported data for changes in inter-layer distance of BGO, solvate phases composition and enthalpy of transitions between low temperature and high temperature phases.Therefore, our study provides so far most comprehensive fundamental knowledge about GO swelling in 1-alcohols.

Results and Discussion
Swelling of BGO in a set of normal long-chain alcohols (C11, C12, C13, C14, C15, C16, C17, C18, C20, C22) was studied here as a function of temperature using XRD and DSC.The study is focused on the phases formed under conditions of saturated swelling when alcohol is added to BGO powder in an amount exceeding the maximal sorption capacity of material at any given temperature.Detailed characterization of the precursor BGO is provided in Supporting Information file (Figures S1,S2, Supporting Information) and in our earlier study. [34]The BGO/alcohol mixture was heated from room temperature up to 363-373K and cooled back to ambient temperature (or somewhat below) with constant ramp while recording XRD images every ≈1s.Most of the alcohols included in this study are solids at room temperature (except for the smallest, C11).Instantaneous swelling of BGO structure occurs at the moment of alcohol melting (or after adding C11 liquid) resulting in a significant increase of interlayer distance and formation of the -phase.The only exception was swelling of BGO in C22 which appeared to be incomplete with some BGO remaining in non-intercalated state, most likely due to the high viscosity of C22 melt (see more details below).Further heating of BGO in all studied here liquid alcohols (C11 to C22) was found to result in similar swelling transitions from -phase to -phase with a step-like decrease of interlayer distance.Figure 1 shows XRD data recorded in the region of this phase transition for BGO immersed in an excess of liquid C11 and C22.XRD data for swelling of BGO in the other alcohols are provided in the Supporting Information file (Figures S3-S10, Supporting Information), including examples of XRD patterns recorded from BGO/alcohol powder mixtures recorded prior to alcohol melting.
The swelling transitions observed in BGO-CX (X = 11-22) are reversible (see Figures S3-S10, Supporting Information).The -phase is recovered after cooling and preserves even after solidification of the molten alcohol.
The properties of the swollen phases of BGO and the transition between and -phases as a function of alcohol chain lengths are discussed in more detail below.
The XRD pattern of precursor BGO shows a strong (001)reflection with d-spacing value of 6.2Å at ambient air conditions and two weak reflections from in-plane graphene oxide lattices.Swelling in all studied alcohols, (C11 to C22) is found to occur similarly to swelling in 1-hexadecanol (C16) described in detail in our earlier study. [34]Since the amount of alcohol is sufficient for saturated swelling, the patterns recorded above the melting point of alcohols show no reflections from BGO proving complete transformation into the new swollen phases (except for swelling in C22).
The low temperature BGO-CX (X = 11-22) phases (-phases) show a set of sharp (00ℓ) reflections up to ℓ = 10-14 corresponding to expanded inter-layer distance and reflections from in-plane BGO lattice, (010) and (011), at the same positions as in precursor BGO (unit cell parameter a = 2.49Å).The -phase shows only one reflection additional to the reflections of BGO lattices (insets in Figure 1A,B).This reflection has a d-spacing of ≈4.2Å corresponding approximately to the width of alcohol molecule considered as a cylinder.This reflection can be assigned to partial ordering of alcohol (C11-C22) molecules intercalated between GO layers.
The interlayer distance of the BGO-CX -phase given by c-unit cell parameter value increases proportionally to the length of alcohol molecules from 36.1Å for swelling in C11 up to 63.1Å in C22.Note that the enormous expansion of the BGO lattice in molten C22 is possibly a record-high for any swollen or intercalated GO materials reported until now.
It is interesting that the distance between GO sheets can be precisely tuned using swelling in 1-alcohols by selecting molecules with progressively longer length.The overall increment in inter-layer distance in the whole set of alcohols (C11 to C22) is 63Å-36Å = ≈37Å or in average 37Å/11 = 3.36Å increase per one carbon in the alcohol chain.
Based on the data, we suggest that the structure of the low temperature -phase of BGO-CX's in the studied set (X = 11-22) is formed by two close packed layers of alcohol molecules with an orientation almost precisely perpendicular to GO planes.This model is in perfect agreement with the increase of c-unit cell parameter, which correlates with increase of alcohol chain length.For example, the BGO lattice expansion by ≈43Å corresponds almost exactly to the length of two C16 molecules (22Å+22Å = 44Å). [35]Similarly, the increase of interlayer distance by ≈30Å is close to double length of C11 molecules (15+15Å), and the increase by ≈57Å is very close to the double length of the C22 molecule (≈29+29Å).See also more detail analysis of -structure for BGO-C16 system reported in our earlier study. [34] step-like decrease of interlayer distance is observed at the point of transition from to -phase.The temperature dependence of d( 001) is plotted for all the studied systems (C11 to C22)  in Figure 2. The change in interlayer distance observed at the point of phase transition is progressively larger for longer alcohols, starting at Δ = 9.2Å in BGO-C11 system, increasing approximately linearly with length of alcohol molecule and reaching Δ = 26Å in the BGO-C22 system.The value of Δ is somewhat smaller compared to the length of the respective alcohol molecule in each system.
