|
21. |
Gel to liquid crystal transitions for vesicles in aqueous solutions prepared using mixtures of sodium dialkylphosphates (R1O)(R2O)PO2–Na+and (R3O)2PO2–Na+, where R1= C10H21, R2= C14H29or C18H37and R3= C12H25, C14H29, C16H33or C18H37 |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2709-2715
Michael J. Blandamer,
Preview
|
PDF (732KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2709-2715 Gel to Liquid Crystal Transitions for Vesicles in Aqueous Solutions prepared using Mixtures of Sodium Dialkylphosphates Michael J. Blandamer,* Barbara Briggs and Paul M. Cullis Department of Chemistry, The University, Leicester, UK LE1 7RH Jan B. F. N. Engberts, Anno Wagenaar and Elly Smits Department of Organic & Molecular Inorganic Chemistry, University of Groningen , Nijenborgh 4, 9747 AG Groningen, The Netherlands Dick Hoekstra Department of Physiological Chemistry, University of Groningen, Bloemsingel 10,9712 KZ Groningen, The Netherlands Anna Kacperska Department of Physical Chemistry, University of Lodz, Pomorska 18,9146 Lodz, Poland The scans recorded by differential scanning microcalorimetry (DSC) for aqueous solutions containing vesicles prepared from mixtures of two sodium dialkylphosphates are complicated where the first surfactant anion (R'O)(R2O)PO2-Na+ has alkyl groups R" and R2 with different chain lengths, and where the second surfactant anion (R30)2P02- has two alkyl groups R" with the same chain length.For mixed solutions where R' = C,,H,, , R2 = C14H29 or C,8H37 and R3 = C12H2,, C14H29, C1&!33 or CiBH37, the DSC traces can be understood in geometric terms. Where the chains can be assembled into bilayers with modest mismatch, the DSC traces show well resolved features. With increase in mismatch, the complexities of the DSC traces are consistent with the presence of bilayers in which the domains differ in composition. Poor chain packing and consequent weak intravesicular van der Waals forces between the alkyl chains favour low temperatures for gel to liquid crystal transitions.The characteristic gel to liquid crystal transitions for vesicles in aqueous solutions' prepared using sodium dialkyl-phosphates, (RO),PO, -Naf, depend2 significantly on the length of the chains in the alkyl group R. In the traces recorded by DSC,3,4 the gel to liquid crystal transition produces an extremum at a melting temperature, T,,analysis of the trace yielding the enthalpy changes for the transitions expressed in terms of surfactant monomer and patchnumber^.^ The latter describe the numbers of surfactant monomers which melt cooperatively. For vesicles formed by the surfactant2 (C,4H290)2P02-Na+, T', = 52.2 "C, the patch number n = 140 and the enthalpy of melting (fusion) Af,, H" = 5.6 kcal (mol monomer)- '.An important contribu- tion to this latter enthalpy is assigned to chain-chain van der Waals interactions within the bilayers. Interest in bilayers formed by these novel surfactants stems from their close relationship with biologically important bilayer systems. However, naturally occurring bilayer systems do not com-prise single substances. Consequently, we examined' the thermal stabilities of vesicles produced in aqueous solutions prepared using mixtures of soldium dialkylphosphates, (R'O)P02-Naf and (R30),P0,-Naf, in which the alkyl chains in each component surfactant are identical. When R1 and R3 differ by C2H4, the temperature T, of an equimolar mixture of the two surfactants is intermediate between the recorded T,s of the solutions prepared from one of the two surfactants.However, when R' and R3 differ by more than C2H4, the DSC traces are complicated. We suggested6 that such complexity is a consequence of a mismatch in chain lengths as the two surfactants associate to form bilayers within the vesicle structures. This conclusion is supported by the results reported here. Solutions containing vesicles pre- pared by mixing two sodium dialkylphosphates, (R'O)(R20)P02-Na+ and (R30),P02-Na+, where R' and R2 differ in alkyl chain length. The DSC traces for these systems are complicated but an underlying pattern emerges based on the lengths of alkyl chains R1,R2 and R3.If a bilayer can be modelled using the building blocks, R1,R2and R3 such that the mismatch is small, the pattern recorded by the calorimeter is simpler than when the mismatch is signifi- cant.If in the matching exercise holes are produced, the tran- sitions identified by the DSC are poorly defined and the enthalpy of transition is small suggesting that the van der Waals interactions between the alkyl chains are weak. Experimental Differential Scanning Microcalorimetry The DSC scans were recorded using a differential scanning microcalorimeter (MicroCal Ltd., USA) as previously de~cribed.~.~At regular intervals during this work (e.g. every 2 days) scans were recorded in which both sample and refer- ence cells were filled with water.The water-water baseline was a gentle concave-downwards curve. It was important to check this baseline at regular intervals because some sur- factant systems produced a deposit on the inner surface of the sample cell. The scans for these systems are not reported, but the presence of such a deposit was signalled by a sharp spike on the otherwise smooth water-water baseline. The deposits were removed by washing the cell repeatedly with hot (e.9. 90°C) concentrated HCl(aq). A clean cell produced a smooth water-water baseline and it was important to re-establish this baseline before the study of surfactant systems was continued. In the experiments reported here the sample cell contained either a single sodium dialkylphosphate (aq; 8.4 x mol dm-3) or an equimolar (4.2 x mol dm-3 for each surfactant) mixture of two sodium dialkylphosphates(aq).The usual scan rate was 60°C h-' and the scans were nor- mally recorded between 5 and 90°C. The exceptions to the latter generalisation involved a more complicated mixed solution prepared using (R'O)(R20)P02-Na+ and (R30)(R40)P02-Naf where R' = C10H2', R2 = C14H29, R3 = C10H2, and R4 = C18H37. In the last case, the DSC scan was recorded over the range 2-90 "C. Surfactants The dialkylphosphates were prepared as described pre- vio~sly.~In all cases, the surfactants were in the form of white solids. Preparation of Vesicles We have emphasised the importance of developing a protocol for the preparation of vesicle solutions if reproducible and repeatable DSC traces are required.Similar considerations apply to the preparation of mixed solutions used in this study. For ease of comparison, we report DSC traces for aqueous solutions containing a common total concentration of surfactant, 8.42 x mol dm-3. For the mixed solu- tions the required mass of each surfactant was weighed out to produce an aqueous suspension, volume, 2.2 cm3, containing equal concentrations (4.2 x mol dm-3) of each alkyl- phosphate. The aqueous suspension was heated to tcm-perature T* (see below) and held at that temperature for 1 h. In detail, temperatures T* for (R'O)(R20)P0, -Na+ and (R30)(R40)P02-Na+ were as follows: (a) T* 2 60°C for (i) R1= C10H2,, R2 = C14H29, R3 = R4 = C12H25; (ii) R' = Cl0HZ1, R2 = C14H29, R3 = R4 = C14H29; (iii) R' = C10H2,, R2 = C16H37, R3 = R4 = C12H25; (iv) R' = Cl0HZ1, R2 = C14H29, R3 = C10H2, and R4= C18H3,; (b) T* 2 70°C for (i) R' = CIOHZ1,R2 = C14H29, R3 = R4 = C16H33; (ii) R' = C10H21, R2 = C18H37, R3 = R4 = C16H33; (c) T* 2 75 "C for (i) R' = C10H2,, R2 = C14H29, R3 = R4 = C18H37; (ii) R' = C10H2', R2 = R3 = R4= C18H37.Protocol for DSC The surfactant solutions were placed in the sample cell in the manner de~cribed.~-~ The solution was cooled in the sample cell of the calorimeter to 5 "C. When thermal equilibrium had been attained, the scan was initiated to a maximum tem- perature of 90°C.After the first scan had been recorded, the solution was allowed to cool slowly in the calorimeter to 5°C. After a predetermined time (see captions to figures), a new scan was recorded from 5 to 90°C. For a few systems described below, the low-temperature limit was set at 2°C because there was clear evidence for an extremum in the recorded trace well below 20°C. Analysis of DSC Scans The scan data were recorded on 3.5 in? discs and later analysed using the Origin (MicroCal) software. The recorded quantity4 was the relative isobaric heat capacity, SC,. In the first stage of the analysis, the water-water baseline (see above) was subtracted from the DSC trace. For most systems reported here, the outcome from this step was a plot showing one or more extrema.A chemical baseline4 produced by a t 1 in = 2.54 an. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 gradual dependence of heat capacity on temperature wassub-tracted. Where an extremum was well defined, the depend- ence of heat capacity on temperature was fitted to an equation describing one or more independent processes of the general form, X(aq) Y(aq), where X(aq) and Y(aq) describe two states in equilibrium. Each such process can be described using eqn. (1) for the molar isobaric heat capacity, C,, m. Here AvH H" is the van't Hoff enthalpy for the transition, cal- culated from (C,, m),= at Tmwhere K is unity. The area under the curve characterised by Tmyields the calorimetric enthalpy of fusion, AfusHzal.The two enthalpies AVHH"and AfusHzal are brought into coincidence using the patch number n.In one limit, n is unity showing that in each vesicle and for each surfactant unit the alkyl chains gain liquid-like freedom independent of the changes taking place in neighbouring chains. With increase in n, the extent of cooperative melting increases within a patch (domain) of the bilayer forming a vesicle. A trace showing several extrema with differences in patch numbers is indicative of domains having very different proportions of the two surfactants. Results In line with our previous practice, we report (i) a comparison of DSC traces obtained for aqueous solutions containing various phosphate surfactants and (ii) repeat DSC traces for the surfactant solutions recorded over an extended period of time.The point of recording the latter set was to test the extent to which the transitions producing the extrema are reversible in the thermodynamic sense. Where the same trace was obtained in repeat scans recorded over a period of several hours it can be reasonably assumed that the gel to liquid transitions are reversible for that range and type of vesicles in solution. Moreover, the validity of using an analysis based on eqn. (1) is confirmed in that this equation assumes complete revqrsibility . Data conforming to these criteria are shown in Fig. 1 for three systems contain- ing (i) the single sodium dialkylphosphate (C10H2 10)(C18H370)P02 -Na+, (ii) the single surfactant (C14H2,0)P02-Na+ and (iii) an equimolar mixture of the two phosphate surfactants.The contrast between the scans for the single surfactants and the mixture was striking [Fig. l(a)J. Nevertheless, the repeat scans [Fig. l(b)] showed that the complicated scans for the mixture were reversible over a period of ca. 14 h. Scans for the surfactant with(CloH210)(C,8H370)P0,-Na+ T, = 20.9"C were analysed [cf:eqn. (l)] in terms of an overall enthalpy change of 5.4 kcal (mol monomer)- ',a patch number of 216 & 8 and two independent transitions. The scan for the surfactant (Cl4H2,0)PO2-Na+ with T, = 52.2"C had a similar enthalpy change of 5.6 kcal (mol monomer)- ', n = 140 and a single transition. In the scan for the mixture (C OH, O)(C ,H 3,O)PO, -Na + + (C 4H2g0)2P02 -Na + [Fig.l(a) and l(b)J, there are two important features at 11.5 and 28.6"C, but they are much less intense than the extrema observed for the component solutions. The feature centred on 28.6"C can be accounted for only by using three or more components having the form described by eqn. (1). The calcu- lated patch number is 222 with an integrated enthalpy of melting of 4.6 kcal (mol monomer)- '. Interestingly, the patch number is almost equal to that for vesicles formed from the pure surfactant, (C, ,H2 O)(Cl ,H3,0)P02 -Na +, although the enthalpy of fusion is lower. The scans for a similar set of solutions in which the surfactant (Cl,H,90),P02-Na+ was replaced by the surfactant (Cl,H,,0)(C,,H370)P02-Na+ (Fig. 2) showed a simpler scan for the mixed system except J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 271 1 0.06 A B 0 0.04 r cI IY Y--iu -0.010 $--. -.. 0.02 k -0.020Ot 20 40 60 0 20 40 60 TPC T/OC Fig. 1 Dependences on temperature of the differential heat capacities (reference = water) for aqueous solutions containing vesicles prepared (aq; 8.4 x rnol dm-3) where T, =from sodium dialkylphosphates. A, Scans recorded for: (a) surfactant (CloH210)(C18H3,0)P02-Na+ 20.9 k0.3 "C; (b) surfactant (Cl,H2,0),P02-Na+ (aq; 8.4 x mol dm-3) where T, = 52.2 0.1 "C and (c) equimolar mixtures (4.2 x rnol dm-3) of (Cl,H2,0),P02-Na+(aq) and (CloH210)(C18H,,0)2P02-Na+(aq).B, Scans (a) to (e) recorded consecutively for the system described in Fig.lA(c); for clarity the scans have been displaced on the heat capacity axis. that the extremum for the mixture was close to that for the single surfactant (C,,H,, O)(C,,H,,O)PO, -Na+, although again with much less intensity. The patch number, 395 f22, is much larger than that for either pure surfactant coupled with a significantly lower enthalpy of fusion, 1.8 kcal (mol monomer)-'. The scan patterns were reversible for five traces obtained over a period of 8 h; cf: Fig. l(b). The extremum could be fitted using three independent transitions having the form given by eqn. (1);Fig. 2(b).The enthalpy change defined by the extremum was much lower at 1.8 kcal (mol monomer)-' than for the two separate solutions; e.g. 3.9 kcal +(mol monomer) -for (C ,H2O),PO2-Na .By contrast, when the chain length of the symmetric anion was increased forming (C,,H,,O),PO,-Na+, the scan pattern of the mixed solutions showed a complex trace with poorly defined extrema, a trend which continued when a further change was made to solutions containing (C,,H,,O),PO,-Na~(aq); Fig. 3. Fig. 3 compares the DSC traces for the individual solutions and an equimolar mixture. The complex scan for the mixture was repeated in five scans recorded over a period of 19 h; Fig. 3(b). It is noteworthy that one of the extrema for the mixture (Cl,H,,O)(CI,H,,O)PO,-Na+-(Cl,H,,O),PO,-Na+was at the same temperature as T, for the solutions containing (C ,H, ,O)(C sH3,0)P0,-Na+, 20.6 "C. Other notable extrema occur near 24,66.5 and 73.6"C.TrC Fig. 2 Dependences on temperature of the differential heat capacities (reference = water) for aqueous solutions containing vesicles prepared from sodium dialkylphosphates. A, Scans recorded for: (a) surfactant (CloH,10)(C18H370)P02-Na+(aq; 8.4 x mol dm-3) where T, = 20.9 k0.3"C; (b) surfactant (Cl,H,,0)2P02-Na+ (aq; 8.4 x lop3rnol drnp3) where T, = 34.8"C and (c) equimolar mixtures (4.2 x mol where T, =dm-') of (C,,H,,O),PO,-Na+(aq) and (CloH210)(C18H370)P02-Na+(aq)21.7 "C. B, Dependences on temperature of the molar heat capacity of an aqueous solution described in Fig. 2A(c);dotted lines show calculated contributions (Origin software, MicroCal Ltd.) from three independent equilibria of the general form, X(aq) sY(aq).2712 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I B 7 0.04 Y-m \ clp O0 0.02 I I 1 1 I 20 40 60 80 20 40 60 80 T/T T/"C Fig. 3 Dependences on temperature of the differential heat capacities (reference = water) for aqueous solutions containing vesicles prepared (aq; 8.4 x lo-, mol d~n-~) from sodium dialkylphosphates. A, Scans recorded for: (a)surfactant (C,oH,10)(Cl,H,70)P0,-Naf where T, = 20.9 f0.3"C (cf: Fig. 1A); (b) surfactant (C,,H,,O),PO,-Na+ (aq; 8.4 x mol drn-,) where T', = 77.1 k0.1"C and (c) equimolar mix- tures (4.2 x mol dm-3) of (C,,H3,0),P02-Naf(aq) and (C,oH,,0)(C,,H370)P0,-Naf(aq).B, Scans (a)-@)recorded consecutively for the system described in Fig. 3A(c); for clarity the scans have been displaced on the heat capacity axis.A similar series of experiments used (C,,H,, O)(C14H,,0)P0, -Na + as a common alkyl-phosphate. In fact, the DSC trace for solutions containing only this surfactant [Fig. 4(a)] showed a much less intense but broad extremum with T, = 14.4 f0.1"C. The calori- metric enthalpy change calculated over the broad transition was small, 1.8 kcal (mol monomer)-'. This contrast with the DSC scan for (C,2H,,0)2P02-Na+(aq) was significant [Fig. qa)] where T', = 34.8 f0.2"C with a calorimetric enthalpy change of 3.9 kcal (mol monomer)-'. Moreover, the tran- sition could be accounted for using eqn (1) in conjunction with n = 168. A single broad extremum was recorded for the equimolar mixture (C ,O)(C 14H290)P02 -Na+-(C,,H,,O),PO,-Na+ at T', = 24.9 f0.1 "C, roughly midway between the T,s for the single surfactants.The calo- rimetric enthalpy for the overall transition is 2.0 kcal (mol monomer)-', n for both component equilibria being 606 & 11. For solutions containing the single surfactant with 3OC A"-"-(b) 200 I Y 7 I-E-100 cp (a) (c) 0 I I 20 40 T/"C longer dialkyl chains, (RO),PO,-Na+ where R = C,,H,, , C,,H,, and C18H37, the T, increases from 52.2 through 66.3 to 77.1 "C. The DSC traces recorded for the equimolar mix- tures containing the surfactant where R = C14H29 and (C ,H ,O)(C 4H,,O)PO, -Na showed a broad extremum + at 37.1"C between T,s or the two component surfactants, a pattern reproduced over five successive scans; Fig.5. The broad extremum was fitted to eqn. (1) using three independ- ent processes but with a common patch number, 148 f6, which is much lower than the number for the pure(Cl,H,10)(C14H,,0)P02~Na+system but is close to that for the (C14H,90)P0, -Na+ surfactant. The calorimetric enthalpy change is 2.6 kcal (mol monomer)-'. With increase in chain length of the surfactant with identical alkyl chains (Fig. 6) the extrema in scans for the mixture covered a broader temperature range and became less intense. At the next level of complexity, the scans were recorded for mixtures of surfactants where the alkyl chains in each sur- B 1 I I I 15 20 25 30 T/"C Fig. 4 Dependences on temperature of the differential heat capacities (reference = water) for aqueous solutions containing vesicles prepared (aq; 8.4 x mol d~n-~) from sodium dialkylphosphates.A, Scans recorded for: (a) surfactant (C,,H2,0)(C,4H,90)P0,-Naf where T, = 14.4 & 0.1 "C; (b) surfactant (C,,H,,O),PO,-Naf (aq; 8.4 x lo-, mol drn-,) where T, = 34.8 f0.2"C and (c) equimolar mixtures (4.2 x lo-' mol dm-,) of (C,oH,,0)(C,4H,90),P02~Naf(aq)and (C,,H,,O),PO,-Na+ where T, = 25.1 "C. B, Dependences on tem-perature of the molar heat capacity of an aqueous solution described in Fig. 4A(c); dotted lines show calculated contributions (Origin software, MicroCal Ltd.) from equilibria of the general form, X(aq) eY(aq). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 0.06 1 7 -c 0.04 I Y-(0 -Y Q"uo 0.02 --I I I I01 I I I I I J I 0 20 40 60 0 20 60 100 T/OC T/OC Fig.5 Dependences on temperature of the differential heat capacities (reference = water) for aqueous solutions containing vesicles prepared from sodium dialkylphosphates. A, Scans recorded for: (a)surfactant (C,oH,10)(C,,H,,0)P02-Na+(aq; 8.4 x lo-, mol drn-,) where T, = 14.4f 0.1"C; (b) surfactant (Cl,H,,O),PO,-Na+ (aq; 8.4 x lop3 mol dm-7 where T, = 52.2 k0.1"C and (c) equimolar mixtures (4.2x mol dm-') of (C,,H,,0),P02-Na+(aq) and (C,,H2,0)(C,,H,,0)P0,-Na+(aq).B, Scans (a)-(e) recorded consecutively for the system described in Figure 5A(c); for clarity the scans have been dispIaced on the heat capacity axis. 0.06 0.04 c I Y-7 0.04 m Y-Y (0 u"Y 0.02 0.02 I0 01 1 I 10 40 80 20 40 60 T/OC T/" C Fig.6 Dependences on temperature of the differential heat capacities (reference = water) for aqueous solutions containing vesicles prepared from sodium dialkylphosphates. A, Scans recorded for: (a) surfactant (C,oH210)(C1,H,,0)P0,-Na+ (aq; 8.4 x lod3mol dm-') where T, = 14.4 0.1"C; (6) surfactant (C,,H,,O),PO,-Na+ (aq; 8.4 x lo-, mol dm-,) where T, = 66.3 k0.1"C and (c) equimolar mixtures B,(4.2x rnol drn-,) of (C,,H,,O),PO,-Na+(aq) and (C,oH,,0)(C,4H,,0)P02-Na+(aq).Scans recorded for: (a) surfactant (C,oH210)(C,,H,,0)P0,-Na+(aq; 8.4 x mol dm-,) where T, = 14.4k0.1"C; (b) surfactant (C,,H,,O),PO,-Na+ (aq; 8.4 x lo-' rnol dm-3) where T, = 77.1"C and (c) equimolar mixtures (4.2x lop3 mol dm-3) of (C,,H,,O),PO,-Na+(aq) and (CloH2,O)(C,,H2,O)PO,-Na+(aq)where T, = 67.3 "C.0.012 B 0.04 I 7 0.008 Y Y-m -Y 0.02 '0.004 0 0 I I I I 1 I 20 40 60 0 10 20 30 40 50 TPCT/OC Fig. 7 Dependences on temperature of the differential heat capacities (reference = water) for aqueous solutions containing vesicles prepared from sodium dialkylphosphates. A, Scans recorded for: (a)surfactant (C,,H2,0)(C,,H2,0)P0,-Na+(aq; 8.4 x rnol dm-') where T, = 14.4 0.1"C; (b) surfactant (CloH,,O)(C,,H,,O)PO,-Na+ (aq; 8.4 x lo-, mol dm-,) where T, = 20.9 k0.3"C and (c) equimolar mixtures (4.2 x lop3rnol dm-3) of these two surfactants where T, = 17.9 & 0.1"C. B, Scans (a)-(e) recorded consecutively for the system described in Fig. 7A(c); for clarity the scans have been displaced on the heat capacity axis. factant differed in length.For the systems described in Fig. 7, T, for the equimolar mixture was approximately midway between the T,s for the individual solutions, the scan being retraced over five consecutive scans. Nevertheless, the scan pattern for each system is complicated but could be accounted for using three independent values of the patch numbers. Discussion The background to the story described here centred on the complex reorganisation which takes place when the tem-perature of lipid bilayers is raised. It is well established that bilayers formed from phospholipids are responsible for bio-chemically important supramolecular structures.The inten-tion was, therefore, to probe the extent to which reorganisation within these layers, particularly the gel to liquid phase transition, depends on the presence of mixtures of surfactants. However, the thermal stabilities of lipid bilayers are complex. An analogy was to be drawn with the structural transitions undergone by vesicles containing mix-tures of surfactants where the chain lengths differ both between and within the surfactants. The differential scanning microcalorimeter has the necessary sensitivity to probe gel to liquid transitions in quite dilute solutions where we can be reasonably confident that intravesicular processes are responsible for the scan patterns and that the contributions from intervesicular interactions are small.The pattern formed by the differential scans for the mix-tures are clearly complicated; Fig. 1-7. However, an import-ant observation concerns the extent to which scan patterns for the mixtures were repeatable. Therefore, we conclude that the melting processes responsible for the extrema are limited to localised domains (patches) in the vesicles. An alternative explanation links the extrema to a massive reorganisation and disruption on increasing the temperature, leading to new structures which, subsequent to cooling back to, say, 5"C, would be characterised by new scan patterns on reheating. Therefore, the data show that an explanation based on the latter model is probably incorrect. Nevertheless, an under-J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 directly to the increasing difference in the total number of carbon atoms in the dialkyl chains; e.g. 24 andfor (C10H210)(C14H250)P02-Na+(aq)36 for (c 1gH3 7O)2'O2 -Na The latter observation prompts us to propose a general explanation based on a model for the bilayers which explores how these dialkyl chains can pack together. We illustrate the point by reference to two examples. In bilayers formed by the symmetric dialkyl surfactants, (C,,H,,O),PO,-Na+(aq) or (C18H370)2P02-Na+(aq),the alkyl chains can be arranged in just one configuration. Hence a single sharp extremum (Fig. 8) is accounted for in terms of bilayers having C24 and C,, carbon number widths. However, for the single dialkyl surfactant (C ,H, , OXC, 4H290)P02-Na (aq) models of the+ bilayers can be drawn having widths with carbon numbers, C,, having short alkyl chains opposite long chains and C,, where long chains are opposite long chains; Fig.9. In Fig. 9 we illustrate two of the four possible arrangements in which the total carbon widths are C24 and c28. The mismatch between the chain lengths is identified by a shaded area in the c28 arrangement. Therefore, van der Waals cohesions in the bilayers are smaller, the cohesion being weaker in the c28 system. In these terms the four components of the scan pattern are separated by small differences in T,; Fig. 4(a). The mismatch in the models for the mixed solutions is more significant. For the mixed system, (CloH,,O) (C14H2,0)P0,-Na+(aq) + (C18H,,0),P0,-Na+(aq)a large number of structures can be constructed with several different widths defined by the carbon number.These models show that if a surfactant can be accommodated within the bilayer of a host surfactant with minimum disruption, the scan shows well defined extrema. Where the chain lengths are incompatible, the melting occurs over a broad temperature I I ; 12 12 ; I I lying trend can be discerned in which extrema for mixttres;-c24 )I I I Ispan a difference in alkyl chain lengths. For example, the 1 I extremum recorded for the mixture comprising I I 18 [('I 2 HZ 5 O)2 -Na and (c1OH, 1o)(c14H290) IPO,-Na+(aq)] is distinct and midway between extrema Irecorded for the individual solutions.We note here that both I 18 I I broader temperature range with increase in the components contain 24 carbon atoms in the dialkyl chains. But with increase in the number of carbon atoms in the ;-c36 W I 1 I symmetric dialkyl surfactant, the scan pattern broadens, Fig. 8 Diagrammatic representation of the arrangements in bilayerscovering a larger temperature range. This trend is linked formed by symmetric dialkyl phosphates I I I ! Id 14 !. .-I I I 14 I I I I I I I I I I I I I * I-i :-I c28Cza 4 I I I I I I I I I I I I Fig. 9 Diagrammatic representation of the arrangements in bilayers formed by (C,,H, ,O)(C,,H,,O)PO, -Na+(aq) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 range and the patches probably differ considerably in com-position. We thank the SERC for their support through the Molecular Recognition Initiative and both the British Council and the CEC for an award under the ‘Go-West’ Scheme to A.K. References 1 T. Kunitake, Angew. Chem., Znt. Ed. Engl., 1992, 13,709. 2 M. J. Blandamer, B. Briggs, P. M. Cullis, J. B. F. N. Engberts and D. Hoekstra, J. Chem. SOC.,Faraday Trans., 1994,90,2703. 3 M. J. Blandamer, B. Briggs, P. M. Cullis, J. A. Green, M. Waters, G. Soldi, J. B. F. N. Engberts and D. Hoekstra, J. Chem. SOC., Faraday Trans., 1992,88,3431, 4 M. J. Blandamer, B. Briggs, J. Burgess, P. M. Cullis and G. Eaton, J. Chem. Soc., Faraday Trans., 1992,88,2874. 5 C. Gutierrez-Merino, A. Molina, B. Escudero, A. Diez and J. Laynez, Biochemistry, 1989,244, 3398. 6 M. J. Blandamer, B. Briggs, M. D. Butt, P. M. Cullis, J. B. F. N. Engberts and D. Hoekstra, submitted. 7 A. Wagenaar, L. A. M. Rupert, J. B. F. N. Engberts and D. Hoekstra, J. Org. Chem., 1983,54,2638. Paper 4/02345E; Received 20th April, 1994, 1994
ISSN:0956-5000
DOI:10.1039/FT9949002709
出版商:RSC
年代:1994
数据来源: RSC
|
22. |
Effects of surfactant charge and structure on excited-state protolytic dissociation of 1-naphthol in vesicles |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2717-2724
Yurii V. Il'ichev,
Preview
|
