摘要:

J. MATER. CHEM., 1994, 4( I), 41-46 Dopant and Impurity Effects in Electrodeposited CdSICdTe Thin FiIms for Photovoltaic Ap p Iications Stephen Dennisont Department of Chemistry, University College London, 20 Gordon Street, London, UK WC1H OAJ The effect of deliberately added CI-, Cu2+ and Ag on the performance of electrodeposited thin-film CdS/CdTe + photovoltaic devices has been studied. It has been found that the incorporation of CI- from the deposition electrolytes into both CdS and CdTe films is essential for the production of high-efficiency devices. The incorporation of Cu2+ into CdTe was found to result in a loss in performance for concentrations in the electrolyte >30 ppb, corresponding to a concentration of ca. lo2' ~rn-~Cu in the film. Below 30 ppb, there is some evidence that Cu2+ may enhance the performance of the device.The incorporation of Ag into CdTe was found to be deleterious at concentrations in the electrolyte as low as the detection limit (5 ppb). By controlling these species within the determined limits, it was possible to produce high-efficiency devices over a period of 4-5 months (> 100 depositions) from a single electrolyte. The fabrication of thin-film semiconductors for photovoltaic cells by electrodeposition is an important commercial goal. Electrodeposition presents a potentially cheap deposition technology which would enable thin-film photovoltaic cells to become competitive with devices based on (single) crystal- line semiconductors. There have been a number of reports of the successful application of electrodeposition in the pro- duction of high-efficiency thin-film devices based on an n-CdS/p-CdTe device, where both CdS and CdTe are elec- trodeposited,'-6 and also where the CdS is chemically dep~sited.~Generally, electrodeposition of CdTe is carried out from aqueous solutions containing high concentrations of Cd2+ (ca.0.5-1 mol dmP3) and substantially lower concen- trations of TeIV(0.3-1 mmol dm-3 as HTeOzf),7-12 although CdTe electrodeposition from non-aqueous solutions has also been rep~rted.'~,'~From aqueous solution the CdTe is deposited as n-type, with conversion to p-type being accomplished by annealing at temperatures in the range 400-450 "C. The production of highly efficient devices requires careful control of both the electrodeposition process and potential dopants or impurities in the deposition electrolyte, in an identical fashion to the controls required of more conventional thin-film deposition processes such as chemical vapour deposition.The effect of dopants and impurities is of critical importance to the properties of the semiconductor. Indeed, the nature of the conductivity type of the CdTe may be critically dependent on what dopant/impurity species are present in the electrolyte. However, there have been few reports of the monitoring or control of dopants or impurities, either in the electrolyte or in the electrodeposited films. Lyons and co-workers11,12 have reported a detailed study of impurities in both the electrolytes used to produce CdTe films and in the electrodeposited CdTe itself.They detected a wide range of impurity elements, both in the electrolyte and CdTe deposit, but no correlation between the elements present with the performance of photovoltaic devices was made. Instead their work concentrated on the analysis and removal of impurities deriving from the materials used to make the deposition electrolytes. On the other hand, there have been some reports of the direct deposition of p-type CdTe,15-17 either by control of the deposition parameter^,'^ or more usually by the addition of a suitable dopant specie^.'^"^ Bhattacharya and co-workers claimed the deposition of p-CdTe by the addition of Cu15 and As16 to the electrolyte.t Present Address: Wallace and Tiernan Ltd., Priory Works, Tonbridge, Kent, UK TNll OQL. However, it was not obvious that p-type conductivity in the CdTe was determined in samples that had not been subjected to an annealing procedure, which also could have been responsible for producing the p-type conductivity. Cocivera and co-w~rkers'~,'~ have also attempted to incorporate group 15 dopants into CdTe, in a non-aqueous electrodeposition procedure. In this case, the dopant was added as an organic phosphine or arsine telluride. Again, no correlation of device properties with dopant concentration was made. This paper describes studies of impurityldopant species in electrodeposited CdTe that occur in the electrolytes either adventitiously or by deliberate addition.Deliberate and con- trolled addition of these species was employed and correlation of the presence of the additive with subsequent electrical and photoelectrical characterisation of the completed photovoltaic device was made. The effects, both beneficial and antagonistic, of the incorporation of Group 11 metals (Cu, Ag) during electrolysis are demonstrated. Also, the crucial importance of including C1 in both CdS and CdTe electrolytes is demonstrated. Experimental CdS films were grown electrochemically from solutions con- taining 0.2 mol dm-3 Cd2+ and 0.009 rnol dm-3 s@-onto tin-doped indium oxide (ITO, Hoya Corporation, Tokyo, thickness: 2500 A; resistivity: 8 square-') at 90 "C. The initial potential for CdS growth was -0.695 V us. SCE.The details of the process have been documented e1~ewhere.I~ Typical CdS filmo thicknesses for 60 min growth were in the range 600-1000 A. CdTe was electrodeposited from aqueous solutions containing 0.5-1.0 mol dm-3 Cd2+ and 3 x mol dmV3 HTeO;. Control of the growth is achieved by monitoring the 'quasi rest potential' (QRP) during deposition according to Panicker et a1." A QRP of -0.50+0.05 V us. Ag/AgCl was used throughout, and required the electrode potential to be set at -0.70k0.3 V. Typical current densities during CdTe deposition were of the order 0.3 0.05 mA cm-2 and a CdTe film of ca. 2 pm thickness was deposited in 2-3 h. Two counter-electrodes were employed: one a spectro-graphically pure carbon rod, the other was cast from Lihigh-purity Te ingot.The electrochemical control system enabled switching between each anode in turn so that [HTeO,:] was kept constant. The reference electrode was a double-junction Ag/AgCl electrode, where the outer compartment was filled with the same concentration of purified CdS04 electrolyte as that in the deposition electrolyte. The reference electrode was J. MATER. CHEM., 1994, VOL. 4 located in a side-arm, connected by a Luggin capillary to the main body of the cell. For both CdS and CdTe deposition, Cd2+ solutions were purified by prolonged (ca. 8 h) pre-electrolysis at a potential close to the Cd deposition potential, such that a small quantity of Cd metal was deposited during the electrolysis.Alternatively, the solutions were stirred with high-purity (> 5 ‘9s’ purity) Cd powder, followed by careful filtration. HTe02f was added to the solution electrochemically, by anodic dissolution of high purity Te (5 ‘9s’ purity) ingot. The as-deposited CdTe was n-type and conversion to p-type was achieved by annealing the CdS/CdTe in air at 400-450 “C for 8-10 min. Additions of dopants/impurities were made as follows: C1-, by addition of purified 0.2 mol dm-j CdC1, solution; Cu and Ag, similarly by addition of more dilute stock solutions (CuSO, and AgNO,, respectively). Electrolyte analysis was carried out using both atomic absorption (AA) and inductively coupled plasma (ICP) emis- sion techniques. The best performance was obtained with the ICP method, and all analytical data quoted are by this method unless otherwise stated.As noted above, the effect of chloride ions on the properties of the CdS/CdTe device was studied. The presence of C1- in the electrolyte rapidly depleted any added Ag’. As a result ICP analysis was carried out immedi- ately following addition of the Ag’, in order to minimise any loss of the ion. Some analysis of C1 and Cu in CdTe deposits was carried out with secondary ion mass spectrometry (SIMS), using an argon ion beam to generate secondary ions from the CdTe surface. Depth profiles were converted from counts to atomic concentrations (atoms ~rn-~) using suitably ion-implanted CdTe crystals as standards. Small-area photovoltaic devices (0.01 cm2) were fabricated by a sequence of chemical etching steps and vacuum depos- ition of Au contacts through a mask.Electrical characteris- ation of devices was carried out using a computer-controlled system that enabled current-voltage characteristics to be obtained both in the dark and under illumination. Illumination was from a 250 W tungsten-halogen source adjusted to give 100 mW ern-,, using a standard silicon cell as reference. A selection of devices from the top, middle and bottom of the electrodeposited CdTe films were analysed to ensure that possible spatial variation in performance did not affect the results. Results and Discussion Previous reports of electrodeposited CdTe photovoltaic cells have demonstrated that device efficiencies in excess of 10% (optical-to-electrical energy) conversion are routinely achiev- able, with appropriate control of electrodeposition pro-cedure~.~,~These results were achieved with devices of area ranging from 0.01 to several hundred cm2.However, in order to make rapid assessment of the device properties, only 0.01 cmz devices were examined in this work. The electrical characterisation of the CdS/CdTe solar cells under illumi- nation gave values of the open circuit voltage (Voc), short circuit current (Isc), fill factor (FF) and overall efficiency of the device. In addition, the series resistance and reverse saturation current density were determined from current-volt- age measurements in the dark and the shunt resistance of devices determined on illuminated devices under reverse biassed conditions.Theoretical considerations of the CdS/CdTe photovoltaic device indicate that the maximum I,, for CdS/CdTe cells should not exceed ca. 25 mA cm-, under 100 mW cm-2 irradiance. In many cases, particularly with small-area devices, current collection may extend beyond the defined contact area and yield I,, well in excess of this value. The V,, is independent of device area and therefore is a more reliable measure of device performance. In addition, the Voc is also sensitive to effects occurring in the p-n junction region and can be used as a measure of these effects on device performance. As will be shown below, a Voc>0.72 V indicates good performance and poor performance is indicated when the V,, falls below 0.70V.Although there may be some unreliability in the absolute values of Isc, important infor- mation on trends occurring in the device can be derived from variations in this parameter. The behaviour of commonly occurring electrically active defects, dopants and impurities in CdTe has been Because the species described in this work cover a wide variation both in chemical properties and behaviour in CdTe, each will be considered separately. Chloride The effect of chloride on the properties of the electrodeposited CdS/CdTe device were studied using a matrix of experiments where the [Cl-] was varied during the preparation of each semiconductor. In the case of CdS, deposition is generally carried out from solutions containing CdCl,, where [Cl-] is always twice that of [Cd2+].In this work [Cl-] was set to either 0, using 0.2 rnol dmP3 CdSO,, or 0.4 mol dm-3 using 0.2 mol dm-, CdC1,.’8 The importance of halide ions in producing CdTe of suitable properties for photovoltaic appli- cations was first alluded to by Basol et d2’,who showed a direct relationship between [Cl-] in the deposition in the electrolyte and device performance in terms of Is,-, for concen- trations in the range 0-0.015 mol dmP3. The effect of adding C1-to the electrolyte in both CdS and CdTe electrodepositions is shown in Fig. 1 and 2. Fig. 1 shows the variation in power conversion efficiency as [Cl-] is varied from 0 to 300 ppm (0-0.0085 mol drn-,) in the CdTe electro- lyte.Curve (a) is for CdTe grown onto chloride-free CdS (grown form CdSO, solution) and (b) for CdTe deposited onto CdS from CdC12 solution. Two important effects can be seen in these data. First, device efficiency shows a strong dependence on the inclusion of C1- in the CdTe electrolyte. If the curve for C1--containing CdS is considered, there is an *-it 0 100 200 300chloride concentration (ppm) Fig. 1 Plot of CdS/CdTe device efficiency as a function of [Cl-] in the CdTe deposition electrolyte, (a) CI--free CdS; (b)CI--containing CdS J. MATER. CHEM., 1994, VOL. 4 43 of the CdTe films. In CdTe films which had been deposited 0.81 0.4v o.3! 0.21 I 1 1 I 1 0 100 200 300 chloride concentration (ppm) Fig.2 Plot of open-circuit voltage as a function of of [Cl-] in the CdTe deposition electrolyte, (a)Cl--free CdS; (b)Cl--containing CdS increase in efficiency from 7.5% at 7 ppm c1- to 10.5 +0.5% at 300 ppm C1-.In other, similar experiments where Cl--free CdTe was deposited onto C1--containing CdS, devices with efficiencies as low as 3% were produced. This also was followed by an increase to 10% conversion efficiency on addition of 300ppm Cl- to the CdTe electrolyte. A similar absolute increase in efficiency is observed over the same range of [Cl-] for cells where the CdS is free of Cl-. In this case initial efficiency was ca. 3% and increased only to 5-670 on reaching a concentration of 300 ppm in the electrolyte (The point at ca.200ppm C1-shows an unexpectedly high efficiency and V,, for Cl--free CdS, for which there was no explanation). The statistical significance of the observed vari- ations in conversion efficiency was tested by analysis of the efficiency data obtained for a large sample (N=70) of CdTe devices deposited under strictly defined and controlled con- ditions (see later). This sample gave a mean conversion of 10.68% with standard deviation (s) of 0.68%. Each point in Fig. 1 was an average of 3 measurements for which 95% confidence limits can be estimated as i-(1.96~/,/3). This value has been included in the figure as an error bar, equivalent to -+0.77%. The variation in conversion efficiency, particularly over the initial stages of C1- addition, is substantially greater than this error bar and therefore must be directly attributable to the addition of C1-.The difference in the behaviour of cells where the CdS does not contain C1- is further emphasised by the variation of V,, with [Cl-] shown in Fig. 2. In the case of the V,, parameter, devices where C1- is present during the deposition of CdS show only a slight variation over the whole range of [Cl-1, but always have a value of 0.72V or more. For cells with Cl--free CdS, much lower V,,s are observed over the whole range of [Cl-] in the CdTe electrolyte. These two observations indicate that C1- contributes in two distinct ways to the operation of the CdS/CdTe cells. First, inclusion of C1-from the electrolyte during CdTe deposition results in an overall increase of device efficiency, irrespective of whether or not C1-was included in the deposition of CdS.Confirmation that C1- was actually incor- porated into CdTe was achieved from Auger depth profiling from Cl--free electrolytes, which showed efficiencies in the 3-7% range, no C1- could be detected throughout the depth of the film above the noise level of the experiment. This corresponded to a [C1-]<0.2% in the CdTe. For films deposited from electrolytes containing 300 ppm C1-, the [Cl-J found throughout the depth profile was 0.5 k0.2 atom%, confirming the presence of C1- in the CdTe films. Considering first the devices with C1--containing CdS, the increasing efficiency is not associated with a significant increase in Voc.Even in the effective absence of C1- from the CdTe electrolyte, the V,, is 0.72 V, which represents behaviour characteristic of good performance. The major improvement in device perform- ance appears in the Isc, as shown in Fig. 3. This shows that I,, increases from 0.18 to 0.24 mA as the [Cl -1 increases from 7 to 300ppm. The effect of C1- inclusion in the CdTe is to improve current transport in the device. This may be by either lowering the bulk resistance of the CdTe, or countering the activity of traps in the polycrystalline semiconductor. Measurements of the series resistance of the devices did show some reduction of the overall resistance, but as this represented the resistance of the whole device and did not separate out the contribution from CdTe alone, it was not possible unam- biguously to determine the origin of the underlying improve- ment in the device.It is important to note from Fig. 3 that similar improvements in Is, are also evident in devices ccuntain- ing Cl--free CdS, indicating that the same mechanism of improvement of material properties is occurring, even though the overall efficiency remains substantially inferior in this case. The second effect of C1- is localised in the junction region of the device. As noted above, in the case of cells fabricated from Cl--containing CdS, improvement in efficiency is not accompanied by a comparable improvement in Voc. Also, with the exception of the odd point for ca. 200 ppm C1- ,V&s for cells with Cl--free CdS are all well below 0.6 V.It can be seen that the presence of Cl- in the underlying CdS layer of the device is important in efficient junction formation between the n-CdS and p-CdTe. This is the result of the annealing step required for the type conversion of the as-deposited n-CdTe to p-type. This step is carried out at temperatures in the range 0.3I I0.1I I I 3 I 0 100 200 300 chloride concentration (ppm) Fig. 3 Plot of short-circuit current as a function of of [Cl-] in the CdTe deposition electrolyte, (a) Cl--free CdS; (b)Cl--containing CdS 400-450°C, and diffusion of [Cl-] from the CdS into the CdTe will be quite rapid under such conditions. These results show that the presence of (an unquantified amount of) Cl- in the CdS film and a carefully controlled Cl--content in the CdTe (0.5 atom%) are a prerequisite for the production of highly efficient CdS/CdTe solar cells.This finding is exemplified in the work of Lyons et al. on the CdS/CdTe solar cell, who made a thorough study of the impurities in electrodeposited CdTe," yet produced devices with V,, and I,, substantially lower than those quoted in this work.12 It is apparent, from their own analytical data, that these workers did not include C1-in their CdTe electrolyte and as a result failed to achieve highly efficient devices, despite the use of rigorous purification employed in their CdTe electrodeposition procedure. One question remains regarding the importance of C1- in this device, with particular reference to CdTe.C1- is known as a shallow donor in CdTe'9,20 and would be expected to promote n-type conductivity in this material. It remains to be seen why the presence of C1- is required to produce good quality p-type CdTe. Basol et only specify that a halide is required in the deposition electrolyte to provide the enhanced behaviour of the CdTe films. A similar study with Br -resulted in CdS/CdTe devices which exhibited no photo- voltaic activity at all. It appears that C1-is a specific requirement for high efficiency CdS/CdTe devices. Copper It was noted earlier that the direct electrodeposition of p-type CdTe had been achieved by the inclusion of Cu2+ in the electr01yte.l~ The actual amounts of Cu2+ required to achieved this were not specified.This prompted the study of deliberately added Cu2' to determine the possible benefits on the perform- ance of the CdS/CdTe device. A further motive for this study was that, in the course of routine electrolyte analysis, Cu2+ was occasionally detected by AA, particularly when deterior- ation in device performance has been observed. ICP analysis was found to be a more sensitive method for the detection of low concentration ionic species in the 0.5 mol dm-3 CdS0, electrolytes and with careful control [Cu"] in the range 5-10 ppb could be detected with an uncertainty of k3 ppb. Prior to the deliberate addition of Cu2+ to the electrolyte, 10 CdTe depositions were made, following which the [Cu2'] was determined.This revealed 7 ppb Cu2+ in the electrolyte prior to any deliberate addition. Sufficient CuSO, stock solution was then added to the electrolyte to increase the [Cu2'] by ca. 20 ppb steps before each of a series of depositions. Samples (ca. 5 cm3) were removed before and after each deposition, to determine initial and final [Cu2']. Fig. 4(a) and (b)shows the variation of device efficiency and V,, as a function of [Cu2+], where Cu2+ was being added to and depleted from the solution. In terms of both efficiency and Voc, the effect of Cu2+ is quite dramatic, with both parameters showing a sharp decrease for [Cu2+] above ca. 30 ppb in the electrolyte. Other device parameters also showed substantial changes due to Cu2+ addition, in particular the shunt resistance of the device, which decreased from ca.500 kR at 5 ppb to < 100 kR at ca. 60 ppb. That Cu2+ had been incorporated into the CdTe film during electrodeposition was confirmed by secondary-ion mass spectrometry (SIMS) and two typical SIMS depth profiles are shown in Fig. 5(a) and (b).Profile (a)was obtained from CdTe deposited from electrolyte containing 11 ppb Cu2+ and (b)was obtained from CdTe deposited with 59 ppb Cu2+. The SIMS data for CdTe films deposited from electrolytes with a range of [Cu2+] are summarised in Table 1, which gives the calculated [Cu] in the CdTe film, based on the response from an ion-implanted CdTe crystal reference. The J. MATER. CHEM., 1994, VOL. 4 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 copper concentration (ppb) Fig.4 (a) Plot of device efficiency as a function of [Cu2+] in the CdTe electrolyte; (b) plot of open-circuit voltage as a function of [Cu"] in the CdTe electrolyte; (H)increasing [Cu"]; (0)decreas-ing [Cu2+] 1o6 I IP-1o5 I ,._.......,.I.. . .. ..........,.. . . ., *.........,..,.... . ... . . . .... . .... i1o4 ,-------A---------_--' -+ lo3 0 Q)c 2-2 lo2 loll1oo 1o -l l, ' 1 I I 0 10 20 30: 0 10 20 30 etch time/min Fig. 5 Ar+-SIMS depth profiles of electrodeposited CdTe films, (a) electrolyte with 300ppm C1- and 11 ppb Cu2+; (b)electrolyte with 300ppm C1- and 59 ppb Cu2+; (..........) Te depth profile; (-) Cd depth profile concentrations of chloride in the films are also in this table.Although the data for [Cu] in the CdTe films do not correlate exactly with the trends in electrolyte [Cu2+], there can be no doubt that Cu has been incorporated into the films during electrodeposition. Comparison of the concentrations of Cu and C1 in the films indicates the different incorporation mechanisms for each species. At the potentials employed for J. MATER. CHEM., 1994, VOL. 4 Table 1 Dopant/impurity concentrations in CdTe films (determined by SIMS) and deposition electrolytes with associated CdS/CdTe device efficiences [Cu2’] in electrolyte [Cl-] ion electrolyte [Cu] in CdTe/ cm3 [Cl] in CdTe/ cm3 efficiency (PPb) (PPm) Po) 5 300 3-4 x1019 6 x1017 10.9 27 300 1020 4-5 x1017 10.3 59 300 1O2O 4 x1017 3.6 38 300 2 x1O2O 5 x1017 4.1 11 300 2-3 x1019 4 x1017 10.8 CdTe deposition, -0.75 &0.3 V us.Ag/AgCl, Cu2+ is reduced under diffusion control, and [Cu] in the film would be expected to be large and proportional to [Cu2+] in the electrolyte. On the other hand, there is no direct electrochemi- cal reaction for C1- at the operating potentials used and C1 incorporation probably occurs by association of Cl- with the Cd2+ ion. As a result the measured concentrations of C1 in the CdTe films is substantially lower than Cu, despite the very large [Cl-] in the electrolyte compared to [Cu2+]. The appearance of a threshold [Cu”] in the electrolyte, above which results in the production of increasingly poorly performing devices, indicates that there is a transition in the behaviour of Cu in the CdTe.At concentrations below ca. 30 ppb, device performance is good, with very large Vocs, in the 0.75-0.76 V range. In this concentration range Cu is probably incorporated substitutionally into the host lattice and acts as an acceptor in the CdTe, or at least as an inactive species. At [Cu2‘1 above the 30 ppb threshold, incorporation interstitially and behaviour as a donor species in CdTe begins. As a result, higher [CU”]~ only serve to compensate the initially ben- eficial contribution from the substitutional Cu. It should be noted also that the SIMS depth profiles show an accumulation of Cu at the CdS/CdTe interface.The importance of this is that Cu is known to diffuse very rapidly through CdS22 and is the major cause of the degradation of the CdS-Cu,S solar cell. The presence of a concentration of Cu at the CdS/CdTe interface may result in Cu diffusion through the CdS layer to the front contact of the device. The overall effect of this is to make the device more susceptible to breakdown, a fact manifested by the decrease in shunt resistance on increasing the [Cu2’] in the deposition electrolyte. It is apparent from the variation of V,, with [Cu”] in Fig. 4(b) that high V&s may be achieved before deliberate additions of Cu2+ were made to the electrolyte. This is not to say that Cu2+ is not necessary to achieve optimum CdTe properties. There may be sufficient Cu2 entering the electro- + lyte by an adventitious route to provide the correct Cu doping level in CdTe.At no time during the ICP analysis of the CdTe electrolyte was the [Cu2+] found below ca. 5 ppb. This probably represents the minimum detection limit for Cu2 in+ this matrix, but also indicates that Cu2+ may always be present at very low concentrations in the CdTe electrolytes. The origin of this very low background level of Cu2+ in the electrolytes is almost certainly the tellurium anodes used to add HTeOz to the electrolyte. Te is produced as a by-product of the electrowinning of copper; therefore, the appearance of minute quantities in even highly purified Te is not surprising. Better process control may be achievable by using TeO, as the tellurium source, and removing the potential source of Cu in the system.Silver Silver, like its Group 11 congener copper, is also defined as an acceptor in CdTe.” Silver is also notable by its presence 45 in the reference electrode used in the control of CdTe elec- trodeposition. In order to determine the potential influence of Ag + contamination from the reference electrode, additions of AgNO, were made to the electrolyte, in the same as for copper. The effects of different concentrations of Ag’ are shown in Table 2. As with Cu2’, analysis of concentrations around 5 ppb became uncertain and a detection limit of 5 ppb was determined for Ag+ also. Although the initial addi- tion of Ag+ was excessive, it can be seen readily that.as it is removed from the solution by electrolysis, the effect of Ag+ on the CdTe is more severe than with copper. At [Ag’] =28 ppb, both efficiency and V,, show large decreases compared to the Ag+-free case. The data in Table2 were found to be repeatable and as with copper additions, the impuritv could be removed by successive CdTe depositions until the high performance condition of the electrolyte was regained. The data in the table show that even at concentrations as low as 10 ppb, Ag’ has an antagonistic effect on the CdS/CdTe device, with some penalty in both efficiency and V,, These results indicate that there is no measurable [Ag’] at which a beneficial presence of Ag’ could be determined and it appears only to act as a poison in CdTe.From these results an upper limit on [Ag+] in the CdTe electrolyte was set at 5 ppb, the detection limit in this electrolyte by ICP. This was found to be readily achievable, since if the reference electrode was removed from the solution after each deposition, then there was no other source from which Ag contamination could occur. Management of CdTe Electrodeposition The control of the dopants and impurities noted above has been shown to be essential in the production of highly efficient CdS/CdTe solar cells. Fig. 6 shows a plot of device effirciency for a series of 100 electrodeposited CdTe films from the same electrolyte. This corresponds to a period of 4-5 months continuous daily use. The filled points correspond to CdTe deposited under conditions where Cu2+, Ag’ and C1- were controlled according to the criteria defined by the findings of this work, i.e.[Cu2+]<30 ppb, [Ag+]<5 ppb and [(TI= 300ppm. The open circles correspond to conditions where depositions were carried out with deliberate alterations to these standard conditions, by either Ag or Cu2+ addition or + other process variables. The plot shows that the efficiency of only one CdTe film drops below the line corresponding to 9% efficiency. This was a value of device efficiency defined as indicating unsatisfactory performance, and would initiate examination of the overall process to determine the source of poor performance. When the deliberately non-standard depos- itions are rejected, this corresponds to a reproducibility of 98% for CdTe films giving an efficiency 9% or ab0t.e and Table2 CdS/CdTe device parameters as a function of [As’] in the electrolyte [Ag’] in electrolyte (ppb) efficiency (YO) VOCP <5 11.0 0.732 400 0.7 0.558 28 0.9 0.557 14 2.7 0.612 10 9.4 0.698 <5 11.0 0.746 <5 11.6 0.744 -51 OI1 B0 0 0 t 0 yo, I II I I 0 20 40 60 80 100 deposition number Fig.6 Plot of CdS/CdTe device efficiency for the series of films 1-100 from a single electrolyte over a ca. 5 month period, (m) standard deposition conditions; (0)depositions with deliberately added Cu2+ and Ag+ indicates that electrodeposition can become a robust process for the production of solar-grade thin-film semiconductors, with reproducible production over an extended period of operation.Further refinements of the process can be envisaged which would increase the robustness of the process. These are: the use of a non-contaminating reference electrode, e.g. Cd/Cd2+ (ref. 11) and Te addition using Te02. The replacement of the carbon electrode might also be considered, since the rods used were somewhat porous and could absorb impurities and release these over a period of time. Care will be required to determine a suitable replacement anode material, since noble- metal anodes are known to dissolve in acidic sulfate solutions, even in the absence of Cl-,23 and would present an additional source of potentially poisoning species.Conclusions This study has made the first correlation between the efficiency of electrodeposited CdS/CdTe thin-film photovoltaic devices and the presence of dopant or impurity species in the depos- ition electrolytes. The results show that C1- incorporation into both CdS and CdTe films is essential for optimum operation of the device. Also, in terms of formation of the active junction between the two semiconductors, it is more important that C1- is incorporated into the CdS film than CdTe. Further work is necessary to determine the mode of activity of C1- in these devices. The correlation of device efficiency with the concentration of Cu2+ and Ag' in the J. MATER. CHEM., 1994, VOL. 4 electrolyte has enabled upper limits for the presence of these species in the CdTe electrolyte. These limits are 30 and 5 ppb respectively.The transition of the behaviour of Cu in CdTe from dopant to impurity at the 30 ppb (ca. 10'' cmP3 in the CdTe film) limit is probably due to a change from substi- tutional to interstitial incorporation into the CdTe lattice. The author thanks Professor D.E. Williams for encouragement and support, in particular for the provision of facilities at UCL. References 1 B. M. Basol, E. S. Tseng and R. L. Rod, Proc. 26th IEEE Photovoltaics Specialists Conference, San Diego. 1982, IEEE, New York, 1982,p. 805. 2 B. M. Basol, E. S. Tseng and R. L. Rod, 4th EC Photovoltaic Solar Energy Conference, 1982, Kluwer, Dordrecht, 1982, p. 719. 3 A. K, Turner, J. M. Woodcock, M.E. Qzsan, J. G. Summers, J. Barker, S. Binns, K. Buchanan, C. Chai, S. Dennison, R. Hart, D. Johnson, R. Marshall, S. Oktik, M. Patterson, R. Perks, S. Roberts, M. Sadeghi, J. Sherborne, J. Szubert and S. Webster Sol. Energy Mazer., 1991,23, 388. 4 A. K. Turner, J. M. Woodcock, M. E. Qzsan and J. G. Summers, in Proc. of 10th European Photovoltaic Energy Conference, Lisbon, 1991, Kluwer, Dordrecht, 1991, p. 791. 5 P. V. Meyers, Sol. Cells, 1988,24, 35. 6 P. V. Meyers, Sol. Cells, 1988,24, 59. 7 M. Takahashi, K. Uosaki and H. Kita, J. Electrochem. SOC., 1984, 131,2304. 8 M. Takahashi, K. Uosaki and H. Kita and Y. Suzuki, J. Appl. Phys., 1985,58,4292. 9 M. Takahashi, K. Uosaki and H. Kita and S. Yamaguchi, J. Appl. Phys., 1986, 60, 2046. 10 M. P. R. Panicker, M. Knaster and F. A. Kroger, J. Electrochem. Soc., 1978,125, 568. 11 L. E. Lyons, G. C. Morris, D. H. Horton and J. G. Keyes, J. Electroanal. Chem., 1984,168, 101. 12 W. J. Danaher, L. E. Lyons and G. C. Morris, Appl. SurJ Sci., 1985,22123,1083. 13 A. Darkowski and M. Cocivera, J. Electrochem. SOC., 1985, 132, 2768. 14 J. von Windheim and M. Cocivera, J. Electrochem. SOC., 1987, 134, 440. 15 R. N. Bhattacharya, K. Rajeshwar and R. Noufi, J. Electrochem. SOC.,1985,132,732. 16 R. N. Bhattacharya, K. Rajeshwar, J. Appl. Phys. 1985,58,3590. 17 J. Llabres and V. Delmas, J. Electrochem. Soc., 1986,133,2580. 18 S. Dennison, Electrochim. Acta, 1993,38, 2395. 19 T. L. Chu, Curr. Top. Photovoltaics, 1988,3,235. 20 K. Zanio, Cadmium Telluride, vol. 13 of Semiconductors and Semimetals, ed. R. K. Willardson and A. C. Beer, Academic Press, New York, 1978. 21 B. M. Basol, E. S. Tseng and D. S. Lo, US Pat. 4,629,820, 1986. 22 F. J. Bryant, A. K. Hariri and C. G. Scott, Absir. 6th Eur. Photovoltaic Solar Energy Conference, London, 1985, p. 281. 23 D. A. J. Rand and R. Woods, J. Electroanul. Chem., 1972, 35, 209. Paper 3/04268E; Received 20th July, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400041