According to our earlier study of BGO-C16 system, the to -phase transition corresponds to a decreased amount of intercalated alcohol.27a] Considering the very similar temperature dependent decrease of inter-layer distance, proportional to the length of alcohol molecules, it can be assumed that the phase transitions in all systems shown in Figure 2 (C11 to C22) correspond to incongruent melting of -phase with forma-tion of -phase and partial release of alcohol from the solid phase into a molten state.
The -phase is less ordered than the -phase in all studied BGO-alcohol systems as evidenced by the disappearance of the d≈4.2Å reflection (see insets in Figures 1A,B) and the increased width (FWHM) of the (00ℓ)-set reflections.
The phase transitions in the BGO-CX (X = 11-22) systems were confirmed and additionally characterized using the DSC method, (Figures 3,4; Figures S15-S23, Supporting Information).DSC was used to determine the compositions of the phases at the temperature of point of alcohol melting and enthalpy of phase transitions in all BGO-CX (X = 11-22) systems.The results from DSC are summarized in Table 1. Figure 3 shows example of DSC traces recorded for BGO in C11 and C22 as well as reference DSC traces recorded for pure alcohols.
Figure 3A shows DSC trace recorded from pure C11 alcohol upon heating.It shows endothermic anomalies due to melting in the heating part and exothermic anomalies due to solidification of the alcohol in the cooling part.The enthalpy of melting recorded in this scan is used as a reference to calculate sorption of C11 by BGO.The DSC scan shown in Figure 3B was recorded when the sample of the -phase (formed by immersing BGO powder into excess of C11 at ambient temperature) was heated over the temperature of transition into the -phase and cooled back to ambient temperature.It shows two anomalies in both the heating part and the cooling part: one anomaly due to melting/freezing of C11 and second due to reversible transition between and -phases.
The anomalies due to melting and freezing of C11 shown in Figure 3B are smaller than expected for the amount of alcohol loaded.That is because part of C11 is intercalated into BGO structure forming solid solvate structure (-phase) and do not contribute to melting anomaly related to the remaining bulk C11.Therefore, the composition of the BGO-C11 -phase at the temperature point of alcohol melting can be determined exactly using the DSC scan.27a,30a,b] Sorption of liquid 1-alcohols at the Table 1.Summary of the results obtained using DSC.From left to right: onset temperature and enthalpy of  to  transition (in kJ g −1 and mol mol −1 units), -phase composition at the temperature point of solvent melting.

System
to  transition  phase composition   temperature point of melting was determined using this method for all studied alcohols in the C11-C22 set (see Table1).
The DSC scan recorded from the BGO-C11 sample (Figure 3B) also shows an endothermic anomaly due to to -phase transition with onset at 305 K (cooling cycle).Figure 3C,D show DSC traces recorded in similar experiments with pure C22 and BGO-C22 samples.The reference DSC trace of pure C22 shows an additional peak during the cooling (with onset temperature 338 K) corresponding to solid-solid (polymorphic) transition.It is known that several long alcohols exhibit this transition. [36]For some alcohols (e.g., C15, C17, and C20), this transition can be observed as a separate feature during melting, for other (e.g., C22) it is only noticeable during crystallization.For BGO-C22 the transition from to -phase is found at 357 K, reverse transition occurs at 350 K. Similar DSC experiments were performed with the other BGO-alcohol systems.Temperature points and enthalpies recorded for the to -phase transitions are listed in the Table 1.
Analysis of the data in Table 1 shows following trends: The amount of alcohol intercalated into BGO (at the temperature close to melting of alcohol) increases (in g g −1 units) proportionally to the length of intercalated molecules (C11 to C20) but remains similar in mol mol −1 units.That is an expected effect of increase in the length of intercalated 1-alcohol molecules.The smaller sorption of C22 is explained by incomplete conversion of all BGO into the -phase.The incomplete transformation was observed using both XRD and DSC methods and most likely explained by the high viscosity of molten C22.For other alcohols the observed trend is in agreement with an assumption of similar structures (-phase) composed of two layers of alcohol molecules intercalated in an orientation perpendicular to the GO planes (Figure 4).
The structural model is shown in Figure 4.In this model the -OH groups of the alcohols are attached to graphene oxide planes.35a] This configuration is more favorable in the structure of pure alcohols due to hydrogen bonding between hydroxyl groups of neighboring molecules.Presence of (rather disordered) hydroxyl groups on GO planes allows to form hydrogen bonds with -OH groups of alcohols.The XRD data does not allow us to distinguish between the two possible orientations of the long alcohol molecules in the -phase but attachment of molecules to GO planes seem to be more likely.