PDF (967KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2717-2724 Effects of Surfactant Charge and Structure on Excited-state Protolytic Dissociation of 1=Naphthol in Vesicles Yurii V. Il'ichev, Kyrill M. Solntsev and Michael G. Kuzmin" Department of Chemistry, Moscow State University, 117234 Moscow, Russian Federation Helge Lemmetyinen Institute of Materials Chemistry, Tampere University of Technology, P.O.Box 589,FIN33101 Tampere, Finland ~ ~ ~ ~~~~~~~~~ The kinetics of protolytic photodissociation of 1-naphthol in the bilayer membrane of cationic vesicles of di- dodecyldimethylammonium bromide (DDAB) and dioctadecyldimethylammonium bromide (DOAB) have been studied in comparison with the same reaction in vesicles of a zwitterionic lipid dipalmitoylphosphatidylcholine (DPPC).Evidence for the existence of two fractions (two types of site) of 1-naphthol molecules in vesicles of cationic surfactants which differ strongly in their rate constants for excited-state proton transfer was found, similar to the case for zwitterionic vesicles. The rate constants of the excited-state proton transfer for both fractions are much higher in bilayer membranes of cationic surfactants than for zwitterionic lipids (DPPC and egg lecithin). The activation enthalpy of excited-state proton transfer (ESPT) for both fractions of ArOH in the membrane of DDAB is ca. 40 kJ mol-', which is much higher than in homogeneous solutions and zwitterionic surfactants. Fluorescence kinetic data for DOAB vesicles allow no reliable conclusions to be drawn as to the temperature dependence of excited-state protolytic dissociation rate constants in these vesicles because the reaction rate is too fast. No significant decrease in the excited-state proton-transfer rate constants at the mem- brane phase-transition temperature of vesicles of cationic surfactants is observed, in contrast to the zwitterionic lipids.All these features characterize distinctions between the properties of the membranes of the vesicles of cationic and zwitterionic surfactants in proton-transfer reactions. Proton transfer, being a fundamental chemical reaction, con- stitutes an important step in many processes in cellular biology.' The elucidation of the dynamics and mechanism of this reaction is the key to understanding how proton-controlled processes are realized in biological membranes.Protolytic photodissociation (proton transfer between an acid in the singlet excited state and water molecules) provides a simple model reaction for proton-transport processes in membranes and could provide a deeper insight into the mechanisms of very complex systems which function in bio- membranes. Steady-state and time-resolved fluorescence measurements have been used extensively to study the kinetics of excited- state protolytic dissociation in proteinsg." and other microheterogeneous systems which can be considered as models of biomembranes [reversed micelles and water: in-oil (w/o) microemulsions,' '-" micelles,' 7-33 and o/w micro emulsion^^^]. The aim of this work was to study the kinetics of protolytic photodissociation of 1-naphthol in a bilayer membrane of cationic vesicles and compare them with the same reaction in zwitterionic lipid vesicles.Our prefious studiedp8 demon- strated some peculiarities in naphthol photodissociation in bilayer membranes of some phospholipids [egg lecithin (EL) and dipalmitoylphosphatidylcholine (DPPC)] when com-pared with aqueous solutions and other microheterogeneous systems. Time-resolved fluorescence data suggested the exis- tence of two fractions (two types of localization site) of naph- thol molecules (ArOH) in lipid membranes. These fractions differ strongly in their rate constants for excited-state proton transfer. Furthermore, temperature-dependent measurements pointed to a strong effect of phase transitions in DPPC bilayer membranes on ArOH photodissociation.8 The physi- cochemical nature of these two localization sites was not identified.Two possibilities, transmembrane and lateral inho- mogeneity of the lipid bilayer, were discussed.8 To obtain additional information on these phenomena we studied the reaction in bilayer membranes with a different structure and charge. We used artificial cationic surfactants with different chain lengths and, hence, different main phase- transition temperatures (T,: didodecyldimethylammonium bromide (DDAB, T', = 17"C) and dioctadecyldimethylam- monium bromide (DOAB, T', = 35°C).35 Data obtained in these systems were compared with kinetic results in vesicles of zwitterionic phospholipids (DPPC and EL).Experimental 1-Naphthol was purified by vacuum sublimation (ca. 10 Pa for 5 days). Agreement of the absorption and fluorescence spectra with literature data and single-exponential fluores- cence decay in several solvents served as criteria of purity. DPPC (99%)was purchased from Sigma. DDAB and DOAB both from Eastman-Kodak, were kindly provided by Prof. F. Menger (Atlanta University, Georgia, USA). All surfactants were proved to be highly free from fluorescent impurities and were used as received. Highly purified deionized water (Millipore MilliQ) was used to prepare solutions for fluores- cence measurements. The overall concentration of the naph- thol in solutions did not exceed mol 1-' in any of the experiments. To prepare vesicles we use the injection method proposed and tested for zwitterionic phospholipid^^^*^' and cationic surf act ant^:^^ an ethanol solution (80 mmol 1-', 0.15 ml) of a surfactant was injected into an aqueous solution (2.85 ml) of 1-naphthol.Aqueous solutions were heated to ca. 50°C before ethanol was injected into the solutions. The pH was controlled by an ionometer with a glass electrode and was ca. 6 in all cases. It was shown previ~usly~~-~~that this method gives monodisperse stable unilamellar vesicles. The size of vesicles is strongly dependent on the initial concentration of the sur- factant in the alcohol, e.g. the diameter of the DPPC vesicles varies from 30 to 120 nm for the concentration range 3-40 mmol 1-'.36 Although we did not specially determine the diameter of the DPPC vesicles, we used a higher concentra- tion of surfactants (ca. 80 mmol 1-') in an attempt to achieve the complete solubilization, thus the size of our DPCC vesi- cles is similar to the largest value reported.36 The diameter of EL vesicles obtained by this method in our previous work' was ca. 150 nm. The diameter of the vesicles of DDAB was shown36 to vary from 220 to 320 nm for an initial concentration of surfactant in ethanol in the range 0.29-0.87 mol 1-'. The diameter of the vesicles obtained in the present work was assumed to be <200 nm since we used a much lower concentration of DDAB (0.08 mol 1-'). The hydrodynamic diameters of the cationic vesicles were determined by quasielastic light scat- tering (QELS) using a 1096 Correlometer (Langley-Ford).? The autocorrelation function of fluctuations in the scattering intensity was analysed by the method of cumulants.Hydro- dynamic diameters were calculated from the directly deter- mined diffusion coefficients of the vesicles. Vesicular solutions were freed from dust by filtration through a 1.2 ,um Millipore filter. No significant loss of naphthol or change in vesicular size occurred after filtration. At 20°C the diffusion coeff- cients and the hydrodynamic diameters of the vesicles were 1.0 x lo-' cm2 s-' and 420 nm for DDAB and 1.9 x lo-' cm2 s-' and 220 nm for DOAB. Unexpectedly, our method of preparation of vesicles produces small unilamellar DOAB vesicles and large DDAB vesicles (we have no direct evidence that they are unilamellar). Ultrasonication of opalescent DDAB solutions using a Branson 1200 sonicator for 30 min at 50°C did not change the opalescence of the solution and fluorescence spectra of 1-naphthol in this solution.To evaluate the possible effects of the size of the vesicles on the temperature dependence of the ESPT rate constants we measured the temperature dependence of the hydrodynamic diameter of the vesicles. The diameter of DPPC vesicles is Inknown not to be very sensitive to temperat~re.~~ our experiments the diameter of DOAB vesicles decreased by only 20-25 nm when the temperature was increased from 20 to 55"C, but for DDAB vesicles a significant decrease of the diameter occurred from 420 to 190 nm when the temperature was increased from 20 to 55°C. This effect was accompanied by a small long-wavelength shift (8 nm) of the ArO- fluores- cence maximum and the disappearance of opalescence.Absorption spectra were recorded on a Shimadzu MPS-2000 spectrophotometer. Fluorescence and excitation spectra were measured with a Shimadzu RF-5000 spectrofluorimeter (both slits being 1.5 nm). Fluorescence decay curves were recorded with a time-correlated single-photon-counting instrument (Edinburgh Instruments 199). A synchronously pumped, cavity-dumped dye laser (Spectra-Physics Model 375; an Nd :YAG laser as a pump source) frequency-doubled with a KDP crystal (model 390; A= 310 nm) was used as an excitation source (pulse width <100 ps).In most experiments the typical time resolution was 24.2 ps per channel. An apparatus function was registered at the excitation wave-length. Decay curves were analysed by the non-linear least- squares method using the Edinburgh program for the IBM PC. The accuracy of fits was estimated by the x2 parameter and by inspecting the weighted residuals. Decay curves were deconvoluted as multiexponential functions with a variable time shift between an exciting pulse and fluorescence curve. Results and Discussion The fluorescence spectra of 1-naphthol in water and vesicle suspensions of various surfactants at the room temperature, 7 The authors are grateful to A.A. Yaroslavov and M. F. Zanso-hova (Dept. of Polymer Sciences, Moscow University) for these mea- surements. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 at the same surfactant concentration (4 mmol 1-and pH x 6 are shown in Fig. 1. An excitation wavelength of 315 nm was used. Fluorescence kinetics in ArOH and ArO -emission bands in DOAB, DDAB and DPPC vesicular solutions at the same conditions are shown in Fig. 2. 1-Naphthol in the S, state is well known4' to be a very strong acid (pK,* = 0.41) and to dissociate in aqueous solu- tion with a very high rate con~tant,4~*~~ k, = (2-3) x loio s-'. Therefore the fluorescence obtained from ArOH excita- tion in water consists almost entirely of the broad emission band of *ArO- formed due to excited-state protolytic disso- ciation.In contrast to water, a distinct short-wavelength band is present in the fluorescence spectra of 1-naphthol in vesicles. In addition, a blue shift of the naphtholate anion emission band is observed in these aggregates. No noticeable variation in fluorescence excitation spectra was observed for different emission wavelengths (360, 440 and 480 nm) in the systems investigated. The excitation spectra (Fig. 1) confirm that naphthol in the ground state, exists only in the proto- nated form and *ArO- formation is a result of ArOH photo- dissociation. At room temperature the ArO- to ArOH fluorescence intensity ratio (Z'/Z) was found to increase in a series of sur- factants: DPPC < DDAB < DOAB.This change in Z'/Z indi-cates an increase in the proton-transfer rate constant in this series. To discover the effect of the nature of the bilayer on excited-state proton-transfer reactions we studied the fluores- cence kinetics of *ArOH and *ArO- in suspensions of these vesicles at various temperatures. Note that because of the very fast (< 0.05 ns40-42) photodissociation of ArOH in the aqueous phase, the fraction existing in the volume phase will not contribute to the ArOH fluorescence kinetics observed in the solution of vesicles. It was previously shown' that the fluorescence spectra of 1-naphthol in a suspension of DPPC vesicles depend on the DPPC concentration because of the incomplete binding of ArOH to bilayer membranes at the low lipid concentrations. With increasing DPPC concentration Z'/Z decreased along with ArO-fluorescence, the maximum being shifted to shorter wavelengths.This indicates more complete solu- bilization of naphthol and retardation of the ESPT efficiency. The fluorescence kinetics of *ArO- at [DPPC] < 3 mmol 1-' was described by a triexponential function with two decay times, the fast decay time being equal to the anion life- time in aqueous solutions and the slow decay time being close to the anion lifetime in non-ionic micelles. In this work in cationic vesicles of DDAB and DOAB we also found that /'/I decreased with increasing surfactant con- h 4-.-C 3 300 400 400 500 L/nm Fig. 1 Fluorescence excitation (A) and emission (B) spectra of 1-naphthol in water (a)and vesicles of various surfactants: DPPC (b), DOAB (c) and DDAB (4; A,, = 480 nm for excitation spectra; Ae, = 3 15 nm for emission spectra; T = 20 "C J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 v)c v) C cC 8 8 channels channels channels !:,),I1 ,,,,(, I .....................I. , , , , , 0.00 100.0 200.0 300.0 400.0 500.0 Fig. 2 Fluoroescence decay curves of 1-naphthol(Aern = 370 nm, ArOH) and its anion (Aern = 480 nm, ArO-) at 20°C and pH z 6 in vesicle suspensions of various surfactants ([surfactant] = 4 mmol 1-l): (a)DDAB, (b)DOAB, (c)DPPC. 24.2 ps per channel. centration, but in contrast to zwitterionic vesicles no varia- tion in the ArO-fluorescence maximum was detected. Moreover, at all the surfactant concentrations studied, the fluorescence kinetics of *ArO- can be fitted by a biexponen- tial function with one rise time and one decay time (Table 1).These effects, together with a noticeable long-wavelength shift and a distinct change in the vibration structure in the excita- tion spectra measured in vesicles (Fig. 1) imply that most of 1-naphthol and its anion is located in bilayer membranes of vesicles (a less polar medium than water). Thus we suggest ki *ArO,-+ H+ -*ArOH, factant concentration can be explained mainly by changing the structure of the vesicles but not by the greater binding of naphthol. Previously6-’ we found that the fluorescence kinetics of naphthols in the membranes of DPPC and EL vesicles show the existence of two fractions of ArOH possessing different reactivities [probably localized in different parts of ‘heterogeneous’ (with respect to proton transfer) membranes] and discussed the kinetics of ESPT reactions in vesicles in terms of the extended pseudophase model: a vesicle suspen- sion with ArOH was considered to consist of a bulk water phase and two different microphases of the lipid membrane. Excited-state protolytic dissociation of these two fractions of ArOH solubilized in the membrane at neutral pH is described by the following scheme: ki 1 *ArOH,, -ArO,,- + H+ Ar0,-+ H+ -ArOH, IArOH,,-ArO,,-+ H’ Scheme 1 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 1 Parameters of the fluorescence decay curves of 1-naphthol and its anion in vesicles of various surfactants ArOH ArO -T/"C cfmmol I-' zl/ns z2/ns 7Jns AJCA, A2/XAi x2 7;fn.s z;/ns AJA, x2 zo/nsa DDAB, phase-transition temperature 17 "C 11 4.0 1.10 2.24 0.69 0.31 1.22 1.02 21.21 -0.86 1.59 5.25 15 4.0 1.11 2.2 1 0.74 0.26 1.86 -' ---5.10 24 4.0 0.60 1.25 0.52 0.48 1.05 0.75 15.56 -0.80 1.34 4.72 35 4.0 0.24 0.73 1.49 0.36 0.59 1.11 0.50 13.23 -0.75 1.30 4.26 45 4.0 0.20 0.53 3.10 0.74 0.26 1.03 0.26 11.89 -0.65 1.31 3.86 55 4.0 0.14 0.32 3.33 0.67 0.33 1.15 0.16 9.90 -1.17 1.24 3.59 25 2.5 0.62 1.30 0.67 0.33 1.29 0.60 14.34 -0.79 1.76 (0.76 18.31 -0.67 1.72)* 25 4.0 0.67 1.36 0.60 0.40 1.14 0.67 20.00 -0.87 1.92 (0.80 18.79 -0.84 1.64) 25 5.5 0.78 1.49 0.69 0.3 1 1.22 0.73 --0.7 1 1.43 (0.95 19.05 -0.77 1.59) DOAB, phase-transition temperature 35 "C 15 4.0 0.15 0.67 1.64 0.29 0.67 1.18 0.49 17.22 -0.69 1.22 5.10 25 4.0 0.14 0.52 1.88 0.32 0.67 0.98 0.43 15.77 -0.64 1.27 4.68 35 4.0 0.17 0.48 2.02 0.33 0.66 1.08 0.36 14.34 -0.70 1.17 4.26 ---4.0041 4.0 0.14 0.46 1.46 0.39 0.58 0.98 -55 4.0 0.2 1 0.61 2.87 0.69 0.28 1.10 0.33 12.1 1 -0.64 1.35 3.59 25 2.5 0.14 0.50 2.63 0.35 0.64 1.02 0.40 14.71 -0.62 1.33 25 5.5 0.19 0.67 2.65 0.30 0.6 1 1.17 0.55 17.70 -0.69 1.05 DPPC, phase-transition temperature 41 "C 22 4.0 1.26 4.78 0.37 0.63 1.51 0.51 11.26 -0.25 1.25 4.8 1 28 4.0 0.87 4.00 0.32 0.68 1.21 0.44 1 1.03 -0.30 1.40 4.55 31 4.0 1.07 4.32 0.36 0.64 1.14 0.48 11.60 -0.30 1.45 4.42 40 4.0 1.29 3.97 0.35 0.65 1.29 1.08 10.87 -0.50 1.32 4.03 55 4.0 1.00 3.18 0.43 0.57 1.17 0.84 11.34 -0.41 1.39 3.59 (I Lifetimes of 1-naphthol in pentanol.' Large values of x2 are due to modulation of decay curve by sine-shaped noise. All missing data are due to unsatisfactory results of fitting.Data in parentheses were obtained at lower time resolution (190 ps per channel) in order to record the longer-lived component more accurately. Here ArOH, and ArOH,, are two fractions of the ground- room temperature we found a biexponential decay. A third state naphthol molecules (solubilized in two different types of exponent with a very small amplitude appeared only at high site); k, and k,, are the excited-state protolytic dissociation temperatures (Table 1, Fig. 3). In DOAB vesicles a much rate constants for these two fractions; zp, zg and z;', zip are better fit was found for the triexponential approximation at the lifetimes of *ArOH and *ArO-, respectively, in the all temperatures, although the amplitude of the exponent absence of protolytic reactions in these sites. This scheme with the longest decay time (1.5-3 ns, depending on the assumes that exchange between these two types of site is too temperature) was <6% (Fig.4, Table 1). The amplitude of slow (compared with the rates of the excited-state reactions) this exponent increased markedly with the surfactant concen- to influence the kinetics of these reactions. At pH x 6 the tration (Table 1). A third exponent can probably be attrib- reverse reaction of *ArO- protonation can be neglected.For uted to the fluorescence of surfactant impurities. This each fraction of ArOH molecules standard expression^^^ for component was neglected in the discussions of excited-state the fluorescence quantum yields of ArOH (6)and ArO (6') proton transfer. should be valid: The fluorescence kinetics of the ArO- anions was fitted with reasonable accuracy by a biexponential function and can be described by a rise time (z;) and a decay time (zi). The and rise time of the ArO- fluorescence is close to the faster decay times of the ArOH fluorescence; however, in all cases the rise l/z = 1/ro+ k (2) time is slightly longer than the decay time. At the same time, the anion decay time is much longer than any of the *ArOH where +o and 4; are the fluorescence quantum yields of decay times.This provides evidence for the irreversibility of ArOH and ArO-, respectively, in the absence of the protoly- *ArOH dissociation at pH x 6. In principle, the fluorescence tic reaction; z is the observed lifetime of *ArOH. kinetics of *ArO- formed via the protolytic dissociation of According to Scheme 1 the fluorescence decay of ArOH (8') ArOH in both the microphases and in the bulk phase should is described by a biexponential function : be described by three rise times and three decay times. 2 However, we can observe only the fastest rise time in one of F(t) = 1 Aiexp(-t/zi) (3) the microphases (the rise time in the aqueous phase, as was i= 1 already mentioned, is too short to be recorded) and one where cli = Ai/C Ai are the relative concentrations of each decay time.The decay times of *ArO- in both microphases fraction, which are equal to the relative initial amplitudes of are probably similar. A substantially shorter decay time of each exponent of the ArOH fluorescence. We fitted the ArOH *ArO- in the aqueous phase (ca. 10 ns) can be hidden owing fluorescence decay kinetics as biexponential and tri-to the relatively small fraction of ArOH present in the esponential functions. Table 1 lists the parameters obtained volume phase (it can be observed as a second decaying com- for the ArOH and ArO- decay curves. For DDAB vesicles at ponent at low surfactant concentrations). Because of these J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 3 n channels Fig. 3 Fluorescence decay curves of 1-naphthol (ArOH) in DDAB vesicles at various temperaturesrc (top): (a) 11, (b) 15, (c) 24, (6)35, (e)45, (f)55. Results of fitting (weighted residuals and values of xz) are presented for each temperature (bottom). xz: (a) 1.18, (b) 1.86, (c) 1.05, (d)1.11, (e) 1.03, (f) 1.15; (a)-(c) biexponential fit, (d)-(f) tri-exponential fit. complications the fitted amplitude of the rising exponent was smaller than the amplitude of the decaying exponent. The excited-state protolytic dissociation rate constants for both fractions of ArOH, k, and k,,, were calculated from the decay times in the following way k, = l/z, -1/zp (4) k,, = l/z, -1/z; (5) where the decay times in the absence of protolytic reactions, zp and zg, for both fractions were assumed to be identical and equal to the lifetime of *ArOH in pentanol, which has a vis- cosity and relative permittivity similar to those in the interior of the membranes (this lifetime is known to be slightly sensi- tive to the nature of the solvent, if not protolytic photoreac- tion occurs, and sensitive to the temperature).Table 2 lists the rate constants obtained for ESPT in DPPC, DDAB and DOAB vesicles at various temperatures. All these calculations of ki assume that radiationless decay processes for all fractions of ArOH* in the bilayer vesicles are the same as those in a viscous solvent such as pentanol. To I - I I 1 0 100 200 300 400 channels 3 0 -4 c I 0 100 200 300 400 channels 13L3 .I 1 0 -2 Fig. 4 Fluorescence decay curve of 1-naphthol (ArOH) in DOAB vesicles at 25°C fitted by biexponential (top, x2 = 1.25) and tri- exponential (bottom, x2 = 0.98) functions. Weighted residuals are also shown. check this assumption and to exclude the possibilities of other sources of the acceleration of decay of ArOH* in the microphase we compared (Fig. 5) the temperature depen- dences of the relative fluorescence quantum yields of ArOH and ArO- obtained from the fluorescence spectra and those calculated from kinetic data obtained by single-photon counting : For each surfactant we used as To the lowest temperature of the experiment: 11 “C for DDAB and 15 “C for DOAB.The coincidence between $J(T)/+(To) measured from the fluorescence spectra and calculated from the kinetic data (zi and a,) at various temperatures reflects only a quantitative consistence between spectral and kinetic data, but the coin- cidence between the measured and calculated values of g5’(T)/g5’(To)confirms that the decrease in zi at higher tem- peratures is really caused by an increase in the proton- transfer rate constants k’; rather than by an increase in any 2722 Table 2 Rate constants of ESPT of 1-naphthol in vesicles of various surfact ants' T/T kJ1OP8 s-l k,,/lO-Bs-l k'/10 -8s-1 DDAB (4 mmol l-'), phase-transition temperature 17"C 11 7.2 2.6 8.0 C15 7.0 2.6 -24 14.6 5.9 11.2 35 39.2 11.4 17.6 45 47.4 16.3 35.9 55 68.7 28.5 59.7 DOAB (4 mmol I-'), phase-transition temperature 35 "C 15 64.7 12.9 18.5 25 69.4 17.1 -35 56.2 8.5 25.5 41 68.7 19.3 -55 44.7 13.6 27.6 DPPC (4 mmol 1-I), phase-transition temperature 41 "C 22 5.9 <0.1 17.5 28 9.3 0.3 20.5 31 7.1 <0.1 18.5 40 5.3 <0.1 6.7 55 7.2 0.4 9.1 ~~ (I All values of rate constants were calculated by using eqn.(4) and (5). k' values were calculated using the rise time of ArO- in In. (5). 1.20 I B 1 1.00 [ I s kY E .. 0.60' s v 0.40 E a 0.20 E 0.000.00 '10.0 20.00 30.00 40.00 50.00 60.00 TI"C Fig. 5 Temperature dependences of relative fluorescence quantum yields of ArOH[(a), (b)] and ArO- [(c), (43 obtained from fluores- cence spectra [(a), (c)] and calculated by using eqn.(6) and (7) from time-resolved measurements [(b),(41in DDAB (A) and DOAB (B). Dashed lines show the phase-transition temperatures (1 7 "C for DDAB and 35 "C for DOAB). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 radiationless decay processes [induced internal conversion or intersystem crossing, quenching by counterion of surfactant (Br-), or by impurities etc.]. The temperature dependence of the measured and calculated +'( T)/+(T) presents an apparent activation enthalpy averaged over all fractions of ArOH. A difference between cationic vesicles and zwitterionic lipid vesicles is evident from the comparison of the data listed in Table 2 and in Fig. 6-8. This is especially true for DDAB and DPPC since the data for DOAB are less precise owing to the very short lifetime of the first fraction of *ArOH.The first observation is that the protolytic dissociation rate constants are much higher in bilayer membranes of cationic vesicles than those in phosphatidylcholine and EL mem-branes. This effect is more pronounced for the fraction of naphthol molecules characterized by the lower proton-transfer rate constant kI,. This is why both ArOH fractions dissociate in DDAB and DOAB bilayers with comparable rate constants even at room temperature. The higher values of the protolytic dissociation rate constant in the cationic vesicles than those in zwitterionic vesicles are in good agree- ment with the data in micelles: in cationic micelles the disso- ciation rate constants are an order of magnitude greater and pK,* values are one unit lower than in non-ionic mi~elles~~ 23.00 A 22.00121 .oo -20.00 1 .oo B 0.80 0.60 E 0.40 I0.20 3.00 3.10 3.20 3.30 3.40 3.50 3.60 103 KIT A, Arrhenius plots of the protolytic photodissociation rate constants of 1-naphthol in DDAB vesicles calculated by using eqn.(4) and (5) from the decay times of *ArOH [(a) and (b)] and from the rise time of *ArO- (c)and of the relative fluorescence quantum yields (4, [In (#'/#), # and #' are in relative units, the ordinate scale is shifted arbitrarily]. B, Dependence of the relative contribution of each ArOH fraction us. inverse of the temperature: (a) APA,, (b) A,/ZA,. The dashed line shows phase transition temperature (17"C).J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 23.0 I 22.01 1 .ooE R I U I I I Io.*oj I I 0.60; 0.40: 0.20 0.00 3.00 3.10 3.20 3.30 3.40 3.50 103 KIT Fig. 7 A, Arrhenius plots of the protolytic photodissociation rate constants of 1-naphthol in DOAB vesicles calculated by using eqn. (4) and (5)from decay times of *ArOH [(a)and (b)] and from the rise time of *ArO- (c) and of relative fluorescence quantum yields (d) [In(+'/+), + and 9' are in relative units, the ordinate scale is shifted arbitrarily]. B, Dependence of the relative contribution of each ArOH fraction us. the inverse of the temperature: (a) A,/ZA,, (b) A,/ZA,. The dashed line shows the phase-transition temperature (35 "C).because of the electrostatic interaction contribution to the Gibbs energy of the reaction. In DDAB and DOAB vesicles no essential change of the rate constants of excited-state protolytic dissociation was found at the main phase-transition temperature in contrast to DPPC vesicles (Fig. 6-8). The amount of the fraction of naphthol molecules that have a lower protolytic dissociation rate constant in cationic vesicles in the liquid-crystalline state (at temperatures higher than the phase transition) decreases rapidly with temperature in contrast to the situation with zwitterionic lipid vesicles, where it is almost constant. Some increase in the fraction with lower rate constants is observed at a phase-transition temperature only for DDAB vesicles (Fig.6). Moreover, a substantial difference was found between DDAB and DOAB vesicles. A very strong increase of both the rate constants with increasing temperature was observed in the liquid-crystalline state of DDAB vesicles, which indi- cates a relatively high activation enthalpy of the proton transfer. Below the phase-transition temperature of DDAB vesicles (17"C) the rate constants were measured only for two temperatures (1 1 and 15 "C) and they are close to each other. 23.00 A C- 21.00 20*oolc 19.00~~""""~""""'~""""3.00 3.10 3.20 !"""3.30 3.40 lo3 KIT 1.oo t , DIDI 0.80 0.201, aI,, ,,,,,,I, ,,,,,( ,';,;,, , ,,,()(,,, 0.003.00 3.10 3.20 3.30 3.40 103 K/T Fig.8 A, Arrhenius plots of the protolytic photodissociation rate constants of 1-naphthol in DPPC vesicles calculated by using eqn. (4) and (5) from the decay time of *ArOH (a) and from the rise time of *ArO-(b). B, The dependence of the relative contribution of each ArOH fraction us. the inverse of the temperature: (a) A,/ZA,, (b) A,/XA,. The dashed line shows the phase-transition temperature (41"C). Fig. 6 shows the temperature dependence of the excited- state protolytic dissociation rate constants in an Arrhenius plot. In the liquid-crystalline state of DDAB vesicles a linear dependence with activation enthalpies of AH' 39 and 40 kJ mol-' was observed for the faster and slower dissociating fractions, respectively. These values are much higher than those in the zwitterionic vesicles of DPPC and EL, where the activation enthalpies for k, were 9 and 21 kJ mol-' for the gel and liquid-crystalline states of the DPPC bilayer, respec- tively.' In homogeneous solutions and in micelles the activa- tion enthalpies of protolytic dissociation are of the same order of magnitude, 10-25 kJ mol-l.'The abnormally high activation enthalpy for DDAB vesicles can be attributed par- tially to the decrease in the size of the vesicles with increasing temperature, as was mentioned in the Experimental section. The temperature dependence of 4'/4 (Fig. 6) yields an apparent activation enthalpy of AH* = 17 kJ mol-'. This is smaller than the activation enthalpies obtained from the kinetic data because of the substantial decrease of the frac- tion of slowly dissociating ArOH molecules with increasing temperature.In DOAB vesicles some modest irregular variations of both k, and k,, with temperature are observed at tem- 2724 peratures below and above the phase-transition temperature. As was mentioned earlier, the deconvolution of the kinetic curves is much less accurate for DOAB vesicles than for DDAB and DPPC vesicles due to the substantially faster decay. The evaluated activation enthalpy is <10 kJ mol-'. The ratio of the total fluorescence quantum yields 4'/4 (Fig. 7) is almost constant over the whole temperature range from 15 to 55"C, which confirms the low value of the activation enthalpy of the protolytic dissociation in DOAB vesicles.All these features characterize differences in the properties of membranes of the vesicles of cationic and zwitterionic sur- factants and demonstrate the possibility of using ESPT lumi- nescent probes for investigations of the structure and properties of membranes. Nevertheless, for a more com-prehensive discussion of the solubilization effects in vesicles on the kinetics of the protolytic reactions, more detailed information on the localization of the probes in the bilayer is necessary. Conclusion Two types of site for 1-naphthol molecules which differ strongly in their ESPT rate constants were found to exist in the membranes of cationic vesicles of DDAB and DOAB. This situation is similar to that seen for vesicles of zwitter- ionic lipids (EL and DPPC). The ESPT rate constants for both sites are much higher in bilayer membranes of cationic surfactants than those in zwitterionic surfactants because of the electrostatic interaction contribution to the Gibbs energy of the reaction. The fraction of the slower-dissociating 1-naphthol molecules in cationic vesicles is significantly smaller than in EL and DPPC vesicles.The temperature dependence of ESPT rate constants for both fractions of ArOH and DDAB vesicles follows the Arr- henius equation above the phase-transition temperature. The activation enthalpy of ESPT for both fractions of ArOH in membranes of DDAB vesicles is significantly greater than in membranes of zwitterionic lipids. In DOAB vesicles these activation enthalpies are, however, much smaller.This can be attributed partially to the effect of the temperature on the size of DDAB vesicles. The change in the ESPT rate constants at the membrane phase-transition temperature of vesicles of cationic sur-factants is much smaller than that for zwitterionic sur-factants. References 1 V. P. Skulachev, in Chemiosmotic Proton Circuits in Biological Membranes, ed. P. C. Hinkle, Addison-Wesley, London, 198 1. 2 M. Gutman, Meth. Biochem. Anal., 1984,30, 1. 3 M. Gutman and E. Nachliel, Biochim. Biophys. Acta, 1990, 1015, 391. 4 M. Gutman, A. Kotlyar, N. Borovok and E. Nachliel, Biochem-istry, 1993, 32, 2942. 5 F. Nome, W. Reed, M.Politi, P. Tundo and J. H. Fendler, J. Am.Chem. SOC., 1984,106,8086. 6 Yu. V. Il'ichev, A. B. Demyashkevich and M. G. Kuzmin, High Energy Chemistry, 1989,23,435. 7 Yu. V. Il'ichev, A. B. Demyashkevich and M. G. Kuzmin, J. Phys. Chem., 1991,%, 3438. 8 Yu. V. Il'ichev, A. B. Demyashkevich, M. G. Kuzmin and H. Lemmetyinen, J. Photochem. Photobiol. A: Chem., 1993,74, 51. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 9 R. Yam, E. Nachliel, S. Kiryati, M. Gutman and D. Huppert, Biophys. J., 1991,59,4. 10 A. Jankowski, P. Stefanowicz and P. Dobryszycki, J. Photochem. Photobiol. A: Chem., 1992,69, 57. 11 M. J. Politi and J. H. Fendler, J. Am. Chem. SOC., 1984,106,265. 12 M. J. Politi, 0. Brandt and J. H. Fendler, J. Phys. Chem., 1985, 89, 2345. 13 H. Kondo, I. Miwa and J. Sunamoto, J.Phys. Chem., 1982, 86, 4826. 14 E. Bardez, B.-T. Goguillon, E. Keh and B. Valeur, J. Phys. Chem., 1984,88,1909. 15 E. Bardez, E. Monnier and B. Valeur, J. Phys. Chem., 1985, 89, 503 1. 16 M. J. Politi and H. Chaimovich, J. Phys. Chem., 1986,90,282. 17 U. K. A. Klein and M. Hauser, 2. Phys. Chem. (Frankfurt am Main), 1975,%, 135. 18 B. K. Selinger and A. Weller, Aust. J. Chem., 1977,30,2377. 19 C. Harris and B. K. Selinger, 2. Phys. Chem. (Miinchen), 1983, 134,65. 20 U. Khuanga, R.McDonald and B. K. Selinger, 2. Phys. Chem. (Frankfurt am Main), 1976, 101,209. 21 A. K. Zaitsev, Yu. V. Il'ichev, 0.Ph. Gorelik, N. K. Zaitsev and M. G. Kuzmin, Khim. Fiz., 1985,4, 1384. 22 S. Abou-A1 Einin, A. K. Zaitsev, N. K. Zaitsev and M. G. Kuzmin, Khim.Fiz., 1986,5, 219. 23 S. Abou-A1 Einin, A. K. Zaitsev, N. K. Zaitsev and M.G. Kuzmin, J. Photochem. Photobiol. A: Chem., 1988,41,365. 24 M. G. Kuzmin, N. K. Zaitsev, A. K. Zaitsev and S. Abou-A1 Einin, Khim. Fiz., 1988,7,492. 25 A. K. Zaitsev, N. K. Zaitsev and M. G. Kuzmin, Moscow Uni- versity Bull., Ser. 2, Chem., 1987,28, 144. 26 M.Bayer, A. K. Zaitsev and M. G. Kuzmin, Khim. Fiz., 1988,7, 172. 27 Yu. V. Il'ichev, A. B. Demyashkevich, M. G. Kuzmin, M. I. Sluch and A. G. Vitukhnovsky, Khim. Fiz., 1992,11, 351. 28 A. K. Zaitsev, Yu. V. Il'ichev, N. K. Zaitsev and M. G. Kuzmin, Dokl. Akad. Nauk SSSR, 1985,283,900 29 Yu. V. Il'ichev, A. B. Demyashkevich and M. G. Kuzmin, Khim. Vys. Energ., 1990, 24, 52. 30 Yu. V. Il'ichev, A.K. Zaitsev and M.G. Kuzmin, Khim. Yys. Energ., 1990,24,141. 31 K. M. Solntsev, Yu. V. Il'ichev, A. B. Demyashkevich and M. G. Kuzmin, J. Photochem. Photobiol. A: Chem., 1994,78,39 32 M. G. Kuzmin and N. K. Zaitsev, in The Interface Structure and Electrochemical Processes at the Boundary between Two Immisci- ble Liquids, ed. V. E. Kazarinov, Springer-Verlag, Berlin, Heidel- berg, 1987, pp. 207-244. 33 N. Chattopathyay, R. Dutta and M. Chowdhury, J. Photochem. Photobiol. A: Chem., 1989,47,249. 34 Yu. V. Il'ichev and V. L. Shapovalov, Izu. Akad. Nauk. Ser. Khim., 1992, 2253. 35 B. C. R. Guillaume, D. Yogev and J. H. Fendler, J. Phys. Chem., 199 1,95,7489. 36 S. Batzri and E. D. Korn, Biochim. Biophys. Acta, 1973, 298, 1015. 37 J. M. H. Kremer, M. W. J. v.d. Esker, C. Pathmamanoharan and P. H. Wiersema, Biochemistry, 1977, 16, 3932. 38 L. A. M. Rupert, D. Hoekstra and J. B. F. Engberts, J. Am. Chem. SOC.,1985,107,2628. 39 A. A. Yaroslavov, A. A. Efimova, V. E. Kul'kov and V. A. Kabanov, Polym. Sci., 1994, 36,215. 40 J. Lee, G. W. Robinson, S. P. Webb, L. A. Philips and J. H. Clark, J. Am. Chem. SOC.,1986,198,6538. 41 S. P. Webb, S. H. Yeh, L. A. Philips, M.A. Tolbert and J. H. Clark, J. Am. Chem. SOC., 1984,106,7286. 42 H. Shizuka, T. Ogiwara, A. Narita, M. Sumitani and K. Yoshi- hara, J. Phys. Chem., 1986,90,6708. Paper 4/027631; Received 10th May, 1994
ISSN:0956-5000
DOI:10.1039/FT9949002717
出版商:RSC
年代:1994
数据来源: RSC
|
23. |