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4( l),47-50 Solid-state Reaction between Molybdena and Alumina: Effect of Water Vapour Pressure on the Dispersion and Nature of the Supported Phases Margarita del Arco: Silvia R. G. Carrazan; Vicente Rives,*a Francisco-Javier Gil-Llambiasb and Pilar Malet" a Departamento de Quimica Inorganica, Universidad de Salamanca, Facultad de Farmacia, 37007-Salamanca, Spain Departamento de Quimica, Universidad de Santiago de Chile, Chile " Departamento de Quimica Inorganica, Universidad de Sevilla, Facultad de Quimica, Sevilla, Spain The dispersion and nature of surface species formed upon calcination of Mo03-y-Al,0, mixtures at 770 K for 10 h and with different molybdena loadings under different water vapour pressures [P(H,O)] have been studied by X-ray diffraction (XRD), zero-point charge (ZPC) measurements and temperature-programmed reduction (TPR); P(H,O) was in the range 25-45 Torr with MOO, loadings up to 2 monolayers (1 monolayer=0.1681 g MoO,/g A1203). The intensity of the most intense XRD peak of MOO, at 326 pm decreased as P(H,O) increased, indicating an increase in the concentration of surface species different from bulk MOO,.Up to 0.7 monolayer MOO, loading, covering of the Al,03 surface with Mo-containing species increases with increasing P(H,O); ZPC at P(H20)=45 Torr was coincident with that of MOO,. P(H20)has no effect on the dispersion of high loaded samples (above 1 monolayer), where all the support surface is covered even for P(H,O) =25 Torr. TPR results indicate the presence of species with different reducibilities, depending on P(H,O) and MOO, loading; their nature has been assessed by comparison with previous results for samples obtained by impregnation; bulk MOO, (responsible for high-temperature reduction peaks) formation has been observed at high MOO, loadings. Catalysts containing molybdenum supported on y-A1203 are widely used in petroleum chemistry for hydrodesulfurization, hydrodenitrogenation and hydrodeoxygenation rea~ti0ns.l~~ The formation of molybdena monolayers or multilayers in Mo03/y-A1203 samples obtained by different impregnation techniques has been widely ~tudied;~?~ however, information on the distribution of molybdena on the alumina surface of samples prepared by mechanical mixing, especially when these samples are calcined in the presence or in the absence of water vapour, is scarce in the literature.With regard to catalysts prepared by mechanical mixing, Knozinger and co-workersc8 have reported that spreading and chemical transformation of the active oxide (MOO,) over the support (e.g. A1203,TiO,) are independent processes. By using ion scattering spectroscopy (ISS), they have demon- strated that the spreading process clearly occurs during ther- mal treatment both in the presence and in the absence of water vapour. Under the experimental conditions used by these authors (720 K and a water vapour pressure of 24 Torr), the spreading process seems to be complete on A120, as well as on TiO, in less than 5 or 10h, respectively. Similar results with respect to molybdena spreading have been reported by us9 from quantitative X-ray diffraction measurements, but electrophoretic measurements have shown that a calcination time of 8-10 h is not enough to achieve a constant percentage of dispersion on the support surface, either in the presence or in the absence of water vapour.On the other hand, by using laser Raman spectroscopy (LRS),Knozinger and co-workers' have shown that a surface transformation of MOO, into polymolybdate takes place only in the presence of water vapour, the rate of this process being strongly dependent on the water vapour pressure. Complete transformation of molybdena into polymolybdate species takes ca. 30 h. Similar results have been obtained by us9p12 in samples prepared by mechanical mixtures by using LRS and TPR techniques. In this paper we report additional results on the spreading process in samples prepared by mechanical mixtures and calcined in the presence of an O2flow saturated with different water vapour pressures (25, 35 and 45 Torr). The structures of the metal oxide overlayers formed by molybdenum on the alumina surface are analysed, taking into account the TPR and the ZPC results for the samples.Experimental An Aluminoxid C y-alumina from Degussa (batch RV005) with a BET specific surface area of 105 m2g-' was used as support, and was calcined overnight at 770K to eliminate adsorbed organic impurities. MOO, was prepared by decomposition of (NH4),Mo,0,4-4H,0 (AHM, from Carlo Erba) at 770 K in air for 5 h.The supported Mo samples were prepared by manually grinding the support and the amount of MOO, required for loading of 0.4,0.7,1 and 2 monolayer. The monolayer capacity (16.81 g Mo0,/100 y-A120,) was calculated from the specific surface area of the support (105 m2 g-') and the area covered by a 'molecule' of MOO,, 15 x lo4pm2,13 that is, the so-called 'geometrical monolayer'.14 These mixtures were hand-ground in an agate mortar for 20 min and then calcined at 770 K for 10h in a quartz reactor under an oxygen flow of 30cm3 min-(from Sociedad Castellana de Oxigeno, S.C.O., Spain, 99.95%, passed through a Superpure gas Purifier model H, from Alltech, to remove impurity traces) saturated with water vapour at 25,35 and 45 Torr.Chemical analysis of the calcined samples indicated that the calcination treatment led to no change in Mo content through sublimation. Naming of the samples, according to the molybdena content and the water vapour pressure used, are summarized in Table 1. Samples are designated as nMA-p, where M =MOO,, A =A1203, n =MOO, content (number of monolayers) and p =water vapour pressure (Torr). XRD patterns were recorded on a Siemens 11-500 diffractometer with graphite-filtered Cu-Ka, radiation ( 154.05 pm) and interfaced to a DACO-MP data-acquisition micro- processor provided with Diffract/AT software. Zeta potential Table 1 Dispersion percentage, isoelectric point (IEP), zero-point charge (ZPC) data and hydrogen consumption during temperature programmed reduction (TPR) of the samples studied sample initial" MOO, lost IEP ZPC H,/Moc -A1Z03 8.8 -MOO, 2.9 -3.10 0.4MA-25 6.72 67 4.5 -7.0 2.64 5.8 -6.6 -0.4MA-35 6.72 87 0.4MA-45 6.72 90 6.0 -6.0 2.29 0.7MA-25 11.76 47 5.5 -6.5 2.52 0.7MA-35 11.76 61 7.2 -3.9 -0.7MA-45 11.76 83 9.8 -3.1 2.37 1MA-25 16.81 64 10.8 -3.1 2.76 1MA-35 16.81 68 11.4 -3.1 -1 MA-45 16.81 74 12.5 -3.1 2.43 2MA-25 33.62 37 12.5 -3.1 2.65 2MA-35 33.62 37 12.5 -3.1 -2MA-45 33.62 37 12.5 -3.1 2.53 "Initial mass of MOO, (g/100 g y-Al,O,); bDispersion (ratio between mass of MOO, lost and initial mass of MOO,); 'Molar ratio between H, uptake and Mo content.(or Zero-point charge, ZPC) measurements were carried out in a Zeta-meter Tnc. instrument (Model ZM-77) using 200 mg of samples (average size ca.2 pm) dispersed in 200 ml of lo-, moll-' KC1 solutions. The pH was adjusted with either mol 1-' KOH or HCl solutions. The zeta potentials were obtained from electrophoretic migration rates using the Smoluchowski equation." TPR diagrams were recorded in a conventional apparatus with a catharometric detector, using a (S%)H,-Ar mixture (from S.C.O.) as carrier gas, with a flow of 50ml min-' and at heating rate of 10 K min-'. Good resolution of the different reduction steps under these experimental conditions was ensured by using sample weights containing ca. 100pmol of MoO,.I6 Results and Discussion The degree of dispersion of MOO, on the A1203 support has been calculated from XRD measurements, following the method previously reported in the literat~re.".'~ Table 1 lists the values of the dispersion percentage and 'lost MOO,' (hereafter LM, corresponding to MOO, species widely dis- persed on the surface and so undetectable by XRD as MOO, crystallites) after calcination of the samples for 10 h in the presence of different water vapour pressures (25, 35 and 45 Torr).The dispersion percentage steadily increases as the water vapour pressure does, indicating a continuous 'loss of MOO,'; such an increase is much more evident in samples with molybdena loadings of 0.4 and 0.7 monolayers. In samples with a higher MOO, content (1 monolayer), the amount of LM shows a slight increase with the water vapour pressure, while in samples 2MA it remains constant as the water vapour pressure changes.It is worth noting that, while for samples 0.4MA and 0.7MA, the amount of LM sharply increases with the water vapour pressure up to a value close to the initial content of MOO, existing in the sample, LM for samples 1MA only show a slight increase with the water vapour pressure, reaching a maximum value of 12.0 g MoO,/lOO g y-Al,O,. This value of LM remains constant for samples 2MA whichever the water vapour pressure, which indicates that the maximum dispersion capacity for these samples has been attained under these experimental conditions. It is also interesting to compare the value of LM for J. MATER. CHEM., 1994, VOL.4 different molybdenum loadings at a given water vapour pressure (Fig. 1). A slight increase in the amount of LM is observed on increasing the MOO, content from 0.4 to 0.7 monolayers for p=25 Torr. When the molybdena loading is increased from 0.7 to 1 monolayer, this value is twice as high, and still slightly increases from 1 to 2 monolayers. A similar behaviour is observed for samples treated at 35 Torr. However, samples prepared under a water vapour pressure of 45 Torr show a different behaviour: a steady increase in LM is observed from 0.4 to 1 monolayer, while the value remains constant from sample 1MA to sample 2MA. Note the high value for LM, even at 25 Torr, for sample 0.4MA; this fact must be attributed to the low MOO, content in these samples (6.72g MoOJ100g y-Al,O,), below the value corresponding to the upmost dispersion on Al,O, (12 g MoO,/lOO g y-A1203).10,17*18On the other hand, if dispersion data for samples containing 0.7 monolayer are compared with those found in MoO,/Al,O, samples prepared by mechanical mixtures and calcined in a static, uncontrolled atmosphere in an open crucible during 24 h,9 it is apparent that a calcination time of 10 h is not enough to attain the dispersion correspond- ing to the geometrical monolayer capacity even for samples treated at the highest water vapour pressure (45 Torr).Electrophoretic migration and TPR measurements have been used for a better understanding of the relationships existing between the dispersion of MOO, on Al,O, and the water vapour pressure.ZPC values for all Mo03/A1,03 samples as well as the isoelectric points (IEP) for MOO, and y-A1203 are included in Table 1; changes in the ZPC values as a function of the water vapour pressure and MOO, loading have been plotted in Fig. 2. The ZPC values for MA-25 samples (Fig. 2) show a steady decrease from the sample containing 0.4 monolayer to sample 0.7MA-25,while a further increase in the MOO, content produces a sharp decrease in the ZPC to 3.1, a value that remains constant for molybdena loadings ranging from 1 to 2 monolayers. In contrast to samples MA-25, those prepared at p=35 or 45 Torr show a sharp decrease in ZPC when the Mo loading increases from 0.4 to 0.7 monolayers.It is interesting to note that at the different water vapour pressures used in this work, values of 3.1 for the ZPC are achieved at Mo loadings of 0.7 or 1 monolayer, and always remain constant when the Mo content is further increased. Taking into account that the ZPC variation in all the samples lies between 7 and 3.1, that the IEP of unloaded y-Al,03 is 8.8, and assuming that bulk and supported MOO, 14 r c. -0 Fig. 1 Dependence of the amount of 'lost MOO,' (g/100 g y-Al,O,) for samples with different molybdena loading with the water vapour pressure during calcination: (0)25 Torr; (0)35 Torr; (V)45 Torr J. MATER. CHEM., 1994, VOL. 4 0.0 0.5 1.0 1.5 2.0 2.5 Mo content (monolayers) Fig.2 Change in ZPC of the samples with different molybdena loading calcined under different water vapour pressures: (0)25 Torr; (0)35 Torr; (V)45 Torr exhibit the same JEP (3.0), the decrease of the ZPC observed for the three series of samples should give a measure of the increasing coverage of the surface of alumina by molybdena.The behaviour observed in these samples is in agreement with the equation which relates the ZPC of a support-supported phase system with the isoelectric points of the pure support and the supported species.lg For samples 0.4-MA, the ZPC values show (Fig. 3) a steady decrease as the water vapour pressure increases, and the value calculated when the sample is calcined at 45 Torr for 10 h coincides with that found in samples with the same Mo loading, but calcined in an uncontrolled atmosphere for 8 h.' This dropping in the ZPC values can be interpreted as follows: the increase in the water vapour pressure favours the continuous formation of surface Mo species which are expected to have a value of IEP <<8.8, as it was evidenced in MoO3/A1,O3 samples with an Mo loading of 0.4 monolayer prepared by mechanical mixtures and calcined for 24 h in an uncontrolled atmosphere.In these samples, only polymolybdate species were detected by LRS.'O-'2 These results are in agreement with those found by XRD, which indicate a slight and continuous 'loss of bulk MOO,' as the water vapour pressure increases. A decrease in the ZPC values is also observed in 0.7MA samples when water vapour pressure is increased, although in 20 30 40 50 water pressure/mmHg Fig.3 Change in ZPC of the samples prepared under different water vapour pressures as a function of the molybdena loading: (0)0.4 monolayers; (e)0.7 monolayers; (V)1 and 2 monolayers this case the variations are sharper than in the case of 0.4MA samples, specially in the 25-35 Torr range. A value of 3.1 for sample 0.7MA-45 indicates that the y-Al,O, surface is com- pletely covered by surface Mo species, the surface concen- tration of which is larger than that of bulk MOO,, as it was put into evidence by the LM values calculated by XRD. The ZPC values for samples 1MA and 2MA remain con- stant for all these samples whatever the water vapour pressure used to prepare them (ZPC =3.1 ).This value of ZPC indicates that the y-A1203 surface is totally covered by surface oxo- molybdenum species and bulk MOO,, but the relative concen- tration of each species cannot be determined from the ZPC results because of the constancy in ZPC for all these samples. However, the differences in the amount of these species (amorphous MOO, and bulk MOO,) in 1MA samples with water vapour pressure can be put into evidence by XRD; the amount of LM increases with increasing water vapour pressure. The nature and relative amounts of dispersed species have been assessed by TPR, since it has been shown that the degree of dispersion changes the reducibility of supported molyb- dena.9 On the other hand, a knowledge of such a reducibility is important since when acting as catalysts in selective oxi- dation processes these systems usually undergo oxidation/ reduction reactions. The TPR profiles of the samples are shown in Fig.4, and the molar H, :Mo ratios, as calculated from the Mo content and the integrated hydrogen consumptions, are included in Table 1. Reduction of bulk molybdena under these experimen- tal conditions starts at 825 K9 with a total hydrogen uptake of 3.1 H,/Mo, thus corresponding to the total reduction of Mo6+ ions to Moo. In all Mo0,/A120, samples studied in this work, reduction starts at lower temperatures (ca. 673 K), mb0 n I I I J 600 800 1000 1200 TIK Fig. 4 Temperature programmed reduction profiles calcined under water vapour pressures of 25 Torr (solid line) or 45 Torr (dotted line): (a)2MA, (b)1 MA, (c) 0.7 MA, (d) 0.4 MA J.MATER. CHEM., 1994, VOL. 4 thus suggesting the presence of well dispersed oxomolyb- denum-containing species, more easily reducible than bulk molybdena. Hydrogen consumptions (2.5 f0.2) are always slightly lower than those expected for a total reduction from Mo6+ to Moo, although it should be noted that the TPR profiles do not recover the baseline at the highest temperature attainable by our experimental system, and so this low consumption can be tentatively ascribed to the existence of a small number of species only reducible at a very high temperature. A very simple TPR profile, with a low-temperature maxi- mum at 750K and a broad high-temperature maximum centred at 1150 K is recorded for sample 0.4MA-45, where a high dispersion degree of supported molybdena was concluded from XRD data.This profile is almost identical to that recorded’ for a 0.7 monolayer Mo03/A1,0, sample calcined at 770 K for 24 h in an uncontrolled atmosphere, where only polymolybdate species were detected by Raman The TPR maximum at 750 K is recorded in the profiles of all MoO3/A1,O3 samples studied here, and its intensity can be taken as a measurement of the amount of surface poly- molybdate species existing in the samples. Thus, when the 0.4 monolayer sample was pretreated under a water vapour pressure of 45 Torr, maxima at 750 and 1150 K have similar intensities to those recorded in the 0.4MA-25 sample.The lower dispersion in this sample, as detected by XRD, gives rise to a new maximum at ca. 1050 K in the TPR profile. This new maximum should be assigned to crystalline MOO,, detect- able by X-ray diffraction. The small changes in the intensity of the maximum at 750 K associated to changes in the water vapour pressure during calcination suggest that surface poly- molybdates are readily formed in samples containing 0.4 monolayer, even at the lowest water vapour pressures used furing pretreatment, in agreement with the small changes above reported in the ZPC values for these samples. On the contrary, the changes in the TPR profile of samples containing 0.7 monolayers can be readily related to changes in the water vapour pressure during calcination.Thus, the intensity of the 750 K maximum in the 45 Torr sample is at least twice than that for the 25 Torr sample. In addition, two maxima clearly recorded at ca. 920 and 1073 K in the TPR profile of the 0.7MA-25 sample vanish when the sample has been calcined under a water vapour pressure of 45 Torr, the latter maximum decreasing its intensity and shifting to 1050 K. The maximum at 920K has been previously ascribed’ to dispersed MOO,, while, in agreement with XRD data (which indicate 53% of crystalline MOO, in this sample) the broad maximum at 1073 K should be ascribed to crystalline MOO,. These data clearly suggest that crystalline MOO,, still present in the 0.7 monolayer samples calcined under a water vapour pressure of 25 Torr, is dispersed and forms surface polymolyb- dates when the samples are calcined under higher water vapour pressures.At the highest molybdena contents (1-2 monolayers), the intensity of the TPR maximum at 750K remains constant and almost independent of the water vapour pressure, thus confirming that in these samples the alumina surface is saturated of polymolybdates even when they have been cal- cined under the lowest water vapour pressure. Maxima at 1073 and 920 K indicate the coexistence of these polymolyb- dates with crystalline and dispersed MOO,, respectively. Small changes in the profiles of the 1 monolayer samples suggest that a small amount of surface polymolybdate is still formed at the expense of dispersed Moo3 when the water vapour pressure changes from 25 to 45 Torr, while the profiles of the 2 monolayer samples remain virtually unchanged at the different water vapour pressures.Authors are grateful for financial support from DGICYT (PB91-425) and Junta de Castilla y Leon (C. Cultura y Turismo, ref. 16/02/9 1). References 1 P. Grange, Catal. Rev. Sci. Eng.. 1980,21, 135. 2 F. E. Massoth and G. Muralidhar, in Proc. 4th Int. Conf. Chem. Uses Molybdenum, ed. H. F. Barry and P. C. H. Mitchell, Climax Molybdenum Co., Ann Arbor, MI, 1982, p. 343. 3 R. Thomas, E. M. Van Dors, V. H. J. de Beer, J. Medema and J. Moulijn, J. Catal., 1982,76,241. 4 J. Sonnemaus and P. Mars, J. Catal., 1973,31.209. 5 F. J. Gil Llambias, A. M. Escudey, A. Lopez Agudo and J.L. Garcia Fierro, J. Catal., 1984,90,323. 6 J. Leyrer, M. J. Zaki and H. Knozinger, J. Phys. Chem., 1986, 90,4775. 7 M. Margraf, J. Leyrer, H. Knozinger and E. Taglauer, Surf. Sci., 1987,189/190,842. 8 J. Leyrer, R. Margraf, E. Taglauer and H. Knozinger, Surf. Sci., 1988,201,603. 9 M. del Arco, S. R. G. Carrazan, V. Rives, F. J. Gil Llambias and P. Malet, J. Catal., 1993, 141,48. 10 M. del Arco, S. R. G. Carrazan, V. Rives and J. V. Garcia Ramos, Muter. Chem. Phys., 1992,31,205. 11 M. del Arco, S. R. G. Carrazan, V. Rives and J. V. Garcia Ramos, Spectrosc. Lett., 1992,25, 73. 12 M. del Arco, S. R. G. Carrazan, V. Rives and J. V. Garcia Ramos, J. Muter. Sci., 1992,27, 5921. 13 T. Fransen, P. C. Van Berge and P. Mars, Stud. Sur- Sci. Catal., 1976, 1,405. 14 G. C. Bond, S. Flamerz and R. Shukri, Furuday Discuss. Chem. SOC.,1989,87,65. 15 R. J. Hunter, Zeta Potential and Colloid Science Principles and Applications, Academic Press, New York, 1981, p. 72. 16 P. Malet and A. Caballero, J. Chem. SOC.,Faraday Trans. I, 1988, 84,2369. 17 Y. Xie, L. Gui, Y. Liu, B. Zhao, N. Yang, Y. Zhang, Q. Guo, L. Duan, H. Huang, X. Cai and Y. Tang, in Proc. 8th Znt. Congr. Catal., Berlin, Germany, Dechema, Frankfurt, 1984, vol. 5, p. 147. 18 P. Dufresne, E. Payen, J. Grimblot and J. P. Bonnelle, J. Phys. Chem., 1981,85,2344. 19 F. J. Gil-Llambias and A. M. Escudey-Castro, J. Chem. SOC., Chem. Commun., 1982,478. Paper 3/04258H; Receizied 20th July, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400047

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4(1), 51-54 Alternative Single-source Precursor for Growth of Indium Pnictide Thin Layers Rydki Nomura,* Takayuki Shimokawatoko, Haruo Matsudat and Akio Baba Department of Applied Chemistry, Faculty of Engineering, Osaka University, Yamada-Oka, Suita, Osaka 565, Japan Novel organometallic precursors for the growth of indium pnictide thin layers are reported. Tributylindium, n -Bu,ln, reacts readily with pnictogen sulfides, Asps3, P,S, and Sb,S, to give sulfur-bridged hetero-binuclear organometallics such as [Bu,ln(SAsBu),],S (l),[Bu,lnSP(S)Bu],S (2), and [Bu,lnSSb(S)Bu],S (3).They are distil- lable oils which can be pyrolysed to give the corresponding indium pnictides without sulfur contamination, under an N, atmosphere at 500-600 "C.Dip-dry pyrolysis of these complexes on an Si(ll1) substrate gave high- quality indium pnictide thin films: lnAs (p300, 1640 cm2 V-' s-'; n300, 1.3 x lOl4 ~rn-~; 600 "C), InPtype, p; Tsub, (p300,1100 cm2 V-' S-'; nsm, 1.2 x 1019cm-3; type, p; Tsub, 600 "C), and lnSb (p300,18400 cm2 V -'S-'; n300, 6.8~10'6cm-3; type, p; Tsub,450 "C).Organometallic precursors provide excellent routes for the fabrication of advanced electronic devices by metal-organic chemical vapour deposition (MOCVD), metal-organic vapour phase epitaxy (MOVPE) and metal-organic molecular beam epitaxy (MOMBE).' Efforts continue towards the improve- ment of organometallic precursors involving production of devices based on 13-15 semiconductor epilayers of high performance, durability and reliability via a reproducible and safe process.2-6 Single-source systems, in which one molecule contains the two or more elements necessary to form the desired semiconductor, represent a rational chemical approach.7-9 Several arsinogallanes, phosphinoindanes and related binuclear organometallics have been used to prepare thin, epitaxial layers of GaAs, InP and other 13-15 com-pound~.~There are some disadvantages in the use of a straightforward single-source ~ystem.~ First, the precursor molecules exist as aggregates and have low vapour pressures and secondly, the requirement for direct bonding between group 13 and 15 elements limits severely the number of organometallic species available.The presence of a direct bond limits the diversification of the precursor molecules.In this paper we propose alternative single-source precursors containing a metal( 13)-sulfur-metal( 15) bridge for 13- 15 semiconductor layers. Experimental Materials Tri-n-butylindium was prepared via literature methods lo and distilled under reduced pressure before use. Phosphorus sulfide (P2Ss)was used after purification by Soxhlet extraction with CS2.11Other metal sulfides (As2S3, Sb2S3, Sb,S5 and Bi2S3, Aldrich or Wako, pure grade) were used as received. A piece of Si( I 11) (10 mm x 10 mm) was cut from a 5 in$ wafer. The surface of the wafer was covered by a thin oxide layer (p-type, Shin-Etsu Semiconductor Co. Ltd., Gunma, Japan) and washed with deionized water; it was then immersed in a synthetic neutral detergent solution and washed with distilled water in an ultrasound bath.Finally the piece of Si was exposed to the vapour above boiling acetone (electronics grade) and dried. t Present address, Department of Applied Chemistry, Osaka Institute of Technology, 0-Miya, Asahi, Osaka 532, Japan. $ 1 in=2.54~10-~m. Analysis Thermal analysis was performed with a SEIKO TG/DTA 20 in a flowing N2atmosphere (5ml min- '). Surface morphology was observed with a Hitachi S-800 scanning electron microscopy (SEM, acceleration voltage 20 kV). Film crystal- linity was assessed with a Rigaku RotaFlex X-ray diffractometer (XRD, Cu-Kcr, 40 kVj80 mA). Mass spectra were recorded with a JEOL JMS-DX303 spectrometer with a JMA-DA5000 data-processing system.Microanalysis for elements other than CHN was performed with a Rigaku 3270 X-ray fluorescence microanalyser. Syntheses All synthetic reactions were carried out under dry and deoxy- genated N2, and used the normal Schlenk and syringe tech- niques. Tri-n-butylindium (4 mmol, 1.14 g) was added dropwise into a suspension of pnictogen sulfide (2mmol; 0.49 g for AszS3; 0.44 g for P2Ss; 0.68 g for Sb2S3; 0.81 g for Sb2S5; 1.03 g for Bi2S3) in absolute benzene (20 ml). After reaction (80 "C, 3 h), volatiles were evaporated off into another Schlenk tube, cooled by liquid N2, by trap-to-trap distillation in U~CUO(ca. 6 Pa). The product (1-3) was obtained by fractional distillation under reduced pressure. The products obtained from Sb2S3 and Bi2S3 were resinous and could not be purified by distillation.Procedure Static pyrolysis was conducted using 50mg of sample in a porcelain crucible heated in a cold-wall vertical quartz tube under an N2 atmosphere. The temperature of the sample reached the prescribed value from room temperature in ca. 1 h. The composition of the pyrolysates was determined by X-ray fluorescence. Details of the dip-dry pyrolysis method are as follows. A small piece (10 mm x 10 mm) of Si( 1 11) was dipped in and out of a solution of [Bu~I~(SASBU)~]~S (1) in benzene under N2, repeatedly, as shown in Fig. 1. Then the substrate was heated in a quartz tube at the required temperature under an N2 atmosphere. Growth of indium pnictide thin films under CVD conditions was carried out under reduced pressure (base pressure = 0.1 Torr; total pressure= 10 Torr; source temperature = 150 "C; growth period =3 h).The CVD system used in this study is illustrated in Fig. 2. Fig. 1 Dip-dry pyrolysis procedure mass flow meter Pirani gauge 0I -vacuum pump substrate heater thermocouple I heater bubbler (source) thermocouple quartz reactor, Q = 40 mm Fig. 2 Horizontal cold-wall CVD reactor employed in this study Results and Discussion We have reported that trialkylindiums have unusual alkylation behaviour,' 3914 especially tributylindium which readily alkyl- ates inorganic and organometallic oxides to give coupled products containing pox0 bridges between indium and another meta1.'5.16 We expected that such reactions should enable us to realize our objective: the production of new precursor systems.We tried the reaction of tri-n-butylindium with all five pnictogen sulfides. The results are summarized in Table 1. Generally organometallics, such as Grignard reagents, alkylate metal sulfides to give alkylmetal sulfides such as R3AsS.'7-'9 Unlike other organometallics, tributylindium provides coupled products (1-3) on reaction with As2S3,P2S5, and Sb2S5 (Scheme 1). These products 1-3 appear as colour- less and viscous oils and were purified by distillation. The two sulfides, Sb2S3 and Bi2S3 gave resinous products, in which the ratio of indium to pnictogen was estimated to be close to unity; these resinous products were not studied in detail.It is interesting that all of the compounds possess In-S-pnictogen linkages, and it should be emphasized that 1-3 can be purified by distillation [boiling points, InBu, + As2S3 --InBu, + P,S, InBu, t Sb2S5 -t I. .zt 0 10 20 30 J. MATER. CHEM., 1994, VOL. 4 &In-(SAsBu),-SlnBu, (1) Bu Bu Bu,ln-S-P-S-P-S-InBu, I (2) : E Bu2h-S-kb-S-Sb-S-tnB~2 (3): : Scheme 1 II (a) 40 50 60 70 80 29fdegrees Fig. 3 XRD patterns of indium pnictide thin films obtained from 1-3 by dip-dry pyrolysis. (a) InAs film obtained by dip-dry pyrolysis at 600 "C for 3 h; (b) InAs film obtained by CVD (T,,,=400 "C);(c) InP film obtained by dip-dry pyrolysis at 650 "C for 5 h; (d) InSb film obtained by dip-dry pyrolysis at 450 "C for 5 h.100 "C/0.5 Torr (l),130 "C/0.6 Torr (2), and 100 "C/0.3 Torr (3)]. Thus, 1-3 have relatively high vapour pressure compared to conventional single-source precursors such as arsinogal- lanes and their analogues.20 These are generally solids at room temperature and have low vapour pressure, which must result in low growth rate and require high vacuum conditions to prepare 13-1 5 semiconductor layer^.^,'^ These new sulfur- bridged organometallics are potentially useful sources if they give indium arsenide free from sulfur contamination. Thermal analysis of 1 showed that the pyrolysis started at 160 "C (onset of weight loss in TG) associated with an exotherm peak which appeared at ca.230 "C(in DTA). InAs was eventually formed with a successive release of the butyl groups and sulfur (up to 600 "C). Static pyrolysis (up to 600 "C)of 1-3 was then investigated further and the pyrolys- ates were analysed by XRD and X-ray fluorescence microanal- Table 1 Preparation of new sulfur-bridged binuclear organometallics pnictogen sulfide product" YO yield' b.p./"C (Torr) [Bu,In(SAsBu),],S (1)' [Bu,lnSP(S)Bu],S (2)d not identified" [Bu,InSSb(S)Bu] ,S (3)j not identified' 98 95 56 100 (0.5) 130 (0.6) 100 (0.3) "For 1-3, satisfactory microanalysis data were obtained. bYields were estimated by the formulae based on the amounts of the sulfides used. '1: 13C{'H} NMR 6, (22.6 MHz; solvent CDCI,; standard TMS) 13.5, 13.7 (CH,), 23.9, 24.0, 24.2, 24.6, 24.7, 25.0, 25.3, 28.4, 29.1, 29.5, 29.7, 30.7 (CH,); MS (EI, 70eV) m/z 1474 (M+,O.l%), 821 (M-Bu,InS-Bu,, 21Y0), 425 (Bu,InSAsBuS+, 55%), 132 (AsBu', 100%).d2:13C(1H} NMR, 6, (22.6 MHz; solvent CDCI,; standard TMS) 13.7, 28.2, 29.4, 23.0 (CH,CH,CH,CH,-In), 13.7, 27.2, 29.2, 23.5 (CH,CH,CH,CH,-P); IR (KRS- 5) vmax/cm-* 590 (P=S); MS (EI, 70 eV) 533 [Bu,InSP(S)BuSP(S)Bu+, 10Y01,38 1 [Bu,InSP(S)Bu+, loo%]. "Resinous products were obtained. '3: 13C{'HJ NMR, 6, (22.6 MHz; solvent CDCl,; standard TMS) 13.3, 25.2, 26.5, 21.8 (CH,CH,CH,CH,-In), 13.5, 24.8, 26.7, 21.5 (CH3CH2CH2CH,-Sb); IR (KRS-5) v,,,/cm-' 435 (Sb=S); MS (EI, 70 eV) 503 [Bu,InSSb(S)BuS', 1%], 471 [Bu,InSSb(S)Bu+, 5"/,], 178 (BuSb', 100%). J. MATER. CHEM., 1994, VOL.4 ysis (Table 2). Formation of a mixed sulfide was detected at temperatures lower than 450 "C. For example, the pyrolysis of 1 at 450 "C for 2 h gave greyish powders that showed the characteristic X-ray diffraction patterns of In2S3 and As2S3. The pyrolysates obtained at 500 "C showed the characteristic X-ray diffraction patterns of InAs. However, they still con- Table 2 Static pyrolysis of 1-3" compound T/"C t/h pyrolysisb,' 1 500 9 InAsSo.ld 600 9 InAs' 2 600 9 InPSo,2dJ 3 500 5 InSbd "Conditions of the static pyrolysis are described in the Experimental; bcomposition was determined with X-ray fluorescence microanalysis and phase was analysed by XRD; 'hydrocarbons, butane and butene, were detected in the exhaust and the corresponding pnictogen sulfides were deposited on the inner surface of the reactor; dno XRD peak was detected; esharp XRD pattern assignable to InAs was observed (JCPDS no.15-869); fbroad XRD pattern assignable to InP was observed (JCPDS no. 32-453); sharp XRD pattern assignable to InSb was observed (JCPDS no. 6-208). tained considerable amounts of both indium sulfide and arsenic sulfide. To obtain sulfur-free indium arsenide, the pyrolysis must be conducted at 600 "C and long pyrolysis times >9 h are needed. In the case of 2, some different features appeared. A significant sulfur contamination remained in InP obtained at 500 "C, and it was difficult to exclude sulfur contamination even at 600 "C.InP thin films contaminated with sulfur had poor morphology.In contrast, indium anti- monide without sulfur contamination was readily available at 500 "C. This may be a result of the low melting point of indium antimonide itself (535 "C). The static pyrolysis results suggest that 1-3 can be readily converted into the corresponding indium pnictides under sufficiently extreme pyrolysis conditions. We attempted to prepare indium pnictide thin layers on an Si(ll1) wafer by the dip-dry pyrolysis technique. InAs layers from 1 with a preferential orientation of (1 11) (Fig. 3), on Si(ll1) at 600 "C, for 3 h (film thickness 0.6 pm) without sulfur contamination. Hall measurements (Van der Pauw) at 300 K indicated that this InAs thin film had p-type conduction and a relatively high carrier mobility ~300= 1640 cm2 V-' s-and low carrier concentration n300 = 1.3 x lOI4 cm-3.Similarly the preparations of InP and InSb thin films from Fig.4 SEM photomicrographs of indium pnictide thin films obtained from 1-3 by dip-dry pyrolysis. (a) InAs film obtained by dip-dry pyrolysis of 1at 600 "Cfor 3 h; (b)InAs film obtained by CVD from 1 (Tub400 "C);(c)InP film obtained by dip-dry pyrolysis of 2 at 650 "C for 5 h; (d) InSb film obtained by dip-dry pyrolysis of 3 at 450 "Cfor 5 h J. MATER. CHEM., 1994, VOL. 4 Table 3 Dip-dry preparation of indium pnictide thin films" source thin film n,,,/cm- ~ ~~~ 1 600 3 InAs 1640 1.3 x 1014 2 650 5 In P 1100 1.2 x 1019 3 450 5 InSb 18400 6.8 x 10l6 "Dip-dry cycles were repeated 10 times and the film thickness was estimated by SEM as ca.0.5 pm; bmobility measured at 300 K; 'carrier concentration at 300 K. 2 and 3 were carried out and the results obtained are summarized in Table 3. Films of InP and InSb obtained here also possessed relatively high mobility and low carrier concen- tration. These results indicate that the thin films obtained from the new type of precursor of this study suffered from a very low level of sulfur contamination. The pnictide films thus prepared by dip-dry pyrolysis showed a smooth surface morphology without any cracks, as shown in Fig. 4.However, only for 1 could the molecular ion peak be detected (rn/z= 1474). Similar stable binuclear fragment ions were detected in all cases: e.g.[Bu,InSAs(S)Bu] + for 1 (rn/z=425, intensity 55%); [Bu21nSP(S)Bu]+ for 2 (rn/z= 381, intensity 100%); +[BuJnSSb(S)Bu] for 3 (m/z=471, intensity 5%) (Table 1). These fragments contain both indium and pnictogen atoms in a 1 : 1 ratio and strongly suggest the formation of the compounds. Furthermore, the release of the sulfur moiety might occur from the immature pyrolysates during the forma- tion of the indium pnictide lattice, because the metal sulfide will readily lose sulfur under reduced pressure or at tempera- tures below the melting point.'2,21,22 The new single-source precursor containing sulfur bridges between 13 and 15 elements are excellent candidates for the deposition of 13- 15 semiconductor layers.These precursors may offer easy access to sulfur-doped semiconductor layers.23 Preliminary experiments showed that growth of InAs thin layers from 1 by CVD is possible but the InAs films thus obtained have a rough grain geometry and poor crystallinity [Fig. 3 (b) and 4 (d), respectively]. Consequently, further optimization of growth parameters is still necessary. Prep- aration of other 13-15 semiconductor layers by MOCVD is also being developed. Part of this work was supported by Grant-in-Aid for Scientific Research, C-04806085 (RN) and the Priority Area of Reactive Organometallics No. 05236102 (AB). R.N. is grateful for the Miyashita Research Grant. We would like to thank Mr. H. Moriguchi of the Analytical Centre of the Faculty for recording the mass spectra.References 1 H. M. Manasevit, Appl. Phys. Lett., 1968, 12, 156; J. Cryst. Growth, 1981,55, 1. 2 W. T. Tsang, J. Cryst. Growth, 1992, 120, 1. 3 K. Adomi, J-I. Chyi, S. F. Fang, T. C. Shen, S. Strite and H. Morkocq, Thin Solid Films, 1991,205, 182. 4 H. Sato, T. Yamada and H. Sugiura, J. Cryst. Growth, 1991, 115,241. 5 M. Kamp, F. Konig, G. Morsch and H. Luth, J. Cryst. Growth, 1992,120,124. 6 H. Sato, Appl. Organomet. Chem., 1991,5,207. 7 A. H. Cowley and R. A. Jones, Angew. Chem. Int. Ed. Engl., 1989, 28,1208. 8 D. C. Bradley, M. M. Faktor, M. Scott and E. A. D. White, J. Cryst. Growth, 1986,75, 101. 9 R. Nomura, H. Matsuda, Invited lecture, 65th Annual Meeting of Chemical Society of Japan, Tokyo, March, 1993, Abstract, p.183. 10 R. Nomura, S-J. Inazawa, K. Kanaya and H. Matsuda, Polyhedron, 1989,8,763. 11 R. Nomura, S-I. Miyazaki, T. Nakano and H. Matsuda, Chem. Ber., 1990,123,2081. 12 R. Nomura, K. Konishi and H. Matsuda, Thin Solid Films, 199I, 198,339;J. Electrochem. Soc., 1991, 138,631. 13 R. Nomura, S-I. Miyazaki and H. Matsuda, J. Am. Chem. Soc., 1992,114,2738. 14 R. Nomura, S-I. Miyazaki and H. Matsuda, Organometallics, 1992, 10,2. 15 R. Nomura, S. Fujii and H. Matsuda, Inorg. Chem., 1990, 28, 4586. 16 R. Nomura, S. Fujii, T. Shimokawatoko and H. Matsuda, J. Polym. Sci., Part A, 1992,30, 153. 17 F. F. Blicke and F. D. Smith, J. Am. Chem. Soc., 1929,51, 1558. 18 F. F. Blicke and E. L. Cataline, J. Am. Chem. SOC., 1938,60,423. 19 N. A. Chadaeva, K. A. Mamakov and G. Kh. Kamai, J. Gen. Chem. USSR, 1966,43,821. 20 A. M. Arif, €3. L. Benac, A. H. Cowley, R. J. Jones, K. B. Kidd and C. M. Nunn, New J. Chem., 1988,12,553. 21 G. B. Samsonov and S. B. Drozdova, ed. Handbook ofSulJdes, Metallurgy Press, Moscow, 1972. 22 R. Diehl and R. Nitsche, J. Cryst. Growth, 1975,28, 306. 23 B. StEpanek, V. Sestakova, V. $mid and V. Charvat, J. Cryst. Growth, 1993,126,617. Paper 3/04236G; Received 20th July, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400051