The temperature point of to transition is found to increase approximately linearly from 305K for BGO-C11 up to 357K for BGO-C22 correlating with the increase in the melting point of the pure alcohols.The enthalpy of to transition also increases in the set C11-C20 (both in kJ g −1 and kJ mol −1 units) but not linearly.More detailed analysis of these trends is provided below.
It is interesting to compare the enthalpy of the to -phase transition with enthalpy of pure alcohol melting.Assuming that the transition enthalpy is only due to melting of alcohol, it is possible roughly estimate for the amount of alcohol released from the solid phase into the liquid phase at the point of transition.According to this estimation, about half of C16 present in phase melts above the point of transition into the -phase and escapes from the structure.Of course, the structural state of C16 intercalated into the BGO structure is not the same compared to bulk C16, which affects the enthalpy of transition.Calculating the composition of the -phase using melting enthalpy of alcohol results in a strong overestimation of alcohol content as verified in our earlier study using example BGO-C16. [34]The overestimation points to a strong change in the structure of intercalated alcohol remaining between the GO layers in the -phase.The data obtained using both XRD and DSC methods allow to conclude that the phase transitions observed in all the studied BGO-CX (X = 11-22) systems are similar in nature.However, exact determination of the -phase structure using XRD is not possible due to absence of any reflections specific to the lattices of intercalated alcohols.Two hypothetical alternatives of high temperature phase structure are presented in the Figure 4.
The approximate correlation of the step change in the interlayer distance at the point of transition with the length of the alcohol molecules could indicate a change of structure from two (vertical)-layer intercalation in the -phase into a pseudo one (vertical)-layer structure (-phase in Figure 4).The meaning of "pseudo" in this case is to note that complete removal of one vertically standing layer of molecules is unlikely.If the alcohol molecules are attached to GO planes, removing some molecules from both sides of inter-layer is more likely to result in structure shown in Figure 4 as -phase.The formation of this structure looks "logical" at first glance assuming the decrease of interlayer distance and partial removal of alcohol molecules at the point of the phase transition.However, our previous detailed study of the BGO-C16 system suggests the swelling transition occurs with formation of another structure with completely disordered layers of alcohol molecules in orientation "parallel" to GO planes.This type of structure was supported by evidence of layer-by-layer removal of alcohol molecules provided by XRD data recorded under conditions of vacuum annealing. [34]he increase in inter-layer distance of the -phase observed for longer alcohols (Figure 2) suggests formation of progressively larger number of intercalated alcohol layers.However, it is not directly clear why the number of layers would increase for longer alcohols.Below we argue that the increase of c-unit cell parameter in the BGO structures intercalated with longer alcohols can be explained assuming a kink-layered structure of -phase.
The -phase shown in Figure 3 was found only if the -phase is cooled down to ambient temperature in an absence of excess alcohol melt.The phase transition between the and phases was confirmed by DSC in our previous study of the BGO-C16 system.A mixture of and -phases is found at ambient temperature if the -phase is cooled down with an amount of melt sufficient only for partial transformation into -phase as confirmed by both DSC and XRD data. [34]Similar results were observed in this study for several other BGO-alcohol systems in experiments where the loading of alcohol was not sufficient to convert all BGO powder into -phase (see Supporting Information file).Slight deficy in amount of alcohol was sufficient in these experiments to form phase at elevated temperature.However, cooling of the phase with an amount of molten alcohol not sufficient for complete transformation into the phase results in a two-phase mixture.The phase is formed by the sorption of all available molten alcohol, the remaining -phase undergoes transition into -phase.Using these experiments, the existence of -phase was confirmed here for BGO systems with C16, C18, C20, and C22 alcohols (see Supporting Information file).The phase is not thermodynamically stable since it is formed only in non-equilibrium conditions when there is no excess of liquid alcohol to be intercalated at near ambient temperatures.Detailed investigation of -phases and other possible alcohol-deficient structures is out of scope of our study focused here on thermodynamically equilibrium phases.
In conclusion of this part, we suggest that the same type of to -phase transition occurs in all studied here BGO-CX (X = 11-22) systems upon heating and reverse transition upon cooling.This suggestion is further confirmed by analysis of general trends which include the temperature dependent swelling transitions in the whole set of 1-alcohols starting from C1 (methanol) and up to C22.This is presented below.