Aggregation, hydrogen bonding and thermodynamic studies on Boc-Val-Val-Ile-OMe tripeptide micelles in chloroform |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2725-2730
R. Jayakumar,
Preview
|
PDF (745KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2725-2730 Aggregation, Hydrogen Bonding and Thermodynamic Studies on Boc-Val-Val-He-OMe Tripeptide Micelles in Chloroform R. Jayakumar, R. G. Jeevan and A. B. Mandal" Chemical Laboratory, Physical and Inorganic Chemistry Division, Central Leather Research Institute, Adyar, Madras 600020,India P. T. Manoharan" Department of Chemistry, Indian Institute of Technology, Madras 600036,India Evidence for micelle formation of Boc-Val-Val-lle-OMe (Boc = tert-butyloxycarbonyl) tripeptide (l),in chloroform has been obtained from IR and Raman scatter fluorescence spectroscopies. The critical micelle concentrations (c.m.c.s) of this peptide, obtained by these techniques, correlate well. It has been found that the micelle forma- tion of the peptide in chloroform is hindered by increasing temperature. The aggregation numbers of the peptide have also been determined to be almost independent of temperature.The AmGe, AmHe, AmSe and AmCp values have been estimated. Results from the above thermodynamic parameters indicate that the driving force for micellization of the tripeptide 1 in chloroform is entirely enthalpic in nature and the aggregates of the peptide in chloroform are ordered. The IR spectra of the peptide in the pre- and post-micellar regions were analysed; there is no change in the intensity of the intermolecular hydrogen-bonding pattern for the peptide in the mono- meric and micellar states. However, the intensity of the solvent-exposed -N-H stretching band increased as a function of peptide concentration after attaining c.m.c.The design of molecular subunits that can self-assemble'*2 into defined structures in dispersion or in the solid state is one of the rapidly growing areas of chemistry and biology. The molecular recognition of the peptides is particularly important during self-assembly processes which result in micelles, vesicles and liquid crystal^.^ Few pep tide^,^.^ most phospholipids and some biologically relevant molecules are known to exhibit such behaviour.6 These surface-active mol- ecules exhibit a wide variety of biological activity.' As membrane-bound receptors, proteins play several important roles in mediating their functions ; the aggregation of membrane-active peptides in apolar media is valuable in the modelling of some interactions.* Ordered aggregates of the peptides in apolar media are also useful for obtaining infor- mation about the physicochemical nature of the interactions operating during self-assembly proces~es.~ Therefore, this study aims to identify interactions that lead to the molecular assembly of peptide moieties.Peptide self-assembly is very similar to the modular assembly involved in protein folding in terms of compactness, the core of non-polar alkyl side- chains and internal architecture. This model explores the contribution to Gibbs energy arising from hydrogen-bonding, van der Waal's interactions, intrinsic conformational propen- sities and solvophobic interactions." The size and number of molecules in the aggregate may be controlled by manipulating the type and orientation of the non-covalent interactions between the monomers.' ' The strong and directional nature of hydrogen bonds between -NH-CO-groups in the peptides contributes to their widespread involvement in self-assembling systems.,,'' Especially in an apolar medium, the solvophobic nature of amide groups is an advantage because of their low solubility in apolar 1iq~ids.l~ This has led to a search for peptides which form persistent packing motifs and defined aggregation numbers. The micelle formation of ampipathic molecules in apolar solvents has been debated in the past because the plot of optical properties vs. ampipathic concentration is not sharp enough to get a clear break point, and as a result this gives variable c.m.c.values. Recently' we have demonstrated that the Boc-Val-Val-Ile-OMe, 1, tripeptide, found in parallel p-sheets of triosephosphate isomerase,' forms micelles in an apolar medium like chloroform. Evidence for this was obtained by using UV-VIS, fluorescence and NMR spectroscopy' where clear break points were observed. Con- formational analysi~'~ of the tripeptide 1 in chloroform has also been carried out using the nuclear Overhauser effect (NOE). These NMR results15 suggested an extended struc- ture for the tripeptide 1 in chloroform. Once micelle formation and the conformation of the tri- peptide 1 in chloroform have been established,15 it is impor- tant to determine its aggregation number.*-14 We have determined aggregation numbers for Boc-Lys(Z)-Tyr-NH-NH, dipeptide and TFA -Tyr-Gly-Phe-Ala-OBz (TFA = trifluoroacetic acid) tetrapeptide micelles4 in aqueous solution and found them to be extremely low.However, the aggregation number of the tripeptide 1 in chloroform is sur- prisingly high, and its proper characterisation has tempted us to study this system further. Re~ently,'~ we have reported aggregation, hydrogen bonding and thermodynamic studies of aqueous tetrapeptide micelles in order to understand the secondary and tertiary structure of the peptides in the light of micelle formation. In this paper, we extend this study to the synthesis and characterisation of the tripeptide 1 aggregate in chloroform and the micellisation of 1.In this paper, we have employed two more independent techniques uiz., Raman scatter fluorescence and FTIR spectroscopy to substantiate our earlier c.m.c. results.' Experimental Materials and Methods Synthesis and Purification of Boc-Val-Val-Ile-OMe (1) Tripeptide All the amino acid derivatives were synthesised by standard procedures. Tripeptide 1 was prepared by a conventional solution phase method. All the intermediates were checked for purity by TLC on silica gel (solvent A: 2% MeOH in CHCl,; solvent B: 3% MeOH in CHCI,) and characterised by 90 MHz and 400 MHz 'H NMR. Specific procedures are described below. Boc-Vul-Zle-OMe. 4.34 g (20 mmol) of Boc-Val was dis- solved in CHCl, (20 ml) and cooled to -5 "C.Ile-OMe -HC1 was added, followed by 2.8 ml of triethylamine (TEA) (2.8 ml). 4.50 g (22 mmol) of N,N'-dicyclohexylcarbodiimide (DCC) was added in fractions for a period of 0.5 h. The reac- tion mixture was stirred at -5 "C for 3 h. After further stir- ring at room temperature overnight, the N,N'-dicyclohexylurea (DCU) was filtered and the filtrate was evaporated in vacuum. The residue was dissolved in CH,CO,Et (30 ml) and washed with 0.5 mol 1-' H,SO, (3 x 20 ml), 0.5 rnol 1-' Na,CO, (3 x 20 ml) and water (2 x 10 ml). Evaporation of CH,CO,Et yielded a solid mass which was dissolved in a minimum amount of CH,CN. The undissolved DCU was filtered and the filtrate was evaporated in vacuum, which yielded a solid mass, homogeneous on TLC (solvent A).Yield: 5 g, 72%. TFA . Val-lle-OMe. 3.44 g (10 mmol) of the above dipep- tide was treated with trifluoroacetic acid (TFA) (30 ml). The removal of the Boc group was monitored by TLC. After 1 h, TFA was evaporated with the help of a water aspirator. The residual TFA was removed in high vacuum. The residue was treated with anhydrous MeOH (3 ml) and evaporated in vacuum. This process was repeated twice to remove traces of TFA. The residue was suspended in water (30 ml), extracted with ether (20 ml) and the aqueous layer made alkaline with Na,CO,. Extraction with CHCI, , followed by drying of the organic layer with Na2S0, and evaporation yielded H-Val- Ile-OMe as a sticky solid. Yield: 2.08 g, 85%. Boc-Val-Val-Ile-OMe.2.08 g (8.5 mmol) of the above free base was dissolved in dimethylformamide (DMF) (10 ml) and cooled in an ice bath. Boc-Val (1.84 g) and 1-hydroxybenzotriazole (HOBT) (1.19 g) were added. To this, 1.85 g (9 mmol) of DCC was added in small portions. The above mixture was stirred at 0°C for 6 h and then at room temperature for 12 h. The precipitated DCU was filtered off and the filtrate was evaporated in vacuum. The residue was dissolved in CH,CO,Et (100 ml). The CH,CO,Et solution was successively washed with 0.5 mol 1-' H,S04 (3 x 30 ml), 0.5 mol 1-' Na,CO, (3 x 30 ml) and water (2 x 30 ml). The organic layer was dried with anhydrous Na,SO, and evapo- rated under vacuum to yield the tripeptide 1as a white solid. The tripeptide 1was purified with solvent system B.Yield: 2.67 g, 80%. Tripeptide 1was fully characterised by 400 MHz 'H NMR (see Fig. 1). Spectroscopic Studies NMR spectra of the peptide were recorded on an JEOL 400 MHz spectrometer. Two-dimensional correlated spectra (COSY) were recorded using a Bruker WH-270 FT-NMR spectrometer. FTIR measurements were performed with a Nicolet 20 DXB spectrometer. The band positions are accu- rate to 0.1 cm-'. Spectral grade chloroform was used for the sample preparation. Raman scatter of chloroform in the absence and presence of the peptide was recorded on a Hitachi Model No. 650-40 fluorimeter using a band width of 5 nm on both excitation and emission monochromators. The c.m.c. of the peptide was determined by plotting the intensity of the Raman scatter of chloroform and the intensity of free NH us.peptide concentration during fluorescence and FTIR spectroscopic measurements, respectively. The c.m.c. determi- nation was also carried out using tert-butylphenol as an external probe. The concentration of the probe was kept low (2 xlo-' mol I-') during the fluorescence spectroscopic J. CHEM. SOC. FARADAY TRANS., 1994,VOL. 90 measurements so that the ratio between the probe and micel- lar concentration is $1. The details for the c.m.c. determi- nation, have been described previously.'* Prior to the measurements, all the solutions were thermostatted for a con- siderable length of time. The temperature fluctuation was in the range of & 0.05-0.1 "C.Determination of the Aggregation Number The aggregation number of the tripeptide 1 in chloroform was determined by measuring the quenching of a micelle- bound fluorescent probe by the binding of a quencher using the following expression :19,20 In(I,/Z) = N[Q]/(C,-c.m.c.) (1) where I, and Z are the emitted light intensities with quencher concentrations of zero and [Q], respectively. fl is the mean aggregation number of the peptide and C, is the total concen- tration of the peptide. Semi-Mg salt of 8-anilino-1-naphthal- enesulfonic acid (ANS) and cetylpyridinium chloride (CPC) were used as a fluorescent probe and quencher, respectively. The utility of ANS as a suitable probe with CPC quencher has already been examined by also performing the above experiment with a pyrene probe.The aggregation numbers obtained by employing these two probes are in good agree- ment with each other. The utility of ANS as a probe and the validity of eqn. (1) have been recently dis~ussed?~"*~' All the experiments were performed in the presence of HPLC grade chloroform and there were no trace amounts of water present in the system. The concentration of the probe was kept SUE-ciently low to prevent exciplex formation. Results and Discussion NMR studies are carried out to characterise tripeptide 1.The complete assigned spectra of tripeptide 1are shown in Fig. 1. Specific assignments of backbone C"H and NH protons are * Im 0 mB I~"'~"~.I~"'~''~~,."......,......I..,I........,..6 5 4 3 2 1 6 Fig. 1 400MHz 'H NMR spectrum of 10mmol 1-'tripeptide 1 in CDC1, at 25°C. Two types of Boc-CH, signals (*) appear in the spectrum. The intensity pattern of these two signals alters in post- and pre-micellar states of the tripeptide. The complete analysis of this pattern will be discussed in the subsequent paper. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 0 1 2 3 6 4 5 6 7 76543210 6 Fig. 2 270 MHz 'H NMR COSY spectrum of 10 mmol 1-' tri-peptide 1 in CDCl, at 25 "C made using a sequential assignment principle (based on a combination of COSY and difference NOE experiments). Examples of COSY spectra for tripeptide 1 with connecti- vities are shown in Fig. 2. The NOE spectra for tripeptide 1 have been reported" recently.The high-field doublet of the NH protons as seen in Fig. 1 is assigned to terminal NH Val because the urethane NH proton chemical shifts always occur at ca. 5.5 ppm in CDCl, solution.22 The CaH, CBH and CYH resonances of Val-1 are obtained by tracing its connecti- vities through cross peaks. The other Val-2, Ile-3 NH, and other connectivities, are obtained by assigning the high-field CBH resonance for Ile-3 as Ile CBH will appear at a higher field than CBH of Val because of the extra -CH2- substi-tution in Iie.23 The c.m.c. values of peptide 1, obtained by UV-VIS, fluo- rescence, NMR, IR and Raman spectroscopies are in good agreement with each other. The Raman scatter fluorescence spectra of chloroform in the absence and presence of various concentrations of peptide are shown in Fig.3. The c.m.c. and aggregation number of the tripeptide 1 at various tem-peratures are depicted in Table 1. The aggregation number of the tripeptide is found to be quite high and almost indepen- dent of temperature (cf: Table 1).The fluorescence intensity of ANS in chloroform increases on interaction with tripeptide micelles suggesting, even in apolar media, that ANS binds to the peptide in the region (either inside or in the interfacial region of the micelle~~~) of lower polarity than the chloro- form alone. The fluorescence emission spectra of ANS in chloroform in the absence and presence of various concentra- tions of tripeptide are shown in Fig.4. Note that there is no r 390 415 465 490 515 540 A/n rn Fig. 3 Raman scatter fluorescence spectra of chloroform at various concentrations of tripeptide 1 at 20°C. Curves 0-6: 0, 0.5, 1, 1.5, 2, 3 and 4 mmol 1-' tripeptide 1, respectively; A,, = 365 nm, A,, = 438 nm. appreciable shift in the ANS emission in the presence of the tripeptide owing to'7324 the decrease in polarity as well as the unaltered microviscosity of the environment.' 7*24 The NH stretching region of the IR spectra of the tri- peptide 1 shows two absorption peaks (cf: Fig. 5): one at 3448 cm-' is attributed to the presence of a solvated N-H and the other at 3300 cm-' may be due to intermolecularly H- bonded N-H groups.25 The IR spectra show that intermo- lecular hydrogen bonding occurs even at pre-micellar levels 40 P 4 400 500 600 700 A/n m Fig.4 Fluorescence spectra of ANS in chloroform at various con- centrations of tripeptide 1 at 20°C; [ANSI = lo-' mol 1-' (fixed)A,, = 346 nm, A,, = 478 nm. Curves 0-4: 0,2, 2.5, 3 and 4 mmol 1-' in tripeptide, respectively. Table 1 C.m.c., aggregation number (m)and some thermodynamic parameters for the tripeptide 1 in chloroform solution at various tem- peratures T/"C ~.m.c./lO-~mol 1-' m A,Ge/kJ mol-' A,H*/kJ mol-' A,Se/J K-' mol-' ACJJ K-' mol-' 5 l.l"Vb 140 -15.7 -19.3 -12.9 15 1.5".b 138 -15.6 -20.7 -17.7 25 2.0a.b.c.d 137 -15.4 -22.1 -22.5 -123 2.1e 30 2.5"*b 136 -15.1 -22.9 -25.7 40 3.1"*b 135 -15.0 -24.4 -30.0 " From IR spectroscopy. From Raman scatter fluorescence.From fluorescence spectroscopy. From NMR spectroscopy. From UV-VIS spectroscopy. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 3500 3400 3300 wavenumber/cm-Amide A regions of IR spectra of tripeptide 1 in chloroform at various concentrationsat 20 "C. Curve 1-5: 1, 3, 4, 6 and 8 mmol 1-of tripeptide 1, respectively. of the tripeptide 1 in chloroform (Fig. 5). However, on micel-lization, there is a steep increase in the intensity of the solvat-ed NH peak suggesting that the onset of micellization accumulates solvated NHs in the peptide aggregate. This is quite surprising because the solvophobic -NH-CO-is supposed to hydrogen bond intermolecularly in a low-dielectric medium.26 The increase in fluorescence intensity of ANS in the presence of tripeptide 1, and the increase in the intensity of the solvated NH stretching band of the tripeptide, suggest that, for this peptide, the solvophobic group might be the isopropyl group of the peptide.Mean peptide aggregation numbers, A, are calculated from the slopes of the plot of ln(Zo/Z)us. [Q] (cf: Fig. 6). Neglecting activity effects and using a biphasic micellar the standard Gibbs energy change for micelle for-mation, A,Ge, of the peptide has been calculated from the following equation : A, Ge = RT In c.m.c. = A, He -TA, Se (2) The standard enthalpy change for micelle formation, A, He, is determined from the slope of the plot of In c.m.c. us. T (Fig.:::I 1.O h% 0.8 0.4 0.2 0 0 2 4 6 8 10 12 [O]/l 0-5 mol dm-3 Fig.6 Results of ANS quenching experiments: ln(Zo/I) us. concen-tration of N-cetylpyridinium chloride for micellar solutions of tri-peptide 1 at 25°C. [Tripeptide] = 12 x mol 1-' (fixed) and [ANSI = moll-' (fixed) in chloroform solutions. Aex = 346 nm; Aem = 478 nm. v E -6.4' v S --6.0' 273 283 293 303 313 T/K Fig. 7 In c.m.c. us. T 7) using the following equation:28 d In c.m.c.A,H~= -RT~ dT (3) To calculate all the thermodynamic parameters, the standard states are chosen as the hypothetical solutions at unit molar concentration. In recent p~blications~*'~we have considered aggregation number as one of the thermodynamic variable~~*'~.~~'in the calculation of the above thermodyna-mic parameters because the aggregation number of such pep-tides is sufficiently low.However, the aggregation number has not been taken into consideration as a thermodynamic variable in the present work because the tripeptide 1 in chlo-roform possesses quite high aggregation numbers which are almost independent of temperature. The A, Ge, A, H0 and AmSe values for the tripeptide 1at various temperatures are given in Table 1. The standard heat capacity change for the micellization, ACF, obtained from the slope of A,He us. T is also given in Table 1. It is assumed that a major factor driving the surfactant molecules into aggregation in water is a positive entropy change, presumably associated with the breakdown of the structured water which surrounds the hydrocarbon chain in the unassociated species.This interpre-tation is relevant to the formamide system in which some structuring by dissolved hydrocarbon also occurs. According to Evans et al.29athe above interpretation is erroneous or, at least, misleading because at high temperatures water loses most of its structural properties and the formation of struc-tured water in the walls of the hydrocarbon cavities is no longer possible. According to Evans et al.,29"it is sensible to attribute the micellization of the tripeptide in chloroform to a negative entropy change which is due to a transfer of the chloroform solvent into the peptide micelles. Table 1 shows that A,Se and AmHe values for the tripeptide are always negative in the temperature range of the investigation.The negative value of AmSOmay arise from the value of A, He, owing, to some extent, to the re-establishment of the hydro-gen bonds in the solvent. Therefore, the results indicate that the driving force for micellization of the peptide is entirely enthalpic in nature. This analysis has also been applied by Evans and Ninham29b to changes in protein conformation and to other biochemically important self-assembly processes. Comparing the results of thermodynamic experiments on model organic compounds, it is apparent that the heat capac-ity changes play a central role in characterising solvophobicinteraction^.^' The negative heat capacity change (see Table 1) is attributed to the disordering of solvent molecules around the exposed solvophilic groups.31 Therefore, the negative heat capacity change indicates disordering of chloro-form molecules around the tripeptide micelles.Convention-ally, solvophobic interactions provide a driving force for J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 micellization, with steric repulsion providing an opposing force.32 Therefore, our results suggest that the micelle forma- tion of the tripeptide l is hindered by the increase in tem- perature, as the c.m.c. value of the peptide increases with increasing temperature (cJ Table 1). In an apolar medium like chloroform, the main interaction in the peptide aggregation is believed to be intermolecular hydrogen-bond formation.33 However, a plot of a large number of literature values for AG* determined in chloroform vs.n (where n is -NH-CO-groups) lead to a straight line with a slope of -5 & 1 kJ mol- due to the stabilization of amide groups of the peptide on interaction with chloroform molecules. In the present case, the three amide groups interacting with chloro- form should result in a minimum of -15 3 kJ mol- ' of energy of stabilisation to the aggregate;34 the AmGe value obtained, -15.4 kJ mol-', is in agreement with this interpre- tation. Thus, chloroform associates with the amide moiety (of association constant, K = 1 dm3 mol-'). Therefore, the only interaction which makes A,,, Ge sufficiently negative is the enthalpic component which suggests the process is entirely enthalpic.If we attribute a negative AmS* to hydro- gen bonding with solvent molecules, then the process does not result in more ordered molecules around the exposed groups (which can interact with the chloroform molecules). Then such a process would also decrease ACF values,35 since release of chloroform solvation leads to a reduction in the number of heat-absorbing bonds. In the present case, the negative value of ACF implies that the chloroform is less ordered around the peptide micelles. Hence, the negative Am€€* may be attributed to the relatively well solvated peptide molecules in the micellar state. The experimental observations leading to these conclusions may be summarized by the following points. (i) The linewidths of the NH signals of the peptide do not change on micellization.' (ii) The ANS fluorescence emis- sions have no appreciable blue shift in the presence of the tripeptide indicating that there is not viscosity change in the interior of the micelles.(iii) The IR intensity of free NH increases upon micellization of the tripeptide in chloroform. (iv) The fluorescence emission intensity of the tert-butylphenol decreases upon micellization' of the tripeptide as the non-polar micellar interior decreases the fluorescence emission of the phenolic moiety.36 (v) The intensity of ANS fluorescence increases in the micellar environment, indicating a less polar environment of the micellar interior. (vi) The large aggregation number of the peptide and its constancy over the investigated temperature range are indicative of solvent-associated monomers in the micelle.(vii) The NOE observed is positive." If 140 molecules of the tripeptide 1 are associated, then the effective molecular weight of the aggre- gate will be ca. 62020. With this large aggregate molecular weight there will be an increase in correlation time which can result in negative NOE as oz, falls in a negative domain. The fact that we observe a positive NOE indicates that the motions of the monomeric peptides are not restricted and the rotational correlation times are short enough in the region oz, 1 at 270 MHz.~' Therefore, the tripeptide 1 micelles formed in chloroform consist of flexible solvated peptides of decreased enthalpy because of their interaction with chloro- form molecules.We are thankful to Dr. G. Thyagarajan, Director, CLRI, Madras for his keen interest in this work. Stimulating dis- cussions with Dr. T. Ramasami, Deputy Director and Head, Chemical Sciences Division are greatly appreciated. The support of the Research Council and the Regional Sophisti- cated Instrumentation Centre facilities, at the Indian Institute of Technology, Madras are gratefully acknowledged. We are grateful to the referees for their constructive suggestions and valuable comments. References 1 J. S. Lindsey, New J. Chem., 1991, 15, 153. 2 G. M. Whitesides, J. P. Mathias and C. T. Seto, Science, 1992, 254, 1312. 3 J. M. Lehn, Angew. Chem., Znt. Ed. Engl., 1990,29, 1304.4 (a) A. B. Mandal and R. Jayakumar, J. Chem. SOC., Chem. Commun., 1993, 237; (b) A. B. Mandal, A. Dhathathreyan, R. Jayakumar and T. Ramasami, J. Chem. SOC., Faraday Trans., 1993,89,3075. 5 (a) G. H. Beaven, W. B. Gratzer and H. G. Davies, Eur. J. Biochem., 1969, 11, 37; (b) W. B. Gratzer, E. Bailey and G. H. Beaven, Biochem. Biophys. Res. Commun., 1967,28,914; (c) D. J. Patel, Macromolecules, 1970, 3, 448; (d) L. C. Craig, J. D. Fisher and T. P. King, Biochemistry, 1965,4, 3 11. 6 S. Hoffmann, 2. Chem., 1987,27,395. 7 Chemistry and Biochemistry of Amino Acids, Peptides and Pro- teins, ed. B. Weinstein, Marcel Dekker, New York, 1974, vol. 3. 8 R. Kishore, S. Raghothama and P. Balaram, Biopolymers, 1987, 26, 873. 9 S. Bonazzi, M.Capabianco, M. M. De Morais, A. Garbesi, G. Gottarelli, P. Mariani, N. G. Ponziborsi, G. P. Spada and L. Tondelli, J. Am. Chem. SOC.,1991,113,5809. 10 K. A. Dill, Perspectives Biochem., 1991, 1. 11 J. D. Wright, in Molecular Crystals, Cambridge University Press, Cambridge, 1987. 12 A. J. Kisby, Adv. Phys. Org. Chem., 1980,17, 197. 13 A. J. Doig and D. H. Williams, J. Am. Chem. SOC., 1992, 114, 338. 14 H. Ihara, H. Hachisako, C. Hirayama and K. Yamada, J. Chem. SOC.,Chem. Commun., 1992, 1244. 15 R. Jayakumar, A. B. Mandal and P. T. Manoharan, J. Chem. SOC., Chem. Commun., 1993,853. 16 J. S. Richardson, Adv. Protein Chem., 1981,34, 168. 17 A. B. Mandal and R. Jayakumar, J. Chem. SOC.,Faraday Trans., 1994,90, 161. 18 (a)A. B. Mandal and B.U. Nair, J. Phys. Chem., 1991,95, 9008; (b)A. B. Mandal and B. U. Nair, J. Chem. SOC.,Faraday Trans., 1991,87, 133; (c) A. B. Mandal, B. U. Nair and D. Ramaswamy, Langmuir, 1988, 4, 736; (d) A. B. Mandal, B. U. Nair and D. Ramaswamy, Bull. Electrochem., 1988, 4, 565; (e) A. B. Mandal and S. P. Moulik, ACS Proc. in Solution Behavior of Surfactants-Theoretical and Applied Aspects, ed. K. L. Mittal and E. J. Fendler, Plenum Press, New York, 1982, vol. 1, pp. 521-541; (f) A. B. Mandal, S. Ray and S. P. Moulik, Indian J. Chem. A, 1980, 19, 620; (9) A. B. Mandal and B. U. Nair, in Advances in Measurement and Control of Colloidal Processes, ed. R. A. Williams and N. C. de Jaeger, Butterworth, London, 1991, ch. 2, pp. 136-149; (h) A. B. Mandal, D.V. Ramesh and S. C. Dhar, Eur. J. Biochem., 1987,169,617. 19 N. J. Turro and A. Yekta, J. Am. Chem. SOC., 1978,100,5951. 20 L. Luo, N. Boens, M. Vander Auweraer, F. C. de Schryver and A. Malliaris, J. Phys. Chem., 1989,93, 3244. 21 B. Geetha, A. B. Mandal and T. Ramasami, Macromolecules, 1993,26,4083. 22 (a) R. Kishore, S. Raghothama and P. Balaram, Biopolymers, 1987, 26, 873; (b) M. Iqbal and P. Balaram, J. Am. Chem. SOC., 1981, 103, 5548; (c) Y. V. Venkatachalapathi, B. V. V. Prasad and P. Balaram, Biochemistry, 1982, 21, 5502. 23 (a)A. Ravi, B. V. V. Prasad and P. Balaram, J. Am. Chem. SOC., 1983, 105, 105; (b) A. Ravi and P. Balaram, Tetrahedron, 1984, 40, 2577. 24 J. Slavik, Biochim. Biophys. Acta, 1982,694, 1. 25 (a) C.Toniolo, G. M. Bonola and S. Salardi, Znt. J. Biol. Macro- mol., 1981, 3, 377; (b) M. H. Baron, C. Beloze, C. Toniolo and G. D. Kasman, Biopolymers, 1978,17,2225. 26 (a) M. K. Gilson and B. H. Hanig, Nature (London), 1987, 330, 84; (b) S. Dao Pin, D. I. Liao and S. J. Remigton, Proc. Natl. Acad. Sci. USA, 1989,86,5361. 27 (a) A. B. Mandal, D. Mukherjee and D. Ramaswamy, Leather Sci.,1981, 28, 283; (b) A. B. Mandal, M. Kanthimathi, K. Govin- daraju and D. Ramaswamy, J. SOC. Leather Technol. Chem., 1983, 67, 147; (c) A. B. Mandal, J. Surface Sci. Technol., 1985, 1, 2730 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 28 93; (d) W. B. Gratzer and G. H. Beaven, J. Phys. Chem., 1969, 73, 2270; (e) D. G. Hall and B. A. Pethica, in Nonionic Sur-factants, ed.M. J. Schick, Marcel Dekker, New York, 1967, ch. R. H.Ottewill, in Surfactants, ed. Th. H. Tadros, Academic 16, pp. 516-557. 33 34 Dekker, New York, 1967, vol. 1, pp. 478-515; (d) P.Mukerjee and K. Mysels, J. Am. Chem. SOC., 1955,77,2937. I. M. Klotz and J. S.Franzen, J. Am. Chem. SOC., 1962,84,3461. (a)H. J. Schneider, R.K. Juneja and S. Simora, Chem. Ber., 1989, 112; (b) H. J. Schneider, Angew. Chem., Znt. Ed. Engl., 1991, 30, 29 30 31 32 Press, London, 1984, ch. 1, pp. 1-7. (a)D. F. Evans, M. Allen, B. W. Ninham and A. Fouda, J. Solu-tion Chem., 1984, 13, 87; (b) D. F. Evans and B. Ninham, J. Phys. Chem., 1986,90,226. (a) J. M. Sturtvent, Proc. Natl. Acad. Sci. USA, 1977, 74, 2236; (b)R. L. Baldwin, Proc. Natl. Acad. Sci. USA, 1986,83, 8669. R. S. Spolar, J. Hu and T. M. Record, Proc. Natl. Acad. Sci., USA, 1989,86,8382. (a)C. Tanford, in The Hydrophobic Effect, Formation of Micelles and Biological Membranes, Wiley, New York, 1973; (b) J. H. Fendler and E. J. Fendler, in Catalysis in MicelZar and Macro- molecular Systems, Academic Press, New York, 1975; (c) P. Becher, in Nonionic Surfactants, ed. M. J. Schick, Marcel 35 36 37 1417. (a) M. H. Abraham, J. Am. Chem. SOC., 1982, 104, 2085; (b) N. Muller, Acc. Chem. Res., 1990,23,28. (a)J. E. Bailey, G. H. Beaven, D. A. Chignell and W. B. Gratzer, Eur. J. Biochem., 1968, 7, 5; (b)H. Greenspan, J. Birnbaum and J. Feitelson, Biochim. Biophys. Acta, 1966, 126, 13. (a)J. D. Glickson, S. L. Gordon, T. P. Pitner, D. G. Agnesh and R. Walker, Biochemistry, 1976, 15, 5721; (b) A. A. Bothner-By, in Magnetic Resonance in Biology, ed. R. G. Shulman, Academic Press, New York, 1979, p. 177. Paper 3/06942G; Received 22nd November, 1993
ISSN:0956-5000
DOI:10.1039/FT9949002725
出版商:RSC
年代:1994
数据来源: RSC
|
24. |
Interaction of water withα,α-trehalose in solution: molecular dynamics simulation approach |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2731-2735
Maria C. Donnamaria,
Preview
|