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4( l), 55-59 Short-range Order in Extended-chain Crystals of Polyoxymethylene from a True Molecular Perspective: An Atomic Force Microscopy Study Daniel Snetivy," Haifeng Yang," Bernhard Glomm,band G. Julius Vancso*= a University of Toronto, Department of Chemistry, 80 St. George Street, Toronto, Ontario M5S 7A7, Canada lnstitut fur Polymere, Eidgenossische Technische Hochschule (Swiss Federal Institute of Technology), CH-8092 Switzerland Atomic force microscopy (AFM) images of extended-chain crystals of polyoxymethylene (POM) obtained in solid-state polymerization have been made to molecular resolution on the surface of the microfibrils which were formed during the topotactic polymerization process. AFM scans with molecular resolution reproduced the expected crystal lattice parameters c =1.72 nm (X-ray data, c =1.739 nm) and a =0.45 nm (X-ray data, a =0.447 nm).The order of the microfibrils within the crystals was analysed and compared with results obtained previously on mechanically oriented polyoxymethy- lene by using AFM and wide-angle X-ray diffraction data. For the well ordered surface of the POM crystals, an AFM imaging-mechanism is suggested which assumes that the contact force is controlled by the outermost methylene groups at the imaged surface. Atomic force microscopy (AFM) was proposed in 1986 to investigate surfaces of materials on an atomic scale.' This new branch of scanning probe microscopy was developed to remedy the limitations of scanning tunnelling microscopy (STM) which, although it yielded images of atoms before the advent of the AFM, is restricted to studies of conducting materials2 Since its discovery, AFM has been used with great success to study surfaces of insulators and metals, as well as the macromolecular architecture of polymeric materials from the Angstrom to the micrometre ~cale.~?~ In AFM, a sharp tip mounted to the edge of a small spring is brought down in close proximity to the sample ~urface.~ The sample is then scanned laterally beneath the tip-spring probe, usually referred to as the microcantilever, by using a piezoelectric positioner.The deflection of the cantilever is measured by various techniques. In the first AFM, the minute deflections of the spring were monitored by an STM attached to the microcantilever as a detector.' In commercial instru- ments (e.g.the NanoScope line by Digital Instruments), the deflection is measured by an optical lever technique. The deflection signal (which is proportional to the measured force acting on the microcantilever as a result of the interatomic interactions between surface atoms and the atoms at the apex of the microcantilever tip) is related to the surface structure of the sample. This signal, which is measured as a function of the tip position, is converted to an optical image of the sample surface. By utilizing commercial piezo scanners, the AFM is capable of covering a magnification range of ca. 108-103x in a single experiment. Unlike other diffraction or scattering techniques (e.g.electron or X-ray diffraction) which result in structural information averaged over typically 1010-1020 atoms, AFM can yield information about the local order and packing of atoms/moIecules at preselected locations of the sample surface.Since the first visualization of a macromolecule, AFM has been used with great success to image polymers. For example, imaging of macromolecular chains in extruded polyethylene5 and oriented polytetrafluoroethylene,6 characteristics of poly- mer lamellar crystals on the micrometre and polymer helix structures in crystalline isotactic p~lypropylene~,~' have been achieved. Our AFM study on fibres of ultrahigh molecu- lar weight polyethylene (UHMWPE) resulted in images of the (010) and (100) crystallographic planes with an excellent reproduction of the expected crystal lattice parameters." In addition, AFM has yielded images with regular nanostructure at surfaces of polymer crystals,12 and has been used to study polymorphism in p~ly(p-phenyleneterephthalamide),'~and the orientation mechanism in ultradrawn p~lyethylene,'~ to mention only a few examples.In an earlier comm~nication,~~ we reported on preliminary AFM results which we obtained on mechanically oriented polyoxymethylene (POM). Images included features of the microfibrillar morphology and molecular resolution of the POM helices packed in the hexagonal crystal structure of POM. Individual turns of the helix and the five turns that make up the repeat unit of the 9/5 helix were visualized.In this paper, we focus on the results we obtained on extended- chain crystals of POM prepared in solid-state (topotactic) polymerization. It has been shown in the 1iteraturel6*" that crystals of cyclic compounds, such as trioxane (CH201, and tetraoxane (CH20)4,can be polymerized directly into crystal- line POM in the solid state. In this study we used POM crystals obtained from tetraoxane which are known to form extended-chain structures.17 If POM is made from (CH,O),, then the macromolecular orientation within the POM crystal is correlated with the c-axis orientation of the precursor monomer crystal (topotactic reaction). It has been demon- strated that short-range molecular order at the scanned surface is important to obtain AFM images using currently available AFM equipment with molecular resolution." Thus, extended- chain crystals of macromolecules are especially well suited for AFM studies.Experimental POM crystals were obtained from trioxane purchased from Aldrich Chemical Company (analytical grade). Standard subli- mation equipment was used for growing large, needle-like single crystals of trioxane. Typical dimensions of the crystals used in polymerization reactions were 10 x 2 x 2 mm. The transparent monomer crystals were irradiated by using a Co6' pray source (Gammacell 220, Atomic Energy of Canada Ltd.) to form POM in a closed container. The estimated dose was 1 MR, and the irradiation was carried out at 32°C.Post polymerization, which immediately followed the irradiation, was carried out in a hot-air oven at 60°C for 48 h. The POM crystals obtained in this process were white, indicating the presence of voids in the material. This was expected, since, as Plate 1 AFM micrograph (image size: 1000x 1000 nm) of the surface of a POM crystal obtained in solid-state polymerization Plate 3 AFM nanograph (image size: 8 x 8 nm) of a microfibril of a POM crystal obtained in solid-state polymerization discussed in the literature,16 the polymer crystals obtained in this process are not truly monocrystalline. Owing to lattice mismatch and strain that develops during the solid-state polymerization process, a fraction of the polymer forms sub- crystallites and twin-structures within the POM crystal.Samples for AFM studies were cleaved from the POM crystals along the expected chain (or crystallographic c) direc-tion by using a Sorvall MT6000 ultramicrotome with glass knives. Microtoming was carefully performed to make sure that the surfaces of the crystalline microfibrils expected to form in the polymerization process remain intact and that the sample will be cleaved along crystal planes. AFM images were taken in air at the freshly cleaved surface using a NanoScopeII instrument (Digital Instruments, Ca.) with an A-type scan head utilizing NanoProbe 100 micron triangular Si,N, canti- levers with wide legs. The effective spring constant of the cantilever was 0.58 N m-l. Images with atomic resolution were obtained in the constant height mode.Wide-angle X-ray diffraction intensities were collected using a Siemens D500 diffractometer equipped with a Cu-Kx X-ray J. MATER. CHEM., 1994, VOL. 4 Plate 2 AFM micrograph (image size: 1000 x 1000 nm) of the surface of an oriented POM fibre obtained by mechanical orientation Plate 4 Two-dimensional autocorrelation pattern of the AFM image shown in Plate 3 source, pin-hole collimator, Huber texture goniometer, graph- ite flat crystal monochromator, and a scintillation detector. The experiments were performed in 8/20 transmission mode, and fibre diagram quadrants were measured in an azimuthal range between 0 and 90". Results and Discussion AFM images of POM on the micrometre scale are discussed first.In Plate 1, a typical AFM scan of an area of 1000 x 1000 nm is captured. The micrograph shows that the crystal consists of well ordered microfibrils with an axis aligned in the crystallographic c direction. For comparison, in Plate2 we show an AFM scan which was obtained on a mechanically oriented POM sample with a high degree of anisotropy. This specimen was prepared from commercially available Ultraform H 2320 polymer (BASF) (for further details see ref. 15). The microfibrils in the mechanically ori- ented POM show imperfect alignment and twisted bundles of J. MATER. CHEM., 1994, VOL. 4 Plate5 Cross-sectional plot (top insert) of the AFM nanograph in the chain direction. The profile was obtained along the line shown in the AFM image captured in the bottom left insert microfibrils. The difference in the degree of order between mechanically oriented POM and POM crystals can also be clearly seen in Fig.1 and 2 which compare the wide-angle X-ray diffraction (WAXD) patterns. Main maxima of the diffraction intensity in the isotropic sample occurred at the following diffraction angles (the corresponding crystal planes are in brackets): 22.9" (100); 34.6" (105); 40.2" (110);48" (115) and (009). The (100) peak of the mechanically oriented POM sample on the equator showed a broader azimuthal width than the corresponding maximum for the POM crystal. This refers to a better alignment of the macromolecules and thus a better order of the microfibrils in the POM crystal.However, scattering intensity for 28=22.9" (100) can be observzd in the pattern of the POM crystal at azimuthal angles distinctly different from the main maximum at 0" (equator). This confirms the presence of twin-crystals and sub-crystallites within the POM sample obtained in solid-state polymeriz- ation. Interestingly, such sub-crystallites are absent from the POM sample obtained by mechanical orientation. AFM images in the nanometre range were obtained on the surface of the microfibrils shown in Plate 1. A typical example (raw data) is displayed in Plate 3 (scan size: 8 x 8 imj, dis- Plate 6 Cross-sectional plot (top insert) of the AFM image in the chain perpendicular direction. The profile was obtained along the line shown in the AFM image captured in the bottom left insert J.MATER. CHEM., 1994, VOL. 4 Fig. 1 A quadrant of the WAXD pattern of mechanically oriented POM playing a periodic structure on the Angstrom scale in two essentially perpendicular directions. Spots with bright tones correspond to high contact forces between the apex of the AFM tip and the imaged molecules, and thus correspond to corrugations 'sticking out' of the sample surface. Dark areas, on the other hand, mean smaller contact force and lower features. Objects parallel to the diagonal that connects the top left and the bottom right corners were identified as images of the polymer chains. This assignment is based on values for the packing distances and on the observation that the direction of the features identified as chains coincides with the crystallo- graphic c (or fibre) direction which was set at 45" with respect to the horizontal scan direction prior to imaging.Plate 4 shows the two-dimensional autocorrelation pattern (2-dAP) of the nanograph captured in Plate 3. This pattern was calculated to reduce experimental noise and to analyse period- icities in the chain direction and perpendicular to it. Profiles of the AFM scans in the chain and chain-perpendicular directions are shown in Plates 5 and 6, respectively, in the top insert of the Plates. Our AFM data showed that the periodicity of the crystal structure in the chain direction, i.e. the length of the c edge of the hexagonal unit ce11'8-21 was c= 1.72 nm as obtained from Plate 5 (expected value from WAXD is c= 1.739 nm).The value of the packing distance in the chain-perpendicular direction was a =0.45 nm, as obtained from Plate 6 (expected value from WAXD is a=0.447 nm). Thus, the agreement between AFM and WAXD data is excellent. This good agreement was obtained by eliminating a systematic error from AFM nanographs. A calibration technique, which takes into account the height of the imaged specimen, was used as described in a previous publication.22 The regular packing of the POM chains in the crystal is obvious in the nanograph shown in Plates 3 and 4. Finally, we briefly discuss a suggested contrast mechanism that would result in an AFM image as in Plate 3. In order to understand the contact forces between the apex of the AFM tip and the imaged molecules, it is useful to look at the computer-simulated ac crystal facet of the POM lattice which was imaged in the AFM experiment.A simulated molecular layer of POM chains packed in the ac facet is shown in Plate 7. Details of the chain, the exposed methylene groups and the underlying oxygen atoms are clearly seen. As men- tioned earlier, the POM chain forms 5 turns within one crystallographic c During AFM imaging, however, the tip will most likely follow the contour of the outermost electrons if the contact force is kept at the least possible value. Fig. 2 A quadrant of the WAXD pattern of a PC)M crystal obtained in topotactic polymerization Thus, on images like the nanograph captured in Plate 3, the bright spots correspond to the exposed methylene groups of the POM chain, and the individual turns of the helix are covered by the methylene units. In Plate4 there are fuzzy regions that separate three bright spots on the image.These 'fuzzy' sections can be identified as parts of the chain where the methyl groups are not directly exposed to the scanned surface, but are turned away to the unexposed side of the macromolecule, and thus are not sticking out of the surface. It should, however, be mentioned, that the exact symmetry of the AFM patterns observed depends somewhat on the spot chosen for imaging at the sample surface and on the material. For example, in our earlier preliminary study15 performed on mechanically oriented POM, we observed five turns within one repeat unit in the crystallographic c direction, which we interpreted as imaging of the individual turns of the POM helix. The AFM nanographs obtained at the surface of POM crystals which are described in this paper, however, can be interpreted as images of methylene groups at the exposed surface of the sample, and thus one would not see the turns of the helical polymer backbone.In these crystals the degree of chain alignment within the microfibrils is significantly better than in the oriented POM samples. which might be one reason for the observed differences. One might speculate that in the POM crystals obtained in solid-state polymeriz- ation, the well ordered ac crystal facet is clearly cleaved and thus the regular pattern of the methylene groups is well exposed.On the other hand, owing to a less perfect chain packing in the mechanically oriented sample, the packing of the methylene groups at the sample surface is more disordered. Thus, the AFM tip for this less ordered sample might, on average, follow the contour of the POM helix rather than the disordered methylene groups. Financial support by the Ontario Centre of Materials Research and the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged. The authors would like to thank Ms. Anne Klemperer for her help with the preparation of the manuscript and the Swiss Federal Institute of Technology for the use of the X-ray apparatus.J. MATER. CHEM., 1994, VOL. 4 Plate 7 Computer-simulated image of a monomolecular layer of the crystal facet visualized in the AFM experiment References 1 G. Binning, C. F. Quate and Ch. Gerber, Phys. Rev. Lett., 1986, 56,930. 2 Scanning Tunneling Microscopy and Spectroscopy: Theory, Techniques, and Applications, ed. D. A. Bonnell, VCH, Weinheim, 1993. 3 G. Binning and H. Rohrer, Rev. Mod. Phys., 1987, 59. 615; D. Rugar and P. Hansma, Phys. Today, 1990,43,23. 4 D. Sarid, Scanning Force Microscopy With Applications to Electric, Magnetic, and Atomic Forces, Oxford University Press, Oxford, 1991. 5 S. N. Magonov, K. Quarnstrom, V. Elings and H-J. Cantow, Polym. Bull., 1991,25,689. 6 H. Hansma, F. Motamedi, P.Smith, P. Hansma and J. C. Wittman, Polymer, 1992,33,647. 7 R. Patil, S. J. Kim, E. Smith, D. Reneker and A. L. Weisenhorn, Polym. Commun., 1990,31,455. 8 D. Snetivy and G. J. Vancso, Polymer, 1992,33,432. 9 B. Lotz, J. C. Wittman, W. Stocker, S. N. Magono\ and H-J. Cantow, Polym. Bull., 1991,26,209. 10 D. Snetivy, J. E. Guillet and G. J. Vancso, Polym. Commun., 1993, 34,429;D. SnCtivy and G. J. Vancso, Polymer, in the press. 11 D. SnCtivy and G. J. Vancso, J. Muter. Chem., 1992,2, 891. 12 D. H. Reneker, R. Patil and S. J. Kim, ACS Polym. Prepr., 1992, 33, 790. 13 D. Snktivy, G. J. Vancso and G. C. Rutledge, Macromoltdes, 1992,25,7037. 14 S. N. Magonov, S. S. Sheiko, R. A. C. Deblieck and M. Moller, Macromolecules, 1993,26, 1380. 15 D. Snetivy and G. J. Vancso, Macromolecules, 1992,25,3320. 16 See e.g. J. Schulz, Polymer Materials Science, Prentice Hall, Englewood Cliffs, 1974, ch. 2.12. 17 Y. Chatani, T. Uchida, H. Tadokoro, K. Hayashi, M. Nishii and S. Okamura, J. Makromol. Sci. Phys., 1968, B2,567. 18 E. Z. Sauter, Phys. Chem., 1933,21B, 186. 19 G. A. Carazzolo, J-Polym. sci, A, 1963, 1,1573. 20 T. Uchida and H. Takodor, J. Polym. Sci. A2, 1967,5,63. 21 Y. Takahashi and H. Takodoro, J. Polym. Sci., Polym. Phys. Ed., 1979, 17, 123. 22 D. Snetivy and G. J. Vancso, Langmuir, in the press. Paper 3/04291J; Received 21st July, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400055

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4( l), 61-69 Rod-like Liquid Crystals of Organic Transition-metal Complexes Part 4.t-Optically Positive Uniaxial Nematic Phase in the Bis[1-(4’-alkoxybiphenyl-4-yl)-3-alkylpropane-l,3-dionato]copper(ii) Complexes Kazuchika Ohta,* Hajime Akimoto, Tetsuya Fujimoto and lwao Yamamoto Department of Functional Polymer Science, Faculty of Textile Science & Technology, Shinshu University, Ueda 386, Japan We report that the long-alkoxy-chain substituted bis(P-diketonato)copper(rI)complexes, bis[l-(4’-dodecyloxybiphenyl-4-yl)-3-alkylpropane-1,3-dionato]copper(11)[alkyl substituent (R’): ethyl, n-propyl and n-butyl], show a nematic (N) phase. The methyl complex has a discotic rectangular ordered columnar (Dro)phase, as reported previously. Therefore, the change from the D, phase to the N phase occurs with changing from R’=methyl to R’=ethyl groups.In order to obtain derivatives having a wider temperature range for the nematic phase, a homologous series of the bis[l-(4‘-alkoxybiphenyl-4-yl)pentane-l,3-dionato]copper(11) complexes (alkoxy =C,H,+,O; n =6,8,10, 11, 14) have been synthesized for R’ =ethyl. The derivative for n =1 1 gave the widest range (AT=16.1 “C).Furthermore, the nematic phase of these complexes was established as an optically positive uniaxial nematic phase. The mesomorphism of the corresponding ligands was also investigated. To date, several liquid-crystalline compounds containing a tative ligand 10 (n=8) and the complex, 12 (n=121 are bis( P-diketonato)copper(II) complex core have been reported described in the following.(Fig. 1). These complexes show a discotic lamellar (DL) phase (1),2 a discotic hexagonal (Dh) phase (2),9,10a discotic 1-( 4-Octyloxybiphenyl-4-y1)pentane-1,3-dione(10, n =8) rectangular (D,) phase (5),” and a nematic (N) phase Sodium hydride (60%; 1.23 g, 30.8 mmol) was placed in a (3-5),12-” It is very interesting that the bis(p-100cm3 three-necked flask in a nitrogen atmosphere: and diketonato)copper(II) complexes, 5, which were recently washed with dry hexane to remove the mineral oil. To reported by Ohta et al., show a D,, mesophase when the this, were introduced 4-octyloxy-4’-acetylbiphenyl (5.00 g, substituent R’ is a methyl group’’ and a monotropic N phase 15.4 mmol) and then freshly distilled tetrahydrofuran (THF) when the substituent R’ is an n-butyl group.16 (ca. 40 cm3).After refluxing for ca. 2 h, the solution was cooled In this paper we focus on phases of the complexes 11 to room temperature (r.t.). Ethyl propionate (7 87 g, (Scheme l), with the substituent R’ ethyl or n-propyl. The 77.1 mo1)-THF(ca. 10 ml) was added dropwise to the solution, new complexes with R’ ethyl or n-propyl have been synthesized and refluxed again overnight (ca. 15 h). After it had been and the effect of the alkyl chain (R’) on the mesomorphism cooled to r.t., the reaction was stopped by adding 1 mol ~im-~ has been investigated. It was found that these complexes show hydrochloric acid (ca. 30 ml). The solvent was evaporated to an enantiotropic N phase, and that the change from the D,, give a yellow solid.The purification was carried out by phase to the N phase critically occurs with changing from the column chromatography (silica gel, chloroform, Rf =0.75) and R’ =methyl group to R’ =ethyl group. recrystallization from ethyl acetate to give 3.42 g of light-We synthesized a homologous series of the complexes 12 yellow powder. Yield 58%. ‘H NMR(CDCI,, TMS) 6, (n=6, 8, 10, 11, 14; Scheme 1) to obtain derivatives with a 0.80-1.90 [m, 18 H, CH3(CH2), CH20- and -COCH,CIH,], wider temperature range for the nematic phase than that for 2.44 (9, 2 H, -COCH,CH3), 3.99 (t, 2 H, -CH,O-), 6.15 (s, complex 11 (m=2). All the complexes 12 showed an enantio- 1 H, enol CH), 6.85-7.98 (m, 8 H, biphenyl), 16.1 (s, 1 H, tropic nematic phase.The temperature range of the nematic enol OH), keto: enol =0:lOO. phase of the n= 11 derivative (AT= 16.1 “C) is the widest for the derivatives. The conoscopic figures indicated that the Bis[ 1-4-present nematic phase is optically positive uniaxial. dodecyloxybihenyl-4-yl)pentane-1,3-dionato]copper(11)( 12, n = We report here the interesting change from the D,, phase 12)to the nematic phase in the complexes 11, and the optically b-Diketone, C,,O-Lig-Et (10, n= 12; 0.16 g, 0.37 mmol) and positive uniaxial nematic phase in the complexes 12. ethanol (ca. 15 ml) were placed in a 30 ml flask, and heated to dissolve. To this solution were added an aqueous solution of ammonium hydroxide (28%, ca. 1 ml) and an ethanolic Experimental solution of copper(I1) chloride dihydrate (0.030 g, 0.19 mmol).Synthesis The solution was refluxed for ca. 1 h. The reaction mixtures The synthetic route for the bis [1-(4’-alkoxybiphenyl-4-y1)-3-were filtered off, washed with hot water and hot ethanol, and copper(~r) complexes (11 and 12) is dried in z)acuo.The residue was purified by column chromatog- alkylpropane-1,3-dionato] shown in Scheme 1. 4-Alkoxy-4’-acetylbiphenyls (8) were raphy (silica gel, chloroform, R, =0.78) and recrystallization obtained by the literature method.I8 The ligands 9 and 10 from carbon tetrachloride to afford 0.11 g of grey-green were synthesized by our previously reported method.” In powder: yield 64%. Table 1 are listed the elemental analysis data and their yields.In Table 2 are listed elemental analysis data and yields of the Measurements complexes 11 and 12. The detailed procedures of the represen- The phase-transition behaviour of the complexes 11 and 12 and their corresponding ligands were observed with a polariz- t Part 3: ref. 1. ing microscope (Olympus BH-2) equipped with a heating J. MATER. CHEM., 1994, VOL. 4 ?R ?R "'"'SR0 \Ry&pR Rown n n o\c(o o0 '0 Rfi RO-Q"o-R OR OR OR R = C,,lih+1 (n= 7.9,11) D?lo R = CH3, C~H~,OCH~,OC~HS, OC3H7 2 monotropic N22-'4 3 w w% R'% R 'oc 12H25 monotropic N" R' = CHa D,ll 4 R' = C,H, monotropic N'~ 5 Fig. 1 Mesogenic bis(fl-diketonato)copper(rI)complex derivatives 6 7 8 RW&t NaH/THF R' v ligand R = C,&,,+ R' = CA2, + 1 9 n=12 m = 1",2,3,416 OR 10 n= 63, 1&12,14,18 m = 2 complex R = C,,Hh+1 R'= Cnhn+1 11 n= 12 m = 1 ?2.3,416 12 n = 6,8, 10-12,14 m = 2 Scheme1 Synthetic route of 1-(4-alkokoxybiphenyl-4-yl)-3-alkylpropane-l,3-diones(9 and 10) and their corresponding copper(r1) complexes (11 and 12) J.MATER. CHEM., 1994, VOL. 4 Table 1 Elemental analysis data and yields of 9 and 10 Table 3 Phase transition temperatures (T)and enthalpy changes (AH) Of 9 (m=1-4) elemental analysis (YO) found (calc.) m m or n formula C H yield (YO) 9 3 79.95 (79.92) 9.39 (9.27) 38 10 6 78.38 (78.26) 8.01 (7.98) 35 8 78.91 (79.13) 8.48 (8.46) 58 10 79.37 (79.58) 8.88 (8.89) 56 11 79.58 (79.66) 9.06 (9.06) 68 12 79.77 (79.92) 9.32 (9.27) 38 14 80.13 (80.34) 9.54 (9.65) 69 18 80.72 (80.95) 10.06 (10.07) 89 Table 2 Elemental analysis data and yields of 11 and 12 elemental analysis (%) found (calc.) phase TC (AHBcal mol-') -phase .IIMMMNYO =relaxation rapid cooling I E 'yea I105.0 t3.51, 135.0 [0.6], 167.5 [l.g] , SAd is0'K1 -HK2--81.6 95.8 17.11 120.0 [1.3] 161.5 p.11E4-SK1--K24-A--is0 rn or n formula C H yield (YO) 12 6 C,6H5406CU 72.08 (71.81) 7.10 (7.06) 66 8 C~OH&~CU73.00 (73.30) 7.60 (7.68) 65 10 C&~OO~CU73.81 (73.44) 8.03 (7.97) 19 11 C,6H,406Cu 74.18 (74.01) 8.23 (8.28) 45 12 C,,H7806CU 74.52 (74.28) 8.41 (8.30) 64 14 C6,H=j,j06Cu 75.15 (75.34) 8.75 (8.81) 75 plate controlled by a thermoregulator (Mettler FP80 and FP82) and measured with a differential scanning calorimeter (Rigaku Thermoflex TG-DSC).The X-ray diffraction measurements on the powder were performed with Cu-Ka radiation, using a Rigaku Geigerflex equipped with a hand- made heating plate controlled by a thermoregulator. Conoscopic observations were made with a hand-made heater with a hole (the diameter is 3.0 mm), Sakaguchi-Dennetsu Samiconheater Super 350, and a temperature controller (Chino KP1250BR00). The sample was sandwiched between two commercially available cover slips (each of thickness ca. 0.15 mm) with a spacer (50 or 100 pm, Teijin Tetron Film). The internal surfaces were without any treatments of silane.For the case of 100 pm thickness, the cell was further sand- wiched between two aluminium plates (thickness: top =ca. 0.3 mm; bottom =ca. 1.5 mm) having a hole (diameter 3.0 mm) in the centre; a d.c. voltage was applied (ca. 200V) using a V-Cstabilizer (Mitsumi Scientific Ind,ustry Co., Ltd, SJ-1051). Results and Discussion Mesomorphism of the Ligands 1-(4'-0odec);loxybiphenyl-4-yl)-3-alkylpropane-1,3-diones(9) In Table 3 are summarized the phase-transition temperatures and phase-transition enthalpy changes for the ligand 9 (R' = CmHzm+l,m= 1-4). The mesomorphism of the derivatives 9 rn =1 and m=4 has already been described in our previous papers.l6>l9Here we synthesized 9 with m=2 and with m=3 and investigated their mesomorphic properties.The derivative 9 (rn=2) has two enantiotropic mesophases, E and S,, similar to the 9 (m=l) derivative. The 9 (rn=3) derivative has monotropic E and enantiotropic SA mesophases similar to the 10 (n=18) compound. The 9 (m=4) derivative has only an SA phase. Thus, for the derivatives 9 the E phase tends to disappear for m>3. It can be thought that the longer the R' group of ligand 9, the more freely the molecules rotate in the layer, because the R groups can disturb the molecular assembly, as shown in Fig. 2. 112.5 [l0.8] 105.2 [&O] 133.8 [1.4]-K sAe -'is0 "Phase nomenclature: K =crystal, E =E phase, SA=smectic A phase and is0 =isotropic liquid. bReference 19. 'This work. dReference 16. 1-(4-AZkoxybiphenyl-4-yl)pentane-1,3-dione(10) Each of these derivatives (n=6, 8, 10-12, 14, 18) has both E and smectic A (S,) phases.Their phase-transition tempera- tures and phase-transition enthalpy changes are summarized in Table 4. In Fig. 3 all transition temperatures for 10 are plotted against the number of carbon atoms in the alkoxy chain. The detailed mesomorphism of the representative com- pounds 10 (n=8) and 10 (n=18), are described as follows. Eh?b R'= Me Wh 1:::o & E QQ monotropicE R'= PP Fig.2 Possible reason for the disappearance of the E phase in 9 when the n-alkyl chain becomes longer Table 4 Phase-transition temperatures (T) and enthalpy changes (AH) of 10. n phase " T/"C(AHlkcal mol-') -phase-= relaxation 108.6 128.1 [1.4] 175.2 [2.q 9626 KIe-Kr-E -7 %-is0 101.0 124.3 [l.t 173.1 p.11~8 K1-* E-SA-is0 98.5 123.4 [1.5L sA 168.3 [2.5] isO 4 -10 K1-100.5K2 11 12 97.1 1060 1 120.5[1.4 158.5[1.9] 14 K~-K~J E -sA-is0 113.0 [13.1] 150.7 [2.6] K is018 "Phase nomenclature: K =crystal, E =E phase, SA=smectic A phase and is0 =isotropic liquid.180 160 140 l2OI100 804(.l.L.I.Q.III.I.0 12 16 number of carbon atoms (n) in the alkoxy chain Fig. 3 Phase-transition temperature vs. number of carbon atoms in the n-alkoxy chain of 10: M, clearing point; 0, E-SA transition; A, melting point; 0,solid-solid transition; x, undefined phase transition J. MATER. CHEM., 1994, VOL. 4 1(-4'-OctyloxybiphenyI-4-yI)pentane-1,3-dione[10 (n=S)], When the K, crystals of 10 (n=8) were heated from room temperature (r.t.), they melted to an E phase at 115.3"C, followed by an SA phase at 124.3"C; on further heating, the SAphase cleared into an isotropic liquid (iso) at 173.1 "C.The textures of the two mesophases are shown in Plate 1. When the isotropic liquid (iso) was cooled to 169.9 "C, a focal-conic fan texture appeared. Plate l(a) was obtained on cooling to 130.0"C;this texture consists of the parts of focal-conic fans and homeotropic arrangement characteristic of the SA phase. When the sample of Plate l(a) was cooled to 119.O"C, the focal-conic fans changed to the arced focal-conic fans and the homeotropic arrangement part changed to a mosaic texture, as shown in Plate l(b); the texture is often observed for the E phase.These E and SA mesophases were confirmed by X-ray powder diffraction. The patterns are shown in Fig. 4 and the data are given in Table 5. The X-ray diffraction powder II 3 (b )I' 3 ~~ 2 20 40 20ldegrees Fig. 4 X-Ray powder diffraction patterns of 10 (n= 8): (a) the SA phase at 155"C, (b) the E phase at 122 "C Table 5 X-Ray diffraction data of 10 (n-8): (a) the SAphase at 155 "C, (b) the E phase at 122°C Miller peak indices no. dobs.lA dCdC.lA (hk I) (a) c=25.9A 1 2 3 24.7 12.9 ca. 4.5 25.9 12.9 - (00 1) (0 0 2)melt of the alkoxy chains (b) a=8.11 A;b=5.47 A, c=26.2 A 25.2 26.2 13.1 13.1 4.54 4.54 4.06 4.06 3.26 3.26 J. MATER.CHEM., 1994, VOL. 4 pattern of the SA phase at 155 "C [Fig. 4(a)] has two narrow reflections in the X-ray low-angle region, which correspond to the (001) and (002) p!anes for a layered structure. The interlayer distance is 25.9 A. It also has a diffuse band around 20=20" in the wide-angle region, corresponding to the melt of the alkoxy chains (peak no. 3). Since it has a layered structure without any order in the layer, this higher-temperature mesophase was identified as either an SA or an Sc phase. The phase has a focal-conic fan texture. It was concluded that it was an S, phase. In Fig. 4(b) the X-ray diffraction powder pattern of the E phase at 122°C is shown. It has two narrow reflections in the low-angle region, which correspond to t,he (001) and (002) planes; the interlayer distance is 26.2 A.It also has three narrow reflections on a diffuse element in the wide-angle region, which correspond to the (110), (200) and (210) planes in a two-dimeqional rec- tangylar lattice; the lattice constants are a=8.11 A and b= 5.49 A. Thus, the lower-temperature mesophase was confirmed as an E phase. 1-(4-0ctadecyloxybiphenyl-4-yl)pentane-1,3-dione [lo (n= 1S)] .This derivative shows different phase-transition behav- iour from the other compounds 10. It exhibits a monotropic E and an enantiotropic S, mesophase. Thus, the E phase of the derivatives 10 with n>18 may have a tendency to dis- appear (Fig. 3). From a thermodynamic viewpoint, the disappearance of the E phase results from a balance of stab- ilities of the E and crystal phases and is attributed to the in- creased stability of the crystal phase with increasing chain length.Here we wish to discuss the disappearance of the E phase from the viewpoint of molecular shape and molecular rotation. Molecules in a layer of the E phase cannot rotate freely around the molecular axis (Fig. 5, top left). The mole- cules can oscillate through angles of <180".20 If the molecules can rotate freely, the mesophase may change to another -u E SA n-.. h \ rc r' Fig.5 Possible reason for the disappearance of the E phase in 10 when the n-alkoxy chain becomes longer mesophase such as an S, phase which has no order in the layer (Fig. 5, top right). Goodby and Gray discussed the structure of the E phase in the n-alkyl 4-n-alkoxybiphenyl-4-carboxylatesand its relationship to molecu- lar structure, in a very similar way to ours.21 The present molecule of derivative 10 can be assumed to consist of two parts, that is a soft hydrocarbon chain contributing to the free rotation (a) and a rigid strip-like plane contributing to make the two-dimensional rectangular lattice of the E phase (b) (Fig.6, bottom). With an increasing number (n) of carbon atoms in the alkoxy chain, the ratio r =a/b increases and the molecules can rotate more easily. In this case, the tight lattice of the E phase is no longer maintained; the E phase may disappear. Mesomorphism of the Copper Complexes Complexes11 In Table 6 are summarized the phase-transition temperatures and phase-transition enthalpy changes for complexes 11 (R' = C,H2,+ 1, rn =1-4).We have already reported in our previous that the rn =1complex shows a discotic rectangular ordered columnar (D,, mesophase, and that the m =4 complex show a monotropic N phase. The complexes with R' =C2H, (rn=2) and R' =n-C3H, (rn =3) synthesized here show an enantiotropic N phase. A temperature range of the N phase for R'=n-C3H9 is extremely narrow in comparison with that for R' =C2H,. The nematic phase of these complexes shows both a Schlieren texture and a marbled texture characteristic of the nematic phase (Plate 2). These complexes show double melting behaviour via the nematic phase.2' For this behaviour, the accurate phase-transition enthalpy changes could not be measured (Table6).As can be seen in Table6, the change from the D,, phase into the N phase critically occurs with changing from R' =methyl group to R' =ethyl group, and the N phase has a tendency to disappear when the substituent R' is longer. We considered the effect of the alkyl chain (R') on meso- Table 6 Phase-transition temperatures (T) and enthalpy changes (AH) of 11 R' phase" T/"C -phaseAHlkcal mol -' .ryww~~~vv)=relaxation CH3b K 135.1 208.8 is0 (decornp)40 171.9 183.6 'N-isoC2H5 'lY.: ,iK2-0.2 159.9 160.0 mC3H7 K, 'N-*is0 111.5 153.3 pCqHgC K1 K2 "Phase nomenclature: K =crystal, D,, =discotic rectangular ordered columnar mesophase, N =nematic, and is0 =isotropic liquid.Reference 1 1. Reference 16. J. MATER. CHEM., 1994, VOL. 4 Plate 1 Textures of the mesophases of 10 (n=8): (a) the focal-conic fan texture of the SA mesophase obtained from the isotropic liquid on cooling to 130"C;(b)the arced focal-conic fan texture and mosaic texture of the E mesophase obtained from the SA on cooling to 119 "C Plate 2 Textures of the nematic phase of 11 (m=3): (a)Schlieren texture and (b)marbled texture Plate 5 Conoscopic observation of the nematic phase of 12 (n= 14)Plate 3 Zigzag disclination in the nematic phase of 12 (n= 14) with the 1/42> plate J. MATER. CHEM., 1994, VOL. 4 Plate 4 Conoscopic -observation showing the uniaxiality of the nematic phase of 12 (n= 14). A, Thickness =50 pm, spontaneous homeotropic alignment; B, thickness = 100 pm, d.c.=200 V. (a) 0"; (b)45"; (c) 90" J. MATER. CHEM., 1994, VOL. 4 Table 7 Phase-transition temperatures (T)of 12 T/"Cphaseb-phase ."vvs) = relaxation 218.9 223.6 =N cis0 (decornp) 196.6 21 1.2 Kl -N -is0 (decornp) morphism comparing the phase-transition behaviours of com- plexes 11. As illustrated in Fig. 6 stage 1, the molecules of 11 (m=1) form dimers in the D,, mesophase." When the R' n group in the vicinity of the core of the complex is substituted by an ethyl group, which is bulkier than the methyl group, this substituent acts as a steric hindrance for forming dimers, and the intermolecular interaction becomes weaker. Therefore, the mesophase of the derivative may change from a D,, phase 6 to a nematic phase (See stage 1-2 in Fig. 6).The bulkier the R' group becomes, the weaker the intermolecular interaction becomes. Hence, the nematic phase may hardly appear and an isotropic liquid phase may appear before the nematic 8 phase (stage 2-3). Complexes 12 In Table 7 are summarized the phase-transition temperatures 182.4 195.310 isoa isoa "1-"4-for 12(n=6,8, 10-12, 14). In Fig. 7,all transition temperatures of derivatives 12 were plotted against the number of carbon atoms (n) in the alkoxy chain. All of these complexes have a nematic phase, and show double-melting behaviour via the 173.2 189.311nematic pha~e.'~,~~ The n=11 derivative has the widest tem- perature range of the nematic phase (AT= 16.1"C)in these complexes.The complexes gradually decompose over ca. 200 "C.Therefore, the derivatives of n d 5 were not synthesized, although they could possibly exhibit the nematic phase above 200°C. The optical textures of these complexes are a Schlieren texture, a marbled texture, and a threaded texture, similar to 12 171.9 183.6 *N-* isoaIL, iK2' 14 163.5 175.1the photographs shown in Plate2. Thus, the mesophase of isoa "4-these complexes could be established as a nematic phase. U stage 2 R' = C~HS,n-C3H7 enantiotropic N U stage 3 is0 (monotropic N) Fig. 6 Possible reason of the n-alkyl chain effect on the mesomorph- ism in 11 a Gradual decomposition for several heating-cooling cycles.Phase nomenclature: K =crystal, N =nematic mesophase and is0 =iso-tropic liquid. 24C t 220 0 g 200 180 164 number of carbon atoms (n ) in the alkoxy chain Fig. 7 Phase-transition temperatures us. number of carbon atoms in the n-alkoxy chain for 12: 0,N-iso clearing point; 0,K,-N melting point; A,K1-N melting point J. MATER. CHEM., 1994, VOL. 4 69 Interestingly, these nematic phases sometimes show a zigzag disclination texture as shown in Plate 3. When the thickness of the sample was thin, this texture could be observed. Chandrasekhar et al. reported that the complex bis[ 1-4-decylbiphenyl-4-y1)-3-( 4-ethoxypheny1)propane- 1,3-dionato]- copper(n), whose structure is extremely similar to that of 12, 2 3 4 5 A. M.Giroud-Godquin and J. Billard, Mul. Cryst. Liq. Cryst., 1981,61, 147. A. M. Giroud-Godquin and J. Billard, Mol. Cryst. Liq. Cryst., 1983, 97, 287. J. Billard, C. R. Acad. Sci. Paris, 1984, 299,905. K. Ohta, A. Ishii, I. Yamamoto and K. Matsuzaki, J. Chem. Soc., Chem. Commun., 1984, 1099; K. Ohta, A. Ishii, 1. Yamamoto and has a biaxial nematic (Nb) and exhibits also the zigzag disclination texture. Therefore, we examined the axiality of the present nematic phase by conoscopic observation. For these observations, it is necessary to align the molecules in a homeotropic structure. Surprisingly, the present nematic liquid-crystalline molecules sometimes align spontaneously in 6 7 8 K. Matsuzaki, Mol. Cryst. Liq. Cryst., 1985, 116, 299. K. Ohta, H. Muroki, A. Takagi, I.Yamamoto and K. Matsuzaki, Mol. Cryst. Liq. Cryst., 1986, 135,247. B. K. Sadashiva and S. Ramesha, Mol. Cryst. Liq. Cryst., 1986, 141, 19. K. Ohta, H. Muroki, A. Takagi, K. Hatada, H. Ema, I. Yaniamoto and K. Matsuszaki, Mol. Cryst. Liq. Cryst., 1986, 140, 131; a homeotropic structure between two cover slips (a thickness G.50 pm) without silane coating on the surface and without applying an external electric field or a magnetic field. Therefore, it is very easy to carry out the conoscopic obser- vations for this nematic state. Plate4A(a) is a conoscopic figure for a spacer thickness of 50 ym. Though the cell was 9 K. Usha, K. Vijayan and B. K. Sadashiva, Mol. Cryst. Liy. Cryst. Lett., 1987, 5, 67; H. Sakashita, A. Nishitani, Y. Sumiya, K. Ohta and I.Yamamoto, Mol. Cryst. Liq. Cryst., 1988, 163, 211. A. M. Giroud-Godquin, G. Sigaud, M. F. Achaid and H. Hardouin, J. Phys. Lett., 1984, L-387; A. M. Giroud-Godquin, M. M. Gauthier, G. Sigaud, F. Hardouin and M. F. Achard, Mol. Cryst. Liq. Cryst., 1986, 132, 35. rotated by 45" [Plate4A(b)] and then by 90" [Plate4A(c)] on the basis of Plate 4A(a), the crossed lines of the conoscopic figures did not change. When the thickness of the spacer was 100 pm, the crossed lines did not change [Plate 4B(a)-(c)], as for 50 pm. In this case, the homeotropic alignment could be achieved only by applying an external electric field (d.c. 200 V). These results supported the fact that the nematic 10 11 12 13 K. Ohta, H. Ema, H. Muroki, I. Yamamoto and K. Matsuzaki, Mol. Cryst.Liq. Cryst., 1987, 147,61. K. Ohta, 0.Takenaka, H. Hasebe, Y. Morizumi, T. Fujimoto and I. Yamamoto, Mol. Cryst. Liq. Cryst., 1991,195, 135. S. Chandrasekhar, B. K. Sadashiva and B. S. Srikanta, Mol. Cryst. Liq. Cryst., 1987,151,93. S. Chandrasekhar, B. R. Ranta, B. K. Sadashiva and V. N. Raja, Mol. Cryst. Liq. Cryst., 1988, 165, 123. phase of the present complexes is uniaxial. Although the tilt angle of the nematic phase was not measured for these conoscopic observations, the homeotropic alignment could be supported because the crossed lines showed no change for rotating the sample. Plate 5 is the conoscopic figure obtained when a 1/41 plate 14 15 16 17 S. Chandrasekhar, V. N. Raja and B. K. Sadashiva, Mol. Cryst. Liq. Cryst., Lett., 1990,7, 65. B. Muhlberger and W.Haase, Liq. Cryst., 1989,5,251. K. Ohta, 0.Takenaka, H. Hasebe, Y. Morizumi, T. Fujimoto and I. Yamamoto, Mol. Cryst. Liq. Cryst., 1991, 195, 123. N. J. Tompson, G. W. Gray, J. W. Goodby and K. J. Toyne, Mol. Cryst. Liq. Cryst., 1991,200, 109. was inserted for the sample of Plate 4B(a). It can be concluded from this figure that the present nematic phase is optically positive. 18 19 20 G. W. Gray, B. Jones and F. Marchon, J. Chem. Soc., 1957,393. K. Ohta, 0.Takenaka, H. Hasebe, Y. Morizumi, T. Fujimoto and I. Yamamoto, Mol. Cryst. Liq. Cryst., 1991, 195, 103. G. W. Gray and J. W. Goodby, Smectic Liquid Crystals, Textures Conclusions 21 and Structures, Leonard Hill, Glasgow, 1984,ch. 5, pp. 83 -88. J. W. Goodby, Mol. Cryst. Liq. Cryst., 1981,75, 179; J.W. Goodby We have investigated the mesomorphic properties of com-pounds 9 and 10. It was found that each of the derivatives 10 for II =6-14 has both enantiotropic E and S, phases but the derivative for n=18 gives a monotropic E phase and an enantiotropic SA phase. On the other hand, the derivatives 9 for m= 1 and 2 have both the enantiotropic E phase and the S, phase, whereas they have a monotropic E phase and an 22 and G. W. Gray, Mol. Cryst. Liq. Cryst., 1978,48,127;G. W. Gray and J. W. Goodby, Mol. Cryst. Liq. Cryst., 1976,37, 157. C. L. Hilleman, G. R. Van Hecke, S. R. Peak, J. B. Winther, M. A. Rudat, D. A. Kalman and M. L. White, J. Phys Chem., 1975, 79, 1566; C. L. Hilemann and G. R. Van Hecke, J. Phys. Chem., 1976, 80, 944; K. Ohta, H. Ema, Y.Morizumi, T. Watanabe, T. Fujimoto and I. Yamamoto, Liq. Cryst., 1990, 8, 311. enantiotropic SA phase for m=3 and only an enantiotropic SA phase for m =4. Therefore, it is apparent that the E phases Paper 3/01024D; Received 19th February, 1993 of both 9 and 10 have a tendency to disappear when the alkoxy/alkyl group is elongated. As reported in the previous papers, complex 11 has a D,, phase'' for R'=methyl group and the complex has a mono- tropic N phase for R' =n-butyl group.I6 The complexes with an R'=ethyl group or an R'=n-propyl group synthesized in this work show an enantiotropic N phase. Therefore, the change from the D,, phase to the N phase in complexes 11 critically occurs with changing from R' =methyl group to R'=ethyl group. The N phase has a tendency to disappear when the substituent R' is longer. All of the derivatives 12 have an enantiotropic nematic phase, and show double-melting behaviour uia the nematic phase. The n = 11 derivative has the widest temperature range of the nematic phase in these complexes. The nematic phase of these complexes was established as an optically positive uniaxial nematic phase, although it showed a zigzag disclin- ation texture. References 1 K. Ohta, Y. Morizumi, T. Fujimoto, I. Yamamoto, K. Miyamura and Y. Gohshi, Mol. Cryst. Liq. Cryst., 1992,214, 161.