Figure 5 summarizes all the available XRD data on the temperature dependent phase transitions in the BGO-CX systems, including the data obtained in this study for X = C11 to C22 and earlier published data for smaller alcohols (X = C1 to C9). [6d,25b,30a,b] The interlayer distances shown in Figure 5 were recorded for BGO immersed in an excess of liquid alcohols at the temperature points just below the phase transition (low temperature LT phase) and just above of swelling phase transition (high temperature HT phase).The swelling phase transitions related to a change of alcohol amount intercalated into BGO structure are found for BGO intercalated with all alcohols except butanol (C4), pentanol (C5), and hexanol (C6).Ambient temperature values of inter-layer distance are used BGO-C5 and BGO-C6.
The data shown in Figure 5 allow to analyse general trends for swelling of BGO in the whole set of normal alcohols C1 to C22.The data suggest two different types of phase transitions between low temperature LT and high temperature HT phases for BGO heated/cooled in excess amount of short and long alcohols.The border between these two types of phase transitions is shown by dashed line at the C10 alcohol.The change of interlayer dis- tance observed in the Type I transitions (≈4-4.5Å)corresponds exactly to the thickness of one alcohol layer with molecules parallel to the GO planes.For BGO in methanol, ethanol and propanol the phase transition between LT and HT phase corresponds to change from 2 layer (2L) to 1 layer (1L) structure, for heptanol the change is from 4 layer (4L) to 3 layer (3L) structure and for octanol and nonanol the change is from 5 layer (5L) to 4 layer structure (4L) structure.
The swelling phase transitions found for BGO in C11 to C22 set of alcohols are of rather different nature and named in following as Type II.The change of inter-layer distance observed in the Type II transitions increases proportionally to the length of intercalated alcohol molecules (Figure 2,5).The difference between Type I and Type II transitions is even more obvious if the value of change in interlayer distance Δ is plotted as a function of 1-alcohol length (number of carbon atoms in chains), Figure 6.The first region corresponds to the Type I transition in BGO intercalated with C1 to C9 alcohols with Δ values corresponding to thickness of one alcohol layer in parallel to GO planes orientation (3.3Å -4.5Å).The second region is found for BGO intercalated with larger alcohols (C11 to C22) and corresponds to the Type II transition.The value of Δ in the Type II transitions increases approximately linearly correlating with increase in the length of 1-alcohol molecules.For example, the step change in inter-layer distance observed in the BGO-C16 system is 19Å which is ≈3Å smaller compared to the length of hexadecanol (C16).
The enthalpies of Type I and type II transitions are also distinctly different (Figure 7a).Insertion/de-insertion of one "parallel" layer of alcohol molecules into BGO structure results in a relatively low enthalpy of ≈0.5 kJ mol −1 (BGO).The enthalpy of transition increases steeply for C11-C15 alcohols reaching ≈7.5 kJ mol −1 (BGO) for C15, C16, C18, and C20.The data obtained for BGO-C22 are omitted here due to incomplete intercalation revealed in XRD experiments (Figure 1b).6d,11,30a] Therefore, the DSC cannot reveal the true enthalpy of transitions for these systems.30a] The difference in the type of temperature driven phase transition is related to rather different structures of low temperature phases.Our studies suggest that small 1-alcohols (C1 to C9) are intercalated into BGO structure as layers parallel to GO planes, while for alcohols larger than C10 intercalation occurs in a form of two close packed layers with orientation of molecules perpendicular to the GO planes (-phase, Figure 4).
The difference in the type of alcohol intercalation can be confirmed by analysis of Figure 7b showing the amount of alcohol intercalated (sorbed) into the BGO structure in the LT phase (in g g −1 ) as a function of alcohol length.The figure shows two distinctly different regions.The sorption of alcohol in g g −1 unit increases almost linearly for C4 to C12 range.The sorption of alcohols in the C12 to C20 range is nearly constant with a value of ≈2.5-3.1 g g −1 .The increase of sorption in g/g units reflects stronger expansion of BGO lattice (inter-layer distance) when intercalated with the longer alcohols.Note that the sorption values for ethanol and propanol are not available in the literature.The DSC method does not provide reliable results for these solvents due to rather complex freezing/melting processes typical for these solvents which include solidification into solid glass phase, crystallization of glass phases etc. [37] The correlation of the LT phase compositions with the interlayer distance revealed by XRD is not straightforward.The layers of all studied alcohols (in parallel to the GO planes orientation) are expected to occupy almost the same volume independently on the chain length.It is also expected that the weight of one layer (in g g −1 units) will be similar for all alcohols (assuming close packing of molecules), possibly changing slightly due to increase of chain length.In other words, the amount of alcohol (in g g −1 alcohol/BGO) sorbed per one layer could be expected to be similar for alcohols in the range C1 to C9.In this case, step-like increase of sorption value would be observed for BGO-alcohols in the range C1 to C9 where the number of intercalated alcohol lay-ers increases up to five (5L).However, it is not the case according to the data shown in Figure 7b.