PDF (609KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2731-2735 Interaction of Water with a,a-Trehalose in Solution : Molecular Dynamics Simulation Approach Maria C. Donnamaria, Eduardo 1. Howard and J. Raul Grigera lnstituto de Fisica de Liquidos y Sistemas Biologicos (IFLYSlB), CONICET, UNLP and Departamento de Ciencias Biologicas, Facultad de Ciencias Exactas, Universidad Nacional de La Plata, C.C. 565, 1900 La Plata, Argentina Molecular simulations of an aqueous solution of a,a-trehalose (a-D-glucopyranosyla-D-glucopyranoside) have been carried out to further the understanding of the effect of a,a-trehalose as a protecting agent against water stress in biological systems. The hydrogen-bond network and water dynamics were found to be only slightly altered compared with pure water (SPC/E model).Some internal hydrogen bonds in trehalose stabilize the conformation that was found to have glycosysidic dihedral angles of 215" and 216". It is found that trehalose can fit into a water structure involving at least ten water molecules per trehalose. Results support the view that the ability of trehalose to protect against water stress is due to the stabilization of biological structures and not to modification of the properties of water. Trehalose (a-D-ghcopyranosyla-D-glucopyranoside) is a disaccharide well known as a natural protector against water stress.' The effect of a,a-trehalose may be due either to modi- fication of the water structure and dynamics or to direct sta- bilization of the biological structures to be protected.Its actual mechanism of action is still unclear. A careful analysis of aqueous a,a-trehalose may, at least, provide new informa- tion to identify the mechanism. Molecular dynamics (MD) simulation has proved to be a highly reliable method to predict carbohydrate properties in solution*-' and, therefore, this technique may help to provide a better understanding of the effect of a,a-trehalose in water. The reliability of the simulation depends on several factors and, until at least some of the simulation data is checked experimentally, they remain only models. Moreover, the starting points of the models have to be experimental data. Unfortunately, some crucial data, e.g. charge distribution, can only be obtained through calculation. This paper deals with the simulation by MD of a,&-tre- halose in aqueous solution.The aim was to gain a better understanding of the effect of a,a-trehalose on biological systems under water stress. We have already published some preliminary data of MD and molecular mechanics results.' Model In a,a-trehalose, two hexapyranose rings are connected via a 1-1 glycosidic linkage. In the model used, both rings were kept rigid in the 4C1 conformation by applying improper torsion potentials that avoid transitions between other pos- sible conformations. All atoms were explicitly included and the tetrahedral geometry of carbon atoms was maintained using improper torsion potentials. Bond lengths were kept rigid and bond angles treated as having harmonic potentials.We did not apply torsion potentials on the glycosidic linkage. The conformation is determined through the dihedral angles only by the atom-atom interaction and solvent effects. We adopt this approach since the inclusion of a predefined tor- sional potential will introduce a bias in the calculated confor- mation~.~.'~ Force-field parameters were used as given in the GROMOS package (Biomos N.V., Groningen-Zurich). Hydrogen atoms interact only through coulomb forces. We use two sets of charge for a,a-trehalose. In one, the charges were computed using the semi-empirical quantum calculation CNIND0/2RF, SET 1. The other set was a modified version of AM1 in which the solvation effect is incorporated ad hoc (AMSOL program) SET 2.The later calculations were carried out by Dr. H. Villar. The determination of charges may be a critical point in this kind of simulation. For this reason, and since the inclusion of solvent screening included in the AMSOL program is more appropriate we shift the results to those obtained by SET 2. Fig. 1 shows a scheme of the model, and Table 1 the atomic charges. Water was modelled using the SPC/E model.'' It consists of a negative charge in the oxygen location and repulsion- attraction potential of a Lennard-Jones 12-6 type. Two posi- tive charges are located on hydrogen atoms. The OH distance is kept at 0.1 nm and HOH angle at 109.47'. Method The MD simulation was performed using the GROMOS package in an IBM RS/6000 32H.Results were analysed by a VAX 11/750 and 486 personal computer. All calculations were made at constant pressure (1.013 x 10' Pa) and con- stant temperature (300 K). The integration time step was 0.001 ps and simulations continued for 100 ps after equi- librium was reached. The criteria for equilibrium were the stability of box length and potential energy. One a,a-trehalose molecule and 270 water molecules were contained in an octahedral box of 1.786 nm x 1.813 nm x 2.283 nm. Results and Discussion Flexibility The relative mobility of both pyranose rings has no restrictions other than those imposed by mutual atom HO \ (b) OH Fig. 1 The a,a-trehalose molecule Table 1 Atomic partial charges for u,u-trehalose using in the simu- lation atom (ring a) charge atom (ring b) charge 0.14 0.10 -0.01 -0.02 -0.04 -0.04 -0.03 0.02 -0.04 -0.02 -0.01 -0.01 -0.25 - -0.32 -0.33 -0.34 -0.34 -0.32 -0.33 -0.30 -0.27 -0.34 -0.34 0.23 0.22 0.23 0.23 0.21 0.23 0.22 0.2 1 0.14 0.15 0.11 0.15 0.10 0.10 0.14 0.10 0.10 0.12 0.09 0.09 0.13 0.13 ~ ~~ a H(ij) Hydroxy-group hydrogen.HC(ij): Hydrogen bound to carbon. interaction. Fig. 2 shows the trajectories of glycosidic dihe- dral angles +[C(a2)-C(a 1)-O(g1)-C(b l)] and Y[C(a 1)-O(gl)-C(b 1)- C(b2)] recorded during the first 100 ps after reaching equilibrium. The average values are: = 216” (RMS = 13.13), and Y = 215” (RMS = 11.98).Both angles fluctuate around almost the same mean value; only one conformation was accessed during the 100 ps studied. These data agree with those obtained by Rees and Thom” using the optical rotation method, although our results are closer to those observed in the crystal. The orientation of the exocyclic groups can be observed through the trajectories of the dihedral angles that fix their 300 (a)f v);200 t loo] 300 v) 8200 U 0 50 100 TIPS Fig. 2 Time trajectories of dihedral angles of the glycosidic linkage of trehalose (a) 4 = C(a2)C(al)-O(gl)-C(bl), (b) Y = C(al)-O(gl)-C(bl)-C(b2) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 200 d 100 t’ I I (bl 1 300 200 0 100 300 200 U 100 0 0 50 100 tlPS Fig.3 Trajectories of dihedral angles of exocyclic hydroxymethyl groups: (a) 0 = C(a4)-C(a5)-C(a6)-O(a6), (b) 0 = C(b4)-C(b5)-C(b6)-O(b6) and (c) hydroxy group a = C(a1)-C(a2)-O(u2)- H(a2) orientation. Fig. 3(a)and (b) show the trajectories of the dihe- dral angle defined by the atoms C(4)-C(5)-C(6)-0(6) for both rings. We see that hydroxymethyl groups prefer certain orientations. The situation is different for the hydroxy group as can be observed in Fig. 3(c) through the movement of the dihedral angle defined by the atoms C(l)-C(2)-0(2)-H(2). The figure shows as an example the ring a, which corre-sponds to the typical behaviour. This result is similar to the one obtained by Brady2 for a-D-glucose in water.Analysis of internal hydrogen bonds (Table 2) shows a number of H-bonds between O(6) and O(2) belonging to different rings. It seems that internal hydrogen bonds are favoured by the particular orientation of the exocyclic methyl groups. Water Diffusion Coefficient To evaluate the effect of a,a-trehalose on the dynamic proper- ties of water, we have computed the water-water coefficient from the atomic mean-square displacement Thus the diffusion coefficient will be proportional to the slope of the line of the mean-square displacement (Ax2) = l/nCn[x(0)-x(t)I2us. time, after the system reaches the diffu- sive regime. As is shown in Fig. 4, the slope corresponding to pure SPC/E water and the simulated aqueous a,a-trehalose solution gives D, = 2.211 x cm2 s-l and D, = 2.163 x lo-’ cm2 s-l, respectively.This implies that, for concen- J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 H sequence (%HB)B H(a2) H(a4) H(a6) H(b2) W4) Wb6) O(a2)-E O(a4)-E O(a6)-€ O(u6)-E O(b2)-E O(b2)-E O(b4)-€ O(b6)-E O(b6)-E (a2)-0(b5)(100) (a4)-0(b6)(100) (a6)4(u4)(25)(~6)-0(b2)(75) (b2)-0(~5)(99)(b2)-0(~6)(01) (64)-0(b6X100) (b6)-0( b4)(66) (u4)-0( u2~3 3) Internal hydrogen bonds detected during simulation of a,a-trehalose ~ ~~~ HB = Ob HB = lb oc/emc (angle/degrees) (distance/nm) 4803 197 0.039 160.697 0.217 62 4999 1 0.0002 151.739 0.196 99 4972 28 0.006 163.033 0.215 24 4717 283 0.057 160.315 0.222 59 4988 12 0.024 147.100 0.199 02 4964 36 0.0072 155.517 0.208 96 %HB indicates the percentage probability of the hydrogen forming a bond in the respective sequence.HB = 0 and HB = 1 indicates the number of times in which the bond was empty and occupied, respectively. 'oc/em is the ratio between the number of occupied bonds and empty bonds. trations up to at least 0.2 mol 1-', a,a-trehalose does not sig- nificantly alter the diffusion coefficient of water. (Note that the experimental water diffusion coefficient is a little higher.) Although the SPC/E model gives a quite reliable value, it is known that truncated octahedral goemetry produces a slight decrease in the computed diffusion coefficient. l4 Hydration Hydration has been studied through the radial distribution functions of water around each atom of a,a-trehalose and their angular distribution functions.We also performed hydrogen-bond statistics. As expected, equivalent atoms in different rings show the same behaviour. The radial distribution function of water around oxygen atoms indicates that ring and glycosidic oxygen atoms have a non-polar behaviour. The other oxygen atoms show a polar feature, an effect that is clearer for O(3) and O(6). Fig. 5 shows a sample of radial distribution func- tions that represent the behaviour. 1.4 I ' I - I ' I ' I ' I I 1.2 1.o 0.8 0.6 <Ax2> 0.4 0.2 0.0 -0.2 G 20 40 60 80 100 tlPS Fig. 4 Average of the mean-square displacement of water molecules us.simulation time for pure SPC/E water (-) and a,a-trehalose solution (---) For the 'hydrophilic' oxygen atoms the first peak is at ca. 0.35 nm. As a rule, the peaks are low compared with similar situations on other carbohydrate^.^>^ Angular distribution functions show how solvent molecules are distributed around the solute atoms. We computed three different angular distribution functions for each a,a-trehalose atom referring to the relative orientation of the line connect- ing the selected atom with water oxygen and the fixed OH, HH and bisector of the HOH angle. The computed angular distribution functions show the expected behaviour according to the polar and non-polar characteristics. Hydrogen-bond Network Additional information lies in the hydrogen-bond network, including the water-ap-trehalose, water-water and a,a-trehalose--a,a-trehalose interactions.Fig. 6 shows the water- water hydrogen bonds for pure SPC/E water and the aqueous solution of a,a-trehalose. The criteria for hydrogen- bond formation are the distance between the proton and acceptor (d < 0.24 nm) and the angle formed by the donor, hydrogen, and acceptor (4 > 145'). For water and a,a-trehal- ose solution the average is d = 0.186 nm and 4 = 162.1'. These numbers for hydrogen bonds correspond to computa- tion over 100 ps for between zero and six hydrogen bonds. We found a slight decrease in the number of water-water hydrogen bonds for a,a-trehalose solution, dropping from an average of 3.183 to 3.141.This difference is hardly significant and the presence of an a,@-trehalose molecule, which forms some hydrogen bonds with water, will require some decrease in the water-water bonds. We can easily explain this differ- ence if 10 water hydrogen bonds are involved with ap-trehal- ose, which is the observed number. Fig. 6 shows some shift for lower numbers of hydrogen bonds, but is so slight as to be not significant. We can conclude that the presence of a,a-trehalose only slightly disturbs the hydrogen-bond network of water. Internal Hydrogen Bonds The average number of internal hydrogen bonds of trehalose observed during the simulation was very small. However, there is a certain number of intramolecular hydrogen bonds that play an interesting role in stabilizing the conformation.The values are shown in Table 2. Note that for all the pos- sible internal hydrogen bonds, the ratio between the occupied and empty state is lower than 6%. There are a small, but noticeable, number of hydrogen bonds between O(6) and O(2) of different sugar rings. The existence of these hydrogen bonds has also been suggested by Rees and Thorn.'' These will inhibit large movements of the glycosidic dihedral angles. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1.5 1.o n v 0,0.5 0.0 0.0 600000r 500 000. n 400000-Q) ~8 300 000 u6 200000 r/nm 4: :I:::::: r/n m 0.0 0.3 0.6 0.9 r/nm 0(6.51 .I..,.,... 0.3 0.6 0.9 0.0 0.3 0.6 0.9 rlnm rlnm Fig.5 Radial distribution functions of water oxygen atoms around different atoms of a,a-trehalose Iwater A few hydrogen bonds in which the ring oxygen was ?! water-trehalose involved have been detected which is unexpected considering the non-polar character of O(5).However, the small number of hydrogen bonds formed is not enough to modify the overall character, although they may make some contribu- tion to the stabilization of the molecular conformation. Conclusions -100 000 -Through the analysis of a molecular simulation of an aqueous solution of a,a-trehalose we have been able to state O! HBO HB1 HB2 HB3 HB 4 HB 5 HB 6 several aspects of the characteristic dynamic behaviour, struc- ture and conformation of the complete solute-solvent system.Fig. 6 Distribution of water-water hydrogen bonds in pure SPC/E Moderate Aexibility in a,a-trehalose Was observed Using a water (W-W) and a,a-trehalose solution (T-W) model of two rigid rings, but allowing them to move freely J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 around the involved glycosidic angles. Since we have not applied torsion potentials on the glycosidic linkage, the con- formation determined through the dihedral angles comes from the atom-atom interaction and solvent effects. Conse- quently we believe that the obtained conformation is not biased by any predefined structure. The relative rigidity of a,a-trehalose seems to be due, at least partly, to the existence of a small number of hydrogen bonds that do not allow large fluctuations around the glyco- sidic linkage.We cannot rule out the contribution of the water lattice structure as a factor which influences the confor- mation. The presence of a,a-trehalose seems to disturb only slightly the hydrogen-bond network of water (diffusion coefficients and statistics of hydrogen bonds). Therefore, we cannot con- sider the changes in the hydrogen-bonding pattern of water as an explanation for the protection of biological structures under water stress. However, the partial folding between the two rings, along the glycosidic linkage, leads to a special structure, where the HO(2) of one residue is close to the HO(6) of the other, increasing the availability of hydrogen bonding, as suggested previously.'2 Such a conformation produces a spatial arrangement of hydroxy groups that maxi- mizes hydrogen-bonding interactions with a putative tri-dymite structure of water.In this context unexpected 0(5)-H(2) hydrogen bonding could also contribute to the stability of the structure. The spacing of hydroxy groups in a,a-trehalose matches the tetrahedrally coordinated oxygen atoms hydrogen bonded into the dynamic network array for water, as depicted in Fig. 7. As seen in the model, most OH groups capable of forming hydrogen bonds with water are oriented towards one surface. The more hydrophilic surface will fit into the water structure. Fig. 7 Scheme of a,a-trehalose molecule over an expanding ice structure. The configuration of a,a-trehalose corresponds to the average confieuration obtained bv MD simulation. It is assumed that the water structure models the hydrophilic surfaces of biological structures (ref.15 and references therein), therefore, we may assume that hydrogen bonds of a,a-trehalose will model biological structures. These results support the view1 that the protective action of a,a-trehalose under water stress is carried out by preser- vation of the bilayer form (against the hexagonal phase) of membrane phospholipids. The inhibition of phase transitions of phospholipids under dehydration is an experimental fact ' that supports the above explanation. We propose that for a system in which water plays a role in the stabilization of the structure, as in biological systems, the a,a-trehalose may replace a number of water molecules bonded to the structure.Upon drying, the stabilizing effect of the water lattice is taken over by the a,a-trehalose, with the help of other carbohydrates that are regularly present in such systems. Therefore, the action of a,a-trehalose as a protector of biological structures under water stress, corresponds to a direct, though weak, action on the biological structures, rather than a modification of the water structure and dynamics. This mechanism may not be unique for a,a-trehal- ose and may be shared with other carbohydrates whose pro- tection properties under water stress are known. This work has been partially funded by the Consejo Nacional de Investigaciones Cientificas y Tecnicas (CONICET) of Argentina.J.R.G. and M.C.D. are members of the Carrera del Investigador of CONICET and Comision de Investigaciones Cientificas of the Province of Buenos Aires (CIC), respec- tively. E.I.H. is a fellow of CONICET. We wish to thank Dr. H. Villar for this thoughtfulness and valuable work in calcu- lating the charges of a,a-trehalose. We are also indebted to Juan Grigera for his help with technical support and Tomas Grigera for his help in the writing. The continued encour- agement of Prof. Felix Franks to continue in the carbo- hydrate field is gratefully acknowledged. References 1 F. Franks in Biophysics and Biochemistry at Low Temperatures, Cambridge University Press, Cambridge, 1985, ch. 6, pp. 112- 124. 2 J.Brady, J. Am. Chem. SOC., 1989,111,5155. 3 B. P. van Eijck, L. M. J. Kroon-Batenbury and J. Kroon, J. Mol Struct., 1990,273, 3 15. 4 J. R. Grigera, J. Chem. Soc., Faraday Trans. I, 1988,84, 1603. 5 J. W. Brady. Curr. Op. Struct. Biol., 1990, 1, 71 1. 6 F. Franks, J. Dadok, R. Kay and J. R. Grigera, J. Chem. SOC., Faraday Trans., 1991,87, 597. 7 E. I. Howard and J. R. Grigera, J. Chem. Soc., Faraday Trans., 1992,88,437. 8 J. R. Grigera, M. C. Donnamaria and E. I. Howard, in Con-densed Matter Theories, ed. L. Blum, Plenum Press, New York, 1993, VOI. 8, pp. 527-534. 9 J. R. Grigera, in Advances in Computational Biology, JAI Press, New York, 1994, vol. 1, pp. 208-229. 10 F. Franks and J. R. Grigera, in Water Science Review, ed. F. Franks, Cambridge University Press, Cambridge, 1990, vol. 5. 11 H. J. C. Berendsen, J. R. Grigera and T. Straastma. J. Phys. Chem., 1887,91,6269. 12 D. A. Rees and D. Thom, J. Chem. SOC., Perkin Trans. 2, 1977, 191. 13 J. P. Hansen and I. R. McDonald, Theory of Simple Liquids, Academic Press, London, 1976, p. 239. 14 K. Remerie, J. B. Engberts and W. F. van Gunsteren, Chem. Phys., 1986, 101,27. 15 H. J. C. Berendsen, in Theoretical and Experimental Biophysics, ed. A. Cole, Marcel Dekker, New York, 1967,59-75. 16 J. H. Crowe and L. M. Crowe, Crybiology, 1981,18,631. PaDer 4100389F: Received 21st Januarv. 1994
ISSN:0956-5000
DOI:10.1039/FT9949002731
出版商:RSC
年代:1994
数据来源: RSC
|
25. |
Brownian dynamics simulation of a multi-subunit deformable particle in simple shear flow |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2737-2742
Maria C. Buján-Núñez,
Preview
|
PDF (663KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2737-2742 Brownian Dynamics Simulation of a Multi-subunit Deformable Particle in Simple Shear Flow Maria C. Bujan-Nuiiezf-and Eric Dickinson" Procter Department of Food Science, University of Leeds, Leeds, UK LS2 9JT The structure and dynamics of a multi-subunit deformable particle are investigated by Brownian dynamics simu- lations. Pairwise hydrodynamic interactions between spherical subunits are described by the Rotne-Prager diffusion tensor and non-hydrodynamic interactions by means of an attractive energy parameter, E. In the absence of shear flow, the average simulated particle structure ranges from 'stringy' (€ <, kT) to compact, with a transition region in between. Time-dependent numerical results are presented for the average subunit coordi- nation number and the average particle perimeter area as a function of increasing shear rate.Particle elon- gation in the flow field leads to a reduction in the effective fractal dimension of the particle. The greatest sensitivity to shear rate occurs in the stringy4ompact transition regime. Substantially different dynamic behav- iour is obtained when the simulation is repeated without the inclusion of hydrodynamic interactions. Recently, we introduced' a new deformable particle model suitable for dynamic simulation consisting of a set of discrete interchangeable spherical subunits. Subject to the connected- ness of the whole multi-subunit structure, the subunits are free to move relative to one another like particles in a loose floc of interacting colloidal spheres which cannot fall apart.This is different from conventional chain-like models of linear or branched macromolecules which have fixed bonds holding the subunits together. Brownian dynamics simulation of the deformable particle with 30 subunits has shown' that, by changing the value of a single energy parameter E, the average configuration of the subunits can be made to change from a stringy structure (E + kT) to a roughly spherical compact structure (E % kT);in the transition region (E = 2.5 kT)the structure is subject to large spatial fluctuations. The present paper is concerned with the influence of simple shear flow on the structure and dynamics of this multi- subunit deformable particle model.The system is simulated by a Brownian dynamics algorithm which includes the effect of intersubunit hydrodynamic interaction^.^-^ Apart from the addition of the flow field, all the other features of the simula- tion are the same as those adopted previously.' The effect of steady shear and elongational flow fields on the structures of deformable macromolecules has previously been investigated experimentally6-" and theoretically.' '-' In particular, Brownian dynamics simulation has been used14-" to calculate the structural and dynamic properties of linear chains of connected subunits in shear flow. However, these previous simulations have either neglected intersubunit hydrodynamic interactions alt~gether'~ or have represented potential where r is the subunit-subunit centre-to-centre separation, and L and E are parameters defining the range and strength of the intersubunit attraction. It is assumed that two subunits are 'bonded' when their pair separation lies in the range 2R < I < L.An individual subunit must have at least one bond, and when E is of the order of, or greater than, the thermal energy (kT)many subunits will tend to have several bonds in order to minimize the overall potential energy of the system. Typically here we shall take E = 0, 2.5 or 4.0 kT with L = 3R and N = 30. For each time interval, At, during the Brownian dynamics simulation, we assume that the displacement of each subunit is given by3v4 Ari = At) + (At/kT)1 DZ + up At + SppAt (2) i where DG is the diffusion tensor, Si is a third-order mobility tensor, Fi is the net force acting on the jth particle, ui is the velocity due to the shear flow, RQ is a stochastic vector, p is the rate of strain tensor, k is Boltzmann's constant and T is the temperature.The superscript O denotes that the quantity is calculated at the beginning of the time interval. The first two terms in eqn. (2) correspond to the algorithm of Ermak and McCammon" with the divergence term omitted since V -Dij them in a very simplistic manner using the Oseen ten~or.'~*'~ is zero for the Rotne-Prager In the present study we represent intersubunit hydrodynamic interactions by a Rotne-Prager tensor" which is well behaved in a Brownian dynamics computer simulation2 and reproduces the essential physics of a multi-particle system in shear flow. 3,4 Simulation Methodology The deformable particle consists of a connected set of N iden-tical rigid spherical subunits (radius R) with intersubunit pair t Permanent address : Department of Physical Chemistry, Uni- versity of Santiago de Compostela, E-15706 Santiago de Compostela, Spain.tensor.2 The term opAt is the displacement arising from the bulk motion of the unper- turbed fluid, and SPpAt represents the effect of interparticle hydrodynamic forces on the shear-induced particle motion." At the level of the Rotne-Prager approximation, the configuration-dependent tensor is given by Dij= Do[Sij/+ (3R/41)(/+ iijFij) -i(R/r)3(3FijFii-r)] (3) where 6, is the Kronecker delta, / is the unit tensor, and i, is the unit vector between subunits i and j.The quantity Do is the scalar diffusion coefficient for an isolated subunit, i.e. Do = kT/6nqR (4) where q is the viscosity of the fluid medium. At the same level of hydrodynamic approximation (i.e.to order I-~),the shear Fig. 1 Close-packed structure of a 30-subunit deformable particle used as a starting configuration for each Brownian dynamics simula- tion tensor Siis given by Si= (5/2) 1(R/r)3FijFij rij (5) j# i At the level of Oseen hydrodynamics (i.e. to order r-’), the shear tensor is zero and the only effect of the shear field on the subunit dynamics is through the velocity ui. This is an unsatisfactory approximation because it neglects the rota- tional couplings between particles which are an important feature of relative particle motion in shear fl~w.’~*~* In this study we consider a single deformable particle (N = 30) subjected to simple shear flow in the x-y plane: u, = Gr,; u, = 0; u, = 0 (6) The shear rate, G, is expressed in dimensionless form as G/Go, where Go = Do/R2.The rate of strain tensor is given by PkI = C(auk/arl) + (aul/ark)1/2 (7) The initial configuration of the deformable particle used at the start of each simulation run is illustrated in Fig.1. This close-packed structure was generated from a simple ballistic aggregation algorithm.’ Results and Discussion In the absence of shear, the average structure of the 30-subunit deformable particle goes from stringy at E x 0 to highly compact at E z5 kT.’ The change in particle trans- lational diffusion coefficient D, calculated from plots of mean- square displacement of particle centre of mass against time, is shown in Table 1.Also shown is the change in inverse radius of gyration, Ri1,for the same values of the subunit-subunit attraction energy, E. We can see that the diffusion coeficient for the compact particle (E = 5kT) is ca. 40% higher than that for the stringy particle (E = 0) and that R,-’ changes by about the same relative amount. This roughly corresponds to the intuitively expected scaling (D-Rg-’). In the presence of shear, the particle becomes less compact to an extent which depends on the value of the shear rate, G, and the interaction parameter, E.We report results here for three different shear rates (G/Go= 0.1, 0.5 and 1.0) and three different interaction energies (E/kT = 0, 2.5 and 4.0). Table 2 Table 1 Effect of energy parameter, E, on the diffusion coeffcient, D, and inverse radius of gyration, Rpl, in the absence of shear flow 0.28 0.21 0.30 0.22 0.31 0.24 0.36 0.28 0.38 0.29 0.41 0.29 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Effect of energy parameter, E, and shear rate, G, on steady-state average subunit coordination number, Tc GIG, E=O E = 2.5 kT E=4kT ~~ ~ 0 2.1 4.4 5.6 0.1 2.1 2.5 4.8 0.5 2.1 2.5 4.6 1.o 2.0 2.4 4.1 lists steady-state values of the average subunit coordination number, rc,as a function of E and G.For zero field (G = 0) and G/Go = 0.5, Fig. 2 shows plots of fc as a function of reduced simulation time t/tR,where t, = R2/6D,. For any particular particle configuration, the coordination number, I,, of an individual subunit is taken as the number of neigh- bouring subunits with r < L. The results in Table 2 show that there is a decrease in the steady-state icwith increasing shear rate. The value of fc for the particle with E = 2.5 kT, which is in the stringy-compact transition regime, is particularly sen- sitive to the flow field: the application of a shear rate of just G = 0.1 Go leads to a reduction in the steady-state value of rc by ca. 40%. Such a low applied shear rate has less effect on fc for very expanded (E = 0) or highly compact (E = 4 kT) par-ticles.For the same simulation runs as in Fig. 2, the plots in Fig. 3 show how the average particle parameter area P,, varies with elapsed simulation time for different values of E and G/Go. The perimeter area, P,, of an instantaneous configu- ration is evaluated by embedding the three-dimensional par- ticle shape in an imaginary cubic lattice of spacing 2R, thereby generating a connected array of filled boxes whose total surface area can readily be calculated in integer units of (2R)’. The data in Table 3 indicate that P, increases with increasing G for E = 2.5 kT or E = 4 kT, but that the change in P, is insignificant for the stringy particle (E = 0).The time taken to reach the steady-state perimeter area is longest for the particle with average structure in the stringy-compact transition regime, i.e. E = 2.5 kT. This is consistent with the longer equilibration time for i, seen in Fig. 2 for this same particle. The compact particle (E = 4 kT) is subject to drifts in P, (and rc)around the steady-state values which are of longer duration than for particles with lower values of E. Table 4 shows how the effective fractal dimension, D,,of the deformable particles changes with shear-field strength. Results are derived from sets of simulations on systems con- taining various numbers of subunits (i.e. N = 4, 6, 8, ... , 28, 30) with D, estimated from a log-log plot of R, against N assuming the scaling relationship The value of D,= 1.9 for E = 0 is similar to that for fractal aggregates produced by diffusion-limited coag~lation.’~ Table 3 Effect of energy par_ameter, E, and shear rate, G, on steady- state average perimeter area, P, Pea GIGO E=O E = 2.5 kT E=4kT 0 160 135 115 0.1 160 150 120 0.5 160 150 120 1.o 160 155 125 Values taken from ref.1. Estimated error & 5 [in units of (2R)*]. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2739 7gl W I," 5 3 .,.. .. 7 , 1 1 t/tR t/tR 9 7 1," 1," gb7 7 n na a s5 s5W v 3 1 t/tR Fig. 2 Influence of reduced shear rate, GIG,, and subunit-subunit interaction energy, E, on average packing density of subunits in a deform- able particle.Each plot shows average subunit coordination number, r,, as a function of reduced time, t/t, :(a)E = 0,G = 0;(b)E = 0, G = 0.5 Go;(c)E = 2.5 kT, G = O;(d) E = 2.5 kT, G = 0.5 Go;(e)E = 4 kT, G = 0;Cf)E = 4 kT, G = 0.5 Go. Application of the shear field reduces the particle dimension- transition region, a shear rate of just 0.1 Go leads to a ality further so that at moderately high shear rate (G/Go= 1) reduction in dimensionality from a rather compact D, = 2.3 the configuration is tending towards a one-dimensional to a stringy D, = 1.9. In contrast, the same low shear rate = 1.4). For the case of the deformable produces a more modest expansion of the compact particle extended structure (0, particle with E = 2.5 kT, which is in the stringy<ompact (E = 4 kT)from D, = 2.6 to D, = 2.4.The preceding simulation results show that the effect of shear flow is to increase the surface area of the deformable Table 4 Effect of energy parameter, E, and shear rate, G, on effec- particle and to reduce its fractal dimensionality as well as the tive fractal dimension, D,,of a deformable particle average subunit coordination number. The compact-stringy transition induced by the hydrodynamic flow field is quanti- D' tatively analogous to the coil-stretch transition of dilute flex- GPO E=O E = 2.5 kT E=4kT ible polymers under the influence of ultrahigh velocity gradients. It is argued by de Gennes" that this latter tran- 0 1.9 2.3 2.6 sition can be expected to be abrupt because of a positive- 0.1 1.7 1.9 2.4 feedback mechanism: as the state of stretching is increased, 0.5 1.5 1.8 2.2 1.o 1.4 1.6 2.0 the inner subunits of the polymer coil become more exposed to the shear flow field, which leads to further chain distortion J.CHEM. SOC. FARADAY TRANS.. 1994. VOL. 90 180 (a)l 180 150 150 60660 R t/tR 150 h h Ks 120 &@ 60 R 150i 60 60 0 500 1000 1500 0 500 1000 15 tltR tltR Fig. 3 Influence of reduced shear rate, GIG,, and subunit-subunit interaction energy, E, on average surface area of a deformable particle. Each plot shows particle perimeter area, P,, as a function of reduced time, t/t, : (a)E = 0,G = 0;(b)E = 0,G = 0.5 Go;(c) E = 2.5 kT, G = 0;(d) E = 2.5 kT, G = 0.5 Go;(e)E = 4 kT, G = O;y? E = 4 kT, G = 0.5 Go.and hence further exposure of hydrodynamically shielded units. The main difference between our deformable particle model and the flexible linear chain is that the latter is much more restricted in terms of the scope for individual subunit motion than is the former. As the shear rate increases and the stringy deformable particle (E = 0) becomes increasingly extended by the flow field (Of + l), the structure of the deformed particle becomes closer to that of the linear flexible chain. To gain further insight into the structures of the deform- able particles in shear flow, it is useful to examine the chang- ing particle configurations shown in Fig. 4-6. ‘Snapshot’ Configurations for the stringy system (E = 0) are illustrated in Fig.