ISSN:0959-9428    

DOI:10.1039/JM9940400061

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4( l), 71-79 Helix Inversion in the Chiral Nematic Phase of a Ferroelectric Liquid Crystal Containing a Single Chiral Centre Christa Loubser: P. L. Wessels: Peter Styringb and John W. Goodby*b a Department of Chemistry, University of Pretoria, 0002 Pretoria, South Africa School of Chemistry, The University of Hull, Hull, UK HU6 7RX An inversion in the cholesteric phase has been found to occur with change in temperature. Additionally, the material under investigation was found to exhibit unusual ferroelectric properties on cooling from the chiral nematic phase. We report the synthesis and physical properties, including pitch and polarization data, for (S)-4-n-octyloxy-2,3-difluorobiphenyl-4’-yl 3-difluorobiphenyl-4’-yl 3-fluoro-4-(2-fluorooctanoyloxy)benzoate. Over the past few years there has been considerable interest it has helical inversions in both the cholesteric and smectic in the synthesis and investigation of chiral liquid crystals, and in particular materials that exhibit ferroelectric smectic C* phases.’ In this context we have investigated several chiral systems where an inversion in chiral properties with respect to temperature has been observed.For example, consider the materials shown below. The first compound,2 (S)-2-methylbutyl4-n-nonanoyloxybiphenyl-4-carboxylate,1, exhibits a smectic A and a ferroelectric smectic C* phase. Upon cooling into the ferroelectric phase a spontaneous polarization develops along the C, axis of the phase, with its value reaching a maximum before falling again.3 The magni- tude of the polarization reaches zero before surprisingly increasing again.Initially, at higher temperatures in the ferro- electric phase, the spontaneous polarization is defined as being negative,’ Ps(-),but as the temperature falls the polarization direction inverts and has the opposite sign, P,(+). CeH174m1 CBHl9O 1 The second compound (S)-2-chloropropyl 4-(4-n-nonyloxyphenylpropioloyloxy)biphenyl-4-carboxylate~ 2, exhibits cholesteric and smectic A phases, and on cooling from the isotropic liquid the cholesteric helix unwinds and then rewinds. Thus, at higher temperatures in the cholesteric phase this material possesses a left-handed helix, but as the temperature is lowered the helix unwinds through a cholesteric or chiral nematic phase which has an infinite pitch to give a cholesteric phase that has a right-handed helix.Compound 3, (S)-2-chloro-4-methylpentyl 4‘-(4-n-hexadecyloxyphenyl-propiolo ylox y) biphenyl-4-carboxylate, behaves somewhat similarly to compound 2, except that the helical inversion takes place in the helical ferroelectric smectic C* phase.5 However, in this case the direction of the spontaneous polariz- ation does not invert with temperature. Finally in this series of chiral substances, compound 4 is quite remarkable in that C* phases, and accompanying the helix transposition in the C* phase the direction of the spontaneous polarization also crosses over.6 We have attempted to interpret these effects in terms of a model of interconverting, but competing, species whose con- centrations are temperature dependent.From this hypothesis we have related the interconverting species to changes in the conformational structures of the molecules.3-6 In this current investigation we have examined the liquid-crystalline properties of compound 5, (S)-4-n-octyloxy-2,3-difluorobiphenyl-4-y1 3-fluoro-4-(2-fluorooctanoyloxy)-benzoate and have found that it too undergoes a helix inversion in the cholesteric phase. However, it is clear that an accompanying helix inversion in structure or a flip-flop in the direction of the spontaneous polarization does not take place in the ensuing ferroelectric C* phase. Hence, this mate- 2 F 4 rial provides yet another example of the first class of inversion phenomena (i) listed below.Thus, so far we have observed the following inversions in the chiral properties/structures of optically active liquid crystals: (i) helix inversion in the cholesteric phase; (ii) helix inversion in the smectic C* phase but no polarization flip-flop; (iii) spontaneous polarization crossover in smectic C* but no helix inversion; (iv) inversions in the helices in both cholesteric and smectic C* phases, and polarization reversal in the ferroelectric phase. The only combinations of inversion effects not yet observed are (v) a helix reversal in the cholesteric phase but not in the smectic C*,accompanied with a polarization flip-flop in the ferroelec- tric phase, and (vi) helix inversions in both smectic C* and cholesteric phases but no crossover in the spontaneous polarization.In the following sections of this article we report on the J. MATER. CHEM., 1994, VOL. 4 EF F 5 synthesis and physical properties of (S)-4’-n-octyloxy-2,3-difluorobiphenyl-4’-yl 3-fluoro-4-(2-fluoro-octanoyloxy) benzoate, and describe results which show clearly that this material undergoes an inversion in its chiral nematic phase. Experimental General Synthetic Procedures The compound, (S)-4-n-octyloxy-2,3-difluorobiphenyl-4’-yl 3-fluoro-4-( 2-fluorooctanoyloxy) benzoate, 5, was prepared according to the synthetic scheme shown in Scheme 1. Initially, 2-fluoroanisole (Aldrich), 6, was brominated using bromine in chloroform to give 4-bromo-2-fluoroanisole, 7 (structure con- (300 MHz, CDCl,, TMS): 7.17 (2 H, m), 6.81 [l H, apparent triplet, (H,H,)= 8.4 Hzz‘J (H,F)], 3.84 (3 H, s), 19F: 139.98 (rel.CFCl,)}. The bromo-substituent was replaced with a nitrile group via 6, firmed by NMR spectroscopy, 6 7 C8Hl7O 11 a cyanation reaction utilising dry copper(1) cyanide in N,N-dimethylformamide (DMF) to yield 4-cyano-2-fluoroanisole, 8. This product was demethylated and the nitrile group hydrolysed in one step by the action of aqueous hydrobromic acid (48% wt./vol.) in glacial acetic acid.7 The hydroxy func- tion of the resulting 3-fluoro-4-hydroxybenzoic acid, 9, was then protected with the use of methyl chloroformate to give the derivative 3-fluoro-4-methoxycarbonyloxybenzoic acid,’ 10.This acid was esterified,’ in the presence of diethylazodicar- boxylate (Merck) and triphenylphosphine (Merck), with 4-n- octyloxy-3,2-difluorobiphenol, 11, to produce the ester 4-n-oct yloxy- 3,2-difluoro biphenyl-4’-yl 3-fluoro-4-methoxycarbonyloxybenzoate, 12. The biphenol, 11, was prepared uia a boronic acid coupling reaction between 2,3-difluoro-4-octyloxyphenyl boronic acid 4-bromoiodobenzene in 1,2-dimethoxyethane (DME)” to give the intermediate 4‘-bromo-2,3-difluoro-4-octyloxybiphenyl which was then oxidised via the 4-boronic acid to the 4’-hydroxy derivative. The Mitsunobu reaction’ was used in the esterification of 10 and 11 because the protecting group was to be retained.Normal esterification conditions involving a base such as pyridine or N,N-dimethylaminopyridinehave a strong tendency to remove the methoxycarbonyloxy group. Subsequently, the protecting group of compound 12 was removed by stirring in a concentrated solution of ammonia in aqueous ethanol to yield the free hydroxy group (compound 13).’ 4-n-Octyloxy-3,2-difluorobiphenyl-4’-yl3-fluoro-4-hydroxybenzoate, 13, was then esterified with optically active 6 91;CH30CCI ~0H 10 + EF 12 5 Scheme 1 Synthesis of 5 J. MATER. CHEM.. 1994, VOL. 4 (S)-2-fluorooctanoic acid, 14, in the presence of dicyclohexyl- carbodiimide (DCC) (Merck) and N,N-dimethylaminopyrid- ine (DMAP) (Merck) to give the final product 5. The chiral fluoro-acid, 14, was prepared using the method reported by Nohira et al.," in which (R)-(+)-1,2-epoxyoctane was initially treated with pyridinium poly( hydrogen fluoride) in ether to give (S)-(-)-2-fluorooctan-l-ol. This alcohol was then esterified with acetic acid and the resulting ester oxidised in the presence of nitric acid to give (S)-2-fluorooctanoic acid, 14.The structure of 14 was confirmed by NMR spectroscopy [SH (300 MHz, CDCl,, TMS): 11.64 (1H, br), 4.98 [l H, dt, 'J (HF)=48.8 Hz, (HH)=5.9 Hz], 2.05 (2 H, m, (HF)= 24.4 Hz), 1.64 (2 H, m), 1.51 (6 H, m), 1.09 (3 H, t)]. Purity and Characterisation of Materials The final product was rigorously purified by flash chromatog- raphy over silica gel (200-400 mesh) using dichloromethane as the eluent.The combined fractions were found to show a single spot by thin-layer chromatography (TLC). After removal of the solvent, the final product was recrystallized sequentially from acetonitrile and light petrol (bp 40-60 "C) until constant transition temperatures were obtained. The chemical purity of the product was investigated by both normal and reversed-phase high-performance liquid chroma- tography; the purity was found to exceed 99% by both methods. Normal phase chromatography was performed over silica gel (5 pm pore size, 25 cm x 0.46 cm, Dynamax Scout Column), reversed-phase chromatography was performed over octadecylsiloxane (5 pm pore size, 25 cm x 0.46 cm, ODS Microsorb Dynamax 18 Column). Acetonitrile was the eluent used in both cases.Detection of eluting materials was achieved spectroscopically using a Spectroflow 757 UV-VIS detector (A=254 nm). The chemical structures of the intermediates in the synthetic route and the final product were determined by a combination of IR spectroscopy (Bomem Michelson 100 FTIR spectropho- tometer), NMR spectroscopy (Bruker AC300 NMR Spectrometer used at 25 "C) and mass spectral analysis (VG 7070H spectrometer operating at 70 eV). The optical rotation (Atago Polax-D Polarimeter) of the final product was deter- mined in solution (using chloroform as the solvent and a concentration of ca. 30 mg ml-l), and monitored carefully to ensure that no racemization in the synthetic procedures had occurred. The optical purity (e.e.) of the final product was determined from NMR spectroscopy using europium D-3-heptafluorobutyrylcamphorate (Lancaster Synthesis) as the chiral shift reagent.The melting points of the intermediates were determined using a Gallenkamp melting point apparatus (UK). The transition temperatures and phase assignments for the final products were determined to an accuracy ofk0.1 "C by ther- mal optical microscopy using a Zeiss Universal Polarizing Light Microscope equipped with a Mettler FP52 microfurnace in conjunction with a FP5 control unit, or by using a Leitz Laborlux Polarized light microscope fitted with a hot stage. Photomicrographs were taken using a Zeiss Universal polariz- ing microscope fitted with an RCA Newvicon video camera in conjunction with an Hitachi VY-200A videoprinter.Temperatures and enthalpies of transitions were investi- gated by differential scanning calorimetry (DSC) using a Perkin-Elmer PC Series DSC7 calorimeter. As a check on instrumental accuracy an indium standard was run at a scanning rate of 10.0"C min-'. The measured latent heat, 28.53 J g-l, compared well with the standard value for indium of 28.45 J g-'. The material was studied at various scanning rates (2, 5 or lO"Cmin-'), for both heating and cooling cycles, after being encapsulated in aluminium pans. Similarly, the measured melting temperature of 156.7 "C compared well with the literature value (156.6 "C). The pitch in the cholesteric phase was measured by determining the distance between the dechiralization lines in the fingerprint texture of the phase using a calibrated Filar eyepiece attached to the Zeiss polarizing microscope.The Filar eyepiece was calibrated against a graduated 1 mm microscope scale (10 pm spacing). The magnitude of the spontaneous polarization of com- pound 5 was measured in 0.25 cm2 indium tin oxide ([TO)-coated test cells that were obtained from Electronics Chemicals High Technology Group, Japan. The internal sur- faces of the cells were coated with polyimide and unidirection- ally buffed. Ac fields were applied using an Advance Electronics AF signal generator J2C in a sine-wave mode. The hysteresis loop was observed on a Dartron Instruments dual trace oscilloscope D17 and the spontaneous polarization (P,) was determined using a Diamant bridge.12 P, of com-pound 5 was evaluated using an applied ac voltage of 1OV p.p.at 60 Hz. The data reported were derived from three different runs and are plotted on a single graph so that the average value could be taken as a curve through the data points. Synthesis of Materials Preparation of 3-Fluoro-4-methoxycarbonyloxybenzoic Acid, 10 A solution of sodium hydroxide (0.25 g) in water (10 ml) was chilled to 0 "C in ice. To this, 3-fluoro-4-hydroxybenzoic acid, 9, (0.5 g, 3.2 mmol) was added. Methylchloroformate (0.606 g, 6.4 mmol) was added slowly to prevent the temperature from rising above 5°C. The reaction mixture was stirred at 0.5"C (3 h) during which time a white suspension formed gradually.The pH was adjusted to 4.5 with addition of hydrochloric acid-water (1:1). The resulting precipitate was filtered off, washed with water and recrystallised from ethanol to give the protected acid 10 as a white solid. Yield, 0.918 g, 67%; m.p. 137-139 "C;aH(300 MHz, solvent CDCl,, standard TMS) 7.90 (2 H, m), 7.34 [1H, dd, 4J (FH,) 8.6 Hz, (H,H,) 7.4Hz1, 3.94 (3 H, s) -COOH not observed, 19F: 131.53 (rel. CFC1,); v/cm-l (KBr Disc): 3200-2600 (br), 1776 (vs), 1694 (s), 1598 (m), 1513 (m),1448 (s), 1285 (br, s), 1197 (s) 930 (s), 769 (s). Preparation of 4-n-Octyloxy-2,3-d~uorobiphenyl-4'-yl 3-fluoro-4-methoxycarbonyloxybenzoate, 12 A solution of 4-n-octyloxy-2,3-difluorobiphenol,11, (0.78 g, 2.33 mmol), 3-fluoro-4-methoxycarbonyloxybenzoicacid, 10, (0.5 g, 2.33 mmol) and diethylazodicarboxylate (0.488 g, 2.8 mmol) in dry THF (50 ml) was prepared under an atmos- phere of dry nitrogen.A solution of triphenylphosphine (0.733 g, 2.8 mmol) in dry THF (15 ml) was then added slowly and the solution stirred overnight at room temperature. The solvent was removed in uacuo and the crude product was purified by flash-column chromatography over silica gel (200-400 mesh) using a mixture of light petroleum (bp 40-60 "C) and dichloromethane (1:1) as the eluent. The solvent was removed from the collected fractions and the residue was purified by repeated recrystallizations from pen- tane to give compound 12 as a white solid. Yield, 0.526 g, 43%; K, 84.0 "C K, 103.8 "C N 177.5 "C I; SH(300 MHz, solvent CDCl,, standard TMS): 8.02 (2 H, m), 7.54 (2 H, m), 7.39 (1H, dd), 7.26 (2 H, m), 7.08 (1 H, td), 6.79 (1 H, td), 4.07 (2 H, t) 3.96 (3 H, s), 1.83 (2 H, pentet), 1.47 (2 H, pentet), 1.32 (8 H, m), 0.88 (3 H, t); v/cm--' (KBr disc): 2949 (s), 2922 (s), 2859 (s), 1775 (vs), 1743 (vs), 1508 (vs), J.MATER. CHEM., 1994, VOL. 4 Plate 1 Microscopic defect texture of (S)-4-n-octyloxy-2,3-difluoro-biphenyl-4'-yl 3-fluoro-4-( 2-fluorooctanoyl-0xy)benzoate at the point where an inversion occurs in the twist direction of the helical structure of the cholesteric phase. In the dark areas of the preparation the homeotropic texture of the nematic phase predominates, whereas in the birefringent fingerprint region remnants of the cholesteric phase remain (x 100) Plate 2 Microscopic defect texture of the contact region between (S)-4-n-octyloxy-2,3-difluorobiphenyl-4'-yl 3-fluoro-4-( 2-fluoro- octanoyloxy) benzoate, 5 (bottom left) and (S)-4-n-hexyloxyphenyl 4-[ 4-(4"-methylhexyloxy)benzoyloxy]benzoate, 15 (top right).(x 100) 1472 (s), 1438 (s), 1305 (vs), 1217 (s), 1192 (s), 1077 (vs), 933 (s), 895 (s), 789 (m), 758 (m), 735 (m); m/z:530 [M'], 197, 153. Preparation of 4-n-Octyloxy-2,3-dijIuorobiphenyl-4'-yl 3-jluoro-4-hydroxybenzoate, 13 A solution of compound 12 (0.5 g, 0.94 mmol) dissolved in a mixture of dichloromethane (20 ml) and ethanol (20 ml) was added to an aqueous solution of ammonia (35%, 20ml) at room temperature. The mixture was stirred (4 h) until TLC showed the reaction to be complete.The solvents were removed in uucuu and the residue purified by column chroma- tography over silica gel (200-400 mesh) using a mixture of light petroleum (bp 40-60 "C) and ethyl acetate (2 :1) as the eluent. The solvent was removed from the collected fractions, and the product, compound 13, was purified by repeated recrystallizations from a mixture of ethyl acetate and hexane to give white crystals. Yield, 0.420 g, 94%; mp 135-137 "C; 6, (300 MHz, solvent CDCl, +10% [2H,]DMS0, standard TMS) 7.80 (2 H, m), 7.46 (2 H, m), 7.18 (2 H, m), 7.02 (1 H, td), 7.00 (1H, m), 6.73 Plate 3 Microscopic defect texture of the contact region between (S)-4-n-octyloxy-2,3-difluorobiphenyl-4'-yl 3-fluoro-4-( 2-fluoro- octanoyloxy)benzoate, 5, (bottom left) and (S)-J-n-decyloxyphenyl 4-[4'-(2"-methylbutyl)benzoyloxy]benzoate, 16, (top right) (x 100) Plate 4 Contact region between (S)-4-n-octyloxy-2,3-difluoro-biphenyL4'-yl 3-fluoro-4-(2-fluorooctanoyl-oxy)benzoate,5, and (S)-2-methylbutylphenyl4-n-octylbiphenyl-4-carboxylate, 17 (x 100) (1H, td), 4.00 (2 H, t), 2.17 (1H, s), 1.76 (2 H, pentet), 1.40 (2 H, pentet), 1.22 (8 H, m), 0.83 (3 H, t); v/cm-' (KBr disc) 3500-3200 (br), 2951 (s), 2921 (s), 2854 (s), 1730 (vs), 1619 (s), 1508 (vs), 1470 (s), 1406 (m), 1305 (vs), 1229 (s),1109 (s), 1087 (s) 894 (m), 869 (m), 799 (s), 751 (s); m/z 472 [M'], 335, 334, 222, 221, 140, 139.Preparation of (S)-4-n-0ctyloxy-2,3-d~uurobiphenyl-4'-yl 3-JEuoro-4-( 2-j7uorooctunoyloxy) benzoate, 5 A solution of dicyclohexylcarbodiimide, DCC, (0.12 g, 0.5.6 mmol) in dichloromethane (5 ml) was added to a solution of biphenol 13 (0.2 g 0.42 mmol), (S)-2-fluorooctanoic acid (0.068 g, 0.42 mmol) and dimethylaminopyridine, DMAP, (0.0154 g) in dry dichloromethane (20 ml) under an atmos- phere of dry nitrogen.The reaction mixture was stirred at room temperature (8 h) after which the dicyclohexyl urea that had formed was filtered off. The solvent was removed in uucuo from the resulting solution and the crude product was purified by flash chromatography over silica gel (200-400 mesh) using a mixture of light petroleum (b.p. 40-60 "C) and dichloro- methane (1:4) as the eluent. The solvent was removed from the fractions collected and the product, 5, was purified by repeated recrystallizations from hexane to give a white solid.Yield, 0.165 g, 64%; Found: C, 68.14 H, 6.47. Calc. for C,, J. MATER. CHEM., 1994, VOL. 4 H4, C, 68.17, H 6.54%, 6, (300 MHz, solvent CDCl,, standard TMS) 8.04 (2 H, m), 7.55 (2 H, m), 7.32 (1H, dd), 7.27 (2 H, m), 7.08 (1 H, td), 6.79 (1H, td), 5.20 (1H, dt, 2JHF48.6 Hz, 5.9 HZ), 4.07 (2 H, t), 2.06 (2 H, 2 X m, 3JHF 25.3 HZ), 1.83 (2 H, pentet), 1.59 (2 H, pentet), 1.45 (2 H, pentet), 1.32 (14 H, .m), 0.89 (6 H, 2 x t); v/cm-' (KBr disc) 2951 (s), 2918 (s), 2861 (s), 1769 (s), 1733 (s), 1508 (vs), 1469 (s), 1308 (s), 1202 (s), 1108 (s), 1079 (s), 892 (m), 867 (m), 793 (m), 749 (m); m/z 616 [M+], 503, 334, 283, 255, 222, 221, 140, 139; +5.1"; e.e.>90% (using europium D-3-heptafluorobutyryl-camphorate as the chiral shift reagent).The proton attached to the chiral centre (double of triplets) was found to shift downfield from 5.2 to 5.76, however, no detectable amount of the other enantiomer was found, therefore the optical purity was assigned a minimum value of 90%. Results In the following sections we detail results that unequivocally prove that compound 5 possesses a helix inversion in the cholesteric phase, and we also discuss the ferroelectric proper- ties of the smectic C* phase of the material. Thermal Polarized Light Microscopy Studies Studies on the Pure Material Thermal polarized microscopy of (S)-4-n-octyloxy-2,3-di-fluorobiphenyl-4'-yl3-fluoro-4-(2-fluorooctanoyloxy)-benzoate, sandwiched between an untreated glass slide and cover slip, showed that this material exhibited cholesteric and smectic C* phases.The following transition temperatures ("C) were determined at heating/cooling rates of less than 1"C min-'. K 89.7 Sp 139.3 N*R 140 N*, 140 N*L 149.6 I The cholesteric phase was easily identified from its Grandjean planar and fingerprint textures, and on cooling to just above the cholesteric to smectic C* transition (140-139 "C) an interesting phenomenon was observed. The helical structure of the cholesteric phase was found to unwind at 140°C so that the mesophase became totally untwisted, and then almost immediately upon further cooling a helical structure reformed, but with the opposite handedness.This is seen clearly for the cholesteric phase (N*) at the point where the fingerprint texture gives way to the homeotropic texture of the nematic phase (N*,), as shown in Plate 1. This plate shows the crossover point in the helical-twist sense; in the black region of the photomicrograph the molecules are essentially ordered so that the direction of observation is along the optic axis of the phase, whereas in the fingerprint region a helical structure exists. As the preparation is cooled further, the fingerprints return over the whole of the specimen and eventually a Grandjean plane texture reforms. However, it should be noted that the lower temperature cholesteric phase does not exist over a very large temperature range as it quickly gives way to the formation of a ferroelectric C* phase.Rotation of the upper polarizer (either side of being crossed to the bottom polarizer) of the microscope, above and below the crossover point, reveals that the sign of the helix inverts with tempera- ture. A slight coloration is observed when the upper polarizer is rotated in the same direction as the helix. This experiment shows that the upper-temperature cholesteric phase is right- handed (1) whereas the lower phase is left-handed (d). This process of unwinding of the helix just before the transition to the smectic C* phase leads directly to the ferroelectric phase being formed with relatively good alignment, which is very unusual for a material exhibiting cholesteric to smectic C transition.For the smectic C* phase normal pseudo-homeotropic and schlieren defect textures are observed. Rotation of the upper polarizer confirms that the helical structure is left-handed (d), which is in agreement with the proposed rules linking twist sense and spontaneous polarization dire~ti0n.l~ Contact Studies Various contact preparations of materials of known helical- twist sense were studied in order to confirm the presence of a helix inversion in compound 5, and to determine the helical twist direction with respect to temperature. The first contact to be investigated was that between the test material 5 and the standard 15.13Compound 15, see structure 5 given below, has been reported to exhibit a laevo rotation of plane-polarized light in its cholesteric phase.Moreover, as this compound has an (S) absolute spatial configuration with respect to its chiral centre, which is itself removed from the rigid aromatic core by an odd number (0)of atoms, then this material is categorised as Sol (RH) by Gray and McDonnell rules.I4 Therefore, the material has a right-handed helical structure. The contact region between the two compounds shows no discontinuity between the two cholesteric phases. Plate 2 shows the texture of a typical region that includes the contact boundary at 147 "C. The green colour appears fairly uniform across the photomicrograph indicating that the two phases have roughly similar pitch lengths in their cholesteric phases, and furthermore the selective reflection of green light indicates that the pitch length is ca. 0.7-0.9 pm.This contact therefore confirms that for the upper temperature region of the cholesteric phase the helix has a right-handed twist. 15 17 A second contact preparation was made between compound 5 and ester 16.15 The standard material, 16, in this case is classified as Sed(LH) by the Gray and McDonnell rules. Therefore the two materials should form a nematic phase where the pitch diverges at the contact boundary as the two helical structures compensate. Plate 3 shows the contact boundary of the two materials at a temperature of 141.6 "C. It can be seen clearly from the rivulet of nematic phase (grey area) running across the preparation that there is a di\ ergence in the pitch.The strong variation in colour across the sample also indicates that the pitch length of the mesophase is changing sharply with concentration. The divergence in the pitch confirms that the upper temperature cholesteric phase does not possess a left-handed helix. When the preparation is cooled down to below the crossover point for the helix J. MATER. CHEM., 1994,VOL. 4 inversion in compound 5, the contact boundary shows no discontinuity confirming that the lower temperature choles- teric phase has a left-hand helical structure. A third contact study was performed with the commercially available material (S)-2-methylbutylphenyl 4'41-octylbiphenyl-4-carboxylate, CE8 (Merck), 17.16At high tem- peratures in the cholesteric phase, the test material, 5, and the standard, 17, were found to have helical structures that compensated, i.e.of opposite twist, and at lower temperatures no discontinuity was found at the contact hence the two cholesteric phase have the same twist sense. Thus, this again confirms that in the upper temperature regime of the choles- teric phase the test material possesses a right-handed helix, and at lower temperatures this inverts to give a left-handed helix. Plate4 gives the most comprehensive view of the inversion phenomenon. This figure shows the texture around the contact region at a temperature of 136.8"C, which is close to the crossover point. On the right-hand side of the figure (high concentration of the test material) the texture appears black, this is due to a transition to the smectic C* phase which exhibits a pseudo-homeotropic texture.A grey-black line, which is a rivulet of nematic runs down the centre of the figure, this nematic region is caused by the inversion in the helix of the cholesteric phase of the test material. To the left there is another black area, this is the actual contact zone between the two materials. Although appearing black, this area in fact has a mosaic texture associated with it which is due to the formation of a blue phase (not resolved by the camera). In the blue-phase region, the pitch of the helical structure is ca. 0.5 pm, i.e. the contact zone shows continuity of the choIesteric phase.Thus, to the left-hand side of the rivulet of nematic in the centre of the figure, which includes the contact zone, the cholesteric phase has a left-hand helix, and on the right of the figure the cholesteric phase has a right-hand helix. 30.0-5.00 3.75-\2 22.5 --5 2.50-.c c. a,c 1.25-15.0-0.0 4 Differential Scanning Calorimetry The phase transitions of the material were also investigated by DSC as shown in Fig. 1. This shows the first heating cycle for compound 5, and it can be seen from the thermogram that the clearing point and cholesteric to smectic C* enthalpies are relatively small and that the peak shapes are broad, suggesting that the phase transitions are weakly first order in nature. It is also interesting to note from the enlargement of the region about the clearing point, shown in the insert, that there are accompanying shoulders on both the cholesteric to isotropic liquid and smectic C* to cholesteric transitions.The two shoulders are on opposite sides of the peaks to one another, indicating that these events are not artefacts of the experimental technique but are real. Further investigations of these phenomena were frustrated, however, because of decomposition of the sample in subsequent heating and cooling runs. Nevertheless, fresh specimens were found to give reproducible behaviour. In addition, no changes in the defect textures of the material, when observed in the polarizing light microscope, could be detected at the temperatures where the shoulders were found to occur.Therefore, the extra peaks/shoulders in the first-heating thermogram remain some- what of a mystery. The cooling cycle shows that these two events in the thermogram become better resolved, but the shoulder at the cholesteric to C* transition does not correspond in tempera- ture to the inversion point in the cholesteric phase. At the present time, therefore, we have no real explanation for the appearance of these shoulders; however, their presence is the subject of further investigations. Pitch The pitch of the helix in the cholesteric phase of compound 5 was measured as a function of temperature from a point I I I I 1 I 125 130 135 140 145 150 Fig. 1 DSC heating cycle for (S)-4-n-octyloxy-2,3-difluorobiphenyl-4'-yl3-fluoro-4-( 2-fluorooctanoyloxy)benzoate, 5.Heating rate 10"C min-'. The second trace shows a blow up of the smectic C* to cholesteric and cholesteric to isotropic liquid transitions. In this trace shoulders on both peaks are clearly visible. J. MATER. CHEM., 1994, VOL. 4 near to the clearing transition down to a temperature close to the phase change to the smectic C* phase. The pitch was found to diverge as the temperature was reduced; however, after the inversion in the sign of the helix had taken place the temperature range preceding the formation of the smectic C* phase was too short for accurate measurements to be made. Therefore, Table 1 gives the pitch (pm) measured up to the point of the inversion, and the values are plotted as a function of temperature in Fig.2. This figure demonstrates quite clearly that the pitch diverges with a reduction in temperature, thereby supporting the view that a change in handedness of the helix occurs. It should be noted in this figure that the temperature over which the cholesteric phase exists is shifted to lower temperature. This is the result of the experimental set-up where a thick cell and a different oven/microfurnace were used in the determination of the pitch. Spontaneous Polarization The spontaneous polarization was measured as a function of temperature in the smectic C* ferroelectric phase of (S)-4-n-octyloxy-2,3-difluorobiphenyl-4'-yl 3-fluoro-4-( 2-fluoro- octanoyloxy) benzoate, 5.The results obtained for three consecutive cooling runs are superimposed in Fig. 3. It can be seen from this graph that P, increases almost linearly with the decreasing temperature which is a little unexpected for a material that exhibits a direct cholesteric to ferroelectric C* transition. For a first-order phase change of this type it is to be expected that P, would show a sharp jump at the phase transition before levelling off rapidly as the temperature is reduced. This more gradual rise in the polarization might be due to the fact that the cholesteric to smectic C* transition is a relatively weak first-order transition as shown by DSC. The direction of the P, in the smectic C* phase was determined according to standard procedures' by poling the material which was held in an ITO-coated electrooptic cell.Throughout the whole temperature range of the phase only a positive spontaneous polarisation, P,(+), was observed, con- firming that no polarisation inversion occurred. Table 1 temperature/"C pitch/pm 136.4 8.5 137.6 6.2 138.1 5.6 140.2 4.4 141.5 3.6 143.2 1.8 + 8-6-Es-+I= .-4-P 4 2-4 136 138 140 142 144 TI"C Fig.2 Pitch of the helix in the cholesteric phase measured as a function of temperature for (S)-4-n-octyloxy-2,3-difluorobiphenyl-4'-yl 3-fluoro-4-(2-fluorooctanoyloxy)benzoate, 5 200-160-cu 'E 120-0 92 80-'Ol 80 90 100 110 120 130 TI"C Fig. 3 P, in the ferroelectric smectic C* phase measured as a function of temperature for (S)-4-n-octyloxy-2,3-difluorobiphenyl-4-yl 3-fluoro-4-( 2-fluorooctanoyloxy) benzoate, 5 The maximum value obtained for the polarization is of the order of 200 nC cm-' which is almost twice as large as that observed for the equivalent 2-methylalkyl-substituted systems, i.e.the fluoro-substituent at the chiral centre exchanged for a methyl group. This increase in size of the polarization can be therefore attributed to the increased effective dipole at the chiral centre. Tilt Angle In the process of measuring P,, the optical tilt angle was also determined as a function of temperature. The results obtained are shown in Fig. 4.The variation of the optical tilt angle with temperature was quite surprising.Initially, at the choles- teric to smectic C* transition, a jump in the value of the tilt angle was found as expected for a first-order phase transition. However, instead of the tilt angle levelling off with the reduction in temperature, it started to fall slowly, dmost halving its value over a 50"Ctemperature range. This behav- iour is not due to a falling value of the spontaneous polariz- ation as it increases steadily over the same temperature regime. Moreover, the tilt angle was found to be relatively small for a material that exhibits a cholesteric to smectic C* transition, for which tilt angles of the order to 45"are not uncommon. As the optical tilt angle is related closely to the positions that the transition moments of the molecules make with respect to the layers, these results appear to indicate that the angle which the aromatic core makes with the layer normal decreases with reduced temperature, i.e.the moleculeb stand up as the sample is cooled. This is the reverse of the normally expected behaviour. 251 ...V. 80 90 100 iio ' 120 ' 130 77°C Fig. 4 Optical tilt angle measured as a function of temperature for (S)-4-n-octyloxy-2,3-difluorobiphenyl-4'-yl3-fluoro-4-( 7-fluoro- octanoyloxy) benzoate, 5 J. MATER. CHEM., 1994,VOL. 4 ‘F L 19 Fig. 5 Zigzag shaped conformer structure of (S)-4-n-octyloxy-2,3-difluorobiphenyl-4’-yl3-fluoro-4-(2-fluorooctanoyloxy)benzoate, 5 The results suggest that the parallel core-core interactions (low tilt) in this particular system are relatively strong in comparison to the situation where they are only partially overlapping (high tilt). This hypothesis is somewhat supported by the fact that the phase transition from the cholesteric phase is only weakly first order, whereas typically for the cholesteric to smectic C* phase change it is strongly first order.Discussion The results obtained show that there is an inversion in the helical twist direction in the cholesteric phase of (S)-4-n-octyloxy-2,3-difluoro biphenyl4’-yl 3-fluoro-4-( 2-fluoro- octanoyloxy) benzoate, 5. The helical order, however, does not invert in the smectic C* phase, and polarization and tilt-angle studies show that the direction of P, remains constant with respect to temperature.In previous investigations of inversion phenomena we have ’1 22 23 Fig. 6 Structures of two conformers where the fluoro substituents have been brought into close proximity suggested that conformational structures play an important role in determining both the twist and polarization directions. In this material the structural architecture is very different from that of other compounds that show inversions. In particular, the chiral centre in compound 5 is closer to the core than in other ‘inversion’ materials which have chiral centres usually removed from the aromatic core by at least two atoms, thereby allowing some degree of free rotation about the chiral centre it~e1f.l~ Similarly, the peripheral ali- phatic chain is longer in compound 5 than in compounds 1 to 4 inclusive.We have suggested in the past that a longer peripheral chain has the effect of rotationally damping the motion of the chiral centre, consequently leading to higher spontaneous polarizations.’* The rotational damping, there- fore, should lead to one principal conformational structure being present, thereby negating the possibility of inversions occurring in chirality-dependent properties. This can be Seen to be clearly the case when the conformational structures of compound 5 are considered. The conformers related to struc- tural changes about the chiral centre are shown together in Fig. 5. In this figure, the structures of the conformers of compound 5 are shown in their most extended forms giving the molecule an overall gross zigzag shape which is conducive to the formation of smectic C* phases.” Structures 18 and 19 show a trans relationship of the carbonyl group and the aromatic fluoro substituent, whereas this arrangement is cis for structures 20 and 21.It is expected that the cis forms will be of much higher energy than the related trans structures because of the increased steric hindrance and strong repulsive polar effects. J. MATER. CHEM., 1994, VOL. 4 In all four structures, assuming that the peripheral aliphatic chain attached to the chiral centre is fully extended and, for reasons of packing constraints, the material retains an overall zigzag molecular shape, then the steric and polar properties about the chiral centre will be similar for each conformer, i.e.the steric bulk and dipole associated with the chiral centre will be on the same side of the molecule in all four conformers. If this is the case the four conformers will have the same associated twist and polarization directions, and therefore there will be no competition between conformers to drive the inversion phenomena. However, it is known2’ in a variety of difluoro-substituted systems that the fluoro substituents prefer to be located adjacent to one another with the fluorine atoms lying at the minimum van der Waals’ distance of approach. If we speculate that this is the case for compound 5, and there is some evidence from NMR studies on closely related materials which suggests this is in fact the then we can examine the conformers in slightly more detail with respect to their steric and dipolar properties about the chiral centre.Fig. 6 shows the structures of two possible conformers in the vicinity of the aromatic ring that carries the chiral group. The fluoro substituents of the chiral centre and the aromatic ring have been brought into close approximation so that the fluoro atom lies to the right of the ring in structure 22 and to the left in 23.It can be seen from the stereochemistry about the chiral centre in 22 that the peripheral chain lies closer to the long axis of the molecule than it does in structure 23.Thus, it might be expected that structure 22 is more conducive to forming liquid-crystal phases because of its lath-like shape, and therefore it may be preferred over structure 23.Nevertheless, the two structures will have different polar and steric properties, and more importantly when the two confor- mers are compared it can be seen that two effects will operate from opposing sides of the molecular structure. The opposing effects for the related fluoro-substituents, the carbonyl groups and the terminal aliphatic chains will put the two conformers into competition. Once we have generated a competition between the two conformer species we can suggest that their concentrations are temperature dependent. If the two species are intercon- vertable via a small energy barrier, then at a given temperature one species will dominate over the other, and when this dominance is reversed at another temperature an inversion in chiral properties will occur as suggested previou~ly.~~ Conclusions In conclusion we have demonstrated that (S)-4-n-octyloxy-2,3-difluorobiphenyl-4’-yl 3-fluoro-4-(2-fluoro-octanoyloxy)benzoate, 5, exhibits a twist inversion in helix of its cholesteric phase.We have shown that a conventional model of competing conformer species cannot be used to explain this inversion phenomenon, and therefore we speculate that the competing species can be created through the fluoro interactions of the fluoro substituent in the aromatic core and the fluoro substituent at the chiral centre. Compound 5 is also shown to possess unusual tilt-angle and thermodynamic properties.The authors would like to thank The University of Pretoria for an Overseas Research Grant for C. L., and Thorn EM1 CRL and Bell Northern Research (Europe) for support of a Lectureship to P. S. We are also grateful to Mr. A. Hassett for mass spectral analysis. References 1 J. W. Goodby, R. Blinc, N. A. Clark, S. T. Lagerwall, M. A. Osipov, S. A. Pikin, T. Sakura, K. Yoshino and H. Zeks, Ferroelectric Liquid Crystals, Gordon and Breach, Philadelphia, 1991. 2 J. W. Goodby, E. Chin, J. M. Geary, J. S. Patel and P. 1,. Finn, J. Chem. SOC., Faraday Trans. I, 1987,83,3429. 3 J. S. Patel and J. W. Goodby, Phil. Mag. Lett., 1987,55,283. 4 A. J. Slaney, I. Nishiyama, P. Styring and J. W. Goodby, J.Mater. Chem., 1992,2,805.5 A. J. Slaney, unpublished results, 1992. 6 P. Styring, J. D. Vuijk, I. Nishiyama, A. J. Slaney and J. W. Goodby, J. Mater. Chem., 1993,3, 399. 7 M. F. Nabor, J. T. Nguyen, C. Destrade, J. P. Marcerou and R. J. Twieg, Liq. Cryst., 1991,10,785. 8 E. Chin and J. W. Goodby, Mol. Cryst. Liq. Cryst., 1986,141,311. 9 0.Mitsunobu, Synthesis, 1981, 1. 10 S. Gronowitz, A-B. Hornfeldt and Y-H Yang, Chem. Scr., 1986, 26, 31 1. 11 H. Nohira, S. Nakamura and M. Kamel, Mol. Cryst. Liy. Cryst., 1990, MOB, 379. 12 H. Diamant, K. Drenck and R. Pepinsky, Rev. Sci. Instr., 1957, 28,30. 13 J. W. Goodby, E. Chin, T. M. Leslie, J. M. Geary and J. S. Patel, J. Am. Chem. Soc., 1986,108,4729, J. W. Goodby and T. hl.Leslie, Mol. Cryst. Liq. Cryst., 1984, 110, 175. 14 G. W. Gray and D. G. McDonnell, Mol. Cryst. Liq. Crjst. Lett., 1977, 34,211. 15 J. W. Goodby and T. M. Leslie, in Liquid Crystals and Ordered Fluids, ed. A. C. Griffin and J. F. Johnson, Plennum, New York, 1984, vol. 4. 16 J. W. Goodby and E. Chin, J. Am. Chem. SOC., 1986,108,4736. 17 K. Yoshino, M. Ozah, T. Sakurai, M. Honma and K. Sakamoto, Jpn. J. Appl. Phys., 1984,23, L175. 18 J. W. Goodby, J. S. Patel and E. Chin, J. Phys. Chem. Lett., 1987, 91,5838. 19 R. Bartolino, J. Doucet and G. Durand, Ann. Phys., 1978,3,389. 20 E. L. Eliel, Stereochemistry of Carbon Compounds, McGraw-Hill, New York, 1962. 21 C. Loubser and P. Wessels, unpublished results, 1993. Paper 3/03665K; Received 28th June, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400071