Indeed, there is step like increase of alcohol sorption from ≈0.55 g g −1 for BGO in methanol (2L structure, ≈0.27 g/g per layer) to 0.95 g g −1 for BGO intercalated with pentanol (C5) and 1.04 g g −1 in hexanol (C6) (Figure 7a).The inter-layer distance observed for BGO intercalated with pentanol and hexanol corresponds to 3L structure (Figure 5).Therefore, total sorption 1.04 g g −1 corresponds to sorption of ≈0.33 g/g per one layer.Sorption of 1.28 g g −1 for heptanol corresponds to 4L structure (Figure 5) according to XRD data and ≈0.3 g g −1 .The "layers" in BGO-octanol (C8) were found to be less dense with 0.88 g g −1 overall sorption for 4L system, 0.22 g g −1 per layer, more similar to methanol and ethanol.30b] The absence of straightforward correlation between amount of sorbed solvent and number of "layers" in alcohols C1 to C9 could indicated deviation from close packing structure of layers (diluted layers) but also could partly be related to some difference between BGO batches used in earlier studies.
Compositions of the LT -phases of BGO intercalated with alcohols C11 to C20 (2.6-3.1 g g −1 ) were determined using the same precursor batch but also do not correlate with the rather strong expansion of inter-layer distance found by XRD.The interlayer distance is almost doubled in the C20 compared to C12 but the change in sorption value (in g/g) is almost absent.It means that the sorption in mol mol −1 is decreasing when moving from C12 to C20 (see Table 1).
It can be suggested that the "layers" of alcohols intercalated into BGO structure are likely to be "diluted" (not close-packed) at least in some systems.That is valid for both types of "layers" with orientation of alcohol molecules parallel and perpendicular to the GO planes.Intercalation of BGO with "diluted" layers was earlier suggested using structural modeling of structures formed in liquid sugar alcohols (sorbitol and xylitol). [38]he change in orientation of intercalated alcohol molecules from parallel (for C<10) to perpendicular in the LT phases is supported by strong increase of sorption for BGO immersed in longer alcohols (C>10).
It can be concluded that the structure of the LT phases of BGO intercalated with long alcohols (C11 to C22) can be confidently described using the model with two "vertically" oriented layers of alcohol molecules (-phase).The composition of these phases is given in the Table 1 as determined using DSC data.
The structure of the -phases found in the BGO-CX systems with X = C11 to C22 is less clear.Below we summarize the main facts about the structure of -phases.According to XRD data the inter-layer distance in -phases is about half of the -phase while the difference in composition (g g −1 ) is ≈2.5-3 times (at least for C16 [34] ), thus indicating rather "diluted" layers.XRD also shows increased disorder of alcohol molecules in the -phase evident from broadening of (00ℓ)-reflections and absence of ≈4.1Å reflection, the only reflection which indicates some ordering of alcohol molecules in phases.
The structure of the -phases is likely to include "layers" of alcohol molecules parallel to the GO planes, according to XRD data obtained under conditions of vacuum annealing of BGO-C16. [34]owever, the alcohol molecules in the -phase are completely disordered as evidenced by absence of any XRD reflections specific to alcohol in all BGO-CX (X = 11-22) systems.The increase in the inter-layer distance observed for the -phases as a function of alcohol chain length is almost linear starting from ≈27Å for the BGO-C11 system to 37.1Å in BGO-C22.Considering the thickness of one alcohol layer of ≈4.5Å the increase over the whole range C11 to C22 correspond roughly to two "layers" parallel to the GO planes.The question remains why the number of "layers" would increase in the structures of BGO intercalated with longer alcohols.
The temperature dependent phase transitions in the BGOalcohol systems can be described as incongruent melting of LT phase with formation of a HT phase plus molten alcohol for all alcohols in the set C1 to C22.However, why the nature of these transitions suddenly changes for alcohols with chain length below and above C10 is still an open question.It is also not completely clear if structures of the HT phases are also significantly different for BGO intercalated with small 1-alcohols (C<10) and long 1-alcohols (C>10).35b] Below we present some simple geometrical considerations, which help to understand how the length of alcohol molecules could affect the type of layered structures in BGO-alcohol systems.35a] (Figure 8).Small alcohols (ethanol in Figure 8) adsorbed on graphene oxide plane in completely random in-plane orientations can easily form layers with approximately close-packed molecules filling all the space on the surface.In contrast, long molecules oriented randomly on the GO plane will unavoidably overlap with each other starting from certain number of added molecules, thus forming kink-layers.Formation of kink layers by overlap of randomly oriented molecules is illustrated in the Figure 8 using example of C12 alcohol.The length of this molecule (≈17Å) is sufficient to form two kinklayers (each ≈4-4-5Å thickness).Assembling two kink-layers on each side of BGO interlayer would result in a 4L phase.Simple geometrical considerations allow us to estimate approximate length of 1-alcohol molecules sufficient for formation of kink-layers, as it is schematically shown in the Figure 8. Considering that the height of one layer is ≈4-4.5Å,only molecules which are approximately twice this distance or longer could form the second kink layer.The length of the molecules also limit the possible number of kinks as it is illustrated in the Figure 4 (image showing structure of -phase) on example of C16 intercalation.The length of the C16 molecule is ≈22Å.Therefore, it can easily form not only kinked two-layers but also three-layered kinks.The 3-layer kink structure can be easily constructed assuming a length of ≈4.5Å in each part of the kinks (4.5Å x5 ≈22.5Å).The structure with 4 kinklayers is not possible since the height of 4 layers is ≈18Å, slightly less than length of whole 22Å long molecule.The experimentally observed 5L structure of BGO-C16 -phase can be imagined by considering completely disordered combination of kinked 2-3 layers formed by C16 molecules on GO planes on both sides of interlayer.