+~)--(d)for G/Go = 0, 0.1, 0.5 and 1.0, respectively. With izaeasirrg shear rate the particle is, as expected, oriented increasingly in the x direction. An interesting feature of the structures in (b)and (c) is that the subunits near the two ends nf the Elongated particle tend to have a higher coordination number, I,, than those near the middle; i.e. there seems to be a tendency towards a loose dumbbell-type structure, similar to what is observed in the early stages of liquid droplet dis- ruption in Couette shear In order to investigate the importance of the subunit- subunit hydrodynamic interactions on the simulated behav- iour, we have repeated some of the calculations with the hydrodynamic interactions switched off, i.e.with D, in eqn. (3) replaced by Do from eqn. (4). What we find is a very differ- ent type of dynamic behaviour as the shear rate is increased. There is a tendency for the system to oscillate between elon- gated and compact states with the lifetime of the compact globular states decreasing with increasing shear rate. This phenomenon is particularly evident for deformable particles in the transition regime as illustrated in Fig. 7(4-(c) for a system with E = 2.5 kT for G/Go = 0.1, 0.2 and 1.0, respec-tively. At the moderate shear rate, G = 0.2 Go,in the absence of hydrodynamic interactions, the particle spends roughly J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Fig. 4 Projected configurations in the x-y plane of a deformable particle with E = 0 at various shear rates: (a) GIG, = 0, (b)GIG, = 0.1, (c) GIG, = 0.5 and (d) GIG, = 1.0 equal amounts of time in compact states (1,x6) and stringy states (1,x2.5).(A similar type of behaviour is exhibited by the parameter P, as a function of simulation time.) At rela- tively high shear rate (G/G, = 1) the system exists predomi- nantly in elongated configurations (f,x2) but makes short excursions into compact globular configurations (F, x 5). In the presence of hydrodynamic interactions, however, these wild fluctuations do not appear (Fig. 2). The behaviour indicated by the data in Fig. 7 seems similar to that reported by Liu14 for a Brownian dynamics simula- tion of a linear Kramers model chain in shear flow.In the Kramers model it is assumed that the segment hydrodynamic interactions can be neglected. In the so-called shear-thinning regime, it was noted14 that the simulated chains undergo a pulsatile motion superimposed on the rotational motion. That is, one moment they are extended and aligned with the n Fig. 5 Projected configuration in the x-y plane of a deformable particle with E = 2.5 kT at a reduced shear rate, GIG, = 0.5 Fig. 6 Projected configuration in the x-y plane of a deformable particle with E = 4 kT at a reduced shear rate, GIG, = 0.5 2741 7-1 0 2800 5600 8400 11200 14(00 tltA Fig. 7 Behaviour of a deformable particle with E = 2.5 kT in the absence of intersubunit hydrodynamic interactions. Average subunit coordination number, I,, is plotted against reduced time, t/t,, for three shear rates: (a)G/G,= 0.1, (b)G/G, = 0.2 and (c) GIG, = 1.0. flow, and the next moment they coil up and rotate in the x-y plane before assuming another extended configuration, but with the two ends reversed from the previous extended con- figuration.The reason for this pulsatile behaviour in the absence of hydrodynamic interactions can be clearly under- stood by considering the form of the shear tensor Si,as dis- cussed on a previous occasion3 for the case of a doublet in simple shear flow. When two or more spheres are aligned parallel to the x axis, the shear flow velocity ui cannot produce any rotation of the aggregate because the line of centres of the spheres lies along the unperturbed streamlines.However, in this orientation there is a small translation con- tribution in the y direction to Ari due to the contribution of the term in Si in eqn. (2). So, the shear tensor Sp has an important role of sustaining doublet rotation when the effect of the bulk flow is zero. If Siis set to zero, which is the case when hydrodynamic interactions are neglected (or just an Oseen approximation is used), once the assembly of spherical subunits has become elongated parallel to the axis this arrangement is likely to become locked in for a considerable time. It is only disrupted by a Brownian fluctuation which is large enough to produce a substantial rotational contribution from the term in uiin eqn. (2).This then triggers a large rela- tive translational motion in the x direction because in the absence of the hydrodynamic interactions there is no aggre- gate rotation in the flow except as caused by the constraints of the intersubunit bonds. This is the origin of the pulsatile motion evident in Fig. 7 and in the paper of Liu.14 However, it is probably not what occurs in most real systems where hydrodynamic subunit-subunit couplings are generally important, especially for large particles. Conclusions The influence of shear flow on structure of an isolated deformable particle has been investigated by Brownian dynamics simulation. This research shows the potential for extending previous simulation studies of deformable particle adsorption’ and particle gel formation26 to situations where hydrodynamic interactions play a role.A specific long-term objective is to develop a statistical model to describe the dynamics of a globular protein monolayer, including effects of unfolding and intermolecular interactions in the adsorbed state. The deformable particle model studied here can be likened to an assembly of mobile Brownian subunits joined together non-specifically by strong flexible invisible links ;this is analogous to how the small ordered mobile domains of a large protein molecule are held loosely together in the molten globule state2’ by the random chain sections of the polypep- tide chain. The dynamics of this model in many ways resem- bles that of a floc of colloidal particle^.^ The only difference is that, whereas particles at the floc edge may break away under the influence of Brownian motion or a shear flow field, the subunits in the model deformable particle are each con-strained by a ‘final’ unbreakable bond to their closest subunit neighbours.The adoption of hydrodynamic as well as molecular interactions between the subunits implies a pen- etration of bulk-phase solvent into the gaps between the sub- units. This latter assumption is obviously much more relevant to the stringy state than to the compact state. The numerical results presented in this paper show the importance of hydrodynamic interactions between subunits on the dynamics of this deformable particle model in shear flow.When hydrodynamic interactions are excluded, the dynamical evolution of the particle structure with time is fun- damentally different from that simulated using the Rotne- Prager tensor to represent the intersubunit pair J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 hydrodynamic interactions. In particular, without the hydro- dynamic interactons, the particle exhibits a pulsatile motion superimposed on the rotational motion. This pulsatile motion appears systematic at rather low shear rates (but above a certain minimum value). At high shear rates, the pul- satile motion still persists but now its character is rather chaotic. References 1 M. C. Bujan-Nuiiez and E. Dickinson, Mol. Phys., 1993,80,431. 2 E. Dickinson, Chem. SOC.Reu., 1985, 14,421.3 G. C. Ansell, E. Dickinson and M. Ludvigsen, J. Chem. SOC., Faraday Trans. 2,1985,81,1269. 4 G. C. Ansell and E. Dickinson, J. Colloid Interface Sci., 1986, 110, 73. 5 E. Dickinson, in The Structure, Dynamics and Equilibrium Properties of Colloidal Systems, ed. D. M. Bloor and E. Wyn- Jones, Kluwer, Dordrecht, 1990, p. 707. 6 F. R. Cottrell, E. W. Merrill and K. A. Smith, J. Polym. Sci., Part A2, 1969,7, 1415. 7 R. N. Dunlap and L. G. Leal, J. Non-Newtonian Fluid Mech., 1987, 23, 5. 8 T. Takebe, T. Hashimoto, B. Ernst and R. S. Stein, J. Chem. Phys., 1990,92, 1386. 9 H-H. Graf, H. Kneppe and F. Schneider, Mol. Phys., 1992, 77, 521. 10 R. de Rooij, A. A. Potanin, D. van den Ende and J. Mellena, J. Chem. Phys., 1993,99,9213.11 P-G. de Gennes, J. Chem. Phys., 1979,60,5030. 12 Y. Rabin, J. Chem. Phys., 1988,88,4014. 13 S-Q. Wang, J. Chem. Phys., 1990,92,7618. 14 T. W. Liu, J. Chem. Phys., 1989,90,5826. 15 J. J. Lbpez Cascales and J. Garcia de la Torre, J. Chem. Phys., 1991,959384. 16 J. J. Lbpez Cascales and J. Garcia de la Torre, Macromolecules, 1990,23, 809. 17 H. C. Otinger, J. Chem. Phys., 1986,84, 1850. 18 J. Rotne and S. Prager, J. Chem. Phys., 1969,50,4831. 19 D. L. Ermak and J. A. McCammon, J. Chem. Phys., 1978, 69, 1352. 20 G. K. Batchelor and J. T. Green, J. Fluid Mech., 1972,56, 375. 21 P. Mazur and W. van Saarloos, Physica A, 1982,115,21. 22 J. Happel and H. Brenner, Low Reynolds Number Hydrody- namics, Noordhoff, Leiden, 2nd edn., 1973. 23 M. Vold, J. Colloid Interface Sci., 1963, 18,684. 24 P. Meakin, in Phase Transitions and Critical Phenomena, ed. C. Domb and J. L. Lebowitz, Academic Press, New York, 1988, vol. 12, ch. 3. 25 E. Dickinson and G. Stainsby, Colloids in Food, Applied Science, London, 1982, p. 187. 26 E. Dickinson, J. Chem. SOC.,Faraday Trans., 1994,90, 173. 27 E. Dickinson and Y. Matsumura, Colloids Surf.B, 1994,2, in the press. Paper 4/01572J;Received 16th March, 1994
ISSN:0956-5000
DOI:10.1039/FT9949002737
出版商:RSC
年代:1994
数据来源: RSC
|
26. |
Surface chemistry and microemulsion formation in systems containing dialkylphthalate esters as oils |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2743-2751
Robert Aveyard,
Preview
|
PDF (1066KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2743-2751 Surface Chemistry and Microemulsion Formation in Systems containing Dialkylphthalate Esters as Oils Robert Aveyard, Bernard P. Binks, P. D. 1. Fletcher* and Paul A. Kingston School of Chemistry, University of Hull, Hull, UK HU6 7RX Alan R. Pitt Kodak Ltd., Headstone Drive, Harrow, Middlesex, UK HA I 4TY The surface chemical behaviour of phthalate esters in oil-water mixtures, both in the presence and absence of a conventional surfactant, has been investigated. We have estimated the surface activity of phthalate esters and the extent to which this type of moderately polar oil can be coadsorbed into concentrated monolayers of conven- tional surfactants at the oil/water interface. In the absence of surfactant, phthalate oils are adsorbed from dilute solution in heptane at the heptane/water interface, exhibiting a surface activity typical of non-aromatic diesters and similar to that of alcohols.However, the surface chemical and microemulsion phase behaviour in mixtures of phthalate oils and aqueous NaCl solutions in the presence of the anionic surfactant sodium bis(2-ethylhexyl) sulfosuccinate (AOT) indicates a low extent of penetration of the phthalate oils into AOT monolayers. This is interesting since the phthalates have similar surface activities (from hydrocarbon oils) to alcohols which are very effective cosurfactants capable of strong coadsorption into surfactant monolayers. The main effects of substituting phthalate oils for alkanes in mixtures of oil-AOT-aqueous NaCl are (i) to increase the critical aggre- gate concentration of surfactant in the oil phase, (ii) to increase the NaCl concentration required for micro- emulsion phase inversion and (iii) to increase the magnitude of the minimum interfacial tension obtained by varying the salt concentration. It is concluded that phthalate oils do not act as strongly adsorbed cosurfactants in AOT-water-oil systems and the extent of phthalate oil penetration of AOT monolayers is less than for alkanes.Surface chemistry and microemulsion formation in oil-water-surfactant mixtures have been widely investigated in recent years. The behaviour of such systems is understood qualitatively for mixtures containing water and apolar oils such as alkanes.’-5 Additionally, mixtures of surfactant-apolar oil-polar solvents other than water have been investi- The effect of varying the oil structure has received attention and systems containing oils consisting of mixtures of an apolar hydrocarbon and an alcohol have been widely Inin~estigated.~.~ these latter mixtures, the alcohol (commonly termed a ‘cosurfactant ’ is generally distributed between the surfactant film at the oil/water interface and the bulk oil and water solvents.Oil species containing polar groups other than hydroxy groups have been less widely studied (e.g. ref. 10-14) although the interaction of such mod- erately polar oils with surfactants is important in a number of technologies including photographic emulsions, foods con-taining triglyceride oils and perfume oil delivery systems.The effect of oil molecular structure on microemulsion for- mation (and associated properties including the oil/water interfacial tension) is thought to be a consequence of the extent of penetration of the surfactant monolayer by the oi1.8,15.16 Apolar oils such as alkanes can penetrate and swell the tail region of surfactant monolayers and the extent to which this occurs is an important factor determining the phase behaviour and monolayer properties such as the bending elasticity.I7 Alcohol oils can penetrate further into the monolayer such that the hydroxy group is located in the plane of the surfactant headgroups. In this study we attempt to determine whether moderately polar oils such as esters (the subject of this work) penetrate concentrated surfactant monolayers and, if so, whether they mix with the surfactant chains only (like alkanes) or with both surfactant chain and headgroups (like alcohols).We have investigated the behaviour of mixtures containing the anionic twin-tailed surfactant sodium bis(2-ethylhexyl) sulfosuccinate (AOT), aqueous solutions of NaCl and dialkyl phthalate esters as the oil component. These oils are inter- esting since they may offer an alternative to environmentally harmful materials currently used and they are of particular interest in the photographic industry as solvents for certain dye materials. We first consider the adsorption of phthalate oils from dilute solution in heptane onto the oil/water interface to establish the surface activity in the absence of surfactant. Secondly, we describe the adsorption of AOT at the phthal- ate oil/water interface and the evolution of low tensions with aqueous phase NaCl concentration. Third, we discuss equi- librium microemulsion phase compositions for AOT-water- phthalate oil-NaC1 mixtures.Finally, some preliminary experiments concerning the interaction of the phthalate oils with AOT monolayers at both air/water and alkane/water interfaces are described. Experimental Materials AOT was obtained from-Sigma and was used without further purification. Water was purified by reverse osmosis and by using a Milli-Q reagent wa,ter system. Heavy water (D,O) was obtained from either Aldrich or MSD Isotopes. The air/ water surface tension at 25°C of the samples used was 71.8 0.1 mN m-’, in good agreement with the literature value of 71.9 kO.1 mN m-1.18 Heptane (Fisons HPLC grade), diisooctyl phthalate (DOP) (Janssen Chimica, 98%) and di-n-butyl phthalate (DBP) (Janssen Chimica, 99%),di-n-heptyl phthalate (Lancaster) and all other chain length phthalate esters used (Eastman Kodak) were passed over an alumina column prior to use to remove polar impurities. The phthalate oils (general structure shown below) showed only a single spot by thin-layer chromatography.The absence of surface-active impurities was checked further by observing emulsions prepared by shaking the oils with pure water. Oil Table 1 Properties of di-ester phthalate oils at 25 "C phthalate ester chain group oil/air tension/mN m -density/g cm -ethyl 37.3 1.1137 butyl 33.5 1 .O423 hexyl 31.8 0.9993 heptyl 31.5 0.9859 isooct yl 31.3 0.9803 nonyl 29.2 0.9653 decyl 30.0 0.9552 samples showing stable emulsions (indicating the presence of surface-active impurities) were rejected.Methods As shown in Table 1, phthalate oils have densities similar to that of water. Low oil-water density differences can cause dif- ficulties in tension measurements. Surface and interfacial ten- sions were determined using one of three methods. Systems with tension values ,>10 mN m-l and density difference values >O.l g were measured by the du Nouy ring method using a Kruss K12 (maximum pull) instrument.Lower tension systems were measured by the spinning drop technique using a Kruss Site 04 instrument. Measurements of systems showing a low density difference ( <0.1 g cm- 3, were made using a sessile drop apparatus constructed in this laboratory. In this method, developed by Padday and Pitt," a drop of the more dense liquid is placed on a solid surface with which it has a contact angle of ca. 180" when immersed in the second liquid. The height of the drop passes through a maximum value with increasing drop volume and this maximum height is recorded. Details of the method of obtaining the tension can be found in ref. 19 and 20. The solid surface used in this work was polished glass on which the oil drops under water showed a contact angle of 170-180".The accuracy of the tension values was estimated to be ca. 5% for systems with a density difference of ca. 0.05 g and it was checked that literature tension values could be reproduced within the estimated accuracy. Ancillary density measurements were made using a Paar DMA 55 instrument. Equilibration of phases was achieved by shaking samples of known composition in stoppered glass tubes followed by centrifugation at ca. 1500 rpm in a Denley BR401 ther-mostatted centrifuge for 1 h. The AOT concentrations in equilibrated phases were determined using a two-phase titra- tion method" with Hyanine 1666 as titrant. Water concen- trations were measured by Karl Fischer titration using an automated Baird and Tatlock AF3 titrator.Chloride ion con- centrations were determined by Mohr titration.22 Results and Discussion Properties of the Pure Phthalate Oils Table 1 shows the measured oil/air surface tensions and den- sities for a range of dialkyl phthalate oils. The oil density matches that of water at a chain length of about six carbon atoms. The density difference between oil and water can be increased by the use of D20 for which the densityI8 at 25 "C is 1.1044 g ~m-~. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Adsorption of Phthalate oils from Dilute Solution in Heptane onto the Heptanewater Interface Adsorption of two of the dialkyl phthalates (DBP and DOP) was measured in order to compare the surface activity of the phthalate oils with that of previously studied species such as alcohols (which can operate as cosurfactants) and with other moderately polar materials such as non-aromatic esters.Adsorption was determined by measuring the heptane solution/water interfacial tension for pre-equilibrated mix- tures prepared by shaking water with heptane containing dif- ferent quantities of either DBP or DOP. It was necessary to check whether the phthalate oils are distributed into the water phase to a significant degree. This was done by measuring the oil/water tension for samples pre- pared by equilibrating water with different volumes of oil containing a constant initial concentration of phthalate oil. If distribution occurs, it leads to a depletion of the oil phase concentration for low oil :water volume ratios and a con- comitant increase in the equilibrium tension.Thus, if the tension is observed to be independent of the oil: water volume ratio it can be concluded that loss to the water phase in this system is insignificant. Fig. 1 shows a plot of tension us. heptane :water volume ratio of a constant initial oil phase concentration of DBP of 7.5 mmol dm-3 (=0.107 mol%). No significant variation of tension is observed and it can be esti- mated that the distribution coefficient P of DBP between heptane and water (defined as P = [DBP],,,,,/[DBP],,,,,,,, , concentrations in mol dm-3) must be 50.2. Data for DOP (not shown) show a similar independence, and hence the equilibrium phthalate concentration can be reliably equated with the (known) initial value for both DBP and DOB under the conditions of the experiments.Fig. 2 shows the variation of the heptane/water tension with ln(phtha1ate mole fraction) for both DBP and DOP; the heptane :water volume ratio was 20 : 1. The solid lines show the best-fit polynomial of third order in each case. If it is assumed that the phthalate solutions in heptane behave ideally in the concentration ranges studied, the area occupied per phthalate molecule A can be obtained as a function of bulk concentration using the appropriate form of the Gibbs equation, A = -kT/(dy/dlnx) (1) where y is the tension, k is Boltzmann's constant, T is the absolute temperature and x is the mole fraction of solute.The gradients of the plots (dyldlnx) were obtained by differentia- 0 5 10 15 20 n-heptane : water volume ratio Fig. 1 Variation of heptane/water tension with heptane :water ratio for a constant initial concentration of 0.107 mol% DBP in heptane at 25 "C J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 r I E z E--. C .-In 0)c (0.-50 C .-t In [mole fraction (%)I 52 I 1 (b) II E z E--. C .-0 cn 0)* -([1.-0 't c .-In[mole fraction (%)I Fig. 2 Heptane/water interfacial tension vs. ln[mole fraction (%)Iof (a)DBP and (b) DOP in heptane at 25 "C tion of the best-fit polynomial equations. The values of surface pressure Il (i.e. the lowering of surface tension caused by adsorption) and A derived from the data of Fig.2 were plotted according to the Schofield-Rideal surface equation23 of state (Fig. 3), which is n(A-A,) = /3kT (2) where A, is the limiting area at high surface pressure and /3 is a constant reflecting intermolecular interactions within the adsorbed monolayer. Eqn. (2) reduces to the Volmer surface equation of state when /3 = 1. This corresponds to a two-dimensional surface gas-like film in which the adsorbed mol- ecules show only short-range excluded area interactions, i.e. they behave as 'hard-disks'. Values of /3 < 1 signify the pres- ence of attractive interactions within the film.24 The tendency for the phthalate oils to be adsorbed can be quantified by estimating the standard Gibbs energies of adsorption (A, po) derived from the limiting low concentration slopes of plots of surface pressure us.solute mole fraction. A, po = RT ln(x/ll) (3) Aapovalues quoted here refer to standard states of unit mole fraction for the solutions and II = 1 mN m- ' for the surface. Values of A,, /3 and Aapo for the two phthalate oils are shown in Table 2. The values of the limiting areas A, appear reasonable for species with two alkyl tails. For DOP, /3 z 1, showing that Table 2 Parameters for the Schofield-Rideal surface equation of state for the adsorption of phthalate oils from dilute solution in heptane to the heptane/water interface at 25 "C oil A,/nm B Aa po/kJ mol -DBP 0.87 0.77 -23.9 DOP 0.77 1.08 -25.2 2745 1.6-1.4-1.2-1.0-E r I z E 0.8-\ h k! 1 Fv -0.6 -0.4 -0.2 //I I I I I 0 1 2 3 4 5 6 area occupied per phthalate molecule at the heptane/water interface/nm2 Fig.3 Variation of l/n us. A for DBP (0)and DOP (0)mono-layers at the heptane/water interface at 25 "C this molecule follows approximately the Volmer surface equa- tion of state. For DBP, /3 < l, indicating that there is a sig- nificantly longer range intermolecular attraction between adsorbed short-chain phthalate molecules in the monolayer. The values of standard Gibbs energies of adsorption are similar for both phthalate chain lengths and can be compared with the following values for adsorption from dilute solution in alkanes for a range of polar oils.(i) The standard Gibbs energy of adsorption for linear alcohols adsorbing from alkane solution to the alkane/water interface is constant ( & 0.5 kJ mol- I) for different temperatures and chain lengths of the alkanes and alcohok2' For the puuposes of this dis- cussion, we quote an average value representative of alcohols of -23.5 kJ mol-'. (ii) The value for methyl dodecanoate adsorbing from dilute solution in octane to the octane/water interface26 at 30°C is -16.4 kJ mol-'. (iii) Diethyl esters of the dicarboxylic acids of structure (CH,),(CO,H), for n = 2, 3 and 8 at 20 and 30°C give a constant value for the standard Gibbs energy of adsorption from solution in octane26 of -25.8 kJ mol-'. It can be seen that phthalate diester surface activity is greater than that of methyl dodecanoate and is comparable to that of non-aromatic diesters and (coincidentally) similar to that of n-alcohols.Note that these data provide a compari- son of the relative extents of adsorption into dilute films from solution in apolar alkanes in the absence of surfactant. The relative adsorptions into a concentrated film of a convention- al surfactant from the pure polar liquids can be expected to be very different. AOT Adsorption at the Phthalate Oilwater Interface In this section we compare the adsorption of AOT at the phthalate oil/water interface with that determined previously at the alkane/water interface. Plots of phthalate oil/water interfacial tension us. equilibrium AOT concentration in either the aqueous or oil phases are shown in Fig.4 and 5 for J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 .; E.. 15 C tl 5 10.-f0 .-E5 0 In( [AOT] .&no1 dm-3 .-5 -16 -14 -12 -10 -8 -6 4 In( [AOT]oi,/mol dm-3) Fig. 4 Variation of y with ln([AOT]/mol dm-3) with DBP as oil. (a) and (b)refer to the equilibrium AOT concentrations in the water and oil phases, respectively; [NaCl] = 0 (O), 0.5 (0)and 1(@) mol dm-3. The arrows in (b)mark the values for c.a.c.,,, . .-0 rn Q)c -16 -14 -12 -10 -8 -6 -4 L.-In([AOT]oi,/mol dm-3) Fig. 5 Variation of y with ln([AOTJ/mol dm-3) with DOP as oil. (a) and (6) refer to the equilibrium AOT concentrations in the water and oil phases, respectively; [NaCl] = 0 (O), 0.5 (0)and 1 (0)mol dm-’. The arrows shown in (b)mark the values of c.a.c.,,, .DBP and DOP, respectively. The break points mark the criti- cal aggregation concentrations (c.a.c.) in the aqueous and oil phases. Values of c.a.c. for the different NaCl concentrations and oils are collected in Table 3. For water, the slopes of the plots for concentrations below the c.a.c. can be analysed using the Gibbs equation to yield areas per AOT molecule at the oil/water interface. For AOT (a 1 :1 electrolyte) in water containing a 1 :1 inert electrolyte such as NaCI, the appropriate form of the Gibbs equation is: A = -kT[x + 2(8 lnf*/a In mD),,,]/(dy/d In mD) (4) where mD and m, are the respective molarities of the sur- factant anion and added salt (NaCI in this case) andf, is the mean ionic activity coefficient of the surfactant.The factor x = 1 + [mD/(mD + m$] varies between two (for zero added salt) and one (for [NaCl] % [AOT]).27 It has been shown previously that the activity coefficient term is eqn. (4) is negli- gibly small when the total ionic strength is small and when [NaCl] % [AOT].7 Thus the term involvingf, is negligible for all systems considered here. The limiting values of A for AOT (obtained at concentrations close to the c.a.c. where the monolayers become ‘close-packed’) are given in Table 3. The variation of the limiting areas with aqueous-phase NaCl concentration can be compared with the behaviour of AOT-alkane-water-NaC1 systems. For alkane systems, the areas decrease with increasing [NaCl] until a certain NaCl concentration is attained after which the areas remain con- stant. According to the simple geometrical ideas discussed by Mitchell and NinhamYZ8 and developed further by Aveyard et aLY8the area at high [NaCl] should be governed (perhaps mainly) by the size of the surfactant tail region including any penetrating oil solvent.Thus, a high limiting area at high salt concentrations for a particular oil implies that oil penetrates the surfactant monolayer to a high degree. Fig. 6 shows the limiting areas of AOT us. [NaCI] for DBP and DOP as oil phases. Although only three area values have been measured, the areas initially fall with increasing [NaCl], but attain con- stant values at high [NaCI], as seen for alkane systems.The limiting values of 0.9 and 0.7 nm2 for interfaces involving DBP and DOP, respectively, can be compared with areas of 0.73 nm2 determined for heptane and 0.61 nm2 for hexa- decane.’ Thus, increasing the chain length of the phthalate oils leads to a smaller area consistent with decreased pen- etration of the AOT monolayer by the longer chains of the oil molecules. Fig. 4(b) and 5(b) show the variation of the oil/water tension with the oil-phase concentrations of AOT. No plots of this type are shown for zero NaCl concentration as no AOT was detected in the oil phase. The plots involving oil- phase concentrations show less-marked break points at the c.a.c. than the corresponding water-phase concentration plots.The oil-phase concentrations which are present in equi- librium with the c.a.c. in water (c.a.c.,,,,,) are designated c.a.c.,,i, and increase with increasing aqueous-phase NaCl concentration (Table 3). If the activity coefficient of AOT in the oil phase remains invariant with AOT concentration [i.e. (a Inf,/a In mD),,,, = Table 3 Summary of parameters for the adsorption of AOT from dilute solution at the phthalate oil/water interface at 25 “C ’ ~ ~~ oil [NaCl]/mol dm - c.a.c.,,,,,/mmol dm - c.a.c.,i,/mmol dm - A/nmz DBP 0 2.5 0 1.23. 