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4( l), 81 -85 Metal-Organic Chemical Vapour Deposition of YBCO using a New, Stable and Volatile Barium Precursor Sarkis H. Shamlian,” Michael L. Hitchman,*” Stephen L. Cookband Barbara C. Richardsb a Department of Pure and Applied Chemistry, University of Strathclyde 295 Cathedral Street, Glasgow, UK GI IXL The Associated Octel Co. Ltd., PO Box 17, Oil Sites Road, Ellesmere Port, South Wirral, UK L65 4HF A new highly volatile Ba-containing precursor which is thermally stable at I atm (101325 Pa) pressure is described. This precursor, which is a bis(P-diketonate) that has only C3F, as substituent groups and which is coordinated with the polyether tetraglyme, is shown to have reproducible carry-over rates and to give repeatable deposition rates for the chemical vapour deposition (CVD) of highly oriented crystalline BaF,.It has also been shown to be suitable for use in the preparation of superconducting YBCO films. Measurements of thermodynamic and kinetic parameters have been made and are compared with those obtained with other Ba-containing precursors. It is concluded that the new Ba-containing precursor is potentially a very promising material for the preparation of superconducting YBCO films. The deposition of inorganic compounds of barium is import- ant for the preparation of high-T, superconducting films or for coatings of, for example, BaF,. Currently, the best quality high-T, films are prepared by laser ablation, but as the demand for large-area samples, high-deposition throughput, good intra- and inter-sample uniformity, and conformality of layers on patterned structures increases, then CVD has the potential to become a powerful technique for preparing thin films of superconductors.However, although a significant number of papers have been published on the CVD of high-T, layers,’ there are still problems associated with reproducibility of growth and the control of layer stoichiometry. These problems often arise from the inadequacies of the precursors, particu- larly those of barium.2 The most commonly used precursors have been based on the fi-diketonate complexes with barium 1 where R and R‘ were initially alkyl groups, but more recently where fluorinated alkyl groups have been used. The main reasons for using fluorinated alkyl substituents is that they are sterically demanding relative to formula weight and that without fluorination intermolecular hydrogen bonding can occur, leading to oligomerisation and decreased volatility. The high electronegativity of fluorine also produces much lower van der Waals interactions.This strategy has been successful in reducing precursor sublimation temperature. For example, with R =R’ =CH, [Le. Ba(ACAC), where ACAC represents pentane-2,4-dionate, commonly known as acetyl- acetonate] it is not possible to sublime the precursor since it undergoes dissociation and only the volatile organic ligand leaves the evaporator,2 but with R =R’=CF, [i.e. Ba(HFA), where HFA = 1,1,1,5,5,5-hexafluoro pentane-2,4-dionate] sub- limation occurs3 at 240-255°C at 1 atm or at ca.165°C at a typical deposition pressure of ca. 10 T~rr.~ Unfortunately, considerable decomposition occurs at both pressures. We have recently rep~rted’.~ on an even more highly fluorinated barium precursor with R =R =C,F,. This compound [Ba(TDFND),.H,O] (where TDFND = 1,1,1,2,2,3,3,7,7,8, 8,9,9,9-tetradecafluorononane-4,6-dione)has an m.p. of 187 “C and, following loss of water, sublimes completely without decomposition at 1 atm. This was the first barium complex to show stable and complete volatilisation at ambient pressure. However, when used for CVD of BaF, it was found that6 whilst there was a very high initial deposition rate there was a marked drop in the rate when the same precursor material was used for a series of depositions.This growth rate variation with time was attributed to the loss of the water and sub- sequent slow structural changes in the molecule leading to stronger inter-monomer forces and decreased volatility. We have subsequently shown that anhydrous Ba(TDFND), gives stable and reproducible growth and that this material is a good precursor for the preparation of high-T, films.7 Anhydrous Ba(TDFND), still has, though, a relatively high melting point, and lower volatility, compared with the yttrium and copper precursors which are usually used for the CVD of YBCO. So to try to enhance the volatility of Ba(TDFND), even further we have coordinated it with a polyether, tetra- glyme (Le.2,5,8,11,14-pentaoxapentadecane).The approach using a polyether complex has been used previously with Ba( HFA),,%” but although the compounds can be completely sublimed at low pressure, at 1 atm decomposition occurs at a temperature near to the sublimation temperature. In this paper we report on the CVD of BaF, and YBCO using [Ba(TDFND), tetraglyme] as a precursor and we show that it gives stable and reproducible growth of high-quality crystal- line films of the former material and, after annealing, super- conducting films of the latter. Experimental Details of the synthesis and purification of the parent barium P-diketonate [Ba( T DFND),] have been given elsewhere.’ The preparation and characterisation of the tetraglyme com- plex have also been reported in detail,” but for convenience they are summarised here.The preparation was carried out under an atmosphere of dry nitrogen using standard Schlenk and syringe techniques. Barium hydride was used as supplied by Strem. Tetraglyme (Aldrich, 99 + %) was dried and stored over activated type 4A molecular sieve. Hexane (Prolab, Normapur grade) was dried by reflux over sodium/ benzophenone in the presence of tetraglyme and was freshly distilled before use. TDFND (5.0 cm3, 20mmol) was added dropwise with caution and starting at ambient temperature to ;I stirred suspension of barium hydride (1.39 g, 10 mmol) in hexane (20 cm3) and tetraglyme (2.2 cm3, 10mmol). An immediate effervescence was accompanied by a rise in temperature and dissolving of most of the solid.After the initial reaction had died down the suspension was heated to reflux in an oil bath for ca. 90 min and a turbid solution formed. The solution was filtered, yielding a clear, faintly yellow solution which, on refrigeration overnight to -20 "C, yielded 9.213 g (78%) of densely packed white crystals. Briefly the NMR analytical data obtained were as follows: 'H NMR dH 5.95 (s, 1 H, COCHCO), 3.73 and 3.53 (m, 8 H, CH,CH,) and 3.32 (s, OCH,). {'H} 13C NMR 6, 175.1 (m, CF,COCH), 117.6 (m, CF,CF,CF,), 109.4 (m, CF,CF,), 89.62 (s, COCHCO), 70.61, 70.22, 69.65 and 69.48 (s, CH2CH2), 58.38 (s, OCH,). Found vs (calculated): C, 28.80 (28.65); H, 1.97 (2.04); N,<0.3 (0); Ba, 11.5 (11.7) wt.%.Thermal analyses were carried out with a Stanton Redcroft STA 100 thermal analyser using Inconel crucibles at atmos- pheric pressure under a nitrogen flow of 40cm3 (standard) min-' (sccm) and at a heating rate of 20°C min-'. The CVD system was an impinging jet reactor (Archer Technicoat Ltd) with a resistive heated substrate platform (US Inc.) and standard gas-handling facilities. Stainless-steel precursor con- tainers were heated by circulating hot air in separate tempera- ture-controlled enclosures. Vapours generated in the containers were then carried by a flow of argon gas through stainless-steel gas lines. The gas lines were fitted with valves which allowed the precursors either to be vented to exhaust, permitting stabilisation of the precursor flows before the start of deposition, or to have direct entry to the reactor.All components of the system between the precursor containers and exhaust, as well as the walls of the reactor, were heated to a temperature of ca. 200°C to prevent condensation of the precursors. The precursors were mixed with 0, immediately prior to impinging on the heated substrate so minimising the possibility of reactions in the gas phase. Waste gases were passed through a cold trap to condense any unreacted precur- sors and were then removed from the system by a rotary vacuum pump. For the Ba precursor the precursor container temperature was either kept constant at 96°C for all the deposition experiments or it was varied in the range 80-107 "C for measurements of the effective enthalpy of vaporisation. The Y and Cu precursors used were Y(TMHD), and Cu( TMHD), (where TMHD =2,2,6,6-tetramethylheptane-3,5-dionate) and the precursor containers were kept at 108 and 101 "C, respectively.These temperatures were chosen to give an approximate ratio of 1 :2 :3 of the three elements in the YBCO films, as determined by EDAX. Measurements of each precursor temperature with a thermocouple probe touch- ing the precursor showed there was a negligible temperature difference between the container and its contents. Deposition of BaF, was performed at a heater platform temperature in the range 400-700 "C corresponding to a substrate tempera- ture range of 350-610°C. Deposition of films containing a mixture of Y,Ba and Cu were carried out at a substrate temperature of 660°C.The pressure in the reaction chamber and in all the gas lines was maintained at 10 Torr throughout for all depositions. The total argon gas flow for deposition was always 600 sccm with 200 sccm being used as carrier gas for each precursor. In the case of BaF, deposition 400 sccm were added just prior to the reactor entry. For an investigation of the order of reaction with respect to the precursor the precursor pot temperature was varied in the range 96-106 "C. The oxygen flow was kept at 200 sccm throughout, but for determining the reaction order with respect to oxygen the flow was varied from 30 to 200 sccm. Substrates for deposition were generally Si( loo), but for investigations of supercon- ducting YBCO cleaved single-crystal MgO tiles (1 cm x 1cm) were used.For BaF, identification of crystalline phases and preferred J. MATER. CHEM., 1994, VOL. 4 orientation in deposited films was achieved using an X-ray powder diffractometer with Cu-Kcc radiation. Film-thickness profiles were measured by a surface profilometer (SLOAN DEKT AK 11) after films had been photolithographically patterned and etched with a suitable mineral acid. Thicknesses, as well as refractive indices, were also obtained by using a two-angle ellipsometer (Gaertner L116B). Films consisting of a mixture of Y and Cu oxides and BaF, were converted to YBCO by postdeposition annealing at atmospheric pressure first of all in a flow of water vapour and then in 0, using conditions and a temperature programme as suggested by Kirlin et a1.12 Measurements of resistance-- temperature characteristics for the YBCO films were made using a standard four-point probe method across the tempera- ture range 10-300 K.Determinations of the quantity of the barium precursor transported into the reactor were made by condensing the precursor vapour in a stainless-steel U-tube (weight ca. 135 g) attached to the line emerging from the precursor pot. The U-tube was cooled by ambient air. The amount of material transported (ca. 3-30mg in 60min) was monitored as a function of precursor temperature, carrier gas flow, and of the time the precursor was held at a given vaporisation temperature.Results and Discussion Fig. 1 shows the results for simultaneous thermal analysis (STA) (combined TG and DTA) for [Ba( TDFND),- tetraglyme], and Table 1 summarises data obtained from this figure together with comparable results for other barium fl-diketonate precursors. The DTA in Fig. 1 clearly shows the sharp melting point of the new barium complex and the TG shows that it sublimes completely without decomposition at a total pressure of 1 atm. Also given in Fig. 1 is an STA for a sample of the complex which had been used for a series of six deposition runs and again no residue remains after complete volatilisation, indicating the thermal stability of the complex. Also since no precautions were taken during handling of the compound to prevent contact with air or atmospheric moist- ure the material would appear to have good resistance to aerial oxidation and not to be particularly moisture sensitive. Samples have been found to have unchanged melting points on prolonged storage in sealed containers and following vacuum sublimation.The slight increase in the melting point of the used sample of the tetraglyme complex compared to that of the fresh sample may be associated with the removal of volatile contaminants, in particular excess tetraglyme, lead- ing to an effective purification of the material. The data summarised in Table 1 show that 1 1 0 50 100 150 200 250 360 3 7°C Fig. 1 Simultaneous thermal analysis of (a) freshly prepared sample of [Ba(T DFND), tetraglyme] and (b)sample after use for CVD.J. MATER. CHEM., 1994, VOL. 4 Table 1 Thermal analysis data for barium P-diketonate compounds water loss onset weightcompound T/"C loss (YO) melt T/"C [Ba( TMHD),H20] 50,90 3,7 -b [Ba( HFODj,H,O] 20,70 2 176 [Ba(DFHD)2H20] 80 2 200 [Ba(TDFND),H,O] 63 2 187 [Ba(TDFNDj21e --186 [Ba( TDFND),lf --196 [Ba(HFA),* tetraglyme] --149 [Ba( TDFND j2* tetraglyme] --70 volatilisation onset T/"C weight loss (%)a decomposition onset/"C total weight loss (YO) reference 200,320' 30 420,580,700' 84 240 38 300 91 220 88 350 91.5 200 99 b- >99 220 98 b- 98 220 99 b- 99.5 160 ca. 609 ca. 29Y 94 160 99 b- >99.5 TMHD, R =R' =C(CH,),; HFOD, R=C(CH,),; R' =C3F7; DFHD, R =CF,; R' =C3F7;TDFND, R =R' =C3F7; HFA, R =R' =CF,."Prior to onset of observable decomposition. bNot observed. 'Occurs in two stages. "Melt possibly detected in residue. 'Freshly prepared (after 0.5 h dehydration at 180°C in the STA). /Aged (after use for > 10 h in growth reactor -recovered sample). gUnder the STA conditions used these events overlap and values are therefore approximate. [Ba( TDFN-D), tetraglyme] has by far the lowest melting point of all the /?-diketonates listed in the table. In addition, it is stable under conditions of ambient pressure, whereas the compound with the next lowest melting point, [Ba( HFA), tetraglyme], is only stable at reduced pressure. Using a higher polyether to complex with Ba(HFA), to form [Ba( HFA),.hexaglyme] (where hexaglyme =2,5,8,11,14,17,20-heptaoxaheneicosane) produces a compound with a lower melting point of 70-72 "C,* but again it is not thermally stable at ambient pressure.Preliminary studies have shown that the hexaglyme complex of Ba(TDFND), has an even lower melting point of 40°C and that it is also thermally stable at 1 atm, but its use as a CVD precursor has not yet been investigated further. An important feature of the polyethers of Ba(TDFND), is that they can conveniently be used in liquid form in a precursor container, and this is highly desirable from the point of view of reproducibility of carry- over rates. The structural form of a solid and the geometry of the container can markedly affect the rate of precursor evaporation., For example, if any sintering of a powder occurs during a deposition run then subsequent runs may give very different results due to changes in the carry-over rate of the precursor.A series of six measurements of carry-over rate made in between deposition experiments gave for a precursor pot temperature of 96 "Cand a pressure of 10 Torr a mean value of (367+38) pg min-' where the error is at the 95% confi-dence level. There was no systematic variation of carry-over rate with time and a significant cause of the size of the error was probably small variations in the temperature control of the precursor pot (see discussion below). The variation of carry-over rate as a function of temperature gave a linear plot of the logarithm of the carry-over rate against reciprocal thermodynamic temperature (Fig.2), and from the slope of this plot a value of the effective enthalpy of vaporisation (AHeff,,)of (99.2+ 11.2) kJ mol-' was calculated. It is import- ant to note that this value of AH,,, will not be a true thermodynamic value since it was determined under non-equilibrium conditions. It may, for example, contain a contri- bution from the energy of activation associated with mass transport and, as such, be dependent on precursor pot geometry. This effect can be shown quite clearly from, for example, the work of Fitzer et aL2 who measured the tempera- ture dependence of evaporation of Ba(TMHD), with two very different geometries for the precursor containers.Values of AHeff,,calculated from their data for the evaporation rates obtained with the two different containers differ by more than 30%. Similarly, literature values of AH,,, for Cu(TMHD), range from ca. 82kJ mol-',13 to 124kJ m~l-','~ again a precursor temperature/'C 115 105 95 a5 757 I I I I 6 31. ' . * ' . -. ' ' 2.5 2.6 2.7. . . ' '. . . .I2.8 2.9 lo3 KIT Fig.2 Temperature dependence of the carry-over rate of [Ba(TDFND)2 tetraglyme]. Total pressure = 10 Torr; carrier gas flow =200 sccm. variation of> 30%, while for Y(TMHD), the variation is >60%.15 Clearly, the AH,,, for [Ba(TDFND), tetraglyme] determined here cannot be meaningfully compared with values reported in the literature for other barium complexes.Indeed, literature values of 'enthalpies of vaporisation' should be treated with great care since the majority of them are far from true thermodynamic values and they should not be quoted as such. However, the relatively high value of AH,*,, does show the need for careful temperature control of the precursor pot; a variation of 1°C would give rise to ca. 10% variation in the carry over rate. Also it can be usefully compared to AH,,, for Ba(TDFND), determined under identical conditions of flow, pressure and precursor pot geometry. For Ba(TDFND),, the value of AHeE,, was found to be (211.8f22.8) kJ mol-'. This significantly higher value than that for the tetraglyme complex is probably indicative of stronger intermolecular bonding in Ba(TDFND),.At a constant temperature of 610°C an average value of (4.69 f0.30) nm min-' was obtained for the growth rate of BaF, from six deposition runs. There was no systematic variation of the growth rate with time. Fig. 3 shows the variation of the growth rate of BaF, with temperature. The presence of kinetic- and transport-controlled regions is evi- dent. However, as has been pointed out elsewhere,I6 there is J. MATER. CHEM., 1994, VOL. 4 : 0 2 1.01 II I 0 -0.5 0 0.01 . " ' 'I. ' " " 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 lo3 WT Fig. 3 Temperature dependence of the deposition rate of BaF,. Total pressure = 10 Torr; precursor temperature =96 "C;carrier gas flow = 200 sccm. not a sharp transition between these two regions and a true kinetic activation energy can only be obtained by allowing for the contribution of transport-controlled growth at low temperatures. This is done using the simple relationship l/jT= l/k + l/j, where j, is the overall or total growth rate, j, is the kinetically controlled growth rate, and j, is the transport-limited con-trolled growth.Taking an average value of j, of 4.86nm min-' from Fig. 3, the kinetic growth rates for T,<43O"C can be calculated. An Arrhenius plot of these j, values gives a kinetic activation energy (E,) of ca. 200 kJ mol-l. This is noticeably higher than the value of E, z 150 kJ mol-' deter-mined for the deposition of BaF, from Ba(TDFND),.7 The higher volatility of the tetraglyme complex probably indicates less intermolecular binding which could be associated with the polyether compound having reasonably strong intramol- ecular binding.Hence a higher energy will be required to break up the molecule. Measurement of film refractive index by ellipsometry gave an average value of 1.474 for T,>430°C. This value is an exact agreement with the literature value for bulk BaFZ,I7 and indicates a high quality crystalline film. The quality of the films was confirmed by X-ray diffraction of BaF, layers deposited at 610 "C which showed that they were highly oriented in the ( 111) direction (Fig. 4). Channelling analysis with Rutherford back scattering (RBS) for 0.5" steps over +5" with respect to the normal to the surface showed no evidence of either planar or axial channeling, again showing the high degree of orientation of the crystalline films.However, such good quality films were only obtained if growth was carried out under mass transport control. For temperatures <ca. 430 "C there was not only a fall off in the growth rate (cf. Fig. 3) but also in the refractive index of the layers (Fig. 5). Layers grown at these lower temperatures were, in addition, very poorly adherent to the substrate. Moving into a depos- ition region where there is some kinetic control means that not all of the precursor reaching the surface will undergo decomposition, and unreacted, or partially reacted, species could be incorporated into the layer, hence leading to a low value of refractive index and a poor-quality film.This concept is supported by the fact that as the partial pressure of the precursor was lowered, at a temperature corresponding to h v) (111).g 10000-2 8000-Y I>-5 6000-c Q)c .E 4000--([I .g 2000--1A. I01 !I:20 40 60 80 100 2tYdegrees Fig. 4 XRD of BaF, on Si. Total pressure= 10 torr; precursor temperature =96 "C; carrier gas flow =200 sccm; deposition tempera- ture =610 "C. mDm 1.4' 8 Q) -0c.-1.21 ' ' . '. ' " . '. I 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 lo3 WT Fig. 5 Dependence of refractive index of barium-containing films on deposition temperature. Total pressure = 10 Torr; precursor tempera- ture =96 "C; carrier gas flow =200 sccm. mixed transport and kinetic control at a higher precursor partial pressure, the refractive index rose until it became equal to 1.474 again (Fig.6) and a good-quality adherent BaF, film was obtained. This is understandable in terms of the shape of a curve for growth rate as a function of reciprocal temperature, where as reactant partial pressure decreases the transport controlled region extends to lower and lower temperatures.I6 It is interesting to note that Ba(TDFND), does not show this 1.6 r X$ 1.4 -.-m. .-9. c.0 e-2 1.2-1.o 0 50 100 150 200 precursor Ar flow/cm3 (std) min-' Fig. 6 Dependence of refractive index of barium containing films on precursor partial pressure. Total pressure = 10 Torr; precursor tem- perature =96 "C;deposition temperature =388 "C. J.MATER. CHEM., 1994, VOL. 4 r of BaF,. It can also be used in the preparation of supercon- ducting YBCO. In this context it is particularly valuable since its volatility is comparable to that of Y(TMHD), and Cu(TMHD), which are generally used as sources for yttrium and copper. Control of the different gas flows then becomes standard procedure for the three cases. Of course, for YBCO preparation, in situ or postdeposition hydrolysis of the fluoride is required and this is a drawback. However, no non-fluorine- containing barium precursors, to our knowledge, show the same degree of stability, volatility and reproducibility that we 1i have reported here for [Ba( TDFND), tetraglyme]. Therefore, this material is an extremely interesting and promising candi- 01date not just for the MOCVD of BaF, but for high-0 50 100 150 200 250 300 TIK Fig.7 Resistance-temperature plot for an YBCO film prepared with the use of [Ba(TDFND), tetraglyme]. Total pressure= 10 Torr; total carrier gas flow =600 sccm; deposition temperature =660 "C. effect; good-quality BaF, films with a constant value for the refractive index can be obtained for deposition well into the kinetically controlled regi~n.~ The presence of the tetraglyme would appear to affect the adsorption properties of the precursor. This is being investigated further. Because of the varying nature of the layer properties at lower temperature for deposition from [Ba( TDFND),- tetraglyme] it was not possible to examine meaningfully kinetic parameters, such as order of reaction.At deposition temperatures >430 "C the variation of deposition rate with partial pressure of the precursor was determined from a log-log plot of growth rate against carry-over rate. The slope of such a plot at 610°C was ca. 0.9, which is consistent with a mass-transport controlled regime. At the same temperature the dependence of growth rate on the partial pressure of oxygen was determined simply by varying the oxygen flow, but keeping the total gas flow and pressure constant. The slope of the plot was ca. 0.1, which means that the order of reaction with respect to oxygen is effectively zero. This is not too surprising since the films being grown were BaF, and not the oxide.No significant participation of the oxygen in the growth process would therefore be expected. The presence of oxygen may be desirable, however, to help to keep the carbon content of the films low. Fig. 7 shows a resistance-temperature plot for an YBCO film prepared, as described above, using [Ba( TDFND),- tetraglyme] as the barium precursor. Semiconducting resis- tivity is seen at high temperatures and the onset of supercon- ductivity is at ca. 80 K with zero resistance being attained at ca. 26 K. These values could undoubtedly be improved upon by optimising the depositions and annealing conditions,12 and the potential value of the barium precursor is apparent. Conclusions We have shown that [Ba(TDFND), tetraglyme] is a stable, highly volatile precursor suitable for reproducible MOCVD temperature superconducting YBCO as well.We acknowledge the support of the Commission of the European Communities under the BRITE/EI JRAM Programme, Contract No. BREU/0438. We also wish to thank Professor D. J. Cole-Hamilton for helpful discussions, Mr. R. P. McGinty and his colleagues for performing the thermal analyses, and Dr. H. Kheyrandish for the RBS results. References 1 M. L. Hitchman, D. D. Gilliland, D. J. Cole-Hamilton and S. C. Thompson, Znst. Phys. Conf. Ser., 1990, 111,305. 2 E. Fitzer, H. Oetzmann, F. Schmaderer and G. Wahl, in Proc. Eighth Eur. Conf. CVD, ed. M. L. Hitchman and N. J. Archer, Les Editions de Physique, Paris, 1991, p.C2-713. 3 A. P. Purdy, A. D. Berry, R. T. Holm, M. Fatemi and D. K. Gaskill, Inorg. Chem., 1989,28,2799. 4 C. O-Gonzalez, H. Schachner, H. Tippmann and F. J. Trojer, Physicu C, 1988,153-155, 1042. 5 S. C. Thompson, D. J. Cole-Hamilton, D. D. Gilliland and M. L. Hitchman, Adu. Muter. Opt. Electron, 1992,1,81. 6 D. D. Gilliland, M. L. Hitchman, S. C. Thompson and D. J. Cole-Hamilton, J. Phys. ZZZ France, 1992,2,1381. 7 M. L. Hitchman, S. H. Shamlian, D. D. Gilliland, D. I. Cole-Hamilton and S. C. Thompson, results to be published. 8 K. Timmer, C. I. M. A. Spee, A. Mackor and H. A. Meinema, Eur. Put. Appl., 1991,405 634, A2. 9 H. A. Meinema, K. Timmer, E. A. van der Zouwen-Assink, C. I. M. A. Spee, P. van der Sluis and A. L. Spek, XXVllIth Znt. Con5 Coord. Chem., Gera, DDR, 1990,Abstr. 6-82. 10 J. A, T. Norman and G. P. Pez, J. Chem. Soc., Chem. Commun., 1991,971. 11 S. C. Thompson, D. J. Cole-Hamilton, S. L. Cook and D. Barr, Eur. Put. Appl., February 1993,92307390. 12 P. S. Kirlin, R. Binder, R. Gardiner and D. W. Brown, SPIE Processing of Films for High T, Superconducting Electronics, 1989, 1187, 115. 13 S. H. Kim, C. H. Cho, K. S. No and J. S. Chun, J. Murer. Res., 1991,704. 14 M. A. V. Ribeiro da Silva, M. D. M. C. Ribeiro da Silva, A. P. S. M. C. Carvalho, M. J. Akello and G. Pilcher, .1. Chem. Thermodynum., 1984, 16, 137. 15 D. D. Gilliland, PhD Thesis, University of Strathclyde, 1993. 16 M. L. Hitchman, Progr. Cryst. Growth, 1981,4,249. 17 CRC Handbook of Chemistry und Physics, ed. D. R. Lide, CRC Press, Boca Raton, F1. 1993 pp. 4-42. Paper 3/04895K; Received 12th August, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400081