This simple model agrees with the XRD data obtained during evaporation of alcohols driven by vacuum.Removing alcohol molecules from the phase of BGO-C16 will lead to less and less overlaps between randomly in-plane oriented molecules and result in local changes of the inter-plane GO distance related to the decreased number of kinked "layers".
The model of kink-layered structure is also in agreement with the observation of phase transition between -phase and -phase observed when the -phase is cooled to ambient temperature without an excess of molten alcohol.This transition is likely related to change of kink-layered structure into the structure with orientation of alcohol molecules nearly perpendicular to GO planes (as in Figure 4).This change occurs with only slight decrease of inter-plane distance of BGO and corresponds to enthalpy of ≈0.1 kJ g −1 for BGO-C16.That is ≈2.5 time's smaller enthalpy compared to the enthalpy of the to -phase transition.Assuming that the change between horizontal and vertical orientations of C16 molecule is similar in these two types of transitions (-to -phase and to -phase), the enthalpy values are proportional to the number of molecules intercalated into BGO in these two phases.If this assumption is true, the amount of alcohol in the -phase in g/g (alcohol/BGO) is ≈2.5 times smaller compared to -phase.
The transitions related to formation of kink-layers are not possible for alcohols with the size smaller than ≈C10.One horizontal and one vertical parts of kink with equal length require total length of molecule to be ≈9Å as in nonanol.The length of octanol and nonanol is on the edge of possibility to form the second layer by kink-model.Assuming that the kinks are unlikely to be formed by molecules strictly perpendicular to each other, the size consideration presented above are only approximate.Nevertheless, it is very likely that low temperature phases of BGO formed with small alcohols are formed by true (not kinked) layers of alcohol molecules.If that is the case, not only low temperature but also high temperature phases of BGO with long and short alcohols (with border case at C10) are different in structure.The HT phases formed in Type I transitions to the layers formed by straight alcohol molecules while for Type II transitions the HT (-phases) include kinked and heavily overlapped longer alcohol molecules.
In conclusion, of this part, we suggest that the change between phase transitions of Type I and Type II observed in BGO-alcohol systems for alcohols with length below and above ≈C10 is related to different orientation of molecules relative to GO planes in the low temperature phases and to straight versus kinked shape of alcohol molecules in high temperature phases.
Finally, some general considerations about properties of long chain alcohols need to be taken into account while considering their interaction with graphene oxide.
The properties of BGO intercalated with longer alcohols are likely to be affected by the strong interaction of the alcohol molecules with each other rather than by the weaker interaction between the GO planes.It is unlikely that any interaction between GO planes preserves after intercalation of long alcohols considering separation by tens of Å.For example, 90% of the volume in the structure of BGO-C22 in the -phase is by occupied by the double (vertical) layer of alcohol molecules.The inter-layer distance of BGO-C22 is ≈60Å and thickness of one graphene oxide layer is ≈6Å.This factor likely contributes to better ordering found in structures intercalated by long alcohols.The better ordering is evidenced by the very sharp XRD reflections and the presence of additional reflections from GO at d = 4.2Å in the phase.
The polar part of alcohol molecule is -OH group, while that carbon chain is hydrophobic.Longer alcohols demonstrate higher melting points and higher enthalpy of melting, thus showing stronger interactions with each other due to van der Waal bonding between the chains.The longer is 1-alcohol molecule, the more hydrophobic it is and the stronger it will be repelled from hydrophilic graphene oxide plane decorated with -OH groups.
The smallest alcohols; methanol, ethanol, and propanol are completely miscible with water.The BGO intercalated with these alcohols also forms a distinct group of materials (Figure 5-7) with a Type I transition between the 1L and 2L phases.Alcohols in the set butanol to heptanol demonstrate progressively smaller solubility in water which drops to an almost negligible level for heptanol (C7).Notably, swelling transitions were found absent for BGO in pentanol and hexanol, and irreproducible in butanol.30a] Longer alcohols are not soluble in water.Nevertheless, BGO still demonstrates Type I transitions in octanol and nonanol [30a,b] thus suggesting structures of LT phases with solvent molecules oriented parallel to graphene oxide plane.