0.5 0.075 0.75 0.88 1.o 0.055 18 0.90 DOP 0 3.0 0 1.68 0.5 0.075 1.o 0.70 1 .o 0.045 3.O 0.70 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90I1.4 , I I 1 I 0 0.2 0.4 0.6 0.8 1 .o [NaCI]/mol dm-3 Fig. 6 Variation of the limiting areas per AOT molecule with WaCl] for DBP (a) as oiland DOP (0) 01, the slopes of the oil-phase plots should be the same as those of the water-phase plot because the area per AOT derived from each data set must have the same value.At the c.a.c. in each phase, the oil-phase slope is CQ. 0.5 that of the water slope for both oils at 0.5 mol dmP3 NaCl. The ratio of slopes is 0.2 for 1 mol dm-3 NaCl for both oils. It can be seen that the oil plot slopes are much lower than for the cor- responding water plots and hence [from eqn. (4)] the value of (aIn f*/a In mD), for the oil phase must be significantly less than zero.Such a decrease in the activity coeficient of the surfactant in the oil phase with increasing concentration is consistent with a progressive aggregation of the surfactant in the oil at concentration less than the c.a.c.,,. The non-ideal behaviour in the oil phase increases with increasing NaCl concentration. This type of pre-c.a.c. aggregation has been observed previously for AOT in systems containing non-aqueous polar solvents with toluene as the oil.' We now consider the variation of the system behaviour with aqueous phase NaCl concentration for AOT concentra- tions above the c.a.c. in mixtures containing comparable volumes of oil and water. With alkanes as the oil the behav- iour is as follows. At low salt concentrations such systems form two phases consisting of an oil-in-water (o/w) micro- emulsion phase in equilibrium with an excess oil phase (a Winsor I system).At high salt concentrations a Winsor I1 system consisting of a water-in-oil (w/o) microemulsion coexisting with an excess aqueous phase is formed. At inter- mediate salt concentrations a three phase Winsor 111 system is formed consisting of a surfactant-rich phase coexisting with excess water and oil phases. The phase progression from Winsor I to Winsor I11 to Winsor I1 is called microemulsion phase inversion and corresponds to a progression in aggre- gated surfactant monolayer curvature from positive in o/w microemulsions, through zero in the Winsor 111 range, to negative in w/o microemulsions. The tendency of the sur-factant monolayer to curve depends on the effective geometry of the surfactant in the monolayer which, in turn, depends on surfactant molecular structure, solvent penetration into the head and tail regions and the strength of electrostatic inter- actions between adjacent surfactant headgroups.Phase inver- sion can be achieved by the variation of any solution parameter which alters the tendency of the monolayer to curve. For example, the addition of salt drives the phase pro- gression towards the Winsor I1 system by screening the elec- trostatic repulsion between headgroups and thus shrinking their effective size. For surfactant concentrations greater than the c.a.c., the interfacial tension of the plane oil/water interface separating the bulk oil and water phases (7,) passes through a minimum value when the system is driven through a phase inversion. The minimum tension occurs within the Winsor I11 range and hence is found under conditions when the preferred sur- factant monolayer curvature is close to zero.The variation of yc with aqueous-phase [NaCl] for DBP and DOP as oil is shown in Fig. 7. The tension passes through a minimum in each case at a salt concentration of ca. 0.5 mol dm-3 NaCl. In comparison, the [NaCl] required for phase inversion of AOT with n-alkanes as oil varies from cu. 0.05 mol dm-3 NaCl for heptane to ca. 0.1 mol dm-3 NaCl for tetradecane.8 This variation for alkanes has been rational- ised on the basis that short-chain alkanes penetrate and swell the tail regions of surfactant monolayers to a greater extent than long-chain alkanes.Hence, both dialkyl phthalates behave like very long chain length (non-penetrating) alkanes in terms of the [NaCl] required to achieve phase inversion. Phase Compositions of Equilibrated AOT-W a t er-Dial kylpbthala te Oil Mixtures Equilibrium distributions of AOT between water and phthal- ate oil phases (DBP and DOP) for AOT concentrations less than the c.a.c. are shown in Fig. 8. No results are shown for zero [NaCl] since 5 mol dm-3 AOT (the detection n [NaCl]/mol dm-3 0.2 (b) [NaCl]/mol dm-3 Fig. 7 Variation of log(y,/mN me')with aqueous phase [NaCl] for (a)DBP and (b) DOP as oil J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 h mI E 0.1 0-zE-.. m i=09 v 0)-0.0' c 11 0.1 1 10 100 log( [AOT]oi,/mmol dm-3) Fig. 8 Equilibrium distributions of [AOT] between water and the phthalate oils for AOT concentrations less than the c.a.c. Note the logarithmic scales. The solid line shows a slope of unity. (a)DBP and (0)DOP (0.5 mol dm-3 NaCl), (W) DBP and (0)DOP (1 mol dm -NaCl). limit of the surfactant titration method21 was found in either of the equilibrated phthalate oil phases. The distribution of a solute may be described by the dis- tribution or partition coefficient P equal to the ratio of equi- librium solute activities in the oil and water phases, eqn. (9, P = Cs~ecieslwaterfwater/C~~e~iesloilf,il (5) where f refers to the activity coefficient and the subscripts refer to the phase. The equilibrium concentration of the species in oil is directly proportional to the concentration in water only if the ratio of activity coefficients in both phases remains constant over the range of concentrations tested.For such a situation, a plot of log[spe~ies]~,,,~ isus. log[specie~]~~, linear and of unit slope (see Fig. 8). For the aqueous phases in the present work, the concentrations of AOT are all < mol dm-3 and hence f,,,,, should be constant for a particular NaCl concentration. If it is assumed that all the deviations from the unit slope behaviour in the data of Fig. 8 are due to changes only inLi, (as for the interpretation of the oil/water tension data), then the results can be used to give a crude estimate of the decrease in activity coefficient of AOT in the oil as the AOT concentration is increased up to the c.a.c.in the oil. It can be seen that the curves for the different oils overlap, but that the 1.0 mol dm-3 NaCl data deviate more from ideal behaviour than the 0.5 rnol dme3 data. For both DBP and DOP as oil, the activity coefficient of AOT decreases ca. five-fold for 0.5 mol dm-3 NaCl and ca. 100-fold for 1.0 mol dm-3 NaCl as the concentration is increased in the c.a.c. in the oil phase. This variation suggests that AOT undergoes a progressive aggregation in the oil phase reaching aggregation numbers in the order of five and 100 for the two different salt concentrations before attaining the c.a.c.at which microemulsion aggregates are formed. Thus, the pre- c.a.c. partitioning behaviour strengthens the qualitative con- clusion concerning non-ideality drawn from the interfacial tension data. For AOT concentrations much larger than the c.a.c., the equilibrium distributions of total (monomer plus aggregated) AOT between the water, oil and third phase (where present) as a function of aqueous phase [NaCl] for both phthalate oils are shown in Fig. 9. For zero [salt], virtually all the AOT IOOr [NaCl]/mol dm-3 2ot t I In 0.5 1.o 1.5 :0 [NaCl]/mol dm -3 Fig. 9 Equilibrium distribution of total AOT between water (O),oil (0)and third phase (0)(where present) as a function of aqueous phase WaCl] for (a) DBP ; (b)DOP is in the aqueous phase, hereas it is all in the oil phase for salt concentrations 21 mol dmV3.Over a range of NaCl concentrations in the region of 0.5 mol dm-3, the AOT is located mainly in a third phase. Thus it can be seen that the post c.a.c. AOT distribution follows the pattern expected for the Winsor 1-111-11 phase progression and that micro-emulsion phase inversion is centred around 0.5 mol dm-, NaCl for both phthalate oils, as indicated also by the varia- tion of yc with [NaCl]. The compositions of the equilibrim surfactant-rich third phases were measured as a function of [NaCl] over the Winsor I11 range. The volume fractions of AOT and water were obtained by Hyamine and Karl Fischer titration, respectively. The oil volume fraction was then estimated by difference.For DBP as oil, the oil volume fraction remains virtually constant at 0.1 over the range of salt concentration from the Winsor-I11 transition to the Winsor 111-11 bound-ary. Over the same salt concentration range, the water volume fraction decreases from 0.73 to 0.30. For DOP as oil, the water volume fraction decreases from 0.55 to 0.20, whilst the oil volume fraction increases from 0.08 to 0.45. Hence DBP is not taken up into third phases to any great extent whereas DOP is extensively solubilised at salt Concentrations close to the Winsor 111-11 boundary. For both oils, the sur- factant volume fractions in the third phases are relatively high (in the range 0.4-0.6). These composition changes in three-phase systems are significantly different to those in systems containing non-polar There are a number of different possible microstructures for surfactant-rich phases of Winsor I11 systems which all have the average surfactant monolayer curvature is ca.zero. Possible microstructures include lamellar liquid-crystalline phases, bicontinuous microemulsions or L, phases.30 All samples of third phases prepared with either DBP or DOP were observed to be optically isotropic and hence are not J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 lamellar phases. For bicontinuous microemulsions, the phase composition is related to the microstructural repeat distance 5 according to eqn. (6),31,32 < = constant 4oil~,,,,,/[surfactant]A (6) where $oil and are the volume fractions of oil and water, respectively, [surfactant] is in units of the number of molecules per unit volume and A is the area occupied per surfactant molecule at the interface.Depending on the exact model assumed for the microstructure, the numerical con- stant can take values of 4,33634 or 5.82.35For the purposes of this discussion we have assumed a value of 6. For both DBP and DOP the volume fraction of AOT ($AOT) in the third phase is of comparable magnitude to $oil and $,at,,. In order to proceed further with the approximate estimation of 5 the oil and water volume fractions have each been taken to include half 4AOT, i.e. 4Later = +water + &$AOT and 6bil = 4oil + $$AoT. If the values of 5 and A are constant throughout the range of Winsor I11 microemulsions, then a plot of [surfactant] us.the product ($bil 4kater)should be linear with the slope equal to (6/5A).The plots for DBP and DOP (Fig. 10) are approximately linear. Taking average values of A for AOT over the range of [NaCl] of 0.9 nm2 in systems with DBP and 0.7 nm2 in the case of DOP (see Table 3), 5 values of ca. 1 and 5 nm are obtained for the DBP and DOP systems, respectively. Owing to the crude approximations used, these values should be viewed only as an indication of the order of magnitude of 5. It has been shown the~retically~~ and experimentally for a variety of microemulsion systems37 that the minimum post- c.a.c. interfacial tension y, (min) scales approximately with the inverse square of the characteristic size of the third-phase microemulsions.To a zeroth-order approximation : y,(min) = k~/r~ (7) Using eqn. (7) and the values of y,(min) (0.67 and 0.48 mN m-' for the DBP and DOP systems, respectively), the values I I I I I0.5 $Later 4bil Fig. 10 Variation of [AOT] with ($Li,$kater) for third phases with DBP (0)and DOP (0)as oil of t are estimated to be 2.5 and 3.0 nm. Thus, the values of the minimum tensions are broadly consistent with the magni- tudes of the calculated repeat distances in the third phases. The partitioning of chloride ion between the bulk aqueous phase and the dispersed water of the third phases was mea- sured by Mohr titration of both the aqueous and third phases.The results can be conveniently represented in terms of the ratio of chloride ion concentration per unit volume of dispersed water in the third phase to that in the aqueous phase (Pdwlaq{[Cl-] in the dispersed water/[Cl-] in the = aqueous phase)). For both DBP and DOP as oil, the value of Pdwlaqdecreases from ca. 0.7 at 0.55 mol dm-3 NaCl to ca. 0.5 at 1 mol dm-3 NaCl. This behaviour, observed previously for AOT in systems with alkanes as oil3* and for a range of other ionic microemulsion is caused by the proximity of the charged surfactant interface which repels the C1- ions and hence causes a reduction in the concentration of NaCl in the dispersed water. For Winsor I1 systems formed at high [NaCl], the extent of water solubilisation in the w/o microemulsion phase is expected to decrease as the preferred monolayer curvature becomes increasingly negative with increasing salt concentra- tion.Fig. 11 shows the variation of R (defined as [water]/ [AOT] in the equilibrium w/o microemulsion phases of the Winsor I1 systems) with [NaCl]. As expected, the water solu- bilisation (and hence the microemulsion droplet size) decreases with increasing [NaCl]. Overall, the amount of water solubilised is rather low, corresponding to little more than the hydration requirements of the AOT molecule.42 In this case, the aggregates formed are probably better con-sidered as hydrated reversed micelles rather than w/o micro- emulsions. Uptake of Phtbalate oils into AOT Monolayers When added to the surface of an aqueous surfactant solution, a small quantity of a non-spreading oil will form a lens in I 12 10 R 8 6 I I ,4 1.o 1.5 2.0 2.5 3.0 [NaCl]/mol dm-3 Fig.11 Variation of molar ratio of solubilised water to AOT (R) vs. [NaClJ for Winsor I1 microemulsions with DBP (0)and DOP (0)as oil J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 4 neptane/water tensions for the addition of DBP and DOP to Winsor I and Winsor I1 systems containing AOT (2.3 mmol dmP3), water, NaCl and heptane at 25 "C [phthalate oil]/mmol dm-3 DBP Winsor I 0 1.9 9.4 Winsor I1 0 1.9 9.4 DOP Winsor I 0 1.3 Winsor I1 6.3 0 1.3 6.3 equilibrium with a mixed monolayer containing both oil and s~rfactant.~~?~~The adsorption of oil into the monolayer results in a decrease in the surface tension by an amount Ay.The surface excess concentration of the oil rOi,in the mixed monolayer is given by the following equation. roil= Ay/kT (8) Eqn. (8) is accurate insofar as Ay is equal to the differential quantity (dAy/d In uoiJ in the limit of the activity of the oil aoiltending to unity. This assumption is true to within ca. 15% for alkane Values of Ay for dialkyl phthalate oils of chain lengths ranging from ethyl to dodecyl were measured for aqueous AOT solutions above the c.m.c. in the absence of NaCl. Ay was found to be <0.1 mmol dm-3 m-' (equal to the accu- racy of the measurements) in every case indicating that little oil penetration of AOT monolayers occurs under these condi- tions.A similar lack of penetration by long-chain alkanes into AOT monolayers has been noted previ~usly.~~The experiment reported here therefore shows that the extent of penetration is not increased to any large extent for the mod- erately polar phthalate oils. Possible coadsorption and cosurfactant properties of the phthalate oils DBP and DOP were examined in the following way. As shown in ref. 8, the addition of strongly adsorbing consurfactant species such as alcohols to oil-water mixtures containing an AOT concentration in excess of the c.a.c. can result in a minimum oil/water interfacial tension under appropriate conditions. For an initial Winsor I system, addi- tion of a cosurfactant which increases the negative curvature of the monolayer will yield a tension minimum whereas a cosurfactant favouring increased postive curvature will cause phase inversion only for an initial Winsor I1 system.The possible cosurfactant effect of DBP and DOP was tested by measuring the heptane/water tension for different phthalate oil concentrations in heptane-water-NaC1-AOT mixtures forming either a Winsor I or Winsor I1 system in the absence of phthalate oil. The results are shown in Table 4; the addition of both phthalate oils to the heptane systems always results in an increase in tension for both Winsor I and I1 systems consistent with the fact that the curves of yc us. [NaCl] for DBP and DOP are shifted to higher tensions relative to the curve for heptane.It can be concluded that DBP and DOP do not act as strongly adsorbing cosurfac- tants in the AOT-alkane system. The lack of penetration of concentrated AOT monolayers by the phthalate oils is con-sistent with the observation that a relatively high NaCl con- centration (compared with the concentrations required for alkanes as oils) is required to phase invert the phthalate oil systems. [NaCl]/mol dm- yJmN m-' 0.0 174 0.35 0.0174 0.39 0.0174 0.48 0.0685 0.14 0.0685 0.16 0.0685 0.25 0.0174 0.35 0.0174 0.37 0.0174 0.35 0.0685 0.14 0.0685 0.21 0.0685 0.32 Conclusions The main conclusions of this study can be summarised as follows. (i) Phthalate oils are adsorbed at the heptane/water interface from dilute solution in heptane.The surface activity for dilute monolayers is typical of diesters and similar to that of alcohols. (ii) AOT is adsorbed at the phthalate oil/water interface to yield saturated monolayers of similar surface con- centrations those found for alkane oils. (iii) At concentrations below the c.a.c., AOT partitions from water to phthalate oils to a significant extent at high salt concentrations. (iv) AOT aggregates progressively in solution in the phthalate oils at concentrations below that required for micromulsion forma- tion. (v) Systems containing AOT and dialkyl phthalates undergo the Winsor 1-111-11 microemulsion phase inversion progression with the addition of NaCl. Phase inversion is centred at ca. 0.5 mol dm-3 NaCl for both DBP and DOP. (vi) The characteristic domain size in the isotropic third phases and the maximum water solubilisation in w/o micro- emulsions are both relatively small (and the minimum oil/ water tension is correspondingly high) for systems with phthalate oils as compared with short chain-length alkane systems. (vii) Phthalate oils appear to show little tendency to penetrate concentrated AOT monolayers at both the oil/ water and air/water interfaces.This is interesting since the phthalates have similar surface activities (from dilute solution in alkane oils) as alcohols and the latter are very effective cosurfactants. We wish to thank Kodak Ltd. (Harrow) and the SERC for the provision of a CASE Studentship to P.A.K. References 1 K.Shinoda and S. Friberg, in Emulsions and Solubilisation, Wiley ,New York, 1986. 2 R. Aveyard, Chem. Ind., 1987,474. 3 D. Langevin, Acc. Chem. Res., 1988,21,255. 4 M. Kahlweit, R. Strey, P. Firman, D. Haase, J. Jen and R. Scho-macker, Langmuir, 1988,4,499. 5 M. Kahlweit, R. Strey, R. Schomacker and D. Haase, Langmuir, 1989,5, 305. 6 A. Martino and E. W. Kaler, J. Phys. Chem., 1990,94, 1627. 7 R. Aveyard, B. P. Binks, P. D. I. Fletcher, A. J. Kirk and P. Swansbury, Langmuir, 1993,9,523. 8 R. Aveyard, B. P. Blinks and J. Mead, J. Chem. SOC., Faraday Trans. I, 1986,82,1755. 9 M. Kahlweit, R. Strey and G. Busse, J. Phys. Chem., 1991, 95, 5344. 10 S. E. Friberg and L. Gan-Zuo, J. SOC. Cosmet. Chem., 1983, 34, 73. 11 H.Kunieda, H. Asaoka and K. Shinoda, J. Phys. Chem., 1988, 92, 185. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2751 12 13 14 15 J. Alander and T. Warnheim, J. Am. Oil Chem. SOC., 1989, 66, 1656. K. Wormuth and E. W. Kaler, J. Phys. Chem., 1989,93,4855. M. El-Nokaly, G. Hiler, Sr. and J. McGrady in Microemulsions and Emulsions in Foods, ed. M. El-Nokaly and D. Cornell, ACS Symposium Series 448, Washington DC, 1991, ch. 3. S. Mukherjee, C. A. Miller and T. Fort, J. Colloid Interface Sci., 29 30 31 32 0. Abillon, B. P. Binks, C. Otero, D. Langevin and R. Ober, J. Phys. Chem., 1988,92,4411. H. Kellay, Y. Hendrikx, J. Meunier, B. P. Binks and L. T. Lee, J. Phys. 11, 1993,3, 1747. L. Auvray, J. P. Cotton, R. Ober and C. Taupin, J. Phys. (Paris), 1984,45,913.B. P. Binks, J. Meunier and D. Langevin, Prog. Colloid Polym. 1983,91, 223. Sci., 1989,79,208. 16 B. W. Ninham, S. J. Chen and D. Fennel1 Evans, J. Phys. Chem., 33 P. Debye, H. R. Anderson Jr. and H. Brumberger, J. Appl. Phys., 1984,88,5855. 1957,28,679. 17 B. P. Binks, H. Kellay and J. Meunier, Europhys. Lett., 1991, 16, 34 J. Jouffroy, P. Levinson and P. G. de Gennes, J. Phys. (Paris), 53. 1982,43,1241. 18 Solute-Solvent Interactions, ed. J. F. Coetzee and C. D. Ritchie, Marcel Dekker, New York, 1969, p. 393. 35 36 Y. Talmon and S. Prager, J. Chem. Phys., 1978,69,2984. M. Kahlweit and H. Reiss, Langmuir, 1991,7, 2928. 19 20 21 22 23 24 25 J. F. Padday and A. Pitt, Proc. R. SOC. London, A, Math. Phys. Sci., 1972,329,421.P. A. Kingston, Ph.D. Thesis, University of Hull, to be published. V. W. Reid, G. F. Longman and E. Heinerth, Tenside, 1967, 4, 292. A. I. Vogel, in A textbook of Quantitative Inorganic Analysis, Longman, London, 5th edn. 1989. R. K. Schofield and E. K. Rideal, Proc. R. SOC.London, A Math. Phys. Sci., 1925, 109, 57. R. Aveyard and D. A. Haydon, in An Introduction to the Prin- ciples of Surface Chemistry, Cambridge University Press, Cam- bridge, 1973, ch. 1. R. Aveyard and B. J. Briscoe, J. Chem. SOC.,Faraday Trans. I, 37 38 39 40 41 42 43 44 R. Aveyard, B. P. Binks, S. Clark, P. D. I. Fletcher, H. Giddings, P. A. Kingston and A. Pitt, Colloids Surf., 1991,59,97. P. D. 1. Fletcher, J. Chem. SOC.,Faraday Trans. 1, 1986,82,2651. P. L. de Bruyn, J. Th. G. Overbeek and G. J. Verhoeckx, J. Colloid Interface Sci., 1989, 127, 244. S. Levine and K. Robinson, J. Phys. Chem., 1972,76,876. J. Biais, M. Barthe, M. Bourrel, B. Clin and P. Lalanne, J. Colloid Interface Sci., 1986, 109, 576. C. A. Martin and L. J. Magid, J. Phys. Chem., 1981,85,3938. R. Aveyard, P. Cooper and P. D. I. Fletcher, J. Chem. SOC., Faraday Trans., 1990,86,3623. J. R. Lu, R. K. Thomas, R. Aveyard, B. P. Binks, P. Cooper, P. D. I. Fletcher, A. Sokolowski and J. Penfold, J. Phys. Chem., 1992,%, 10971. 1972,68,478. 26 27 R. Aveyard and J. Chapman, Can. J. Chem., 1975,53,916. D. G. Hall, B. A. Pethica and K. Shinoda, Bull. Chem. SOC.Jpn., 1975,48,324. 28 J. D. Mitchell and B. W. Ninham, J. Chem. SOC.,Faraday Trans. 2, 1981, 77, 601. Paper 41029576; Received 18th May, 1994
ISSN:0956-5000
DOI:10.1039/FT9949002743
出版商:RSC
年代:1994
数据来源: RSC
|
27. |
Solutions of aluminium in liquid lithium: electrical resistivity of liquid alloys |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2753-2755
Richard J. Pulham,
Preview
|
PDF (396KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2753-2755 Solutions of Aluminium in Liquid Lithium :Electrical Resistivity of Liquid Alloys Richard J. Pulham," Peter Hubberstey and Petra Hemptenmacher Chemistry Department, Univeristy of Nottingham, Nottingham, UK NG7 2RD The electrical resistivity of solutions of aluminium (up to 15.76 mol% Al) in liquid lithium has been measured by a four-point capillary method from the melting point of lithium, 180.5"C,to 440"C.Aluminium increases linearly the resistivity of liquid lithium by a relatively large amount; 5.9 x lo-' Q m (mol% Al)-' at 320°C to 5.7 x lo-' Q m (mol% AI)-' at 440°C.These relatively large values make the resistivity technique potentially useful for monitoring the concentration of aluminium during chemical reactions in these metallic solutions.The use of electrical resistivity measurements has played a part for a long time in following the chemical reactions of salts and metals when dissolved in the liquid alkali metals. This can be seen in the reactions of the salts Li2C2 and Li3N in liquid lithium to form Li2NCN in the crystalline state,' and the reaction of solutions of silicon in liquid lithium with gaseous nitrogen2 to form Li,SiN, . Moreover, the absolute value of the resistivity provides some help in identifying the nature of the solute species in solution. Correlations between resistivity and solute solvation energy3 and solute core potential/size4 have been developed to rationalise trends in the resistivity increases caused by a variety of solutes in liquid lithium.More recently there has been a role reversal in that it has been the more reactive lithium that has been titrated5 out of liquid Pb-Li (17 mol%) by nitrogen, hydrogen, oxygen and water vapour using an electrical resistivity monitor6 specifi- cally developed for continuous measurement of the lithium concentration in this dense eutectic alloy. This liquid is a potential coolant/tritium-breeder for some nuclear fusion reactor designs. Similarly, should the need arise to plate steel with aluminium compounds at high temperature to reduce tritium diffusion in such systems, then this might be done in situ from liquid lithium selutions using resistance monitoring. This paper measures the increase in resistivity of lithium on adding aluminium to study further the reactions of this Group 13 element in this unusual solvent.Experimental The basic apparatus and procedure for the determination of the resistivities of metallic solutions have been described in detail elsewhere.' It was from this design and accompanying procedure that all the many other versions, including the industrial monitor, are derived. The present work, however, incorporated two specific additional features : a secondary liquid-metal pump to ensure homogeneity of the alloys (aluminium is about 5.5-fold more dense than lithium and could take up to 40 min to dissolve) and a multisample inser- tion device both to hasten the procedure and avoid the chance of contamination from repeated loadings in a glove box.The apparatus is shown in Fig. 1. Briefly, the bulk of the liquid alloy (40 g, ca. 80 cm3) was contained in a cylindrical steel (AISI 321) reservoir, V (diam. 25 mm, height 100 mm) to which was attached a thermocouple pocket, TC, and two vertical (to prevent gas pockets) electromagnetic pumps, P1 and P2, for mixing the alloy by circulation and for contin- uous sampling of the alloy into a capillary (0.d. 3 mm, i.d. 1 mm) for resistance measurement, respectively. Both pumps drew liquid from the bottom of the reservoir and returned it to the top. The capillary loop was short-circuited by two discs, S, so that the resistance of the intervening section, R, (length 160 mm) could be measured by means of silver electri- cal leads, two on each disc, through which a constant current (ca.3 A) was passed. The resistance of the capillary and alloy were determined by measurement of the potential difference across the filled capillary and across a standard resistance (0.01 Q) connected in series. These resistances were converted into resistivities, p, for the alloy from a knowledge of the resistance of the empty capillary and its known dimensions. All necessary calculations and calibrations are described else- where.7 After calibration and filling with lithium (Koch-Light, 99.98%;ca. 30 g) in a glove box under argon, the top of the reservoir was sealed to a glass system to allow both evac- uation of gas through tap T1 and addition of consecutive, known weights of aluminium (Goodfellow, 99.999%; 1 mm thick foil, ca.1 g) through a greaseless tap, T2, from a rotat- able glass dispenser, D. The metal section of the apparatus was enclosed in a thermostatically controlled, fan-assisted, air oven. Two independent sets of experiments were performed. In the first set, successive weights of aluminium were added to lithium to give compositions of 0.98, 2.86, 4.76, 6.63, 8.46, 10.31, 12.19, 14.26 and 15.67 mol% Al. In the second set, the compositions were 1.15, 2.04, 2.97, 3.74, 4.56, 5.39, 6.34 and D T1 P1 7 P2 Fig. 1 Resistivity apparatus 2754 7.12 mol% Al. For each composition the solution was cooled and the resistivity was determined at specific temperatures.Results Experiment 1 The effect of decreasing temperature on the resistivity of the various compositions is shown in Fig. 2 for experiment 1. The resistivities are equilibrium values because on reducing the temperature, measurements were recorded only when they became constant. Each resistivity-temperature graph is of the same basic shape. For example, the homogeneous liquid 6.63 mol% A1 solution has a resistivity which decreases smoothly from 71 x lo-' R m at 460°C to 68 x lo-' R m at 324 "C; below this temperature the resistivity would no longer be linear but would drop sharply due to precipitation of Li,Al, and consequent depletion of aluminium from the solution. Thus the precipitation of solids generally sets the lower temperature limit.The set of near horizontal lines are thus the resistivities of the various homogeneous solutions. The data are summarised in Table 1 as equations portraying the temperature dependence of the resistivity. Experiment 2 The resistivities of these homogeneous solutions are again nearly parallel straight lines, and the data, in the form of 140 I -15.76120 -tct--c 14.26 --+-12.19 I ---.+-. 10.31 ---8.46p 80 MY-6.63 4.76 40 20 150 200 250 300 350 400 450 500 T/T Fig. 2 Resistivity 0s. temperature (experiment 1) for solutions of aluminium (concentrations in mol% Al) in liquid lithium Table 1 Parameters in the resistivity-temperature equations ~/10-~R m = a + bT + cT2 mol% A1 a b c R2/ofor p T/T 0.98 1.15 20.48 23.41 0.056 0.048 -3.99 -3.29 0.997/0.15 0.99910.05 189-441 213-434 2.04 2.86 2.97 3.74 4.56 4.76 5.39 28.20 29.96 36.47 40.87 46.1 1 48.15 51.67 0.051 0.064 0.029 0.030 0.029 0.027 0.026 -3.71 -4.80 1.00/0.04 0.992/0.28 0.996/0.13 0.989/0.18 0.994/0.10 0.994/0.20 0.99710.06 246-451 260-437 265-460 278-435 297-434 310-387 308-435 6.34 6.63 7.12 8.46 57.02 58.36 62.1 1 71.18 0.026 0.029 0.025 0.023 0.995/0.09 0.959/0.14 0.993/0.10 0.9 5910.1 4 318-435 326-441 325-434 340-423 10.31 12.19 80.16 92.53 0.024 0.022 0.8 39/0.2 5 0.53410.27 363-432 376-416 14.26 100.25 0.029 0.76510.21 402-432 15.75 111.12 0.026 0.74710.1 6 4 10-438 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 resistivity-temperature equations, are combined with those of experiment 1 in Table 1.Discussion Temperature Dependence of the Resistivity Over the relatively small concentration range covered in these experiments, the curvature of the resistivity-temperature lines (experiment 1) in Fig. 2 is similar (experiment 2 also) but not identical to that of pure lithium' (lowest solid line, no points) for which the resistivity is given by p/R m = 16.476 x lo-' + 4.303 x 10-"T -2.297 x 10-13T2 This can be shown mathematically at 4OO"C, for example, where 6p/6T values R m "C-') derived from the equations at the various concentrations (mol% Al) are 0.024/ 0.98, 0.022/1.15, 0.021/2.04, 0.026/2.86, 0.032/2.97, 0.029/3.74, 0.026/4.56, 0.027/5.39, 0.024/6.34, 0.024/6.63, 0.024/7.12, 0.033/8.46, 0.026/10.31, 0.022/12.19 and 0.029/14.26.The value for lithium is 0.025 x lo-' SZ m "C-' at 400°C and falls within the range of values for the solutions. There is no discernible trend in coefficient with increasing concentration. Values at other temperatures can be calculated from the data in Table 1. At a given concentration, the temperature coefficient, 6p/6T, decreases with increasing temperature (Fig. 2 and equations in Table 1). A quantitative measure can be derived from the data in Table 1. Thus, for 0.98 mol% Al, for example, 6p/6T values decrease from 0.040 x lo-' SZ m"C-' at 20°C to 0.021 x lo-' SZ m0C-l at 440°C. Values at other compositions are available from the equations (Table 1). Composition Dependence of the Resistivity The more useful parameter is the concentration dependence of the resistivity, 6p/6x, which is given by the resistivity of the various solution concentrations at a given temperature.For this purpose the results from the two experiments are com- bined in Table 2. Over the relatively small concentrations studied, there is a linear increase in resistivity with increasing concentration of aluminium, the slope of which gives the composition dependence, 6p/6x. The coefficients from 320 "C to 440°C only are shown. The values range from 5.9 x lo-' SZ m (mol% Al)-' to 5.7 x lo-' R m (mol% Al)-'. There are fewer measurements available at 260 "C, 280 "C and 300 "C, and too few at 220 "C and 240 "C to derive any reliable coeffi- cient. The linearity implies that each aluminium atom added to the solution exhibits the same extra electron scattering as its predecessor, but no trend in coefficient with increasing concentration is evident.This is not expected to hold over the entire composition range, of course, because the resistivity of alloys, at a temperature high enough to maintain homoge- neous solutions, should describe a parabola rising to a maximum between lithium and aluminium and then falling to the value for pure liquid aluminium. The slope should vary accordingly. The composition dependence is the feature that is most useful in following chemical reactions in liquid lithium. It provides a simple measure of the concentration of aluminium during, say, a titration of aluminium from the solution by nitrogen to form the insoluble compounds AlN or Li,AlN,.Similarly, a potential use of the technique might be in follow- ing the extraction of aluminium at high temperatures from the solution by steels or nickel alloys as they become plated J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2755 Table 2 Resistivities/lO-' i2 m and resistivity-composition coeff-potential and different size (volume) has on the periodic cients, (Sp/6x)/lO-' i2 m (mol% Al)-' of solutions of aluminium in potential of the lithium solvent matri~.~ The group of solutes liquid lithium Al, Ga, In and T1 are intermediate in the Periodic Table and T/OC it will be informative to determine in the future which of the ~~ two theories best fits this group. mol% A1 320 340 360 380 400 420 440 The determination of the solvation enthalpy of aluminium in liquid lithium from a thermochemical cycle3 requires a0 27.89 28.45 28.99 29.51 30.00 30.49 30.96 0.98 34.3 1 34.9 1 35.47 36.00 36.50 36.96 37.40 knowledge of both the enthalpy of solution, Aso,H, of alu- 1.15 35.40 35.93 36.43 36.90 37.35 37.77 38.16 minium and enthalpy of formation, A,H", of the phase that 2.04 40.72 41.25 41.75 42.22 42.66 43.08 43.46 precipitates, Li9A14.The gradient of the solubility plotI2 of 2.86 45.52 46.17 46.78 47.35 47.88 48.37 48.83 In xAlus. T-(xAl = mol fraction of Al) gives a value of 39.0 2.97 45.75 46.33 46.9 1 47.49 48.07 48.65 49.23 kJ mol- for AsolH, according to the ideal solution equation 3.74 50.47 51.07 51.67 52.27 52.87 53.47 54.07 4.56 55.39 55.97 56.55 57.13 57.71 58.29 58.87 In xAl = -AsolH/RT + AsolS/R4.76 56.79 57.33 57.87 58.41 58.95 59.59 60.03 5.39 59.99 60.5 1 61.03 61.55 62.07 62.59 63.1 1 but the solvation enthalpy cannot be derived until a reliable 6.34 65.34 65.86 66.38 66.90 67.42 67.94 68.46 experimental value of Af H" for Li9A14 becomes available.6.63 -68.22 68.80 69.38 69.96 70.44 71.12 Similarly, it appears that a value of 6p/6x is needed for 7.12 -70.61 71.1 1 71.61 72.1 1 72.61 73.1 1 gallium in lithium before trends in the aluminium group can 8.46 78.54 79.00 79.46 79.92 80.38 10.3 1 ---89.28 89.76 90.24 90.72 be compared with the behaviour of Group 2 elements.12.19 100.9 101.3 14.26 ----11 1.9 112.4 113.0 15.76 -----122.0 122.6 ~ ~~~~ Conclusions 6p/6~= 5.90 5.89 5.88 5.80 5.74 5.73 5.74 k 0.04 0.03 0.03 0.03 0.04 0.03 0.03 The electrical resistivity, p, of liquid lithium is substantially increased, e.g. 6p/6x = 5.7 x lo-* R m (mol% Al)-' at 400"C, on dissolving aluminium (up to 15.76 mol% and 460 "C),but the temperature coefficient of resistivity, 6p/6T, ofwith Ni-Al, Fe-A1 or Ni-Fe-A1 alloy intermediates in corro- the solutions is not much different from that of pure lithium. sion protection. The composition coefficient, 6p/6x, is near the middle of the The resistivity of lithium is increased by adding any solute known range of coefficients for other solutes, e.g. 2.1 for so the technique is not specific. Each solute on its own, oxygen, 11.2 for germanium, in liquid lithium.The present however, increases the resistivity by a characteristic amount, results can be used to follow chemical reactions of aluminium and there is now a substantial bank of values (Table 3) on in metallic lithium solvent. which to draw. For electronegative solutes, such as oxygen, hydrogen, We thank Kim Harper and Susan Smith for practical assist- nitrogen, silicon, germanium, tin and lead, the coefficient ance. 6p/6x has been correlated with the solvation enthalpy (anionic charge/radius) of the solute in liquid lithi~rn.~ For electropositive solutes such as magnesium, calcium, strontium and barium in liquid lithium 6p/6x has been corre- References lated with the effect that a solute atom of different ion core 1 R.J. Pulham, P. Hubberstey, M. G. Down and Anne E. Thunder, J. Nucl. Muter., 1979,854,299. 2 A. T. Dadd and P. Hubberstey, J. Chem. SOC., Dalton Trans., Table 3 Resistivity-composition coeffcients, (6p/6x)/lO-' R m 1982,2175.(molo/oX)-',for solutes in liquid lithium 3 P. Hubberstey and A. T. Dadd, J. Less-Common Met., 1982, 86, 55.solute X 6PI6X T/T 4 P. Hubberstey and P. G. Roberts, Physica B, 1994,198,307. 5 P. Hubberstey, T. Sample and M. G. Barker, J. Nucl. Muter., sodium" 0.3 310 1991,179-181,886.magnesiumb 1.6 402 6 P. Hubberstey, T. Sample and M. G. Barker, Fusion Eng. Des., calciumb 0.3 402 1991, 14, 227.strontiumb 0.7 402 7 C. C. Addison, G. K. Crefield, P. Hubberstey and R. J. Pulham,bariumb 4.8 402 J. Chem. Soc. A, 1971, 1393. aluminium 5.7 400 8 G. K. Creffeld, M. G. Down and R. J. Pulham, J. Chem. SOC.,indium' ca. 6 650 Dalton Trans., 1974, 2325. thalliumd :a. 11 800 9 M. G. Down, P. Hubberstey and R. J. Pulham, J. Chem. SOC., silicon' 10.4 400 Faraday Trans. I, 1975,71, 1387. germaniume 11.2 400 10 C. van der Mare1 and W. van der Lugt, J. Phys. (Paris), 1980,tin' 11.3 400 C8, 516.lead' 9.0 400 11 V. T. Nguyen and J. E. Enderby, Phiios. Mag., 1977,35,1013.hydrogen' 4.9 400 12 R. J. Pulham, P. Hubberstey and Petra Hemptenmacher, J. oxygene 2.1 300 Phase Equilib., in the press. nitrogene 7.0 400 Ref. 9.' Ref. 4. Ref. 10. Ref. 11. Ref. 3. Paper 4/027 151;Received 9th May, 1994