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4( l), 87-97 Physical Properties of Polypyrrole Films containing Dicyanoaurate(1) An ions, P Py-Au( CN)P Raoul Cervini," Robert J. Fleming: Brendan J. Kennedy" and Keith S. Murrayfa a Department of Chemistry, Monash University, Clayton, Victoria 3168, Australia Department of Physics, Monash University, Clayton, Victoria 3168, Australia Department of Chemistry, University of Sydney, Sydney, New South Wales, 2006, Australia Polypyrrole dicyanoaurate, PPy-Au(CN),, has been synthesized in thin-film form by electrochemical oxidation of pyrrole in aqueous or non-aqueous media. The film samples have been characterized by elemental analysis, IR spectroscopy and X-ray photoelectron spectroscopy (XPS). The room-temperature conductivity of one of the films (303 S cm-') is one of the highest reported to date for any PPy-anion system.X-Ray diffraction studies show that the preferred orientation of the PPy chains in PPy-Au(CN), is parallel to the anode surface. The films show excellent stability in air at room temperature, their conductivity decreasing by less than 2% over 50 days. The electrochemical properties of PPy-Au(CN),-coated platinum electrodes have been probed in detail using cyclic voltammetric (CV) methods. Electrically conductive organic polymers have been the subject of intense interest over the past few years and a wide range of materials have been de~eloped.'-~ Environmental stability is essential if these polymers are to be used for commercial application^.^,^ In this context polypyrrole (PPy)has attracted considerable interest."" Electro-oxidation of the monomer yields a cationic polymer, charge neutrality usually being achieved through the incorporation of a counterion such as toluene-p-sulfonate, ClO;, NO, or BF; .'' Synthesis con- ditions, e.g.counterion concentration, choice of solvent, elec- tric potential of the anode on which the polymer is deposited and solution temperature, can affect profoundly the electrical and physical properties of the film.'2,'3 Recently the effect of incorporating transition-metal com- plexes in PPy films has been These anions vary in shape and size and in their inherent spectroscopic, magnetic and redox behaviour. Thus large planar anions such as the tetrasulfonated phthalocyanines of manganese,', iron,I5 ~obalt'~,~~ have been incorporated.PPy-and ~opper'~>'*-~~ hexacyanoferrate(11) films and powders have been obtained through chemical and electrochemical oxidation of pyr-role.22,23 Takak~bo~~ synthesized PPy films containing metal-oxalate and metal-EDTA counterions, with electrical conductivities reported of up to 100 S cm-'. Belanger et have prepared a PPy-MoS, conducting polymer by careful manipulation of the reaction conditions. Underhill and co-w~rkers~~,~'have recently reported the synthesis of PPy films containing small planar anions such as [Ni(mnt)J -or CPd(mnt),l-anions (mnt= 1,2-dicyanoethylene-1,2-dithiolene)and the [Pt(CN),]'-dianion; the electrical conductivities of these materials ranged from 15 to 30 S cm-'.Tourillon and Garnier2' earlier obtained similar conductivities when re-doping neutral poly( 3-methylthiophene) with Pt(CN)z- and Au(CN), . It is interesting to note that the latter types of planar and linear cyanometallate anions have been used extensively in the area of conducting charge transfer salts, the crystal structures of which generally display sheet-like network stacking.31 We have recently reported3' the electrosynthesis of films containing the planar tetracyanonickelate dianion, Ni(CN)i-, and these films were shown to display high conduc- tivities of ca. 120 S cm-' and had good environmental stability at room temperature. In this paper we present data on PPy films containing the linear dicyanoaurate anion, Au(CN), .Some of the films obtained here display high conductivities in the vicinity of 300 S cm-', and they have excellent room- temperature environmental stability. The principle behind using such linear anions was to examine the effect of changing the geometry of the dopant from the planar geometry used in our previous st~dies.'~*~**~~ Experimental Synthesis Pyrrole monomer (Aldrich) was purified by distillation under reduced pressure and stored at 0°C under nitrogen. The supporting electrolytes K2Ni(CN)4,32 (Et4N),[Ni(CN)4]33 and (Et,N)[Au(CN),I3" were prepared by published methods. The potassium salt of Au(CN), (Aldrich), the sodium salt of toluene-p-sulfonate and the solvent acetonitrile (Aldrich) were used as received.The pyrrole concentration in the electrolyte solution was maintained at 0.2 mol drnp3. The electrochemical cell in which the films were grown consisted of a single compartment. Parallel platinum (Pt) electrodes measuring 2 cm x 2.5 cm and spaced 1.5 cm apart were used. Before deposition of each film the surfaces of both electrodes were polished sequentially with silicon carbide paper (fine grade) and Al,03 powder and thoroughly washed. The electrolyte solution was deoxygenated by bubbling with high-purity nitrogen and a nitrogen blanket was maintained in the cell during deposition. Most films were deposited at room temperature. Film 7 was grown at 0°C. A BAS Model SP-2 potentiostat was used to maintain the anode, on which the films were deposited, at a constant voltage (0.70-0.95 V) relative to SCE.Current densities were usually in the range 0.10-3.0 mA cm-, with a deposition time of 2 h yielding film thicknesses in the range 15-60 pm. The initial current observed at the onset of electropolymenzation stabilized after a few seconds and remained at this value during the 2 h growth period. Longer periods than ca. 2 h led to a decrease in current on account of anion depletion effects; see Results and Discussion section. Blue-black PPy-Au(CN), films were washed thoroughly in 50% ethanol-water, peeled from the anode with the aid of a scalpel, dried in a nitrogen stream and stored in a vacuum desiccator. The film growth conditions are given in Table 1. Analysis and Characterization Microanalytical data were obtained on a selection of the films by the National Analytical Laboratories, Melbourne, and the Chemical and Microanalytical Services (CMAS), Essendon J.MATER. CHEM., 1994, VOL. 4 Table 1 Growth conditions for several [Au( CN),] --doped polypyrrole films, and individual PPy-Ni (CN),, PPy toluene-p-sulfonate and PPy- Pt(CN), films. Films 1 and 2 grown on indium tin oxide electrodes, the rest on platinum E(vs. current density/ [anion]/mol conductivity" ~sample SCE)/V mA cm-' dm-3 solvent ( 290 K)/S cm 1 PPy-Au(CN), 2 PPy-Au(CN), 3 PPY-AU( CN), 4 PPy-AU (CN), 5 PPy-Au(CN), 6 PPy-Au(CN), 7 PPy-Au(CN), 8 PPy-Ni(CN),' 9 PPy-p-tolSO, 10 PPy-p-tolSO, 11 PPy-CIO, 12 PPy-Pt (CN)," 0.95 0.75 0.025 0.80 0.10 0.025 0.85 0.50 0.05 0.90 0.63 0.025 0.90 0.80 0.025 0.90 0.65 0.025 0.90 1.62 0.025 0.70 2.7 0.05 0.70 2.6 0.05 0.85 2.9 0.08 0.95 2.1 0.08 0.60 - 0.25 acetonitrile + 1YOwater acetonitrile + 1Yo water water acetonitrile + 1% water acetonitrile + 1% water acetonitrile + 1% water acetonitrile + 1% water 50% ethanol-water water acetonitrile + 1% water acetonitrile + 1YOwater water 99 6 54 222 159 198 303 110 54 110 38 17 "The range in conductivity values for films 4-7, synthesized more than once, was +25 S cm-'.bThis film was labelled 11 in reference 30. 'Data from reference 27. North, Melbourne. Good agreement was obtained by the two analytical laboratories when investigating the same film sample.Analytical data are given in Table 2. Before analysis each film was dried at 120°C and Torrt for 24 h over P205. Thermogravimetric analysis was carried out using a Setaram C592, TG-92 instrument at a heating rate of 10°C min-' over the range 20-500 "C, under a nitrogen atmosphere. IR spectra of crushed films incorporated in KBr pellets were obtained using a Perkin-Elmer model 1750 FTIR spec- trophotometer. Diffuse reflectance IR spectra of very thin films (ca. 0.25 pm) still attached to the Pt anode were obtained using a Bio-Rad Digi-Lab FTS-60 FTIR. These spectra were corrected for the background of the Pt substrate. X-Ray diffractograms (XRDs) were obtained using a Scintag Pad( V) diffractometer and Cu-Ka radiation.Scattered intensity data were collected over the range 3-100". A Hitachi S570 electron microscope operating at 10 kV was used to obtain scanning electron micrographs (SEMs) of the growing and anode faces of the films. A BAS-100 electrochemi- cal analyser was used for cyclic voltammetric studies. The electrolyte solutions in acetonitrile (HPLC grade) and water were checked for any spurious redox active species. The electrical conductivity of the films was measured using the van der Pauw four-probe method. Four small gold elec- trodes were vacuum deposited on the edge of one surface of the film, and thin copper wires were attached to these elec- trodes using silver epoxy resin (Allied Products Corp.).The conductivities of samples cut from the same sheet of film agreed to within 10%. In order to measure the electrical conductivity of the film in the direction normal to its flat surfaces, it was removed from the electrode and a gold t 1 Torr= 133.322 Pa. electrode was vacuum deposited completely covering one side. A gold counter electrode was also vacuum deposited on the other side of the film but of smaller area. The conductivity, 6,was then calculated from the equation a=t/(AR), where t is the film thickness, A the area of the smaller electrode, and R the measured resistance. In order to investigate the influence of contact resistance on the (r values, the measurements were repeated on samples of nearly double the thickness.Surface resistivity measurements were made by evaporating two straight parallel electrodes of length d on the surface of the film, separated by a distance L The surface resistivity was evaluated from the equation, p =Rd/L. X-Ray photoelectron spectra (XPS) were recorded on a Kratos X-SAM 800 with an Mg-Ka (1253.6 eV) X-ray source operating at 10 or 15 mA and 15 kV. For survey spectra the pass energy was 80 eV, while individual lines were collected with a pass energy of 20 eV. The operating pressure was below 5 x Torr. No evidence for beam-induced reduction of the samples was observed. Double-sided adhesive was used to attach a self-supporting disc of PPy-Au(CN), and powder sample of (Et,N)Au(CN), to the stainless-steel disc.The growth surface was used for the film sample. Samples were transferred through air to the spectrometer and were not pretreated in the spectrometer. The transfer method causes some contamination of the surface. Spectral analysis employed Vision software in which a linear background was included and individual lines were taken to have a mixed Gaussian/Lorentzian (90 : 10 or 50 : 50) shape. The surface composition was calculated for the appropriate sensitivity factors. Results and Discussion Influence of Synthesis Conditions on Conductivity Table 1 gives the growth conditions and the electrical conduc- tivity for seven PPy-Au(CN), films and individual Table 2 Microanalytical data for three [Au(CN),] --doped PPy films, and individual PPy-Ni(CN),, PPy-toluene-p-sulfonate and PPy-Pt(CN), films average %C %H %N %O charge on sample no.formula calc. PPy--Au(CN), 3 (C,H3N),,5[Au(CN),][0]0,534.2 PPy-Au(CN), 4 (C,H,N),,,[AU(CN),][O]~~~34.2 PPy-Au(CN), 7 (C,H".5 [Au(CN),I 31.7 PPy-Ni(CN), 8 (C,H,N),,,[Ni(CN),][O],,, 56.7 PPy-p-tolSO, 9 (C,H,N)3,,[p-t0lSO3][O]o,5 59.0 PPy-Pt(CN)," 12 (C,H3N),,5[Pt(CN),][H,O],,o 43.6 "Data taken from reference 27 pyrrole conductivity anal. calc. anal. calc. anal. calc. anal. unit ( 290 K)/S cm -34.9 2.0 1.8 14.9 14.3 1.9 1.8 0.4 54 34.6 2.0 1.8 14.9 13.7 1.9 2.0 0.4 222 32.7 1.6 1.7 14.8 13.5 --0.5 303 56.5 3.0 3.1 23.6 23.0 6.8 6.8 0.33 110 59.3 4.4 4.1 11.5 11.4 --0.33 54 43.6 2.9 2.8 19.0 18.4 --0.4 17 J.MATER. CHEM., 1994, VOL. 4 PPy-Ni(CN),, PPY-P~(CN),,,~and PPy-toluene-p-sulfonate films. The measured conductivities of PPy-toluene-p-sulfonate are similar in size to values reported earlier for this system. Perusal of the data for films 1-7 reveals a large spread in conductivity values, depending on synthesis condition^.^,^,^^ The growth conditions giving PPy-Au(CN), films with high- est conductivity consist of an applied voltage of 0.9 V, an anion concentration of 0.025 mol dmP3, an acetonitrile + 1% water medium and a Pt substrate. Films of PPy-Au(CN), were mechanically robust when grown from either aqueous or organic media. The surface of the film which was in contact with the electrolyte had a blue-black matted appearance, while the surface contacting the anode had a brilliant dark- blue appearance.On the other hand films of PPy-Ni(CN), grown from organic media (acetonitrile + 1YO water) were found to be inflexible and easily torn. As indicated in the introduction, Tourillon and Garnier29 have previously pre- pared polythiophene films from water with Au(CN), as the dopant, and the conductivity value for a pressed pellet was 10 S cm-', much lower than the values found here (see later). PPy-Au(CN), films 3-7 had thicknesses in the range 45-55 pm while films 1 and 2 were ca. 20 pm thick. Films grown on IT0 electrodes did not adhere well to the electrode surface and spontaneously detached from the anode surface when reaching a thickness of ca. 20 pm. The highest value of conductivity obtained for the various PPy-Au(CN), films was 303 S cm-' (film 7) and a synthesis temperature of 0°C was employed in this case.Several groups have reported that lowering polymerisation temperature in this way can lead to such a conductivity in~rease.,~,~~ Presumably the lower syn- thesis temperature causes a decrease in the frequency of branching along the PPy backbone and hence the degree of crosslinking, which in turn increases the mobility of the charge carriers. The conductivity values reported here for PPy-Au(CN), are some of the highest reported for PPy films, and exceed the values observed for other PPy-cyanometallate films. Some interesting effects were observed during and after the electrosynthesis of PPy-Au(CN),.Depending on the synthesis conditions, the current density began to fall gradually after approximately 2 h of growth. The extent of this drop was greater when water was used as the solvent than when acetonitrile was used. Secondly, examination of the cathode electrode after the electrosynthesis revealed deposition of metallic gold. Clearly, anion depletion of the solution had occurred to some degree as the electrosynthesis proceeded. While we have not investigated these chemical changes in any detail, we note that early has suggested that they are complex and may depend on the pH of the solution. Kinetic studies of the reduction of Au(CN), solutions have proposed that the neutral intermediate species, AuCN, exists in equilib- rium with Au(CN); prior to a slow electron-transfer process.i.e. Au(CN), + AuCN+ CN- AuCN + e-+Au + CN -Despite this anion depletion process we have firm evidence that Au(CN), is the sole anionic dopant in the PPy matrix. Electron microprobe data for PPy-Au(CN), revealed the expected three gold peaks, and microanalysis showed, by difference, that 50% of the film mass was gold in the form of Au(CN);. The IR spectrum of PPy-Au(CN), showed a very sharp stretching frequency at 2154 cm- ',characteristic of the coordinated CN group. The corresponding band for K [Au(CN),] occurs at 2141 cm- '. Furthermore, experiments designed to prepare a PPy-CN film were unsuccessful, leading to the conclusion that free CN- is not incorporated readily in the film, as would be expected from the above equations. The most compelling evidence for Au(CN), being the chemi- cal form of the dopant was provided by XPS data and described below.Table 2 gives elemental analysis data for three of the PPy-Au(CN), films, numbered as in Table 1, and for PPy- toluene-p-sulfonate and PPy-Ni(CN), films. Previouslj pub- lished data for PPy-Pt(CN), are also presented for compari- son. In general the measured C, H and 0 percentages are close to those calculated from the assumed empirical formula, but the N percentage is a little lower. Great care was taken to dry all films thoroughly prior to analysis being made. Generally films grown from aqueous or organic solvents suffered a 2-5% weight reduction when investigated by thermogravimetric analysis under an atmosphere of nitrogen.This corresponds to oxygen being incorporated in the films the amount of which is consistent with the findings of several other workers, who have synthesized PPy films containing a range of anions.3843 This extraneous oxygen most likely exists either as covalently bound oxygen, present as hydroxy or carbonyl groups, or as chemically adsorbed oxygen, rather than as water.,' It is expected that conductivity will be influenced by doping level and hence by the degree of charge on the PPy backbone. Specifically, the greater the backbone charge the greater should be the conductivity. However films 3, 4, and 7 do not show this expected relationship. As the conductivity decreases, the average charge on the pyrrole ring decreases to 0.33.Edge et a/.27have reported a charge of 0.4 on each ring in their PPy-Pt(CN), film but the conductivity value is much lower in that case. Thus films 3 and 4 have similar elemental compositions and backbone charges, but very different con- ductivities. Similar backbone charges would be expected to yield similar charge carrier concentrations. Since the conduc- tivities differ markedly, so must the charge carrier mobilities, presumably because the polymer network structures are different. X-Ray Photoelectron Spectra The survey spectrum of the reference compound, (Et,N)Au(CN), is shown in Fig. l(b) whilst the Au 4f XPS is given in Fig. 2(a). Other than weak lines from a Ag contami- nation on the X-ray run no extraneous lines are observed in the survey spectrum.The Au 4f spectrum is a simple spin- orbit doublet. The Au 4f7/,-4f5,, binding energies (EB) are 85.0 and 88.6 eV, respectively. The Au 4f7/, EB demonstrates this to be Au', the values for Auo and Au"' being 83.8 and ca. 87.5 eV, re~pectively.~~ The half-widths are 1.55 eV. The sur- face Au: C ratio 1:6 was different from the expected 1 : 10, although we note the N 1s signal was poorly resolved and there is a larger than usual error in this e~timation.,~ The survey spectrum of the PPy-Au(CN), film [Fig. l(a)] showed the expected C Is, N Is, 0 1s and Au lines as well as those due to an Ag impurity in the spectrometer. Apart from an unidentified line near 440 eV there is no evidence for any unexpected impurities.The Au 4f signal, shown in Fig. 2(b), is a well resolved doublet with E, of 85.1 and 88.7 eV. It is obvious from this that one valence state of Au is present, and the relatively narrow half-widths (1.47 eV) suggest that a single species is present. The N 1s spectrum is a well resolved singlet at 399.2 eV [Fig. 2(c)]. That a single N 1s line is observed is rationalised by noting that the N 1s EBof CN- is ca. 399.5+ 1 eV which is comparable to that found in polypyrrole. In a number of polypyrrole samples the N 1s XPS appears as an asymmetric doublet owing to (partial) localization of the charge on one of the two pyrrole N sites?6 In the present PPy-Au(CN), film, the fact that a single N 1s J.MATER. CHEM., 1994, VOL. 4 iC fc Is Au 4f '9% \(%"-,i__ I 600 500 400 300 200 100 binding energylev Fig. 1 XPS of (a) PPy-Au(CN),, (b) Et4N[Au(CN),]. Peaks marked (*) are due to contamination in the spectrometer at the time of measurement 90 aa 86 a4 peak is observed, demonstrates that there is complete delocal- ization of the positive charge. The C 1s manifold for PPy-Au(CN), can be analysed in terms of three or four lines. The peak positions and relative intensities are listed in Table 3. The weak, higher EBlines are indicative47 of oxidised groups such as C-0 or C-OH as postulated from the microanalytical data. The two stronger lines have been assigned to a and p carbon atoms on the pyrrole group. When the C 1s photoemission manifold was considered to contain only three peaks, the separation between the a and B carbon peaks was unrealistically low indicating that four lines are present.The apparent intensity of the B binding energy/eV carbon line is higher than that reported by Atanasoska et ~22.~~ for C10;-doped PPy and this possibly reflects differences in Fig. 2 XPS of (a)Au 4f levels in Et4N[Au(CN),], (b)Au 4f levels in the amount of hydrocarbon impurities present in the two samples, such impurities being intrinsic to the films or originat- ing from the atmosphere or from spectrometer oil. The 0 1s peak is a poorly resolved singlet with EB=532.6 eV. The surface Au :N atomic ratio is 1 :3 :6. In this case both the Au 4f and N 1s XPS are well resolved and some confidence is placed on the accuracy of this ratio. The similarity of the Au 4f XPS of pure (Et,N)Au(CN), and PPy-Au(CN), demon- strates the electron density on Au in the two samples is very similar and suggests the Au(CN), moiety has remained intact upon formation of the polypyrrole film.Moreover the Au :N ratio is in reasonable agreement with that expected from the analytical data. In short, the XPS strongly support the hypoth- esis that Au(CN), is present as the only anion in the polypyrrole film. sample mA c 1s ~~~~ ~ Et4N [Au(CN)zI 15 284.6 (78) PPy-Au(CN), 10 286.1 (17) 288.3 (5) 284.0 (25) 284.7 (56) 285.7 (16) 287.1 (3) Crystallinity In order to determine whether PPy-Au(CN), films adopt a preferred orientation with respect to the anode surface, X-ray diffraction measurements were performed in both the reflec- tion and the transmission modes.These operational modes correspond to the scattering vector being perpendicular and PPy-Au(CN),, (c) N 1s level in PPy-Au(CN),; [note that labels (a) and (b) are the reverse of Fig. 11 Table 3 Summary of XPS results. The figures in parentheses are the relative intensities, normalised to 100 for each PES current/ Au 4f" 85.0 (57) 88.6 (43) 84.9 (56) 88.5 (44) "For Au 4f the two lines are the spin-orbit components of the 4f signal of a single species. parallel to the film surface respectively. The X-ray diffraction data for PPy films are summarised in Table 4. The reflection diffractogram for PPy-Au(CN), (4) shown in Fig.3(a)exhibits on!y one peak at 20 =26.6", corresponding to a d spacing of 3.35 A, in the direction normal to the surface J. MATER. CHEM., 1994, VOL. 4 Table 4 XRD data in reflection and transmission modes for PPy-Au(CN), and PPy-ClO, reflection d spacing transmission d sp$cing film 2OIdegrees /A 2OJdegrees /A q/S cm-' a,/S cm-011:GI PPy-Au(CN), 26.6 3.35 17.3 5.12 222 ~~ ~ 0.0033 67000 (4) 35.0 2.56 PPy-c10, ( 11) 23.0 3.87 22.0 4.04 38 0.05 760 I 'I I 200 150 100 0 50 25 1 I I 1 0 12 22 32 12 22 32 angle (28)/deg rees Fig. 3 XRDs for films (a) PPy-Au(CN), (4) in reflection mode, (b) PPy-Au(CN), (4) in transmission mode, (c) PPy-C10, (11) in transmission mode and (d) PPy-C10, (11) in reflection mode on which the X-rays are incident.This peak has been assigned to scattering from face-to-face PPy chains at spacings close to the van der Waals distance for aromatic groups.35 It is thus associated with crystalline order in the polymer backbone and is in the 28 range 20-28" in the XRD of all PPy-anion systems. In the transmission mode, the X-ray beam is again incident on one of the broad surfaces of the sample but the detector is located on the opposite side of the sample from the X-ray source. If the PPy chains were arranged with some form of long-range order in the direction normal to the broad surface in which the X-ray beam is incident, a clearly defined peak would be expected, as in the reflection mode data.However the peak at 26.6" in Fig. 3(b) is weak and broad, and further weak features exist around 28 =35". A very diffuse and weak peak appears around 28 =17.3", suggesting some very weak ordering may be occurring with a d spacing large compared to that observed in the reflection mode. It is clear that the preferred direction of structural ordering is parallel to the broad surfaces. The molecular structure of Au(CN), in crystalline salts such as K[Au(CN),] consists of layers of linear so that when used as a dopant anion Au(CN), would be expected to promote alignment of the PPy polymer chains parallel to the anode surface. This appears to be the case, in view of the differences noted in the transmission and reflection diffractograms of PP~-Au(CN)~.The scattering intensities around 17 and 35" in the transmission mode may correspond to scattering from the random packing of the PPy chains. In order to examine the effect on crystallinity of changing the spatial geometry of the anion, a film of PPy-C10, (1 1) was grown from acetonitrile + 1 % water and X-ray diffracto- grams were again obtained in both transmission and reflection modes. In reflection mode [Fig. 3(d)], where the incident beam is probing the structure within the film, a scattered intensity peak w~s observed at 28=23", corresponding to a d spacing of 3.86A. A slightly broader peak was observed in the transmission mode at an angle of 22", corresponding to a d spacing of 4.04 A [Fig. 3(c)].Thus the two curves were very similar, indicating that PPy-ClO, possesses a very nearly isotropic molecular organisation. Amongst previous structural studies Mitchell and co-worker~~~~~have carried out extensive studies focussing on the molecular organisation of PPy films. They used a diffractometer operating in a symmetrical mode, maintaining a fixed angle between the plane of the sample and the scattering vector, over the complete scattering vector range. This arrangement enabled them to examine the structure within the plane of the sample and through its thickness. Their results indicated that PPy films doped with toluene-p- sulfonate and naphthalene disulfonate exhibited significant molecular anisotropy arising from the planes of the pyrrole rings having a preferred orientation with respect to the anode surface. Incorporation of planar-type dopants would be expected to enhance this effect.Since toluene-p-sulfonate has both planar and non-planar features, the unit may act as a template during the growth of the film. However the use of approximately spherically shaped counterions such as ClO, or BF, tends to give a largely disordered molecular structure. In related studies Buckley et reported the X-ray diffraction patterns for a number of PPy films doped with alkylsulfonate anions. The peak at 28 =20-28" in reflection mode assigned to interplanar PPy scattering, was absent in transmission mode, thus implying that the PPy chains lie principally in the plane of the film.Conformation energy calculations5' also confirm that the lowest energy confor- mation for linked pyrrole units is the planar configuration. It is therefore likely that anions having essentially planar shapes can, in general, play the role of a template and influence the orientation of the growing PPy chains. In an attempt to see if the anisotropic structural features exhibited by PPy-Au(CN)* related to electrical properties, the conductivities of the bulk, parallel and perpendicular to the broad surfaces of the films, were measured and the results are shown in Table4. The conductivity of PPy-Au(CN), parallel to the film surface, q,is 222 S cm-', an above-average value for PPy films, The surface conductivity contri- bution to the oIl value of 0.5 S cm-1 is negligible.The perpen- dicular conductivity, oL,was 3.3 x lop3S crn-', averaged over two films of different thicknesses (3.7 x lop3S cm-I for a 50 pm film and 3.0 x S cm-1 for an 85 ym film). Thus the conductivities are highly anisotropic with o!,/:~=6.7 x lo4. Although there is some indication of conductivity normal to the flat surfaces decreasing with increasing thickness, the variation is small compared with the ratio ol,/a,. Assuming that the PPy chains in PPy-Au(CN), are preferentially aligned parallel to the electrode surface, then transport of charge should be easier along the pyrrole chains than between them. In the bulk material the charge carriers move in random directions in the absence of an external electric field. Application of such a field along a particular direction will generate a steady drift velocity in that direction.Thus a field applied between two electrodes on opposite broad surfaces of a planar sample will yield the conductivity normal to these surfaces (neglecting contact resistance). The four-probe tech- nique, with four electrodes near the periphery of one surface of a planar sample, will yield a conductivity which in principle is influenced by charge flow both along the surface and in the bulk parallel to the surface. If it is assumed that the surface component is negligible, then the bulk conductivity parallel to the surface can be obtained.The more anisotropic a film the more all/aLwill depart from unity. There have been a few previous reports in the literature dealing with such conduc- tivity mea~urements.~',~~-~~ Mitchell et aL5' reported consider- ably lower values of 0, compared with oII for PPy-toluene- p-sulfonate. Kaneto et al.57 and Ito et a1.58separately reported oll/o, values of 103-105 for a series of polythiophene films. Garcia-Camarero et reported a q/oLvalue for PPy-C10, of 760 the same as that obtained here. They concluded that the intra-chain conductivity is higher than the inter-chain conductivity, because the PPy chains are arranged in chains parallel to the electrode. However, no XRD data were pro- vided to support these claims. The present X-ray diffracto- grams for PPy-ClO, would suggest that it is an essentially isotropic material, despite the oll/ol value of 760, the latter being lower by a factor of 10, compared with the value obtained for PPy-Au (CN ),.Morphology Scanning electron micrographs for a series of PPy-Au(CN), films grown on a platinum substrate are shown in Fig. 4. The PPy-Au(CN), sample (3) grown from water displays a rough globular surface, and there is evidence of tube-like structures. In this case the tubes are of different sizes. Some are open and others closed; however all these structures are composed of nodular aggregates. An interesting feature of this sample is the surface of the anode face, which is usually reported as being featureless" but which here appears blistered.A some- what similar description of the anode surface morphology of J. MATER. CHEM., 1994, VOL. 4 PPy-Co-sulfophthalocyanine in terms of a 'interconnected fibrillar array' has been reported recently by Chiu et The micrograph of PPy-toluene-p-sulfonate, grown in water and shown in our earlier paper,30 displays a regular array of tubular structures with only some having nodular aggregates on the walls of the tubes. Burford et a/.8 have observed that such growths may result in the generation of oxygen gas at the surface. In contrast to films grown from water, films of PPy-toluene-p-sulfonate grown from acetonitrile possess a regular array of overlapping hemispheres which are tightly packed and dense [Fig.4(f)]. The anode surface of PPy- Au(CN), grown from acetonitrile, (d), is featureless and a cross-section of the same film, (e),shows evidence of a layered structure.The latter feature is possibly associated with the growth of the PPy chains parallel to the anode surface as predicted from the XRD data. It is clear from the SEM data that the choice of electrolyte and solvent can affect the surface morphology of PPy films, especially when the solvent used is water. Environmental Stability of Free-standing Films The variation of conductivity of some of the PPy-Au(CN), films during a 50 day exposure to air at room temperature, is shown in Fig. 5. The PPy-Au(CN), film with an initial conductivity of 303 S cm-', 7, showed a less than 2% decrease in conductivity.Films 3 and 4 showed decreases of 15 and 7%, respectively. These films all follow a trend, namely that the higher the initial conductivity of a sample the less prone it is to attack by oxygen. We have observed this phenomenon previou~ly,~'as have a number of other author^.^ In general terms, as the polymer chain becomes more highly oxidised, as a result of greater doping levels, and therefore more conjugated, then its stability is increased. Thus there appears to be a correlation between the anion doping level and the magnitude of conductivity in PPy. It is interesting to note for the sequence pyrrole, bipyrrole, terpyrrole and polypyrrole, the oxidation potential for the oxidation reaction shifts towards more negative values.62 This relates to the reactivity of each species towards air, which increases as the oxidation potential becomes more positive and the conjugation length correspondingly becomes shorter.Thus for PPy films which contain the same dopant anion, yet differ by several orders of magnitude of conductivity, the prediction of long-term stab- ility can be made with confidence. The initial increase in conductivity for PPy-Au(CN), (7) when exposed to air has also been observed previously for PPY-N~(CN),~' and for toluene-p-~ulfonate.~~ The effect of oxygen on polyacetylene samples also led to an increase in conductivity and this was attributed to the formation of charge carriers on the polymer chains due to removal of electrons by oxygen.63 It was reported that upon further exposure the conductivity of polyacetylene decreased, due to irreversible damage to the polymer chains by formation of hydroxy and carbonyl groups.The same phenomenon may apply to the present PPy-Au(CN), films. Degradation studies of PPy-Au(CN), films under high- temperature conditions are currently being made and reflec- tance FTlR spectroscopy is being used to monitor the films. The thermal ageing at elevated temperatures, in dry air, mimics the room temperature effect of air, but at a much accelerated rate. In general the peaks broaden with increasing ageing time and shift toward higher frequencies. Important IR bands occur at ca. 3300 cm-' (probably owing to hydroxy or hydroperoxo groups) and ca. 1720 cm-' (carbonyl groups).These results suggest that incorporation of oxygen into the PPy backbone leads to a decrease in conjugation and hence, also, in conductivity. J. MATER. CHEM.. 1994, VOL. 4 Fig. 4 SEMs of PPy-Au(CN), grown from water (3); (a) growth surface, (b) anode facing surface and from acetonitrile+l% water (4), (c) growth surface, (d)anode facing surface, (e) cross-section. SEM of the growth surface of PPy-toluene-p-sulfonate grown in acetonitrile + 1'41 water is shown in (f) J. MATER. CHEM., 1994, VOL. 4 cycle, which cause the microscopic pores or channels within the polymer to straighten and/or open up, allowing free .--movement of ions and solvent on subsequent cycles. Pickup c .-> et have reached similar conclusions. 03 1.o By cycle 10 (not shown) good reversibility is evident, withUc 0 fairly symmetrical cathodic and anodic peaks, suggesting rapid-.-2 charge transport is occurring.A useful measure of electro-chemical reversibility is the ratio ip(,) :&), which should ideallyTS) L 0 be unity. For system (i) this was 1.24 at cycle 10, falling to 1.2 after the 20th cycle and thereafter remaining unchanged to the 35th cycle [Fig. 6(b)].Thus it appears that degradation of the film is negligible after the first 10cycles. The voltammog-rams of PPy-Au(CN),/MeCN-TBAP, system (ii), are very0.81similar to system (i). The ip(c) ratio for cycle 10 was 1.49,0 20 40 60 time/days Fig.5 Effects of ageing in laboratory air on the room-temperature conductivity of (W) PPy-Au(CN), (7); (0) PPy-Au(CN), (4); (A)PPY-AU(CN), (3) Cyclic Voltammetry of PPy-Au(CN ), coated Pt Electrodes A series of thin films of PPy-Au(CN),, 0.5 pm thick, were grown on a platinum disc electrode (area=0.02 cm2) by passing a charge of 2.5mC.Prior to deposition the electro-chemical solution was purged with nitrogen for 10 min, and a nitrogen blanket was maintained over the solution during the deposition. The modified electrodes were transferred to a fresh solvent/electrolyte and cycled over the range -1.0 to + 1.0 V relative to SCE at a scan rate of 50 mV s-’; 25 full cycles (i.e. 50 instrumental sweep segments) were completed for each PPy/electrolyte/solvent system. The cyclic voltammo-grams are shown in Fig.6 and the electrochemical data are summarised in Table 5. The midpoint potential Emwas calcu-lated for all films as the average of the anodic and cathodic peak potentials (E,,,, and Ep(+ respectively). The aims of this CV study were to examine the effects of the gradual deterioration of the PPy-Au(CN), films under electrical stress, and to observe the effects on the electrochemi-cal properties of changing the solvent electrolyte medium. In the latter case the acetonitrile-LiC10, and acetonitrile-TBAP systems were used to investigate the exchange of cations having very different ionic radii. (TBA),-Ni( CN), was chosen in order to study both the electrostatic influence of an anion having a 2-charge and the effect of using a planar anion compared to the more spherical ClO, ion.The PPy-Au(CN), films grown in acetonitrile + 1% water, or in water alone, exhibit strikingly different waveforms on the first scan, compared with those recorded on subsequent scans. It can be seen in Fig. 6(a) that, in general, the first reduction peak differs in voltage by 0.3 V compared to sub-sequent scans. In the ideal case, CVs of redox-active films should show symmetrical mirror-image cathodic and anodic wave^.^,,^^ In PPy films, however, charge transport is depen-dent on a number of factors including electron exchange reactions between neighbouring oxidised and reduced sites, and the diffusion of counterions, cations and solvent into and out of the film. For PPy-Au(CN),/MeCN-LiClO, where the PPy-Au(CN), film was grown in acetonitrile + 1YO water, designated as system (i), the initial cathodic peak at -0.55 V is quite sharp.The cathodic peak is broader on the second sweep and increases a little in height during the first five cycles. The position and shape of the first reduction peak is believed to arise from structural movement of the polymer chains, which causes distortion of the PPy backbone from an aromatic to a quinonoidal structure.66 Albery et have suggested recently that morphological changes occur during the first increasing to 1.54 at the 20th cycle. The difference between the two systems is the electrolyte cation, one being much bigger than the other. Thus it would appear that the dop-ing-undoping process in PPy is dominated by anion uptake and expulsion with little influence being exerted by the cation. In the CV curve for PPy-Au(CN),, recorded in MeCN-(TBA),Ni(CN), [system (iii)], there are two closely spaced reduction peaks in the first sweep, and two oxidation peaks, yet on the subsequent cycle one of the reduction peaks has disappeared.The CV sweeps shown in Fig. 6(c)are typical of the electrochemical response from cycles 15 to 25. The cathodic and anodic peaks at -0.38 and -0.09 V, which are characteristic of PPy, are more sharply defined than in either system (i) or (ii), indicating that insertion of [Ni(CN),]’-ions into PPy and elimination of the dopant [Au(CN),]-ions occur rapidly. Thus the PPy-Au(CN), film can better accommodate the planar [Ni(CN),I2-ion during cycling rather than the spherical-like [ClO,] -anions, on the basis that there is less disruption to the PPy chain structure.The origin of the second oxidation peak, occurring at 0.95 V, may be a consequence of specific interactions between the cationic polymer and the ion-pair form of the salt [in this case (TBA),Ni(CN),], causing a second PPy oxidation.68 In a further experiment, made after completion of 25 cycles, the PPy-Au(CN), film was rinsed in water, dried and returned to the MeCN-(TBA),Ni(CN), solution. The cycling was recommenced, and the first reduction wave showed two peaks, similar to that described above for the virgin PPy-Au(CN), film. On the next cycle, however, only one reduction peak was observed, [see Fig.6(d)].These results suggest that the initial reduction peaks are associated with cation uptake (TBA+),65which can be removed from the film during washing, since only one peak was observed without the washing process. Thus cation incorporation in the polymer matrix during the first reduction does not appear significantly to affect the redox behaviour of PPy on subsequent cycles. The CV curves for PPy-Au(CN), films grown from water were recorded in MeCN-LiClO, [system (iv)] and also in MeCN-TBAP [system (v)] . The electrochemical response of these two systems is similar, that of (iv) being shown in Fig. 6(e). The waveform shape and peak voltages differ from those of the PPy-Au(CN), film grown in acetonitrile.Microanalytical, XRD, and morphology data, described above, also indicate that the physical and chemical composi-tion of the two films are different. The polymer structure of PPy-Au(CN), grown in water is less conjugated than PPy-Au(CN), grown in acetonitrile, as deduced from the charge on the PPy chains, and the film quality deteriorates as the electrochemical cycling continues. A comparison of the behaviour of the films during cycles 23-25 reveals that the redox couple of PPy-Au(CN), grown in water is at more positive potentials than that of PPy-Au(CN), grown in acetonitrile consistent with shorter conjugation chain lengths. In addition, it is known69 that species with more extensive J. MATER. ('HEM., 1994, VOL. 4 I I I I I I I I I iA I I I I I I I I 1 I +1 .o 0 -1 .o +0.5 0 -0.5UV vs.SCE UV vs. SCE Fig. 6 Multiple sweep CVs for 0.5 pm thin films of PPy-Au(CN), prepared in organic or aqueous media. See Table 5 for the various conditions used. The starting voltage, 0 V, is shown as a cross. (a) System (i) cycles 1-5; (b)system (i) cycles 23-25; (c) system (iii) cycles 23-25; (d)system (iii) cycles 26-27; (e) system (iv) cycles 23-25; (f) system (vi) cycles 23-25 Table 5 CV data for PPy-AU(CN), films system type of polymer film" electrolyteb first cycle reduction peak EP(C)lV sweep cycle 2-5 EP(CP sweep cycle 2-5 Ep(0 sweep cycle 23-25 EPW /v sweep cycle 23-25 Ep(,)fV Em for sweep cycles 2?-25/v (i) (ii) (iii) PPy-Au( CN), (0) PPy-Au (CN), (0) PPy-Au(CN),(o) MeCN-LiC10, MeCN-TBAP MeCN-(TBA),Ni(CN), -0.55 -0.70 -0.47 -0.65 -0.25 -0.18 -0.38 -0.10 -0.06 -0.09 0.85 -0.25 -0.15 -0.38 -0.10 -0.03 -0.06 0.95 --0.18 --0.09 .-0.22 (iv) PPy-Au(CN),(aq) MeCN-LiC10, -0.18 -0.42 -0.13 -0.09 0.0 0.10 0.05 (v) PPY-AU (CN )2 (aq) MeCN-TBAP -0.15 -0.68 -0.10 -0.13 0.15 0.25 0.20 (vi) PP y-Au (CN), (0) H,O-LiClO, -0.35 -0.06 -0.02 0.07 -0.47 0.34 -0.30 0.21 (vii) PPy-Au(CN ),(aq) H,O-LiClO, -0.30 -0.19 -0.05 0.12 0.07 0.10 "(0)=grown in acetonitrile + 1% water, (aq)= grown in water.bMeCN =acetonitrile, TBAP =tetrabutylammonium perchlorate. conjugation exhibit a lower oxidation potential, e.g. Ep(a)is 2.14 V for cyclohexene, and 1.74 V for 1,4-~yclohexadiene. It is therefore possible that the PPy structure can contribute to differences observed in the redox potentials.The fact that the peaks are quite broad for systems (iv) and (v) may be evidence for numerous redox domains within the film, each domain having its own redox potential. Thus it would appear that the PPy-Au(CN), film grown in acetonitrile is more electro- chemically stable than the same film grown in water, and that it contains a more uniform distribution of anions. Other features such as difference in doping levels and surface mor- phologies may also play a part. Changing the solvent from acetonitrile to water has a significant effect on the behaviour of PPy films. Thus, the film PPy-Au(CN),/MeCN-LiClO,, grown in acetonitrile and cycled in acetonitrile, exhibits little change in waveform up to 25 cycles. However, as shown in Fig.6(f),cycling in water generates much broader peaks and smaller maximum currents. Multiple peaks also develop after a number of cycles. The electron transport process is therefore more efficient in aceto- nitrile than in water, a phenomenon noted by other~.~'.~~ 96 J. MATER. CHEM., 1994, VOL. 4 In system (vii), an aqueously grown film of PPy-Au(CN), was cycled in water/LiCIO,. The waveform (not shown) is similar to system (vi), multiple peaks develop immediately and are more sharply defined. It is possible that the oxidation process occurs Giu two steps. In the first the PPy surface is 13 14 15 16 J. Heinze, Top. Curr. Chem., 1990, 152, 1.F. Bedioui, C. Bongas, J. Devynk, C. Heid-Charreton and C. Hinnen, J. Electroanal. Chem., 1986,207, 87. A. Elzing, A. van der Putten, W. Visscher and E. Barendrecht, J. Electroanal. Chem., 1987,233, 113. T. Skotheim, M. Velaziquez-Rosenthal and C. A. Linkous, partially oxidised by attracting anions, and in the second the anions enter the film and oxidation of the film occurs simul- 17 J. Chem. SOC.,Chem. Commun., 1985,612. B. R. Saunders, R. J. Fleming and K. S. Murray, Synth. Met., 1992, taneously. In these ClO, -containing electrolytes the starting PPy film should eventually be converted to PPy-CIO,. To confirm this, a PPy-Au(CN), film grown in water was success- ively reduced and oxidised in water-KC10,. Examination of 18 19 20 47, 167. R. Cervini, R.J. Fleming and K. S. Murray, ‘Polymer 91’ IUPAC Int. Symp., R.A.C.I. Polymer Division, Melbourne, 1991, 607. C. S. Choi and H. Tachikawa, J. Am. Chem. Soc., 1990,112, 1757. S. Kuwabata, K. Okamoto and H. Yoneyama, J. Chem. Soc., the microprobe data revealed that while gold was still present in the film [as Au(CN);], potassium and chloride ions were also observed, presumably due to incorporation of the electro- lyte anion and cation. Multiple CV waves of the type shown in Fig. 6(f) have recently been observed by Albery et when studying PPy-polystyrene sulfonate cycled in 21 22 23 24 Faraday Trans. I, 1988,84,2317. D. J. Walton, P. M. Hadingham, C. E. Hall, I. V. F. Viney and C. Chyla, Synth. Met., 1987,41, 295. M. Zagorska, A. Pron, S. LeFront, Z. Kucharski, J.Suwalski and P. Berrier, Synth. Met., 1987, 18,43. S. Dong and G. Lian, J. Electroanal. Chem., 1990,291,23. M. Takakubo, Synrh. Met., 1987, 18, 53. water-K,SO,. The insertion of SO$-anions into the film occurred at low oxidation levels of the polymer, owing to the high charge density of the sulfate ion. It was postulated that the flood of such anions swamps the egress of cations (K’) thus generating a second oxidation peak. 25 26 27 28 0. Belanger, F. Girard, S. Ye and G. Laperrikre. J. Electround. Chem., 1992,334,35. B. Kaye and A. E. Underhill, Synth. Met., 1989,28, C97. S. Edge, A. E. Underhill, P. Kathirgamanathan, P. O’Connor and A. I. Dent, J. Muter. Chem., 1991, 1, 103. K. M. Cheung, D. Bloor and G. C. Stevens. J. Muter. Sci., 1990, 25, 3814.Conclusions 29 30 G. Tourillon and F. Garnier, J. Electroanal. Chem., 1984, 161,407. R. Cervini, R. J. Fleming and K. S. Murray, J. Muter. Chem., 1992, Incorporation of Au(CN), anions into polypyrrole films can, under appropriate conditions lead to properties such as enhanced environmental stability and electrical conductivity 31 32 33 2, 1115. J. Williams and K. Carneiro, Adr. Inorg. Chem., 1985, 29, 249. W. C. Fernelius and J. J. Borbage, Inorg. Synth., 1946,2, 227. W. R. Mason (1x1) and H. B. Gray, J. Am Chem. Soc., 1968, when compared with other metal anion systems. The linear structure of the Au(CN), ions plays a significant role in determining the growth pattern of PPy. XRD and electrical conductivity measurements show that the preferred orien- tation of PPy in PPy-Au(CN), is planar with respect to the anode surface.34 35 36 37 38 39 90, 5721. W. R. Mason (111) and H. B. Gray, Inorg. Chem., 1967, 7, 55. K. J. Wynne and G. B. Street, hfucromolecule~,1985,18, 2361. D. S. Maddison and J. Unsworth, Synth. Met. 1989,30,47. J. A. Harrison and J. Thompson, Electrochim Acta, 1973, 18, 329. R. Qian, J. Qui and J. Shen, Synrh. Met., 1987. 18, 13. M. J. Ribo, A. Dicko, J. M. Tura and D. Bloor, Polj-mer, 1991. CV studies on a polymer-coated platinum electrode show that PPy-Au(CN), films are electrochemically stable after the initial few cycles. The electrochemical response is influenced by the nature of the solvent, electrolyte and growth conditions of the film. Morphological studies show that PPy-Au(CN), grown in water possesses tubular surface growths, and that 40 41 42 43 32,728.J. T. Lei, R. Cai and C. R. Martin, Synth. Met., 1992,46, 53. H. Mao and P. G. Pickup, J. Am. Chem. Soc., 1990,112, 1776. L. S. Curtain, M. McEllistrem and W. J. Pietro, J. Phys. Chem.. 1989,93,1637. E. T. Kang, K. G. Neoh, Y. K. Ong, K. L. Tan and B. T. Tan, Macromolecules, 1991,24,2822. the anode face has a ‘blistered’ surface. Further work is 44 Practical Surface Analysis., ed. D. Briggs and M. P. Seah, Wiley, planned to test the films in electrocatalytic and electrochemical sensing applications. 45 46 Chichester, 1983,2nd. edn., vol. 1. Vision Reference Manual, Kratos Analytical, Manchester, 1992. L. Atanasoska, K. Naoi and W. H. Smyrl, Chem. Mater., 1992, This work was supported by a grant from the Australian Research Council (to R.J.F.and K.S.M.). 47 48 4, 988. K. K. Kim, Y. W. Mai and B. J. Kennedy, J. Muter. Sci., 1992. 27,6811. A. G. Sharpe, The Chemistry of Cyano Comple yes of the Transition References 49 Metals, Academic Press. London, 1976. G. R. Mitchell, Polym. Commun., 1986, 27, 346. 1 2 3 4 5 6 7 8 9 10 11 12 M. G. Kanatzidis, Chem. Eng. News, 1990, Dec. 3, 36. C. N. Billingham and P. 0.Calvert, Adv. Polym. Sci., 1989,90, 1. Handbook of Conducting Polymers, ed. T. A. Skotheim, Marcel Dekker, New York, 1986, vol. 1 and 2. R. S. Potember, R. C. Hoffman, H. S. Hu, J. E. Cocchiaro, C. A. Viands and T. 0.Pochler, Polym. J., 1987, 19, 147. J. T. Lei, W. B. Liang, C. J. Brumlik and C. R. Martin, Synth.Met., 1992, 47, 351. D. Curran, J. Grimshaw and S. D. Perera, Chem. Soc. Ren., 1991, 20, 391. D. S. Maddison and C. M. Jenden, Polym. Int., 1992,27, 231. B. K. Moss, R. P. Burford and M. Skyllas-Kazacos, Muter. Forum, 1989, 13, 35. F. P. Bradner, J. S. Shapiro, H. J. Bowley, D. L. Gerrard and W. F. Maddams, Polymer, 1989,30,914. Y. H. Park, S. H. Choi, S. K. Song and S. Miyata, J. Appl. Polym. Sci., 1992, 45, 843. M. Salmon, A. F. Diaz, A. J. Logan, M. Kroonbi and J. Bargon, Mol. Cryst. Liq.Cryst., 1982,33, 265. G. G. McLeoch, K. Jeffreys, J. M. R. MacAllister, J. Mondell, S. Affrossman and R. A. Pethrick, J. Phys. Chem. Solids, 1987, 48, 921. 50 51 52 53 54 55 56 57 58 59 60 61 62 63 G. R. Mitchell and A. Geri, J. Phys. D: Appl. I’hys., 1987,20, 1346. G. R. Mitchell, F.J. Davis, R. Gywenski and W. S. Howells. J. Phys. C: Solid State Phys., 1988,21,411. G. R. Mitchell, F. J. Davis, C. H. Legge, Synth. Met., 1988,26, 247. G. R. Mitchell, F. J. Davis and M. S. Kiani, Br. Polpm. J., 1990. 23, 157. M. S. Kiani and G. R. Mitchell, Synth. Met., 1992,48,203. L. J. Buckley, D. K. Roylance and G. E. WneL J. Polpm. Sci.,Part B, Polym. Phys., 1987, 25, 2179. B. J. Orchard. B. Freidenreich and S. K. Tripathy, Poljwer. 1986, 27, 1533. K. Kaneto, K. Yoshino and Y. Iniushi, Jpn. .I. Appl. Phys., 1982, 21, 567. M. Ito, H. Shloda and K. Tanaka, J. Poljvn. Sci., Part C, Polym. Lett., 1986, 24, 147. E. Garcia-Camarero, F. Arjona, C. Guillen, E. Fatas and C. Montemayor, J. Muter. Sci., 1990,25,4914 R. Qian and J. Qiu, Polym. J., 1987,19, 157. H. T. Chiu, J. S. Lin and C. M. Huang, J. .4ppl. Electrochem., 1992, 22, 358. A. F. Diaz and T. C. Clarke, J. Electroanal. Chrm., 1980, 111, 115. N. C. Billingham, P. D. Calvert, P. J. S. Foot and F. Mohammad, Polym. Degrad. Stab., 1987, 19, 323. J. MATER. CHEM., 1994, VOL. 4 97 64 J. Heinze and M. Dietrich, Muter. Sci. Forum, 1989,42,63. 69 T. Shono, A. Ikeda, J. Hayashi and S. Hakozaki, J. Amer. Chem. 65 G. L. Duffitt and P. G. Pickup, J. Chem. SOC., Faraday Trans., 1992,88,1417. 70 Soc., 1975,97,4261. J. M. KO,H. W. Rhee, S. M. Park and C. Y.Kim, J. Electrochem. 66 67 J. L. Bredas and G. B. Street, Acc. Chem. Res., 1985,18, 308. F. Li and W. J. Albery, J. Chem. Soc., Faraduy Trans., 1991, 87, 2949. 71 Soc., 1990,137,905. S. Kuwabata, H. Yoneyama, H. Tamura, Bull. Chem. SOC.Jpn., 1984,57,2247. 68 A, F. Diaz, J. I. Castillo, J. A. Logan and W. Y. Lee, J. ElectroanaL Chem., 1981,129, 115. Paper 31047925; Received 9th Augus;, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400087