It can be assumed that -OH groups of the alcohols are attached to the graphene oxide planes by hydrogen bonding.The longer is hydrophobic chain, the stronger the repulsion from the hydrophilic graphene oxide plane.The longer the hydrophobic chain, the more likely it will lie over some -OH groups on graphene oxide planes.It can be assumed that the change in orientation of intercalated alcohol molecules from parallel to vertical (moving from shorter to longer molecules) is then related to increased hydrophobicity of longer molecules.

Conclusion
In conclusion, we have found that inter-layer distance of BGO can be tuned using intercalation of normal alcohols in a rather broad interval of values.The structure of BGO intercalated by 1alcohols C11 to C22 (-phases) just above the melting point was analyzed using synchrotron X-ray diffraction and composition of -phases determined using DSC sorption tests.An extraordinary large inter-layer distance of ≈63Å was found in the -phase of BGO-C22 at ambient temperature.
We also report a set of swelling transitions found for BGO in all studied here alcohols C11 to C22.The Type II to -phase transitions occur at elevated temperatures and correspond to a significant decrease of BGO interlayer distance (≈2-fold), ≈2.5-3 fold decrease in amount of intercalated alcohol and stronger disorder of alcohol molecules.It is suggested that the change of structure in the to -transitions also corresponds to a change in orientation of the intercalated molecules from nearly perpendicular to the graphene oxide planes to kink-layered parallel.This kind of phase transition is named as Type II to distinct it from Type I transitions related to insertion/de-insertion of one layer of alcohol molecules in parallel to GO orientation.
Analysis of the general trends for intercalation of all 1-alcohols in the set from C1 to C22 shows that Type I transitions occurs for alcohols with chain length smaller than C10, whereas Type II transitions are found for BGO in alcohols with chains longer than C10.One of the remaining questions is why swelling transitions are not found for BGO in alcohols with certain length (e.g., pentanol and hexanol).
Finally we note that Type I and Type II transitions are found here for intercalated BGO at the conditions of saturated sorption in an excess of liquid solvents.Other "kinetic" structures can be obtained in solvent-deficient systems, the systems where liquid solvent is removed from the system or alcohols are allowed to evaporate from solid solvate phases of BGO-alcohols.Some properties of alcohol-intercalated BGO materials might potentially be interesting for applications.The distance between GO planes can be tuned using insertion of alcohols as molecular pillars rather precisely and in a very broad interval of values (3Å to 60Å).Short alcohols (e.g., methanol and ethanol) evaporate from GO rather rapidly at ambient conditions thus making solvate structures rather unstable when taken out of solutions.However, BGO materials intercalated by alcohols with melting point above ambient (C>11) are rather stable due to negligibly small evaporation rate of alcohols at ambient conditions.These materials could be interesting for membrane applications.Combination of hydrophilic GO planes with long hydrophobic chains of 1-alcohol molecules in the same structure might result in unusual membrane properties of these materials.Therefore, multilayered GO foils or films intercalated with long alcohol molecules might provide a new type of applications as membranes or coatings.
However, the study presented here is mostly of fundamental value providing comprehensive summary of graphite oxide swelling in 1-alcohols, one of the most common type of polar solvents.It adds to fundamental understanding of structures and structural transitions in unusual materials composed by 2D sheets of GO and short/long chain 1-alcohols (C1 to C22).We demonstrate that relative orientation of these two structural units at the GO-alcohol interface is controlled by 1-alcohol chain length and related also to increasingly hydrophobic nature of longer molecules.

Experimental Section
35b] Starting material was 10.25 g natural Graphite (325 mesh) purchased from "Alfa Aesar".The final yield was 14.5 g brown colored powder.All alcohols were of high purity (> 97%), purchased from Sigma Aldrich.
Characterization: The BGO batch used in this study was characterized after preparation using XRD, XPS, FTIR, TGA/DSC (Figures S1,S2 in Supporting Information file).18a,30b,38] The batch studied here showed C/O = 2.65 and negligible component due to C = O in C1s XPS spectra indicating rather low number of hole defects typical for HGO. [39]The averaged interlayer distance provided by XRD was given by d(001) = 6.0-6.2Å.Note that it depends on humidity at ambient air conditions (see Supporting Information file).
XPS spectra were recorded with a Kratos Axis Ultra electron spectrometer equipped with a delay line detector.A monochromatic Al K source operated at 150 W, a hybrid lens system with a magnetic lens, providing an analysis area of 0.3 × 0.7 mm, and a charge neutralizer were used for the measurements.The binding energy scale was adjusted with respect to the C1s line of aliphatic carbon, set at 285.0 eV.All spectra were processed with the Kratos software.TGA was done using a Mettler Toledo TGA/DSC1 STARe System.Experiments were performed from room temperature up to 700 °C at a heating rate of 3 K min −1 under nitrogen or air flow (40 mL min −1 ).