ISSN:0956-5000
DOI:10.1039/FT9949002753
出版商:RSC
年代:1994
数据来源: RSC
|
28. |
Electrical properties of an ethanol–dodecane mixture near the upper critical solution point |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2757-2763
Kazimierz Orzechowski,
Preview
|
PDF (879KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2757-2763 Electrical Properties of an Ethanol-Dodecane Mixture near the Upper Critical Solution Point Kazimierz Orzechowski Institute of Chemistry, Wrocra w University, Poland The electric permittivity and conductivity of an ethanol-dodecane critical mixture have been investigated as a function of frequency (100 Hz-1 MHz) and temperature. Dispersion of the electric permittivity and conductivity were found and these phenomena were interpreted in terms of Maxwell-Wagner polarization. The temperature dependence of E for high frequencies was described according to Senger’s model and the theoretically predicted critical exponent (0.89) agrees with the experimental results. The temperature dependence of the conductivity was analysed both for low (f -= f, ; f, is the characteristic frequency for Maxwell-Wagner polarization) and high (f > f,) frequencies.The non-linear temperature dependence of the background term makes the estimation of the conductivity critical exponent very difficult, but for low frequencies the exponent cp = 1 -a is the most probable one. The Maxwell-Wagner permittivity increment defined as AE,, = E,<,, -E,,,, diverges in the vicin- ity of T, with an exponent of -0.39. The conductivity increment defined as do,, = -c,<,~diverges in the vicinity of T, with an exponent of 0.22, close to the value predicted on the basis of the Wagner model, b-exponent value. The maximum of the miscibility curve of two liquids called the ‘upper critical solution point’ is an interesting case of a continuous phase transition.Large and long-lived concentra- tion fluctuations affect the macroscopic properties of mix-tures close to the critical point. The temperature dependence of the dielectric permittivity E is predicted to be:’v2 E/P = A, i-A2t i-A3te+ A4teiA where p is the density, t = (T -TJT, is the reduced tem- perature, T, is the critical temperature, 8 is the critical expo- nent and A is the correction to the scaling e~ponent.~ The. three-dimensional Ising model predicts A = 0.5.4 According to theoretical expectations’ and e~perirnent,’~~ 8 = 1 -a, where a is the critical exponent for the heat capacity at con- stant pressure and composition. The 3D Ising model predicts a = 0.11.4The relative permittivity is divided by the density, p, which has a critical anomaly described by the functional form of eqn.(1). It was shown in our previous paper7 that neglecting the anomaly of p leads to a small decrease in 0. Sometimes, as in case of perfluoromethylcyclohexaneecarbon tetrachloride mixtures, the critical anomaly of the electric permittivity was assumed to reflect essentially the behaviour of the density.8 Thoen and co-workers’-’ ’ showed that the critical behav- iour of the electric permittivity depends on the frequency of the electric field. These authors observed the low-frequency dispersion of the dielectric permittivity and interpreted it in terms of the Maxwell-Wagner polarization caused by the ionic impurities in the mixture giving rise to concentration fluctuations.At low frequencies a pronounced increase in E close to the critical point was observed, while for high fre- quencies (f >f,, where f, is the characteristic frequency of the Maxwell-Wagner polarization) smooth behaviour of E and decreasing 1 ds/dT 1 were found. The authors concluded that the theoretically predicted temperature dependence of E [eqn. (l)] describes E(T)only if f>f,. f, depends on the system investigated and increases rapidly when the electrical conduc- tivity of mixture increases.’ According to Senger’s theory’ the critical amplitude [A, in eqn. (l)] should be opposite in sign to dT,/dE2. However, this derivative is extremely difficult to obtain experimentally.Values of dT,/dE2 were reported by Debye and Kleboth for nitrobenzene-2,2,4-trimethylpentane1’ and by Beaglehole for cyclohexane-aniline mixtures.’ In both these cases nega- tive values of the derivatives were found, leading to positive A, values and a decrease in I d(E/p)/dT I near T,. The predic- ted critical anomaly is usually consistent with experimental results for f >f, . The electrical conductivity of critical binary mixtures has been investigated many times, but neither theory nor experi- mental results led to a generally accepted picture. The tem- perature dependence of the electrical conductivity is approximated by a function similar to that for the electric permittivity :’ o = B, + B,t + B3tq+ B4tq+A (2) where cp is the critical exponent for the electrical conductivity.Experiments give a spectrum of cp values ranging from 0.3 to 0.9; nevertheless, the most recent experiments and recalcula- tion of the older results seem to suggest cp = 1 -a. Stein and Allen’ linked the electrical conductivity anomaly with the critical behaviour of viscosity and predict- ed cp = 0.3. Jasnow et adopting the proton-hopping model, predicted that the electrical resistivity has a 1 -a anomaly. Shaw and Goldburg, on the basis of the percolation approach, found that cp = 2b, where B = 0.33 is the order parameter exponent. l7 Some authors suspect that the electri- cal conductivity anomaly has the same critical exponent as that of the electric permittivity, namely 1 -a.18 The same exponent was predicted by Wheeler,” who took into account an anomaly with respect to acid dissociation. Ramakrishnan et aLzoproposed cp = 1 -vl, where v is the correlation-length critical exponent.As for the electric permittivity, the electrical conductivity depends strongly on the frequency. The frequency depen- dence of the conductivity was explained in terms of the Maxwell-Wagner effect caused by microinhomogeneities in the vicinity of the critical point.”,’ ‘v2’ In this paper we report measurements of the electric per- mittivity, conductivity and density in the ethanol-dodecane system close to its critical composition. The measurements were performed over broad temperature and frequency ranges and allowed us to investigate the critical anomalies of the electric properties and the low-frequency phenomena described already by the Maxwell-Wagner polarization.Because of the long-range specific interactions between 2758 alcohol molecules, the alcohol-hydrocarbon system is inter- esting not only in the vicinity of the critical point, where uni- versal properties are expected, but also in the precritical region where the correlation length is of the order of several intermolecular distances. Alcohol-hydrocarbon mixtures do not easily yield precise dielectric measurements, especially at low frequencies; however, the large difference in the electric permittivities of the components and non-ideality of the mixture (the non- linear dependence of the electric permittivity us.concentration) should increase the critical anomaly of E,’* while similar densities of pure compounds protect the system from the density gradient. A relatively large electric conduc- tivity, even after careful purification of the components, will allow to obtain precise conductivity data. These properties of the selected system make it convenient for our measurements. Experimental The electric properties were measured using a Hewlett Packard HP 4284A bridge in the frequency range 0.1-103 kHz. The capacitor was made of stainless steel and fused quartz. The air capacity was 1.5 pF. The relative accuracy of E was +0.001 and absolute accuracy kO.1. For conductivities in the frequency range 0.5-10 kHz the relative accuracy of CT was 0.02 pS rn-l and the absolute accuracy 0.5 pS m-’.At higher frequencies the error in CT was much greater. The tem- perature was stabilized using a water-air thermostat with precision up to _+0.002 K. The temperature was read by means of calibrated thermistor. The absolute accuracy was fO.l K but resolution was better than k0.002 K. The density was measured using an automatic MG-2 densi-tometer manufactured by ECOLAB, Poland. The precision was f0.2 kg m-3; the temperature during density measure- ments was stabilized up to kO.01 K. Ethanol was obtained by purification of the 95% material. The alcohol was refluxed over freshly ignited calcium oxide for 6 days. The calcium oxide was changed daily. Finally, the alcohol was dried using magnesium activated by iodine and then distilled.Dodecane was refluxed over sodium and frac- tionally distilled at reduced pressure. The purity of the liquids was controlled on the basis of their boiling temperatures and electrical conductivities. All operations were performed under a dry nitrogen atmosphere. Measurements were performed for the close to critical mixture x, = 0.686 (mole fraction of alcohol). The phase-separation temperature was accompanied by an inflection of the permittivity which allowed us to detect T, simultaneously with the electrical measurements. The T, value obtained was 284.594 K. This is consistent with the value reported by Francis (285 K),23and is slightly smaller than that obtained by us previously (286.7 K).24 Results Fig.1 shows the temperature dependence of the electric per- mittivity of a near-critical mixture (x, = 0.686)for frequencies ranging from 5 to lo3 kHz. For temperatures far from the dependence of E is almost linear. For T < 287 K a change in IdE/dT I is observed; for low frequencies I dE/dT I increases and for high frequencies it decreases. The dependence of the permittivity on frequency for different temperatures is shown in Fig. 2. For temperatures far from the critical one (T -T, > 10 K), the ~cf)dependence is almost linear and, prob- ably because of electrode polarization effects, E increases slowly with decreasing frequency. When the temperature J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 9.50’ TrC Fig.1 Temperature dependence of the electric permittivity of the close to critical mixture (xc = 0.686) for frequencies ranging from 5 to lo00 kHz: (a) 5 kHz, (b) 10 kHz (+0.05), (c)50 kHz (+0.10), (d) 100 kHz (+0.15), (e)500 kHz (+0.20), cf) 1 MHz (+0.25); ordinate dis- placements in brackets reaches T,,dispersion of E is observed. The characteristic fre- quency cf,)for the ethanol-dodecane mixture is ca. 150 kHz. The temperature dependences of the electric conductivity obtained at different frequencies are presented in Fig. 3. As for the permittivity, measurements were performed in the one-phase region only. The electric conductivity is a smooth function of temperature and decreases as the temperature approaches T,. For frequencies between 0.5 and 100 kHz the experimental points could be described by a single curve. For frequencies > 100 kHz a gradual increase in c as well as qual- itative changes in the precritical CT(T) dependence were found.The frequency dependence of the electrical conductivity is shown in Fig. 4. As concluded on the basis of Fig. 3 in the frequency interval 0.5-100 kHz, the conductivity is approx- imately constant. For larger frequencies an increase in the conductivity is observed. The electric permittivity measure- 9.50t\8.90 9.30 8.70 E E 8.50 8.30 1 10 100 1000 frequency/kHtz Fig. 2 Experimental values of E us. frequency for different temperaturesrc: (a) 11.628, (b) 12.033, (c) 13.068, (4 14.082, (e) 15.020,(f) 30.025 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I 11 14 17 20 23 26 TpC Fig. 3 Temperature dependence of the electric conductivityobtained for different frequencies: (a)0.5 kHz, (b)5 kHz, (c) 100 kHz, (d)500 kHz, (e) 1 MHz ments in an ethanol-dodecane mixture give a critical relax- ation frequency of ca. 150 kHz. In this frequency range the conductivity increases (Fig. 4). Discussion Dispersion of E and Q When the frequency of an electric field decreases, dispersion of the electric permittivity and conductivity in a close to criti- cal mixture is observed. The origin of this low-frequency relaxation is usually explained by the Maxwell-Wagner pol- arisation caused by large precritical fluctuations. This concept has been explored in many recent papers; however, it should be stressed that the Maxwell-Wagner polarisation concerns the case of a real inhomogeneous system containing stable phases of definite shape (layers, droplets, rod-like etc.) and boundaries between them.The critical mixture could be described as an inhomogeneous one only in the limit T + T,, whereas according to Fig. 4, the a dispersion is observed even 8 K above T,. Consequently the model describing the system 19 18 c I E2 17 --. 16 15 0.1 1 10 100 1000 frequency/k Hz Fig. 4 Frequency dependence of electrical conductivity for different temperatures/”C: (a) 11.628, (6) 13.068, (c) 15.071, (6)20.033, (e) 25.394, cf)30.025 as really inhomogeneous could give only a qualitative description of the observed phenomena.According to Wagner” the complex relative permittivity (E*) and conductivity (a*)of a system containing spherical particles (droplets) immersed in a continuum are frequency- dependent quantities : &I -&h O1 &* = &h + -+-(3)1 + ioz a* = a,+ ioz(a, -a,) 1 + ioz + icmO&h (4) where E, and a, are the electric permittivity and conductivity at low frequencies cf<fMw), &h,ah at high frequencies, cD and aDare the permittivity and conductivity inside the droplets, gB and uB in the surrounding medium, 4 is the volume frac- tion of a droplet. In the vicinity of the critical point we assume that the electric permittivity of droplets and the sur- rounding medium are similar. Consequently : &D -&B 6 &D + EB ; &BED % 6; bD -OB 4CD + (TB ; bB bD x 0; Neglecting small terms and assuming 4 = 0.5, eqn.(5)-(8) give : Eqn. (9) and (10) predict that if the electric permittivity and conductivity inside the droplets are larger than in the back- ground medium, the dielectric experiment will yield an increase in E and decrease in u when the frequency decreases. The Maxwell-Wagner increments in the electric permit- tivity and conductivity are defined as follows: AEMW = and AaMw= bh -ol. According to eqn. (9) and (lo), the increments can be written: It is reasonable to suppose that E~ -E~ zxD -xBz t” and aD-oBz xD-xBz tS, where xDis the concentration of the polar component inside the droplet and xB is the concentra- tion of the polar component in surrounding medium.Conse- quently, the expected temperature dependence of the &,, - 2760 Maxwell-Wagner increments can be written : (13) Eqn. (13) and (14) were derived assuming that the diameter of the inhomogeneities is temperature independent. In the case of a the assumption adopted is not very important, but the influence of the change in the diameter of the inhomoge- neities on a Maxwell-Wagner permittivity increment should not be neglected. Taking into account the analogy between a critical mixture and a truly inhomogeneous system, the polar- ization of 'droplets' containing charge carriers depends both on the diameter of the 'droplets' and the concentration of the charge carriers inside them.Following this concept, Thoen et al., on the basis of a water-in-oil emulsion model, obtained the temperature dependence of AEMW(t):l0 where v is the correlation length exponent. For the three- dimensional Ising model p = 0.326 and v = 0.630 and AEMw(t) is expected to diverge with an exponent (-v + p) = -0.305. Fig. 5 shows the experimental Maxwell-Wagner increments defined as follows: as a function of temperature. The experimental values of the Maxwell-Wagner increments contain not only the critical terms but also non-critical ones. The difference in permit- tivities and conductivities measured at low and high fre- quencies were observed not only in the vicinity of the critical point, but also at much higher temperatures.We suspect here that additional electrode polarisation effects disturb the mea- surements at low frequencies. The temperature dependences of the experimental Maxwell-Wagner increments were approximated by the equations: where t is the reduced temperature. We used the fitted value of the critical temperature. The method of obtaining T, is described in the next paragraph. Results of the fitting are pre- v)2 h---. 1.50 -0.40 2 &5 F00 v 0z 1.00 -0.30 v Nb I $ -0.20 2 v 0.00 I I 0.10 0.00 0.02 0.04 0.06 0.08 reduced temperature Fig. 5 Permittivity and conductivity increments defined in eqn. (16) plotted us. reduced temperature. As,, = 0.121 + 0.0097~-'.~;x2 = 5.64. AsMw= 0.132 + 0.0045t-0.39(best fit); x2 = 4.75.ACT,, = x2 =2.16 -4.01~O.~~~;3.04. AoMW= 2.55 -3.88t0.22 (best fit);x2 = 1.425. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 sented in Fig. 5. For the conductivity increment the best fit was obtained for t,b = 0.22; we found, however, that the proper approximation would be obtained with the exponent equal to its theoretical value (~ = fl) also. The data collected in Fig. 5 also show that the theoretically predicted exponent for AE~,,, describes the experimental values properly (9 = p -v); nevertheless, the best fit was found for an expo- nent of -0.392.It seems that the fitting results confirm the predicted temperature dependences for AEMW and AoMW. Temperature Dependence of Electric Permittivity According to theoretical expectations, the critical properties of the electric permittivity should be approximated by eqn.(1). However, the mentioned theoretical expectations do not take into account Maxwell-Wagner phenomena and conse- quently only measurements performed at frequencies larger than that characteristic of the Maxwell-Wagner effect could be described by eqn. (1). Because the relaxation frequency in an ethanol-dodecane mixture was ca. 150 kHz, in the fitting routine we used results obtained at 1 MHz. Eqn. (1) contains the product ~/p,which means that in order to obtain the inherent anomaly of E, the critical proper- ties of density should be taken into account. Suitable mea- surements of the temperature dependence of density in the vicinity of the critical mixing point have been performed and the p( T)dependence is found to be linear : p/g cmV3 = 0.7768 -8.09 x lop4T/"C The accuracy achieved does not allow observation of the critical anomaly of p, and we expect that the anomaly of p would not influence the critical properties of the electric per- mittivity.Nevertheless, including the temperature dependence of the density allows compensation for the effect of the expansibility on E. The results presented in Fig. 1 show that the anomaly of E should be observable not only close to the critical point, but also in a wider temperature interval. This is the reason why in the fitting routine we decided to use the results obtained within a wide temperature interval. Thus, it is important to select a correct representation of the background permit- tivity.The measurements of E performed in pure ethanol over a very large temperature range led to a polynomial tem-perature dependence containing a square term.26 Our experi- ments covered only 20 K,but because of the high resolution of our measurements, the linear temperature dependence of the background could not be accepted and the additional square term describing the background was included in eqn. (1). Consequently in the fitting process we used the following equation: &//I= A, + A2t + A3t2+ A4te + (18) A correct and precise value of the critical temperature is necessary to obtain proper values of the critical parameters. In our measurements the critical temperature was estimated from the inflection in the E(T)dependence and consequently T, was found with limited precision (T,d 11.55"C).In order to improve the precision of measurement of T,, the critical temperature has to be regarded as an adjustable parameter. Estimation of seven parameters (A, -A,, T,, 0) in a free fitting routine should result in a strong correlation between them. Taking into account the above, we decided to include the additional requirements in the fitting process. We sup- posed that for the properly selected critical temperature the critical exponent should have the theoretically predicted value and consequently the fitting error should have a J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 0.50 I 0.46 1 %*\ minimum for 8 = 0.89 I 0.44 xz 0.42 0.40 0.38 0.36 0.80 0.85 0.90 0.95 1.oo critical exponent Fig.6 Estimation of critical temperature. Eqn. (1) was fitted to experimental points obtained for 1 MHz. The temperature was regarded as a critical one, when the minimum of x2 was obtained for 8 = 0.89. T,/"C: (a) 11.400,(b)11.42O,(c)11.444,(d) 11.460,(e)11.480. minimum for 8 = 0.89. The quality of fitting was tested usingx2 defined as follows: where Var(E), = a(&)? + (ae/aT)?o(t)z; and o(T)~are the standard deviations of the electric permittivity and tem-perature, respectively, N is the number of experimental points; p is the number of adjustable parameters. The stan-dard deviations were 0.001 for E and 0.002 for T. Fig.6 demonstrates that the minimum of x2 for 8 = 0.89 was obtained when T,= 284.594 K. The critical temperature obtained is very close to the value reported by Francis.23 The adopted method of determination of the critical tem-perature contains an assumption that the critical exponent describing permittivity is equal to the theoretically predicted value. Taking into account the above, the adopted fitting routine cannot be used to test the critical exponent value. Fig. 7 shows the results of the fitting of eqn. (18) (6 = 0.89, T, = 284.594 K) to the experimental values of E obtained at 1 MHz. The differences between the experimental points and the fitted curve do not exceed 24~)~.The parameters obtained by the fitting are: A, = 12.072; A, = -136.73; A, = -128.23; A, = 66.27; A, = 122.05.x2 was equal to left scale 12.0 1 I 0.005 0-11.8 -;'i 0 m0 w ~ % .* .* 0 t-. . .. 0.000 II 0 *0 r -0 -11.6 ? 0. D 0,0w 0 0,0 right scale -0.005 11.4 xx 11.210 14 18 22 26 30-0.0 10 0.391. Examination of the fitting results shows that the criti-cal amplitude (A,) has a positive value, which according to the Sengers model' gives a negative value for dT,/dE2. The negative value of dT,/dE2 in the ethanol-dodecane mixture was deduced by us on the basis of NDE (non-linear dielectric experiment) results27 and was also obtained in other critical mixture^.'^*'^ Temperature Dependence of Conductivity As was mentioned in the Introduction, the temperature dependence of a can be approximated by eqn. (2).The critical anomalies of conductivity are observed over a wide tem-perature interval and a proper estimation of the temperature dependence of the background term is very important to obtain the correct value of the critical exponent cp. For this purpose we performed electrical conductivity measurements in a ethanol-benzene reference mixture (0.68 mole fraction of alcohol). The concentration of the mixture was similar to the critical one and the temperature dependence of the conduc-tivity is shown in Fig. 8. The linear approximation of aBcan be accepted only in a limited temperature range. In order to describe the background over the 20 K range, the square equation for aB(T)should be used. Hence the temperature dependence of a in a critical ethanol-dodecane mixture should be written: a = B, + B2t + B,t2 + B4tP+ BStQ+* (20) Eqn.(20) contains seven adjustable parameters (B,-B5, T,, cp) describing a rather smooth a(t) dependence and the free fitting routine will decrease the credibility of the estimated parameters. In order to reduce the number of adjustable parameters, the critical temperature was fixed at constant value obtained from &(t)measurements. The background term contains two adjustable parameters (B2 and B,). Unfor-tunately it was not possible to use the parameters obtained in the reference non-critical mixture instead of B, and B, in eqn. (20). The conductivity of the ethanol-benzene system is a few times larger than that of ethanol-dodecane.Nevertheless, on the basis of the measurements performed in the reference mixture, some qualitative conclusions can be drawn : we expect that B, should be positive and B, ,describing the cur-vature of a(T), negative. This qualitative observation drawn on the basis that the reference mixture does not reduce the number of adjustable parameters, but allows elimination of physically meaningless fitting results. In the fitting process the 70 I 10 13 16 19 22 25 T/"C T/"C Fig. 8 Temperature dependence of the electrical conductivity Fig. 7 Results of the fitting of eqn. (18) to experimental points obtained in the reference ethanol-benzene mixture: (a) 500 Hz, (b) 1 obtained for 1 MHz (0 = 0.89, T, = 284.594 K) MHz 2762 critical exponent cp was fixed at some discrete values, predict- ed by different theories.For the electric permittivity we used the results measured at sufficiently high frequencies, where the Maxwell-Wagner effect does not disturb the experimental results. For the con- ductivity measurements the choice of a suitable frequency is not as evident. To avoid the Maxwell-Wagner polarization, a frequency > 150 kHz should be used. On the other hand, taking into account the symmetry of eqn. (9) and (lo), the high-frequency limit of E (E~ zE~)is comparable with the low- frequency limit of a (a, x og),which points to an analysis of low-frequency data. However, only some of the theories take into account the similarity of inhomogeneous and critical systems.Many theoretical analyses explain the observed phe- nomena as a consequence of the critical anomalies of other quantities, such as transport phenomena, viscosity' and asso~iation.l~*~~Such theories should explain the r~ anomaly observed at high frequencies, where the Maxwell-Wagner effect is negligible. Hence we performed the fitting both for low (0.5 kHz) and high (1 MHz) frequencies. Results of the fittings are presented in Table 1. Low-frequency Results (0.5kHz) Examination of the results collected in Table 1 shows that the quality of fitting (tested by x2) increases with cp. On the basis of the reference mixture we learn that B, should be positive and B, negative. When the complete equation was fitted, a positive B, coefficient was obtained for exponents of 0.66 and 0.89 and B, was positive in all cases.It is characteristic that neglecting the correction to scaling term in eqn. (20) yields the correct temperature dependence of the background for most of the fitted exponents except the 0.89 one. In this case, however, the absolute value of B, is very close to B,, which probably arises from the difficulties in estimating parameters describing terms having a similar temperature dependence. Taking into account the error obtained in the fitting routine it seems that the low-frequency temperature dependence of cp can be described by an exponent equal to 0.89 (cp = 1 -a). High-frequency Results (1 MHz) As observed for low frequencies, the quality of fitting of the complete equation (with a correction to scaling term) Table 1 Results of fitting eqn.