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4(1), 99-104 Reaction of Molten Sebacic Acid with a Layered (Mg/AI) Double Hydroxide Simon Carlino and Michael J. Hudson* Department of Chemistry, University of Reading, P.0. Box 224, Whiteknights, Reading, Berkshire, UK RG6 2AD Molten decane-1 ,1 0-dicarboxylic acid, sebacic acid, reacted with the layered double hydroxide (LDH) of initial composi- tion MgsA13,~(0H),8.82(C03)1.51(N03)0.364.5H*0,at a temperature of 150 "C which is ca. 24 "Cabove the melting point of the acid. The reaction was not strongly exothermic and hence may be controlled by a suitable heating programme. X-Ray powder diffraction showed that the products were polyphasic and microcrystalline materials, the compositions of which were related to the original molar ratios of acid :LDH.Typically, the dominant phases were unintercalated LDH up to an acid :LDH ratio of 0.75 :1; intercalated dianion (ratio ca. 1:1); salt, e.g. magnesium sebacate (for ratios 4 :1 and higher). Fourier-transform infrared (FTIR) and 13C solid-state magic-angle spinning nuclear magnetic resonance (MAS NMR) spectroscopies of the reaction products confirmed the presence of the dianion -02C(CH2),C0,-. In no case was there evidence for undissociated guest acid molecules. The mechanism appears to involve the sorption of the acid, initially at the crystallite edges, followed by reaction of the molten acid with the basic solid host with evolution of CO,. The separated LDH layers then allow the acid to diffuse in. The product from the 1 :1 mixture was similar to the intercalation compound obtained using a modified version of Drezdzon's method, which had previously been thought to be ineffective for sebacic acid.Interest in layered double hydroxides, LDHs, and their inter- calation compounds stems from their potential as catalysts and as highly selective sorbents.'Y2 An early study on the synthesis of LDHs suggested that coprecipitation reactions could easily be applied to the preparation of hydrotalcites intercalated by a,w-dicarboxylic acids using aqueous solutions of their disodium sah3 The intercalation compounds formed were not well defined microcrystalline materials. There is improved crystallinity when reformed LDHs,~ which have been calcined to the oxide and then rehydrated with carbon- ate-free water, are used.However, this method is only useful for guests which are soluble in the chosen solvent, which is normally carbonate-free water. One limitation, therefore, of the above wet methods is that it is often necessary to exclude carbon dioxide from the reaction apparatus. In order to obtain products with improved crystallinity and to develop methods by which the rate of intercalation may be controlled, the reaction between the molten dicarboxylic acid and an Mg/A1C03 hydrotalcite-like material was investigated. In this particular study, the intimately mixed reactants were heated slowly in air to a temperature which was 20-30°C above the melting point of the acid and the mixture was allowed to come to equilibrium over several hours.It was considered that such a method would allow the reaction to proceed at the interfaces between the reactants and would provide sufficient time for the guest moieties to diffuse through the interlamellar regions of the LDH host. In addition, the LDH interlayer water and that formed by dehydroxylation of interlayer hydroxy groups, may diffuse out through the chan- nels and pores of the interlamellar regions without being forced through the crystal layers, thus preserving the layered structure of the host. The host chosen for this study was the hydrotalcite, Mg6A1,(OH),,(C0,).4H20, which is one of the anionic min- erals known as layered double hydroxides (LDHs)' or anionic clays. The structure of LDHs is similar to that of brucite, Mg(OH),, which consists of layers of Mg" ions coordinated octahedrally by six oxygen anions (as OH-)., These octahedral units form infinite sheets via edge sharing, and then stack by means of hydrogen bonding between the hydroxy groups of adjaceni sheet^.^ The height of one Mg-(OH), layer in brucite is 4.77 A.8 In hydrotalcite, some Mg" ions in the octahedral sheets have been replaced by All1'.The Mg-A1-(OH), layers then carry a net positive chargeg which is balanced by anions such as Cog- and NO, (ref. 10)in the interlayer region. The interlayer region also contains water molecules which are able to move randomly and to rotate about their CZvaxes." Layered double hydroxides are known to be able to intercalate polyoxometallates'2~'3 but the thermochemistry of these remains to be studied.The changes which occur in hycirotal- cites themselves upon heating are well d~cumented.'~"' Below 200"C, which applies to this study, only the interstitial water is lost and the water molecules diffuse along the subsurface pores in the sheets. MAS NMR (27Al) has indicated that all the aluminium atoms are octahedrally coordinated in the hydrotalcite. The structural model of the LDH involyes the brucite-like layer containing magnesium and hydroxy groups and a loosely stacked layer containing aluminium, hj droxy groups and the carbonate anion. In the oxide, which is obtained by heating the LDH above 450°C, some of the aluminium atoms (ca. 25%) have tetrahedral geometry, as indicated by 27Al MAS NMR studies." These are thought to exist principally at the surface of the LDH octahedral layers and probably account for the increase in catalytic/sorbent activity shown by the oxide materials relative to their unheated parent LDHs.In order to distinguish the synthetic method used in this study from the various wet methods4.', the phrases 'thermal reaction' or 'thermal intercalation' have been used. There have been few previous studies of the thermal intercal- ation of acids or amines into clays. KatoI7 studied the selective solid-state intercalation of cis and trans isomers into montmor- illonites using maleic and fumaric acids. Both maleic and methylmaleic acids were intercalated into montmorillonite whereas their geometrical isomers (trans forms) fumaric and methylfumaric acids were not intercalated.This selectivity was thought to derive from the ability of the cis isomers to form chelate complexes with the A1'" of the clay. The same group has studied the solid-state intercalation of organoam- monium ions" which was considered to proceed by an ion- exchange mechanism. In addition, reactions involving the sorption of guest molecules such as a~rylamide,'~n-alkylamine2' and 2,2'-bipyridine2' onto montmorillonite have been studied. The corresponding reactions with the layered double hydroxides have not, to the best of our knowledge, been studied previously. In a prior study we have intercalated phthalocyanines22 into layered phosphates by means of in situ synthesis using molten 1,2-dicyanobenzene.Experimental Materials Both of the metal nitrates [Mg(NO3),-6H2O and Al( N03)3*9H20] together with the sodium hydroxide [as 50% (w/v) (ca. 12.5 mol drn-,) solution] were standard GPR grade reagents (BDH). The anhydrous sodium carbonate was an AnalaR grade reagent (BDH). The sebacic acid used (Aldrich) was more than 99.5% pure. The solvent used was doubly deionised water which had a conductivity of less than 0.05 pS cm-' at room temperature and a pH of ca. 8. Preparation Mg-Al-CO, Hydrotalcite-like Materia2 The hydrotalcite-like material used in the thermal reaction was prepared according to the method of Reichle3 with extra cycles of washing and centrifugation of the slurry.An aqueous solution of Mg(N03),*6H20 (85.47 g, 0.33 mol) and A1(N03),.9H20 (62.52 g, 0.17 mol) in doubly deionised water (233 cm3) was added dropwise to an aqueous solution of NaOH (380 cm3; 23.35 g, 0.58 mol) and Na2C0, (anhydrous) (33 g, 0.32 mol). This addition was carried out with fast stirring (ca. 300 rpm). Found: Mg, 16.28; Al, 8.95; C, 2.49; H, 3.66; N, 0.69% [which were the values calculated for Mk413.4 (OH)18.82(03 )1.51(NO3 )o.36 4.5J-4201. The total percentage for water (surface-sorbed and interlayer) was esti- mated using thermogravimetric analysis (TG) as 14%. TG also gave the overall mass loss for COZ- as 24%, in agreement with the above formula. The ratio X for the aluminium substitution [X =Al/( Mg +Al)] into the brucite [Mg(OH), layer] was found to be 0.35 for the LDH and the product from the 1 : 1 reaction mixture.23 These ratios imply that the LDH and the product from the 1:1 reaction mixture had the highest A1 content from the experimentally observed limits of solid solubility.Sebacic Acid Disodium Salt The sebacic acid disodium salt for the wet coprecipitation reaction used as a comparison with the thermal reaction method for XRD, FTIR and MAS NMR was synthesized by dissolving sebacic acid (32.36 g, 0.16 mol) in NaOH (57.5 g, 1.44 mol) in doubly deionised water (320 cm'). This solution was then used in the coprecipitation reaction using a modified version of Drezdzon's method16 which was found by that author to be ineffectual for sebacic acid.The successful method developed in this study involved the slow addition of a solution of magnesium nitrate and aluminium nitrate to the basic sebacic acid solution with fast stirring overnight. One additional modification to the method of Drezdzon was that the reaction was carried out at room temperature. Thermal Reactions The thermal reaction of molten sebacic acid with the LDH was carried out in air by slowly heating the intimately mixed LDH and sebacic acid in a silica crucible at a ramp rate of 1 "C min-' up to 150°C. The reaction mixture was held at this temperature for 8 h in order for equilibrium to be attained and maintained. The mixture was finally cooled to room temperature at a rate not exceeding 10 "C min-l. The reaction products were washed in hot ethanol (100cm3) to remove J.MATER. CHEM., 1994, VOL. 4 excess sebacic acid plus other soluble materials and filtered using a grade 4 glass sinter. The products were finally air- dried at 40°C for 18 h. Found for the 1 : 1 mixture: Mg, 13.60; Al, 7.37; C, 14.68; H, 4.85; N, 0.14%. Owing to the polyphasic nature of the reaction products (see XRD section) it was not possible to calculate a single molecular or empirical formula for each of these materials. Physical Measurements FTIR spectra were measured on a Perkin-Elmer 1720-X FTIR equipped with IRDM data management system. All samples were run as 13 mm diameter dry KBr discs with the ratio of sample to KBr as 1:lOO by mass.24 The X-ray powder diffraction was carried out on a Spectrolab Series 3000 CPS-1200 diffractometer using Cu-Kr radiation (i= 1.54059 A).Samples were run on a plate orientated at an angle of 3-5". Typical accumulation times for each powder pattern were in the range 650-700 s. Analysis Thermal analysis of the samples was carried out using a Stanton Redcroft STA 1000 TG-DSC Simultaneous Thermal Analyser referenced against recalcined alumina. The thermal analyses (TG/DTA) of the LDH-sebacic acid intercalation compounds were run at a ramp rate of not more than 5°C min-'. The samples were run in nitrogen (with a flow rate of 56 cm3 min-l) in order to limit the oxidation of the organic groups and then in air. Carbon, hydrogen and nitrogen analyses were carried out in duplicate by Medac Ltd., Brunel University.Inductively coupled plasma emission spectrometry (ICP-OES) using an ARL 3500 machine along with scanning electron microscopy using a JEOL 840 were carried out by staff at the Postgraduate Research Institute for Sedimentology, University of Reading. The 27Al and 13C MAS NMR were run by the SERC Solid- state NMR Service at the University of Durham. For the 13C spectra, a frequency of 75.431 MHz was used with cross-polarization and a spin rate of 4540 Hz. The 27Al MAS NMR were run at a frequency of 78.152 MHz again with cross-polarization and a spin rate of ca. 12 kHz. Evolved gas analysis (EGA) for the evolution of carbon dioxide from the LDH and 1 : 1 mixture of the reactants was carried out by members of the Catalysis Research Group, University of Reading, using a Perkin-Elmer 8410 gas chroma- tograph with a 2 m Porapak column.Results and Discussion Thermal Properties For the host, Mg6A13 .4 (OH)18.82 (co3)I .5 I ( NO310.36 * 4.5H20, there are ca. 19 hydroxy groups plus the carbonate and a small amount of nitrate, with which the sebacic acid dianion may exchange. The TG data for the LDH alone have been previously The loosely bound (surface) water is lost up to 100°C and bound (interlamellar) water at ca. 150°C. For sebacic acid itself, there are two thermal events at 136 (melting) and 290 "C (boiling) but there is little loss of mass before 280°C. Fig. 1 shows the thermal degradation of mixtures of the acid and the hydrotalcite in comparison with the TG/DTA curves for the LDH (host) itself.It can be seen that there is a gradual mass loss as the LDH (a) is heated. The reaction between the acid and the hydrotalcite is not strongly exothermic but the endotherm, Fig. l(h) or DTA curves, corresponding to the melting of the acid can be clearly seen. For this reason, the controlled heating of the reagents does not produce locally high temperatures so that this type of reaction has the potential of being controllable. J. MATER. CHEM., 1994, VOL. 4 displace some of the interlayer carbonate groups in the LDH at a temperature of 115-150°C. This evolved CO, can be identified with the interlayer CO, of the LDH interlayer because sebacic acid does not decompose at such a low temperature (i.e.just above its melting point of 135-137 T). However, there is extensive decomposition if the sebacic acid is heated above 275 "C. X-Ray Powder Diffraction Results Fig. 3 shows the X-ray powder diffraction pattern obtained for the reaction product formed by the thermal reaction of LDH with molten sebacic acid at the initial molar ratios (dicarboxylic acid: LDH) of 1:1. One major reflection in all of the ppwder patterns obtained for the ratios of up to 1 :1 is 10 0 70- -1 0 60. -20 - 0 100 200 300 400 500 IF 1004 Bl 3. the 7.7 A (11.5" 28) do,, rtflection of the host LDH. Another 1-11.0 major reflection is at 19A (4.5"28) for the LDH host with the intercalated dianion.This implies that the intercalation products produced by the thermal reaction of the dicarboxylic acid with the LDH with low (i.e. 1:l or lower) initial acid :LDH ratios are biphasic and contain an unintercalated LDH phase. Th,e products show a gradual increase in the reflection at 19 A with the increase if! the initial acid: LDH ratio up to 1 :1. This reflection at 19 A is assigned as the dOo3 W 80 -3.0 0 50 100 150 200 250 300 7°C Fig. 1 The heating profiles for (a) the host LDH (b) 1:1mixture with sebacic acid Fig. 2 shows the EGA results for loss of carbon dioxide from the LDH host and the 1 :1 mixture. It can be seen that, for the LDH itself, there is little loss of carbon dioxide below 225°C but for the mixture there is a significant evolution of CO, in the region of 125-150°C.Clearly, this loss of carbon dioxide is associated with the reaction of the LDH with the sebacic acid guest. There is a further evolution at 210°C for the mixture and then the remaining CO, is lost as the boiling point of the sebacic acid is reached. In the LDH host the evolution of C02 decreases until the temperature reaches ca. 300°C when all the remaining carbonate in the LDH is lost. It appears, therefore, that the molten sebacic acid is able to 1oc 7.7 (003') 3.8 (006*) . 2.6 11' . 0 10 20 30 40 50 60 70 28ldegrees Fig. 3 X-Ray powder diffraction pattern for the thermal reaction product with initial sebacic acid: LDH ratio: 1:1 heated to 150°C (indexed using Jones 4) ~~ ---_ ___TI"C Fig.2 Evolved gas analysis profiles showing the evolution of C02 in the LDH host and the 1 : 1 mixture of LDH and sebacic acid reflection for the LDH intercalated with the sebacate diani~n.~,'~From previous studies of intercalated smectite clays2' or layered phosphates," it is known that organic diamines having terminal functional groups are strongly bound to the layered host and impede the diffusion of more guest molecules. Consequently, a similar case may occur here in that the guest sebacate dianions bind to A1 or Mg at the edges of the layers and the bound dianion may impede the further diffusion of guest species into the host. However, the reaction at the layer edges serves to separate the layers.When the initial ratios of acid to LDH were significantly in Excess of 1: 1 (over 4:1) the dominant reflection was at 14 A which is typical for magnesium sebacate. There were no additional overtone reflections which might indicate a layered intercalation compound. It appears, therefore, that the reac- tion of the LDH itself and the high concentration of molten sebacate leads to sebacate salts. Thus the reaction of mag- nesium hydroxide with sebacic acid may be a useful prepara- tive route for magnesium sebacate. The X-ray powder diffraction pattern obtained for the coprecipitation reaction product is shown in Fig. 4.This can be indexed from values quoted in the paper by Chibwe and Jones.4 The overall shape of the powder pattern resembles that in Fig.3 obtained by the thermal intercalation route f9r the 1: 1 thermal reaction product. The dOo3reflection of 17.6 A, Fig. 4,represents a slight difference from that obtained for the thermal reaction products but can be explained as a change in the angle (from ca. 90 to ca. 60") of orientation of the dianion and the LDH layers. FTIR Spectra J. MATER. CHEM., 1994, VOL. 4 Solid-state MAS NMR 27 A1 MAS NMR Spectra The 27 A1 MAS NMR spectra of the 1: 1 thermal reaction product together with that of the coprecipitation reaction are shown in Fig. 5(a) and (b), respectively. The resonance of octahedrally coordinated aluminium occurs in the range -10 ppm to +20 In the parent LDH all of the aluminium is present in an octahedral en~ironment.~ The peak in the 1:1 thermal reaction product shows a single resonance at +8.3 ppm which is also assigned to aluminium in an exclusively octahedral environment.The same is also true for the 0.25: 1, 0.50: 1 and 0.75: 1 thermal reaction products. The coprecipitated material, Fig. 5(b), shows the single resonance of octahedral aluminium at +7.8 ppm, although the peak is broader than that of the thermal reac- tion product. 13CMAS NMR Spectra Typical I3C MAS NMR are shown in Fig. 6 and some structural correlations of the peaks are given in Table 1. An important feature of 13C NMR is that carbon atoms in two different molecules, which have the same environment, have the same chemical shift in solution NMR within kO.2 For solid-state NMR, correlation of chemical shifts is also possible between compounds with similar structures.By using the 'additive shift' technique34 it was possible to assign the observed resonances of the C2 and C3 methylene carbons of the dicarboxylic acid chain to either the undissociated acid or to the dianion, although the assignment of other methylene groups (ix.C4 or C5) was tentative. There are three principal resonances observed in the NMR I3C spectra of the reaction products. These occur in the regions of +182ppm for carboxy groups; f170ppm for carbonate and +29 to +39 ppm for the methylene carbons The FTIR spectra of the reactants and the produ~ts~~,~~ in the sebacic acid skeleton. The peak at 170.6 ppm wasshowed that the principal change on reaction of sebacic acid assigned to carbonate because the 13C peak for sodiumwith the LDH is the disappearance of the 1725 cm-' band in sebacic acid, which is characteristic of [v,,,,(CO)] of undis- sociated carboxy groups, and the appearance of several bands I at 1560 cm-' in all of the products which are characteristic of the ionised [vasym(CO)]carboxy group.The multiplicity of v,,,,,(CO) bands is indicative of the existence of a diversity of locations for the COO- groups in the LDH interlayer as proposed in an earlier study on the intercalation of fatty acids into LDHs by Borja and D~tta.~' These probably include COT (Cog-) and CO,M+(OH) (1542 cm-'). There was no evidence for undissociated carboxy groups.1001 n I+ 7.8 U' . -,,,,+ 0 10 20 30 40 50 60 70 200 100 0 -100 -20028ldegrees 6 Fig.4 X-Ray powder diffraction pattern of the LDH-sebacic acid Fig. 5 "A1 MAS nmr of the reaction products from (a)the 1 : 1 and coprecipitation product4 (b)the wet coprecipitation method J. MATER. CHEM., 1994, VOL. 4 I-250 200 150 100 50 0 .__;..II h 250 200 150 100 50 0 6 Fig. 7 Scanning electron micrographs: (a) LDH, (b) 1: 1 thermal reaction product Fig. 6 I3C MAS NMR of (a) sebacic acid, (b) sebacic acid disodium salt and (c) 1 : 1 LDH-sebacic acid thermal reaction product shows that there is an increase in the average size of the carbonate is at the same shift whereas that for sodium crystallites compared with the LDH itself; i.e.for 1: 1 the hydrogen carbonate is at + 164.5 ppm. As the ratio of sebacic crystallite size is ca. 15 pm. This increase in size may have acid in the original mixtures was increased, so there was a been caused by aggregation brought about by cross-linking decrease in the carbonate peak at + 170.6 ppm and the by the sebacic acid dianions. Thus, the sebacic acid dianion emergence of a peak at + 173.4 ppm. It is suggested that this may act as a binder between adjacent crystallites. shift may be due to partial relocation of the carbonate groups.31 Conclusions Scanning Electron Microscopy It has been demonstrated that sebacic acid has been interca- Fig. 7 shows the scanning electron micrographs obtained for lated into the host LDH by means of a thermal reaction.(a) the host LDH and (b)the 1: 1 thermal reaction product. With acid: LDH ratios of 4:1 or more the salts such as The average size of the crystallites in the host LDH is ca. magnesium sebacate are the dominant products. EGA analysis 3 pm, which value is larger than the literature value (ca. for C02 of the reaction profile indicated that the reaction The hexagonal plates of individ- proceeded with the displacement of some of the interlayer 0.2 pm) quoted by Rei~h1e.I~ ual crystallites are just visible. The SEM of the 1: 1 product carbonate anions of the host LDH. Table 1 Assigned I3C MAS NMR data (ppm) for LDH-sebacic acid thermal reaction products sebacic acid 0.25 : 1 0.50: 1 0.75 : 1 1:l sebacic acid disodium salt product product product product assignmcnt -182.50 182.00 182.47 182.31 182.24 coy -----CO,H -170.69 170.60 170.60 170.56 c0;--38.90 38.41 38.38 38.56 38.38 c2of co, -----C2of COzH 29.40 28.7 28.57 27.82 27.86 c3of coy 104 J.MATER. CI-IEM., 1994, VOL. 4 We would like to thank Dr Malcolm Buck at Laporte Research and Development for his help and encouragement with this paper. Mr Roger Taylor of Solvay is also thanked for his help in the early stages of this study. The authors are 13 14 15 T. Kwon, G. A. Tsigdinos and T. J. Pinnavaia, J. Am. Chern. Soc., 1988,110,3653. S. Miyata, Clays Clay Mineral., 1975,23,369. W. T. Reichle, S. Y. Kang and D. S. Everhardt, J. Catal., 1986, 101, 352. also grateful to Laporte Research and Development for a grant for SC.Mr P. Loader and Mr M. Hayes for running the EGA. Dr D.C. Apperley of the SERC NMR service at Durham University is thanked for his valuable help with the NMR. 16 17 18 19 M. A. Drezdzon, Inorg. Chem., 1988,27,4628. C. Kato, M. Ogawa, M. Hirata and K. Kuroda, Chem. Lett., 1992,365. C. Kato, T. Handa, M. Ogawa and K. Kuroda, Chem. Lett., 1990, 71. C. Kato, M. Ogawa and K. Kuroda, Chern. Lett., 1989, 1659. 20 C. Kato, M. Ogawa, T. Hashizume and K. Kuroda, Inorg. Chem., 1991, 30, 584. 21 C. Kato, K. Kuroda, M. Ogawa and K. Kato, Clay Sci., 1990, References 22 8, 31. M. J. Hudson, X. Hu and P. C. H. Mitchell, Solid State Ionics, 1 2 3 4 5 6 7 8 T. Nakatsuka, H. Kawasaki, S. Yamashita and S. Kohjiya, Bull. Chem. SOC. Jpn., 1979,52,2449.F. Cavani, F. Trifiri, and A. Vaccari, Catal. Today, 1991,11, 173. W. T. Reichle, J. Catal., 1985,94, 547. K. Chibwe and W. Jones, J. Chem. SOC., Chem. Commun., 1989, 926. M. Chibwe and W. Jones, in Pillared Layered Structures-Current Trends and Applications, ed. I.V. Mitchell, Elsevier Applied Science, London, 1990. A. F. Wells, Structural Znorganic Chemistry, Oxford University Press, Oxford, 5th edn., 1984,p. 259. H. W. F. Taylor, Mineral. Mag., 1973,39, 377. J. R. Smyth and D. L. Bish, Crystal Structures and Cation Sites of 23 24 25 26 27 28 29 30 1993,61, 131. C. Misra and A. J. Perrotta, Clays Clay Minerd., 1992,40, 145. D. H. Williams and I. Fleming, Spectroscopic, Methods in Organic Chemistry, 3rd edn., McGraw-Hill, London, 1980, p. 37. G. W. Brindley and S.Kikkawa, Clays Clay Mineral., 1980,28,87. P. G. Rouxhet and H. W. F. Taylor, Chirnia, 1969,23,480. R. Tettenhorst, C. W. Beck, G. Brunton, c'luys Clay Mineral., 1961,9, 500. M. S. Agashe and C. I. Jose, J. Chem. Soc., Faradaji Trans. 2, 1979, 75,733. M. J. Hernandez-Moreno, M. A. Ulibarri. J. L. Rendon and C. J. Serna, Phys. Chem. Mineral., 1985, 12, 34. C. J. Serna, J. L. Rendon and J. A. Iglesias, Clays Clay Mineral., 1982,30, 180. 9 10 the Rock-forming Minerals, p. 69, Allen and Unwin, London, 1988. W. T. Reichle, Solid State Ionics, 1986,22, 135. L. Pesic, S. Salipurovic, V. Markovic, D. Vucelic, W. Kagunya 31 32 M. Borja and P. K. Dutta, J. Chem. Phys., 1992,96,5434. G. Engelhardt and D. Michel, High-Resolution Solid-state NMR of Silicates and Zeolites, Wiley, Chichester, 1987,p. 142. and W. Jones, J. Muter. Chem., 1992,2, 1069. 33 L. P. Lindeman and J. Q. Adams, Anal. Chem., 1971,43, 1245. 11 G. Marcelin, N. J. Stockhausen, J. F. M. Post and A. Schutz, 34 D. M. Grant and E. G. Paul, J. Am. Chem. Soc., 1964,86,2984. J. Phys. Chem., 1989,93,4646. 12 T. Kwon and T. J. Pinnavaia, Chem. Muter., 1989,1, 381. Paper 3104695H; Received 4th August, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400099