Part of the experiments with heating and cooling was performed at the Rossendorf Beamline (BM20), [40] ESRF, using a Cryostream 800 System (Oxford Cryosystems) and a radiation wavelength of  = 0.7381Å (16.798 keV).Experimental data were collected in transmission geometry using a DECTRIS PILATUS3×2 M Si area detector and extracted with the Bubble software. [41]The detector geometry parameters were calibrated with PyFAI using the a NIST LaB6 standard reference material.The intensity of X-ray beam was adjusted using filters to avoid beam-induced damage.
Main part of data was recorded at the DanMAX beamline at the MAX IV Laboratory.The temperature was controlled using a Cryostream 800+ (Oxford Cryosystems).The diffraction data were collected up to a maximum 2 of 28 degrees (Q max ≈3.75 Å −1 ) using a wavelength of 0.8266 Å (15 keV) on a DECTRIS PILATUS3×2 M CdTe area detector.The wavelength and geometry of the instrument were refined using a LaB 6 standard (NIST SRM660c) using the program PyFAI. [42]The raw area detector data were azimuthally integrated to intensity versus 2 using the MatFRAIA algorithm. [43]The beam was attenuated to avoid radiation induced damage to the samples.
Glass capillaries with diameters of 0.7-1 mm and transmission geometry were used in both sets of XRD experiments at ESRF and at MAX IV.Typically, the powder samples of BGO and C16 were ground together in different proportions, loaded into glass capillaries which were sealed to prevent solvent evaporation at higher temperatures.The mixture was then heated above the melting point of C16 using a controlled temperature ramp with simultaneous recording of XRD images (typically 1s per diffractogram).Some 10s exposures were also recorded at selected temperature points.

Figure 1 .
Figure 1.XRD patterns recorded in situ during heating (0.5 K min −1 ) in the temperature interval of to -phase transition.A) BGO-C11 system and B) BGO-C22 system.Insets show angular region for the reflection with d ≈4.19Å suggestively assigned to disordered lattice of intercalated alcohol molecules.The XRD data collected during the cooling half cycle in these experiments are provided in Supporting Information file.Radiation wavelength  = 0.8266 Å.

Figure 2 .
Figure 2. Temperature dependence of inter-layer distance d(001) for BGO samples immersed in excess of liquid 1-alcohols.

Figure 3 .
Figure 3. DSC traces of A) pure C11 and C) C22 and DSC traces recorded from BGO powder immersed in an excess of B) C11 or D) C22 with heating/cooling rates of 2 K min −1 .C11 is a liquid at ambient temperature and C22 is solid.Therefore, the BGO/C22 powder mixture was first heated over the melting point of alcohol in order to achieve saturated sorption with formation of -phase and cooled back to ambient temperature.The DSC scan shown in (D) was recorded in the second heating/cooling cycle.

Figure 4 .
Figure 4. Hypothetical structural models illustrating change of BGO-C16 structure in to -phase transition.The transition into the alcohol-deficient -phase occurs only when the -phase is cooled down in absence of liquid alcohol where additional intercalation is impossible.

Figure 5 .
Figure 5. Inter-layer distances observed for BGO-C X intercalated with 1alcohols (number of carbon atoms X = C1-C22).Low Temperature (LT) phases (black symbols) correspond to inter-layer distances observed in BGO-C X systems just above the melting point of alcohol.Interlayer distance of High Temperature (HT) phases (red symbols) is given for temperatures just above the point of phase transition.Data for C11 to C22 alcohols (red and black circles) are from this work, while the data for alcohols with carbon numbers C1-C9 are from previous studies. [6d,25b,30a,b] Note that phase transitions are absent for BGO intercalated with C4, C5 and C6 alcohols.

Figure 6 .
Figure 6.Difference in inter-layer distance at the point of phase transition observed in BGO-alcohol systems upon heating/cooling.The data points of 1-alcohols with carbon number 11-22 are from this study.For the smaller alcohols (methanol to nonanol) the data are from refs.[6d,25b,30a,b].

Figure 7 .
Figure 7. a) Enthalpy of phase transitions found in BGO-alcohol systems (C1 to C20) in J mol −1 (BGO).The grey dashed line is a guide to the eye.b) Composition of low temperature phases for BGO immersed in an excess of liquid alcohols in g g −1 (alcohol/BGO) determined using DSC method at the point of alcohol melting.The data points for C10 to C20 alcohols are from this study (black symbols), for smaller alcohols (red symbols) data are form refs. [27a,30a,b,37].