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. MATER. CHEM., 1994, 4( 1), 105-1 11 Organo-soluble Segmented Rigid-rod Polyimide Films Part 5.t-Effect of Orientation Fred E. Arnold Jr., Dexing Shen, Frank W. Harris and Stephen 2. D. Cheng* Institute and Department of Polymer Science, The University of Akron, Akron, Ohio, 44325-3909, USA A series of semi-rigid aromatic polyimides have been synthesized via a one-step polymerization in which the poly(amic acid) precursors were not isolated. The polyimides were synthesized from 3,3’,4,4’-biphenyltetracarboxylic dianhydride (BPDA), pyromellitic dianhydride (PMDA), and 2,2’-bis(trifluoromethyl)-4,4’-diaminobiphenyl(PFMB) in refluxing rn-cresol. Uniaxially oriented films were utilized in order to investigate the chain orientation in the ordered state. Wide-angle X-ray diffraction (WAXD) patterns showed an increase in the crystal orientation and the crystallinity for oriented CPI(1OO/O) films with increasing draw ratio.Dichroic ratios obtained under polarized Fourier transform infrared spec- troscopy (FTIR) represented overall orientation factors attributed to both ordered and non-ordered regions, and they were a function of the incident angle and draw ratio. This indicates that the conformational arrangement and chain packing changed during deformation. The orientation along the molecular chain axis and the lateral orientation of the imide planes became increasingly enhanced with increasing draw ratio, an indication of planar alignment of the imide planes as well as uniaxial orientation of the molecular axis.The surface morphology of unoriented and uniaxially oriented films was also studied via transmission electron microscopy (TEM). The lateral crystal dimension of the ordered regions observed from TEM corresponded well to the apparent crystallite sizes of the (310) crystalline plane in BPDA-PFMB crystals obtained from WAXD experiments. For many years it has been recognized that ‘in-plane orien- tation’ of wholly aromatic polyimides exists in thin films, namely the crystallographic c-axis preferentially aligns itself parallel to the surface of the film. The majority of work carried out thus far concerns poly( 4,4‘-oxydiphenylene- pyromellitimide) synthesized from pyromellitic dianhydride (PMDA) and 4,4’-oxydianiline (ODA),’,2 which is commer- cially available under the trade name of Kapton produced by Du Pont.and widely used in electronic packaging applications. In recent years this kind of structural anisotropy has been observed in other aromatic polyimide films such as a polyimide synthesized from 3,3’,4,4’-biphenyltetracarboxylicdianhydride (BPDA) and p-phenylene diamine (PDA).3 In the first two publications of this series we reported a detailed study of the structure formation and macroscopic properties of the polyimide film synthesized from 3,3’,4,4’-biphenyltetracarboxylicdianhydride (BPDA) and 2’2’-bis(trifluoromethyl)-4,4’-diaminobiphenyl (PFMB) in refluxing rn-~resol.~.~ BPDA-PFMB thin films with a thickness of 10-30 pm were found to exhibit an in-plane orientation of the crystallographic c-axis, which preferentially orients itself parallel to the surface of the film.This was determined through ~ t Part 4:F. E. Arnold et al., J. Mater. Chem., 1993, 3, 353. B PD A-P FM B both transmission and reflection modes of WAXD experi- ment~.~The in-plane orientation of the crystals can also be characterized by finding a correspondence between a highly oriented fibre pattern scanned along both the equatorial and the meridianal directions and the pattern obtained from unoriented films using these two geometrical modes. It has been speculated that this structural anisotropy is the result of the rigidity and linearity present in aromatic polyimides. Part 3 in this series was concerned with the effect of forming copolymers by the incorporation of pyromellitic dianhydride (PMDA) into the BPDA-PFMB system.The focus was on the effect of varying the molar percentage of PMDA-PFMB on the thermal and dynamic mechanical properties of BPDA-PFMB-based copolyimides. Chain rigidity and lin- earity were two crucial parameters which determined the thermal mechanical and dynamic mechanical properties of the copolymers studied.6 Part 4 in this series discussed the relationship between the structural anisotropy and the resulting anisotropic properties in the unoriented copolyim- ide films.7 In this paper we focus on the anisotropic structure of the BPDA-PFMB-based copolyimides in the oriented state as opposed to unoriented films. The chemical structures can be expressed by PMDA-PFMB J.MATER. CHEM., 1994, VOL. 4 and are designated as (BPDA-PFMB),-( PMDA-PFMB),. It can be further simplified to CPI(X/Y), where X and Y represent the molar percentages of the comonomers ranging from 0 to 100. CPI(lOO/O) represents the homopolyimide BPDA-PFMB, while CPI( 50/50) represents a statistically alternating copolyimide of BPDA-PFMB and PMDA-PFMB. This family of copolyimides is soluble in hot phenolic solvents and was prepared by a one-step polymeriz- ation in which the poly(amic acid) precursors were not At high temperatures above 140 "C, the copolyimides show a homogeneous solution state up to a concentration of 12% (m/m). Upon cooling, the copolyimide solutions undergo a sol-gel transition to form a gel-like state as well as an ordered structure.The transition temperatures and kinetics depend upon the concentration, molecular weight and the chain rigidity of the polyimides.'O.ll Aromatic polyimide films are of great interest because they possess excellent electrical and mechanical properties together with high thermal, thermo-oxidative and dimensional stabilit- ies. High-modulus aromatic polyimide fibres have also been spun from the isotropic state by a dry-jet wet-spinning pro- c~ss.'~,'~In order to understand the localized order existing in unoriented films it is advantageous to study oriented samples (films and fibres). WAXD and polarized infrared spectroscopy (FTIR) were employed to determine the effect of orientation on unoriented and oriented samples and to investigate the change in the three-dimensional crystalline order as the molar percentage of PMDA-PFMB was increased.The resulting morphology for both oriented and unoriented films was also studied by TEM. Experimental Materials The copolyimides were synthesized from 3,3',4,4'-biphenyltetracarboxylic dianhydride (BPDA), pyromellitic dianhydride (PMDA), and 2,2'-bis( trifluoromethy1)-4,4'- diaminobiphenyl (PFMB). The diamine monomer synthesis was first reported by Rogers et for polyamide synthesis. A macro-monomer was first synthesized from 1mol of BPDA and 2 mol of PFMB to form soluble blocks within the m-cresol. For CPI( 50/50)another mole of PMDA was added to the polymerizing solution, and statistically one obtains an alternating copolyimide of BPDA-PFMB and PMDA-PFMB.Owing to the difference in the chemical reactivities of the two dianhydrides they must be added separately to avoid phase separation, which can be observed optically and through dynamic mechanical experiment^.'^ Detailed polymerization procedures are discussed in other publication^.^.^ The intrinsic viscosities of CPI(100/0), CPI(85/15), CPI(70/30) and CPI(50/50) are 4.5, 3.5, 3.5 and 3.0 dl g-', respectively.6 Film Preparation CPI(X/Y) films were prepared by spreading a 2% (m/m) rn-cresol solution on a glass plate and drying at 150°C for 5 h and 250°C for an additional 2 h under reduced pressure in a vacuum oven. In this manner, films were prepared with thicknesses of 10-30 pm.Precise control of the solution concentration and the thickness of the casting dope on the glass plate was necessary to control the final thickness. Oriented films of CPI(lOO/O) were prepared to study the orientation of the chain molecules with respect to the draw direction. The films were drawn under tension at ca. 400-450°C up to 500% extension. Oriented CPI(X/Y) films were prepared in a similar manner. Wide-angle X-ray diffraction WAXD experiments were conducted using a Rigaku X-ray generator with a 12 kW rotating anode. The point focused beam was monochromatized by a graphite crystal sensitive to Cu-Ka radiation. In order to study the orientation in the films, a Siemens 2-dimensional area detector was used for the oriented CPI(X/Y) films.Crystal orientations in oriented CPT(X/Y) films were meas- ured based on the Hermans equation: 2f, =3(C0S2 bc)-1 (1) where fc is the orientation factor along the draw direction, and #c represents the angle between the centre of diffraction of this plane and the c-axis of the crystal unit cell. The numerical values of the mean-square cosines in this equation can be determined from the fully corrected intensity distri- bution reflected from the (003) crystalline plane, Ic(bc,a), averaged over the entire surface of the orientation sphere: Polarized Fourier Transform Infrared Spectroscopy Polarized FTIR were carried out on a Mattson galaxy model 5200 FTIR. Polarization experiments were performed using a Harrick Ge polarizer coupled to a variable Brewster angle holder.Uniaxially oriented films [CPT (Xi Y I] were mounted onto the Brewster holder. Film thicknesses were in the range 5-10 pm. Two principal vibrational bands were studied, the 1778 cm-' and 738 cm-'. The transition moment vectors of the carbonyl symmetric and asymmetric stretching (1778 cm-') mode and in-plane and out-of-plane bending (738 cm-l) mode are shown in Fig. 1.16 The intensity of the absorption is proportional to the dot product of the electric- field vector (E)and the transition moment vector (M)accord-ing to: IKE.M=E~cos 0 (3) The intensity is thus maximized when the two vectors are perfectly aligned (6 =0). The stretching (1778 cm- ') mode possesses a transition moment vector lying along the imide plane and thus parallel to the chain direction (in-plane).The bending mode of the carbonyl unit (738 cm-') possesses a transition moment vector that is perpendicular to the imide plane (out-of-plane). l7 For the FTIR experiments, both vertically and horizontally (b ) 0 +out of plane 0 -into plane Fig. 1 Two vibrational modes for in-plane and out-of-plane vibrations of imide rings. (a) 1778 cm-' absorption band; (b)738 cm-' absorp-tion band. J. MATER. CHEM., 1994, VOL. 4 polarized infrared radiation was used. The orientation in uniaxially oriented CPI(X/Y) films can be measured from the dichroic ratio of these two bands discussed previously in order to obtain the relative orientation with respect to the draw direction: (4) The dichroic ratio (R)is defined as the ratio of absorbance of radiation polarized parallel to the draw direction (All)to the absorbance of the that polarized perpendicular to the draw direction ( L41).For a film with random orientation R =1. It is often more convenient to determine the ratio of the optical densities (II,and IJ rather than the absolute absorbance. Dichroic ratios will be calculated as a function of both the draw ratio as well as the angle of incidence relative to the plane of the film. The experimental set-up is shown in Fig. 2. Eqn. (4) was used to calculate the dichroic ratios of the 738 cm-' band, while the dichroic ratios of the 1778 cm-' band were calculated in an inverse manner. Transmission Electron Microscopy TEM was used to study the surface morphology of unoriented and oriented films.The morphological studies were conducted in a JEOL 1200 EX11 transmission electron microscope. A single replica technique was employed.18 The oriented samples were coated with Au-Pt (40-60, 30" oblique to the sample surface) and carbon (90' to the sample surface) uia vacuum deposition. The film was then soaked in methylene chloride-- trifluoroacetic acid (CH,Cl,-TFA) to dissolve the polyimide and the replicas were recovered for analysis. The unoriented samples were subjected to a potassium permanganate etching prior to the vacuum deposition." Results and Discussion Orientation in the Films As shown in Fig.3, the WAXD fibre patterns for CPIs clearly indicate that the copolyimides do not possess three-dimensional order since no quadrant diffraction spots can be seen. The addition of PMDA-PFMB thus disrupts the three- dimensional order which exists in CPI( 100/0), and leads to a two-dimensional mesophase order. WAXD fibre patterns for oriented CPI( 100/0) films with different draw ratios reveal that the overall crystallinity is increased by observing the further development of the diffraction spots located in the quadrant regions.', The crystal orientation has to be investi- gated since the macroscopic properties depend on both the overall orientation and the ~rystallinity.'~"~ Hermans orien- tation factors were calculated for CPI(lOO/O) based on eqn.(1) and (2) from WAXD results which is an indication of the crystal orientation. It is clear that with increasing draw ratio, the crystal orientation increases drastically. Fig. 4 shows the orientation factors as a function of the draw ratio. At low draw ratios the crystallites are less aligned. As the draw ratio \I IU film Fig. 2 Experimental set-up (top view) for polarized FTIR experiments at different angles between the incident beam and film surface increases one observes a drastic increase in orientation with a plateau being reached at a draw ratio of ca. 5. Dichroic ratios were measured for different draw ratios of CPI( 100/0) films. Fig. 5 illustrates the spectra for CPI( 100/0) films drawn to 500%. The 1778 cm-' absorption band is very prominent when the electric-field vector of the polarized IR beam is parallel to the draw direction.This is not surprising since the transition moment vector is oriented parallel to the imide planes (in-plane). If one investigates the 738 cm-' absorption band when the electric-field vector is parallel to the draw direction, the intensity is close to zero. This indicates a planar alignment of the imide planes which is more or less parallel to the surface of the film. On the other hand. when the electric-field vector is oriented perpendicular to the draw direction, one observes a drastic decrease in the intensity of the 1778cm-' band and an increase in the 738cm-' band. The decrease in the 1778 cm-' band is expected owing to the orientation of the transition moment vector (parallel to the imide plane) while an increase of the 738 cm-' band reflects the imide planes are not perfectly parallel to the film surface, and instead, they are tilted relative to the film surface.The tilting of the transition moment vector is, however, always perpendicular to the draw direction. Fig. 6 shows the variation of the dichroic ratios as a function of the draw ratio for CPI( 100/0)films. The dichroic ratios for the 738cm-' band were calculated using eqn. (4) while the dichroic ratios for the 1778 cm-' band were calculated in an inverse manner in order to compare the degree of order on the same scale, and in this case R=l represents random orientation, while R =0 represents perfect alignment of the transmission moment vectors.It is evident that upon increas- ing the draw ratio the alignment of the chain molecules is drastically enhanced. The in-plane alignment of the imide planes is also improved as the draw ratio is increased. It is clear that the value of dichroic ratio of the 738 cm-' band is small compared to that of the 1778 cm-' band at the same draw ratio. Moreover, the variance (the rate of change) in the dichroic ratios of the 738 cm-' band is larger than those of the 1778 cm-I band. Note that the magnitude change of the dichroic ratios is a measure of the alignment of the imide planes along two perpendicular directions. As the draw ratio increases the change of dichroic ratio of the 738 cm-I band is an indication that the planar alignment is increased owing to a reduction of the tilting of the imide planes.On the other hand, the variance of the 1778 cm-' absorption band reflects the enhanced alignment of the molecular axis relative to the draw direction. This variance is, furthermore, expected to be smaller than that of the 738 cm-' band since for the 738 cm-' band the inplane orientation of the molecular chain axis originally existed in the unoriented films. This experimental observation manifests that with increasing draw ratio, the alignment of the imide planes parallel to the film surface is relatively easier than the alignment of those along the molecu- lar chain axis. This can be understood by the fact that the tilting of the imide planes is only associated with the torsional angle change of the covalent bonds and the change of conju-gational length in CPI, while the change of alignment of the imide planes along the molecular chain direction has to involve a larger scale molecular motion related to the chain conformational change.The former largely belongs to an intramolecular, non-cooperative motion, while the latter to an intermolecular cooperative motion. From the dichroic ratios (Fig. 6) for oriented CPI( 100/0) films a plateau region is not observed at higher draw ratios, and linear relationships for both 1778cm-' and 738cm-' absorption bands as a function of the draw ratio are evident. However, from Fig. 4, WAXD experiments illustrate a plateau which is reached at a draw ratio of five.Note that WAXD J. MATER. CHEM., 1994, VOL. 4 Fig. 3 WAXD fibre patterns for uniaxially oriented CPT copolyimides with compositions of (a)CPI( 100/0), (h)CPI(85/15), Ic) CPI(70/30) and (d)CPl(50/50) 1.oo-0.92.-f0-5 0.84.-I0-.-5 0.76:; 0/c. (I)cCal.-Z 0.68.-/ Fig. 4 Crystal orientation factor changes with different draw ratios for CPI( 100/0) films experiments provide only the degree of crystal orientation, and FTIR analysis is a representation of the overall orien- tation, which is attributed to the orientation of both crystalline and non-crystalline regions. So at high draw ratios the overall orientation is improved mainly by the alignment of the chain molecules in the non-crystalline regions.In order to study detailed chain conformation and packing changes during the deformation of the films, a systematic investigation on the intensities of the two vibrational absorp- tion bands (1778 cm-' and 738 cm-') at different incident angles under a polarized IR beam was conducted. Note that I!!!! !!I!/ I ] ' " ' ! !';I 1900 1640 1380 1120 860 600 wavenumberkm-' Fig. 5 Polarized FTIR spectra of CPI(lOO/O) films at a draw ratio of five along the directions (a) parallel and (h) perpendicular to the elongation when the R values were calculated for the 1778 cm-' band, the electric-field vector was parallel to the draw direction, while for those of the 738 cm-l band, the electric-field vector was perpendicular to the draw direction.The experimental set-up is shown in Fig. 2. Fig. 7 and 8 show the dichroic ratios J. MATER. CHEM., 1994, VOL. 4 magnitude and the variance (the rate of change) of the0.701 738 cm-' band is relatively high. This indicates that at low incident angles the transition moment vector of this absorption band is tilted with respect to the incident beam, and there is a larger component of the vector parallel to the vibrational direction of the incident beam. This component contributes R much to the absorbance of the polarized IR beam at low 0.28 incident angles. With increasing incident angle, the R values of the 738 cm-' band drastically decrease, revealing that this component contribution to the absorbance decreases hecause the tilting angle between the transition moment vector and incident beam becomes smaller.One can thus conclude that 0.141+-LLL0.001 2 3 4 5 6 in the highly oriented CPI( 100/0) films the imide planes are nearly parallel to the film surface. draw ratio Furthermore, for both absorption bands, the R values Fig. 6 Dichroic ratio changes with the draw ratio for CPI( 100/0) decrease with increasing draw ratio. This is an indication that films for (0)1778 and (0)738 cm-I with increasing draw ratio, both orientations of the molecular chain axis along the draw direction (observed from the change of 1778cm-') and lateral orientation of the imide planes (observed from the change of 738 cm-') are enhanced. This kind of chain conformation and packing leads to a conceivable physical picture of uniaxial planar orientation. Fig.9 and 10 show the changes of the R values as a function 0.66 0 of the incident angle for the 1778 and 738 cm-' bands for both CPI(70/30) and CPI(50/50)films at a draw ratio of two. R 0 Greater sensitivity of the R values to the incident angle in the 738 cm-' band for CPI(70/30) films and CPI(50/50) films was observed compared to those of the 1778 cm-' band. Furthermore, if one looks at the three CPI films at the same 0.421 draw ratio of two, the absolute R values for both the Fig. 7 Relationship between dichroic ratio at 1778 cm-' absorption and incident angle for CPI(100/0) films with different draw ratios: 0,2x; 0,3x; A,4x; A,5x .oo 0.e 1.o 0.42 0 0 00 0 0 0 0.6 0 0 F20 36 52 68 84 100R 4 0 A 0 incident angle/degrees 0 0 Fig.9 Relationship between dichroic ratio at (0) 1778 and f. 0 (0)738 cm-' and incident angle for CPI(70/30) films at a draw ratio 4.0 0 of 2 0.2 4 A 0.90O.Ot! ............. .........10 27 44. 61. 78 95 incident angle/degrees Fig. 8 Relationship between dichroic ratio at 738 cm- 'absorption and incident angle for CPI( 100/0) films with different draw ratios: 0,2x; 0,3x; A,4x; A,5x R 0 (R)of CPI( 100/0) films as a function of both the draw ratio and the incident angle. With increasing incident angle the values of dichroic ratios decrease, and it seems that this is a 0.24 0 general trend for every draw ratio. Since the transition moment vector for the 1778cm-' band is parallel to the imide planes, and thus to the molecular chain axis, the change of R with respect to the incident angles is not as prominent as in the case of the 738cm-' band.This change obtained from the Fig. 10 Relationship between dichroic ratio at (0) 1778 and 1778cm-I band ranges between 0.1 and 0.14 at each draw (0)738 cm-l and incident angle for CPI(50/50) films at a draw ratio ratio. On the other hand, at low incident angles both the of 2 1778 cm-I and 738 cm-' absorption bands decrease with increasing PMDA-PFMB composition. At an incident angle of 30" the R value of CPI(lOO/O) films was 0.64 for the 1778 cm-I band, and it decreased to 0.49 for CPI(50/50) films; that of CPI( 100/0) was 0.82 for the 738 cm-' band, and decreased to 0.53 for CPI( 50/50)films.Again, the decrease of the R value in the 738 cm-' band based on the composition is more drastic compared to that of the 1778 cm-' band. On the other hand, the incident angle changing from 20 to 90" leads to a 0.4 decrease for the 738 cm-' band, and less than 0.15 for the 1778 cm-' band for all the three cases. These experimental observations may be explained by confor-mational and chain packing changes during the deformation. In particular, these changes indicate that with increasing PMDA-PFMB composition, the chain rigidity and linearity increase. As a result, both orientations along the molecular chain axis and lateral orientation of the imide planes are enhanced.Therefore, the microscopic structure in orientated CPI thin films is conceivably closer to a uniaxial planar orientation with increasing PMDA-PFMB composition. Surface Morphology Fig. 11 and 12 represent surface morphologies of unoriented and oriented CPI(lOO/O) thin films at different draw ratios (two and four) obtained by TEM. For unoriented films, the etching method is necessary in order to observe the ordered texture. Fig. 11 shows that in the unoriented polyimide film some ordered regions exist, which are fibril shaped with a lateral size of ca. 5 nm. The fibril direction in this figure is randomly distributed. If one assumes that the chain direction is parallel to the fibril direction which is true based on the electron diffraction in very thin films," this TEM observation reveals a random distribution of the c-axis of the ordered crystal regions in the film plane, which is recognized as 'in- plane orientation'.With increasing draw ratio, as shown in Fig. 12(a) and (b), the fibril type of texture becomes clearer and more oriented along the draw direction. It is interesting that the lateral size of the fibril region seems to decrease with increasing draw ratio, and corresponds well with the apparent crystallite size of the (310) crystalline plane observed from WAXD experiments using the Sherrer equation.12 Note that the (310) crystalline plane is one of the (hkO) planes in a monoclinic crystal lattice of the CPI( 100/0)crystals, and it is perpendicular to the c-axis [(OOl) crystalline plane].For example, a draw ratio of two leads to a lateral size of the fibril texture of ca. 4.5 nm (upper limit), and at a draw ratio of five, this lateral size decreases to ca. 3.5 nm (lower limit). Fig. 11 TEM surface morphology of unoriented CPI(100/0) film after etching J. MATER. CHEM., 1994, VOL. 4 Fig. 12 TEM surface morphology of CPI(100/0)film at a draw ratio of (a) 2 and (b)4 Furthermore, the fibril textures start clustering along the lateral direction. Fig. 13 includes the results obtained both from WAXD (points) and from TEM (vertical bars) obser- vations. A similar surface morphological observation with fibril texture can also be found in other CPI films at a draw ratio of two. These morphological observations indicate that under deformation the CPI films undergo a transition from an in-plane orientation to a uniaxial planar orientation of the chain molecules with a formation of the ordered regions which 4.4 .$ 3.8 cd c v)x 3.2 c m 2.0t : : : ! : ; 1 ! : : ' ! ! ! : : 1 ,L,1 3 5 ~ 7 ~ 9 draw ratio Fig.13 Apparent crystallite size of (310) crystalline plane observed from WAXD and crystal size perpendicular to the draw direction observed from TEM J. MATEK. CHEM., 1994. VOL. 4 is either crystalline [for CPT(100/0)] or mesophase type in nature (for other CPT copolyimides). Conclusions Oriented CPI films have been studied through WAXD and FTIR experiments to investigate the orientation effect on chain conformational and packing changes.As the draw ratio increases, both the orientation along the molecular chain axis (and therefore, the draw direction) and the lateral orientation of the imide planes with respect to the film surface are enhanced. Furthermore, with increasing PMDA-PFMB com-position. the imide planes also become increasingly parallel to the film surface in oriented films. The surface morphology investigation indicates that the ordered regions in both unori- ented and oriented films are fibril in nature. The fibril direction becomes increasingly parallel to the draw direction of the films. In oriented CPT( 100/0)films the lateral size of the fibril texture decreases with increasing draw ratio, which corre- sponds well to the apparent crystallite size change of the (310) crystalline plane observed from WAXD experiments.This work was supported by SZDC’s Presidential Young Investigator Award from the National Science Foundation (DMR-9157738) and the industrial matching funding from Hercules Inc. Support was also provided by the National Center of Science and Technology for Advanced Liquid Crystalline Optical Materials (ALCOM) from the National Science Foundation (DMR-8920147) at Kent State University, The University of Akron and Case Western Reserve University. References 1 H. Lee, D. Stoffey and K. Neville, New Linear Polymers. McGraw-Hill, New York, 1967, pp. 183 and 224. 2 C. E. Scroog, J. Polym. Sci., Macromol. Rec.. 1976,11,161 3 D. Y.Yoon, W. Parrish, L. E. Depero and M. Ree, ,Vuterial Science of High Temperuture Polymers for Microelectronic s, MRS Symposium Proceedings, Pittsburgh, PA, 1991, vol. 227. 4 S. Z. D. Cheng, F. E. Arnold Jr., A. Zhang. S. L. C. FIsu and F. W. Harris, Macromolecules, 1991,24, 5856. 5 F. E. Arnold Jr., S. Z. D. Cheng, S. L. C. Hsu, C. J. Lee and F. W. Harris, Polymer, 1992,33, 5179. 6 F. E. Arnold Jr., D. Shen, C. J. Lee, F. W. Harris, S. Z. D. Cheng and H. W. Stark-weather Jr., J. Muter. Chem., 1993,3, 185. 7 F. E. Arnold Jr., D. Shen, C. J. Lee, F. W. Harris, S. Z. D. Cheng and S.-F. Lau, J. Mater. Chem., 1993, 3, 353. 8 F. W. Harris and S. L.-C. Hsu, High Perform. Polym., 1980, 1, 1. 9 F. W. Harris, in Polyimides, ed. D. Wilson, H. D. Stenzenberger and P. M. Hergenrather, Chapman and Hall, New York, 1990, pp. 1-37. 10 S. Z. D. Cheng, S. K. Lee, J. S. Barley, S. L. C. Hsu and F. W. Harris, Macromolecules, 1991, 24, 1883. 11 S. K. Lee, S. Z. D. Cheng, C. J. Lee, F. W. Harris, T. kyu and C. Yong, Polym. Znt., 1993,30,215. 12 S. Z. D. Cheng, Z. Wu, M. Eashoo, S. L. C. Hsu and F. W Harris, Polymer, 1991,32, 1803. 13 M. Eashoo, D.-X. Shen, Z.-Q. Wu, C. J. Lee, F. W. Harris and S. Z. D. Cheng, Polymer, 1993,34, 3209. 14 H. G. Rogers, R. A. Gaudiana, W. C. Hollinsed, J. S. Manello, C. Mcgowan and R. Sahatjian, Macromolecules, 1985, 18, 1058. 15 F. E. Arnold Jr., Ph.D Dissertation, University of Akron Akron, Ohio, 44325-3909, 1993. 16 R. W. Snyder and C. W. Shen, Appl. Spectrosc., 1988,42, i03. 17 J. L. Koenig, Spectroscopy of Polymers, American C hemical Society, Washington D.C., 1992. 18 R. H. Olley, D. C. Bassett and D. J. Blundell, Polymtr, 1986, 27,344. 19 S. Z. D. Cheng, J.-Y. Park, C. J. Lee and F. W. Harris Polym. Prep., ACS, Dic. Polym. Chem., 1993,34(1), 774. Paper 3/04449A; Received 26th July, 1993

ISSN:0959-9428    

DOI:10.1039/JM9940400105

出版商:RSC    

年代:1994

数据来源: RSC