(20) to conductivity data; was held fixed during each fitting; x2 is defined in the text 500Hz;[T = B, + B2t + B3t2+ B4t'+'+ B5t'P+A 0.3 15.04 -166.3 163.9 -3.70 137.8 0.49 0.66 14.84 478.1 528.4 -13.81 -645.4 0.42 0.89 14.82 290.0 907.4 -49.61 -641.3 0.40 500 Hz; [T = B, + B,t + B3t2+ B4tV 0.3 14.41 61.50 -310.3 3.57 -1.61 -0.5 14.62 42.50 -227.3 9.04 1.20 0.66 14.69 115.1 -144.9 23.72 -0.95 0.89 14.74 -189.2 11.02 185.16 -0.67 1 MHz;[T=B,+B,~+B,~~+B,~'+'+B,~~+~ 0.3 16.75 13.84 -135.4 -1.826 26.108 0.58 0.66 16.67 276.9 72.00 -19.89 -297.4 0.53 0.89 16.65 453.4 448.4 -200.8 -498.5 0.48 1 MHz; = B, + B,t + B3t2+ B4t' 0.3 16.64 56.98 -225.2 -0.448 -0.61 0.5 16.61 58.90 -232.1 -1.052 -0.62 0.66 16.60 61.92 -238.2 -2.606 -0.63 0.89 16.59 80.94 -248.4 -18.37 -0.65 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 increases with cp. However, the requirements obtained in the reference mixture hold only when cp = 0.3. The simplified equation changes the fitting results considerably. In all cases positive B, and negative B, were found. In contrast to the results from fitting the complete equation, the simplified equation yields an increase in the fitting error when the criti- cal exponent increases. It seems that the equation adopted to describe q(t) for high frequencies is incomplete. This same conclusion could be drawn up on the basis of conductivity dispersion results presented in the previous paragraph.According to the Maxwell-Wagner model, the high-frequency conductivity data should contain the term tS [eqn. (lo)], reflecting the increase of conductivity inside alcohol- rich fluctuations. Taking into account the small error in the fittings (x2 -= 1)we could not discriminate between any of the theoretically predicted exponents describing the high-frequency conductivity data. Both the fitted equations (with and without correction to scaling term) led to negative values of the critical amplitude B, . Conclusions The ethanol-dodecane mixture is a convenient system for the investigation of both permittivity and conductivity anomalies that occur in the vicinity of the critical mixing point.The conductivity of the system arises not only from ionic impu- rities, but most probably from the dissociation of alcohol molecules. Consequently the charge accumulation, in the area of fluctuations rich in the polar component, is not a diffusion-controlled process. In our system, because of the dissociation of the alcohol molecules, the local charge density is pro- portional to the alcohol concentration. This property of the mixture allows exploration of the simple assumption that the electric properties inside and outside 'droplets' are pro-portional to the difference in the concentration of the alcohol. Such an assumption leads to the prediction that do,, should diverge in the vicinity of T, with an exponent equal to p (order parameter exponent), while the exponent for AE~, should have a value of (B -v) (v is the correlation length exponent)." The experimental results confirm the predicted divergences.The Maxwell-Wagner polarization model adopted was formulated for a real inhomogeneous mixture. The question why, even in the precritical region, where the fluctuations are small and short-lived, the model works prop- erly is still open. The character of the electric permittivity anomaly depends on the frequency. For low frequencies an increase in ld~/dTI and for high frequencies a decrease of this quantity is observed. The anomalies observed for high frequencies were described according to Senger's model. The method adopted for estimating the critical temperature does not allow testing of the critical exponent describing the E anomaly.We found, however, that a critical exponent equal to 1 -a could describe the obtained ~(t)results properly. The temperature dependence of a was described according to eqn. (20). The extensive temperature interval where the anomalies were observed and the non-linear temperature dependence of the background results in a large number of adjustable parameters. The credibility of the obtained expo- nent is low it seems; however, for low-frequency measure- ments the a(T) dependence could be described by the exponent cp = 1 -a. The a(T) dependence obtained at high frequencies is a smooth function and an attempt to fit the equation with many adjustable parameters does not allow discrimination between any of the fitted exponents.It was found that the critical amplitude is positive. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2763 The conductivity data obtained, especially at high fre-quencies, seem to be characteristic of systems containing a component with a tendency to dissociate. We suspect that the observed high-frequency conductivity properties should reflect the increase in conductivity inside alcohol-rich fluctua- 10 M. Marabet and T. K. Bose, Phys. Rev. A, 1982,25,2281. K. Orzechowski, Ber. Bunsenges. Phys. Chem., 1988,92,931. R. H. Cohn and S. C. Greer, J. Phys. Chem., 1986,90,4163. J. Thoen, R. Kindt and W. Van Dael, Phys. Lett. A, 1980, 76, 455; Phys.Lett. A, 1981, 87, 73. J. Thoen, R. Kindt, W. Van Dael, M. Merabet and T. K. Bose, tions arising from the dissociation of alcohol molecules. This effect will not be observed at low frequencies because of trap- ping of the ionic species inside alcohol-rich fluctuations. For the other critical mixtures, without components having a pro- pensity to dissociate, the conductivity is promoted by ionic impurities. In such mixtures the increase in conductivity in 11 12 13 14 15 Physica A, 1989,156,92. J. Hamelin, T. K. Bose and J. Thoen, Phys. Rev. A, 1990, 42, 4735. P. Debye and K. Kleboth, J. Chem. Phys., 1965,42,3155. D. Beaglehole, J. Chem. Phys., 1981, 74, 5251. E. M. Anderson and S. C. Greer, Phys. Rev. A, 1984,30,3129. A. Stein and F. G. Allen, J. Chem.Phys., 1973,65,6079. fluctuations rich in the polar component is also expected, but 16 D. Jasvow, W. J. Goldburg and J. S. Semura, Phys. Rev. A, 1974, this process has to be diffusion controlled, and the critical anomaly is more complicated. 17 18 9, 355. G.-H. Shaw and W. J. Goldburg, J. Chem. Phys., 1976,65,4906. N. Kumar and A. M. Jayannavar, J. Phys. C, 1981,14, L785. I am very grateful to Professor Dr. L. Sobczyk, Professor Dr. H. Kolodziej and Dr. J. Glinski for their interest and helpful 19 20 J. C. Wheeler, Phys. Rev. A, 1984,30,648. J. Ramakrishnan, N. Nagarajan, A. Kumar, E. S. R. Gopal and G. Ananthakrishna, J. Chem. Phys., 1978,68,4098. discussions. This work was supported by programme 22679- 21 M. K. Gunasekharan, S. Guha, V. Vani and E. S. R. Gopal, Ber. 9 1-02. Bunsenges. Phys. Chem., 1985,89, 1278. 22 J. Goulon, J. L. Greffe and D. W. Oxtoby, J. Chem. Phys., 1979, 70,4742. References 23 A. W. Francis, Adv. Chem. Ser., 1961,31. 1 J. V. Sengers, D. Bedeaux, P. Mazur and S. C. Greer, Physica A, 1980,104,574. 2 L. Mistura, J. Chem. Phys., 1973,59,4563. 3 F. Wegner, Phys. Rev. B, 1973,5,4529. 4 J. C. Le Goulou and J. Zinn-Justin, J. Phys. Rev. B, Condens. 24 25 26 27 K. Orzechowski, Physica B, 1991, 172, 339. K. W. Wagner, Arch. Electrotech., 1914, 2, 371. C. P. Smyth and W. N. Stoops, J. Am. Chem. Soc., 1929, 51, 3312. K. Orzechowski, results presented at Statphys 18, Berlin, 1992. Matter, 1980,21, 3976; J. Phys. Lett., 1985,46, L-137. 5 S. C. Greer and M. R. Moldover, Annu. Rev. Phys. Chem., 1981, 32, 233. Paper 4/016621;Received 2 1 st March, 1994
ISSN:0956-5000
DOI:10.1039/FT9949002757
出版商:RSC
年代:1994
数据来源: RSC
|
29. |
Directional symmetry of the time lag for downstream absorptive permeation studied by the matrix method |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2765-2767
Jenn Shing Chen,
Preview
|
PDF (322KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2765-2767 Directional Symmetry of the Time Lag for Downstream Absorptive Permeation studied by the Matrix Method Jenn Shing Chen Department of Applied Chemistry, National Chiao Tung University, Hsinchu, Taiwan, 30050,Republic of China The directional symmetry of the downstream time lag for absorptive permeation, accompanying a first-order reaction, across a heterogeneous membrane has been proven by the matrix method based on the theory of Siegel. Owing to the heterogeneity the partition coefficient (K),diffusivity (D)and rate constant (k) are all depen-dent on position. The directional symmetry was first shown for multiple laminates with each laminate having a distinct K, D and k. The transmission matrix of the system is then a sequential product of those of the individual laminates. The proof is for the unit value of the determinant of the transmission matrix for each laminate and, hence, for the whole system.Directional symmetry is thus proven for the heterogeneous membrane, since it can be visualized to be an assembly of an infinite number of infinitely thin, homogeneous laminates. A typical experiment for permeation across membranes involves maintaining, at upstream and downstream faces, the penetrant activity at constant levels a, and ad, usually with a, > ad = 0, respectively. The penetrant activity in the mem- brane is initially adjusted to ad for running absorptive per- meation or to a, for running desorptive permeation. The'v2 accumulated amount of penetrant release is then measured as a function of time at the downstream or upstream faces.A plot of total release us. time gives the steady-state per-meability from the slope and the time lag from the intercept with the time axis. Thus four time lags are given from various combinations of upstream/downstream and desorptive/ absorptive permeation. 'v2 In addition, the four time lags are also associated with reverse permeation resulting from the exchange of the upstream and downstream compartments.'T2 In conjunction with the permeability, appropriate time lags are used to determine the diffusivity and solubility (or parti- tion coefficient) of the penetrant inside the membrane if diffu- sion is Fi~kian.~-~ They are also employed to identify and study the non-Fickian time-lag increments due to the time and/or position dependence of diffusion parameter^^.^ and a consistency check of the experimental determination of various time lag^.^,^,^ In practical applications, however, the diffusion time lag can be visualized as a gauge of the stability of colloid flocculation,* or as a measure of the induction time of crystalli~ation.~~' These are possible because the above- mentioned processes can all be modelled as diffusion under the influence of a potential field.Consider under what conditions the downstream absorp- tive time lags for forward and reverse permeation are equal, i.e. the directional symmetry holds. Jaeger' ' was probably the first author to address this point in the context of heat con- duction across a multi-laminate slab.The extensive theoreti- cal studies of Petropoulos et have revealed that the a1.1*5912913 time-lag symmetry holds for any type of heterogeneity if D and K are functions of position, x, only. In the more general case where D and K are functions of both x and concentra- tion, p, the directional symmetry applies only if there is sym- metry about the midplane of the membrane. For K(p, x) and D(p. x) without symmetry about the midplane, a distinction between the two cases of separable and non-separable p, x, can be made by checking the symmetry or asymmetry of flux, respectively. Those results are summarized in Table IV of ref. 5, and serve a practical, diagnostic purpose.Up to now, however, only permeation across heterogeneous membranes in the absence of chemical reactions has been considered. Of increasing practical importance are membranes with associated reactive moieties, such as catalysts or enzymes, to enhance the productivity of chemical and biological processes in catalytic membrane reactor~,'~to enhance the per-formance of biosensors,' or to simulate active transport using uneven distribution of enzyme activities. A theoretical study' has also suggested that the directionality of products and substrate fluxes and the separation of product and sub- strate can be drastically improved by using an appropriate, non-uniform distribution of the reactive moieties. Hence, it is desirable to extend the discussion of directional symmetry to include the case where the reaction takes place inside the membrane.Only first-order reactions with position-dependent rate constants are considered here. This simplifica- tion will, of course, not cover all cases, especially for the enzyme catalytic reaction which usually has a Michaelis-Menten type rate constant. However, as pointed out by Kubin and Spacek,16 in the case of a linear gradient distribu- tion of the enzyme with high or low concentrations of sub- strate, the accompanying first-order reaction will have a position-dependent rate constant. The matrix method will be employed to prove the direc- tional symmetry of the downstream time lag for absorptive permeation across a heterogeneous membrane accompanying a first-order reaction.In a previous publication'8 the direc- tional symmetry of the time lag for the same system was proven using the symmetry property with respect to the exchange of two coordinate arguments of the Green's func- tion associated with the diffusion equation. A review of the matrix formulation of the mass diffusion (or heat conduction) problem can be found in ref. 19-21. Transmission Matrix of a Homogeneous Membrane Consider one-dimensional absorptive permeation across a homogeneous membrane extending from x = xu to x = xd, with xd > xu. The penetrant is also involved in a first-order reaction with rate constant k inside the membrane. The con- centrations of penetrant in the upstream compartment (at the left side of the membrane) and downstream compartment (at the right side of the membrane), p,(t) and pd(t), respectively, are time-dependent.The permeation experiment starts with zero concentration of penetrant inside the membrane. The J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 mathematical formulation is then where T(s)is the resultant transmission matrix of the whole membrane. T&), the transmission matrix of the ith com--a P(X, t) = D -a2 (x, t) -kp(x, t); ponent laminate, is in the same form as the transmission at ax2 matrix in eqn. (6) except for D, K, q and h being attached by a subscript i. where D is the diffusivity of the penetrant in the membrane; p(x, t), p(x, , t), p(xd, t) represent, respectively, the concentra- tion inside the membrane and at the upstream and down- stream sides.K, K,, & are partition coefficients in the membrane, upstream and downstream compartments. The boundary conditions take into account the requirement that the activity of the penetrant (=p/K) should be continuous at interfaces. We will analyse the problem in the Laplace domain. Expressing eqn. (1) in terms of s, the Laplace variable, and ?(x, s), the Laplace transform of p(x, t), gives d2 S?(& s) -p(x, 0) = D 2?(x, s) -kfi(x, s);dx With p(x, 0) = 0, the solution to eqn. (2) is given by K sinh[q(x, -x)]fi(x, s) = sinh[q(x, -xu)] {F fid(S)+ -sinh[q(x -xu)] (3) Kd 1 with q = (s + k/D)'I2. The permeation fluxes at the upstream side (x = xu) and downstream side (x = xd) are then j,(s) = DqK{Y coth(qh) -where h (= xd -xu) is the thickness of the membrane, and the positive direction of the fluxes is defined to be from left to right.Eqn. (4)and (5) can be rewritten in a matrix form (6) The 2 x 2 array in eqn. (6) is the transmission matrix for absorptive permeation across a homogeneous membrane. Directional Symmetry of Downstream Time Lags for Absorptive Permeation accompanying Chemical Reactions across Multiple-laminate and/or Heterogeneous Membranes Consider a membrane which consists of a series of n different homogeneous laminates and extends from x = xu to x = xd. A typical ith slab (i = 1, 2, . . . ,n), extending from x = xi-to x = xi, is characterized by a thickness di , a diffusitivity Di, a partition coefficient Ki and a first-order rate constant ki.Thus we have di= xi -xi-1, xo = xu and x, = xd . The initial and boundary conditions are the same as in eqn. (1). In terms of transmission matrices the permeation problem is formu- lated as [z] ..., T,(s)[-] [-I= T,(S)T,-~(S), K, = T(s) K, (7) The transport equation, eqn. (7), can be converted to [ = T-'(s)[K] (8) This is possible because the determinant of every component matrix T(x) (i = 1, . . . ,n) is unity, and hence the determinant of T(s)is also unity, i.e. T(s)is non-singular. Let the 2 x 2 matrix, T(s),be represented by a general form as (9) where a(s), B(s), y(s), 6(s) are real numbers. With unit determi- nant of T(s),its inverse becomes As mentioned before, exchanging upstream and downstream compartments results in flow reversal.It is natural to define, in reverse permeation, the positive direction to be from right to left. Schematically where the superscript R is used to signify the reverse per- meation. A combination of eqn. (8), (10)and (11) yields This is the transport equation of reverse absorptive per- meation formulated in matrix form. According to Siege121 the downstream absorptive time lag can be calculated from the transmission matrix element as where T,,(s)is the element of first row and second column of T(s).For both forward and reverse permeation TI&) is iden- tified to be P(s). Thus it is readily seen that the downstream absorptive time lags for forward permeation, L, and those for reverse permeation, I!, are identical, i.e.(14) A heterogeneous membrane, in which D, K and k are con- tinuously varying with position, can be visualized as an assembly of n infinitely thin, homogeneous laminates with n + 00, keeping the total thickness equal to the original one. Since the preceding proof is valid irrespective of the value of n, and the thickness of each laminate, the directional sym- metry obviously holds for absorptive permeation accompany- ing a first-order reaction across a heterogeneous membrane. The work was funded in part by the National Science Council, Taiwan, Republic of China under the project NSC 82-0208-M-009-0 19. References 1 J. H.Petropoulos and P. P.Roussis, J. Chem. Phys., 1967, 47, 1491. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2767 2 3 J. H. Petropoulos, Adu. Polym. Sci., 1985,64,93. K. Tsimillis and J. H. Petropoulos, J. Phys. Chem., 1977, 81, 12 J. H. Petropoulos, P. P. Roussis and J. Petrou, J. Colloid Inter- face Sci., 1977,62, 114. 4 2185. C. Savvakis and J. H. Petropoulos, J. Phys. Chem., 1982, 86, 13 J. H. Petropoulos and P. P. Roussis, J. Chem. Phys., 1967, 47, 1496. 5128. 14 T. Naka and N. Sakamoto, J. Membrane Sci., 1992,74,159. 5 J. H. Petropoulos and P. P. Roussis, J. Chem. Phys., 1969, 50, 3951. 15 W. R. Vieth, Membrane Systems: Analysis and Design, Hanser Publishers, Munich, 1988. 6 J. H. Petropoulos and P. P. Roussis, J. Chem. Phys., 1969, 51, 16 M. Kubin and P. Spacek, Polymer, 1973,14, 505. 1332. 17 T. Ciftci and W. R. Vieth, J. Mol. Catal., 1980,8,455. 7 P. P. Roussis and J. H. Petropoulos, J. Chem. SOC., Faraday Trans. 2, 1976,72,737. 18 J. S. Chen and F. Rosenberger, Chem. Eng. Commun., 1991, 104, 41. 8 9 10 R. D. Vold and M. J. Vold, Colloid and Inteflace Chemistry, Addison-Wesley, London, 1983. H. L. Frisch, J. Chem. Phys., 1957,27,90. H. L. Frisch and C. C. Carter, J. Chem. Phys., 1971,54,4326. 19 20 21 A. H. van Gorcum, Appl. Sci. Res. A, 1951,2,272. B. Bunow and R. Ark, Math. Biosci., 1975, 26, 157. R. A. Siegel,J. Phys. Chem., 1991,95,2556. 11 J. C.Jaeger, Quart. Appl. Math., 1950,8, 187. Paper 3/067141; Received 9th November, 1993
ISSN:0956-5000
DOI:10.1039/FT9949002765
出版商:RSC
年代:1994
数据来源: RSC
|
30. |
Molecular conformation ofn-alkyloligo(oxyethylene)s in the solid state studied by Raman spectroscopy. Effect of the end group |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 90,
Issue 18,
1994,
Page 2769-2774
Sei Masatoki,
Preview
|
PDF (671KB)
|
|
摘要:
J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2769-2774 Molecular Conformation of n-Alkyloligo(oxyethy1ene)s in the Solid State studied by Raman Spectroscopy Effect of the End Group Sei Masatoki and Hiroatsu Matsuura* Department of Chemistry, Faculty of Science, Hiroshima University, Kagamiyama , Higashi-Hiroshima 724,Japan The molecular conformation of a-hexadecyl-o-methoxyoligo(oxyethy1ene)s CH3(CH,),,(OCH,CH,),0CH3 (C,,E,C,) (rn = 1-4) in the solid state has been studied by Raman spectroscopy. The conformations of these methoxy group terminated compounds were compared with those of the corresponding hydroxy group terminat- ed compounds (C16E,,,). The conformational change of the inner oxyethylene segments from the extended to the helical structure takes place at rn = 4 for C, 6E,C,, while it takes place at rn = 3 for C,,E,. These observations, together with those for other homologous compounds, indicate that the end group of the oxyethylene chain affects the conformation of the molecule.The difference in the conformational behaviour between the com- pounds ending in a methoxy group and those ending in a hydroxy group results from the difference of the layer structures of crystals. The conformational variety of n-alkyloligo(oxyethy1ene)s in the solid state is associated primarily with the high flexibility of the oxyethylene chain which assumes the most appropriate conformation by adapting itself to its environment. The structure of block oligomers consisting of the conforma- tionally contrasting n-alkyl and oxyethylene chains has attracted much attention in recent years; the former chain prefers the extended conformation, while the latter prefers the helical conformation.The molecular conformation of a series of 7 1 a-n-alkyl-o-hydroxyoligo(oxyethy1ene)s CH,(CH,),-,(OCH,CH,),OH (C,E,) with n = 1-16 and rn = 1-8 in the solid state has been investigated systemati- cally in our Raman spectroscopic studies, and a number of interesting conformational features have been revealed. '-' These studies have shown that the molecular conformation of the C,E, compounds with n < 4 is basically helical, while the conformation of those with n > 5 depends significantly upon the oxyethylene-chain length. For the latter compou'nds, the conformational transition takes place at rn = 3-4 from the highly extended form to the helical/extended diblock form as the number of oxyethylene units increases. The polymorphic conformational behaviour of C1& in the solid state has been reported in detail in a separate paper.5 The effect of the end group of the molecular chain on the conformation of the whole molecule is another important structural feature of the block oligomeric compounds.The study of the compounds ending in a methoxy group, a-n-alkyl-o-methoxyoligo(oxyethy1ene)s CHJCH,), -1(OCH2CH2), OCH, (C,E,C,), is thus relevant to this problem for comparison with the C,E,, ending in a hydroxy group. In our recent work,, we studied the conformation of C,E,C, (rn = 1-4) in the solid state by Raman spectroscopy and discussed the difference in the conformational behaviour between the compounds ending in methoxy and hydroxy groups.Booth and c~-workers,~-~ on the other hand, have studied the morphology and crystallinity of a variety of C,E,C, compounds by means of Raman spectroscopy, X-ray diffraction and differential scanning calorimetry and dis-cussed the molecular structure of these substances. In the present work, we have extended the conformational studies of diblock n-alkyloligo(oxyethy1ene) compounds in the solid state to another series of compounds terminated by a methoxy group, C, ,E,C, [a-hexadecyl-o-methoxy-oligo(oxyethylene)s] with rn = 1-4. These compounds contain a longer alkyl chain than the C,E,C, compounds studied previously.6 The results of the two series of compounds, C,E,,,C, and C16EmC1, provide further information on the effect of the terminal methoxy group on the conformation of the molecule, in comparison with the effect of the hydroxy group in C,E, and C,,E, .3*4 Experimental Materials C1,E,Cl with rn = 1-4 was synthesized in the present work.C,,E3Cl and C,,E4Cl were prepared by the conventional method of Williamson ether synthesis, while C16ElCl and C,,E,C, were prepared by an improved method using a phase-transfer catalyst.".' 'The materials thus prepared were purified by repeated distillation in vucuo. Raman Spectroscopy The samples of CI6EmC1 were contained in sealed glass ampoules, and their Raman spectra were measured in the solid state at liquid-nitrogen temperature.The spectra were recorded on a JEOL JRS-400D Raman spectrophotometer equipped with a Hamamatsu R649 photomultiplier. The 514.5 nm line of an NEC GLG3200 argon ion laser was used for Raman excitation. A bandpass filter was used to eliminate the laser plasma emission. For spectral calibration, the neon emission lines were utilized. The solid phases of Cl6EmC1 (rn = 1-4) for the Raman measurements were obtained by three different methods of solidification : (1) The liquid sample was cooled rapidly, reaching near liquid-nitrogen temperature in <3 min. (2) The liquid sample was cooled more slowly, reaching the same temperature in >30 min. In this case, the solidified substance was then annealed by warming it to a temperature slightly below the melting point and maintaining it at around this temperature for > 1 h.It was cooled to liquid-nitrogen tem- perature before recording the spectra. (3) When the liquid sample of C16E,Cl was cooled to just below the melting point, a waxy solid was obtained. This waxy solid was cooled rapidly to near liquid-nitrogen temperature. 2770 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Results and Discussion obtained by methods (l), (2) and (3) will be called solids 1, 2 Spectral Analysis and Conformation Determination of and 3, respectively. The observed spectra were analysed in relation to the Cl ,E,Cl molecular conformation by utilizing the conformational key The Raman spectra of C,,E,C, (rn = 1-4) in the solid state bands that had been established previously for C,E,C,.are shown in Fig. 1. For C,,E,C, and C,,E,C,, the spectra Most of these key bands were derived from the con-obtained by methods (1) and (2) are the same, but differ from formation-spectrum correlations for oxyethylene chains. '' was also utilized for the determi- the spectrum obtained by method (3). The compound The LAM-1 ~ibration'~''~ C,,E,C, gave three different spectra depending upon the nation of the molecular structure. Normal-coordinate calcu- solidification method, whereas C,,E,C, gave essentially the lations were performed on a number of possible same spectra irrespective of the method. The solid phases conformations of C,,E,C, molecules in order to confirm the spectral interpretation. In the calculations, the force con- stants established previously for the alkyl and oxyethylene chains' 5.l6 and a programme, MVIB,' were used. The conformational key bands observed for various solid phases of C,,E,C, are listed in Table 1, where the conforma- tional and vibrational assignments of the respective bands are indicated.The distinctive conformational fragments of the C,E,C molecules are those of the oxyethylene-adjoining CH,CH,CH,O' 'Isegment, the inner OCH,CH,O segment and the terminal OCH,CH,OCH, segment. On the basis of the observed key bands, the molecular conformations of C,,E,C, in the solid phases have been determined as sum- marized in Table 2, where the conformations of the corre- sponding C,&, corn pound^^-^ are also given.The skeletal molecular models of C,,E,C, representing the conforma- tions thus determined are shown in Fig. 2. The conformations of C,E,C, and C,E,, as determined in the previous w~rk,~.~ are given in Table 3 for comparison of the molecular confor- mations of the relevant compounds. A brief discussion is given below on the conformation determination of the C,,E,C, compounds studied. C16E1C1 This compound gives the same type of spectra when the solid phase is obtained by the three different methods. The observed key bands indicate that both of the alkyl and oxy- ethylene chains adopt the extended conformation [Fig. 2(a)]. The wavenumber of the LAM-1 band, 116 cm-', is consis- tent with the fully extended molecular structure consisting of 21 coplanar backbone atoms, in conformity with the LAM-1 wavenumber (1 15 cm-') for solid CH,(CH,),,CH, .18 The Raman band at 1415 cm-' corresponds to the band at Y i 1 iIf 0 1200 800 400 0 wavenumber/cm-' (a1 (b) (c) (a (elFig.1 Raman spectra of C,,E,C, in the solid state: (a) CI6E,C, (solids 1-3); (b) C,,E,C, (solids 1 and 2); (c) C,,E,C, (solid 3); (d) Fig. 2 Skeletal molecular models of C,,E,C, : (a) C,,E,C, (solids C,,E,C, (solids 1 and 2); (e) C,,E3C, (solid 3); (f)C,,E,C, (solid 1-3); (b) C,,E,C, (solids 1-3); (c) C,,E3C, (solids 1 and 2); (d) 1); (9)C,,E,C, (solid 2); (h) C,,E,C, (solid 3) C,,E,C, (solid 1);(e) C,,E,C, (solid 2) Table I Conformational key bands" and assignments for C,,E,C, (m = 1-4) in the solid state Raman wavenumber/cm -c v,v,"A 16E2C 1 C16E3C1 1 6E4C 1 C16E1C1 solids 1-3 solids 1, 2 solid 3 solids 1 and 2 solid 3 solid 1 solid 2 solid 3 conformational assignmentb vibrational assignmentc 6 r \oinner OCH,-CH,O in t CH, scissors 0 1497 m 1498 m 1499 m 1415 m 1415 m 1416 m 1414 mw 1411 m extended alkyl chain in orthorhombic CH, scissors or monoclinic cell 1252 w 1253 w 1254 vw terminal (C)O-CH,-CH,-OCH, in tgt CH, twist 1238 w, br 1232 mw 1238 w, br inner (C)O-CH,-CH,-O(C) in tgt CH, twist 1170 w 1166 w 1170mw 1168 w 1170 mw 1170 w 1166 w oxyethylene-adjoining (C)CH,-CH,-CH,-O(C) CH, rock in ttt 954 w 955 w 957 w terminal OCH,-CH,0CH3 in t C-O(CH,) stretch 939 vw OCH,-CH,-O-CH,-CH,O in gt-tg CH, rock 918 vw oxyethylene-adjoining (C)CH ,-CH,-CH,-O(C) C-0 stretch, C-C stretch in tgt 893 mw 891 mw 890mw 890mw 891 mw 890 mw 888 m 889 mw terminal CH,CH,-CH,CH, in t CH, rock, (CH,)C-C stretch 852 mw 854 mw, br 852 mw 850 mw 851 mw, br terminal OCH,-CH,OCH, in gd C-O(CH,) stretch, CH, rock 294 mw oxyethylene chain in tgt helical conformation helix breath 116 vs 105 s 105 s 90 s 83 s molecular chain, fully or partly, LAM-1 (accordion) in the extended conformation a Approximate relative intensities: vs, very strong; s, strong; m, medium; mw, medium-weak; w, weak; vw, very weak; br, broad.* t and g denote trans and gauche conformations, respectively. Conformational assignment is based on the conformation-spectrum correlations for oxyethylene chains" and the normal coordinate analysis. Vibrational assignment is based on the normal coordinate analysis.Weak bands at 852 cn-' forC,,E,C, (solids 1 and 2) and at 857 cm-' for C16E,C, (solid 3) are assigned to the alkyl chain (CH, rock). N44CL J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 compound' Cl6ElCl rapid cooling (solid 1) slow cooling (solid 2) C16E2C1 rapid cooling (solid 1) slow cooling (solid 2) C16E3C1 rapid cooling (solid 1) slow cooling (solid 2) C16E4C1 rapid cooling (solid 1) slow cooling (solid 2) rapid cooling slow cooling C&2d rapid cooling slow cooling C16E3d rapid cooling slow cooling 1 6E4d rapid cooling slow cooling Table 2 Conformations" of C,,E,C, alkyl chain: oxyethylene-adjoining (C)CH,-CH,-CH,-O(C) ttt ttt ttt ttt ttt ttt ttt tgt alkyl chain,b oxyeth ylene-adjoining (C)CH2-CH,-CH,-O(C) ttt ttt ttt ttt ttt tgt and C,,E, (rn= 1-4) in the solid state oxyethylene chain inner terminal (C)O- CH -CH,-O(C) (C)O-CH -C H,-OCH t t t t t t ttt t t t ttt t t t ttt t 9 t ttt t 9 t ttt t 9 t tgt t 9 t oxyethylene chain inner terminal (C)O-CH,-CH,-O(C) (C)O-CH,-CH,OH t g+t t g+t " t and g denote trans and gauche conformations, respectively.Conformation of the alkyl chain is successive trans (t),with the conformation of the oxyethylene-adjoining segment as indicated. ' For the molecular structure of solid 3, see text.Ref. 3-5. similar wavenumbers for orthorhombic and monoclinic n-alkanes.lg This band is interpreted as a crystal-field split component of the CH, scissoring mode of the extended alkyl chain; the other component is observed for C16ElC, at 1444 cm-l. Cl6E2Cl The essential difference between the two spectra, one observed for solids 1 and 2 and the other for solid 3, is present only in the 1400-1510 cm-l region. The former spec- trum shows prominent bands at 1497 and 1503 cm-', while the latter shows a band at 1490 cm-'. A distinct band at 1415 cm-' is observed only for the latter. An examination of the conformational key bands leads to a fully extended molecular structure for the three solids [Fig. 2(b)]. This struc- ture, consisting of 24 coplanar backbone atoms, is further evi- denced by the LAM-1 wavenumber of 105 cm-', which coincides substantially with 104 cm-' for solid CH3(CH,),,CH,.'8 C16E3C1 The two spectra of this compound, one observed for solids 1 and 2 and the other for solid 3, are noticeably different in two ways: thz appearance of a distinct band at 1416 cm-' for salid 3 iiad the broad spectral feature in the region below JWCI crn for solid 3.The key bands of solids 1 and 2 give the molecular structure in these solids phases, in which the dkyl chain is fully extended and the oxyethylene chain is i" :tended except for the terminal oxyethylene segment in the gauche conformation [Fig. 2(c)]. In the spectrum of solid 3, the bands characteristic of the oxyethylene chain are broad or ill-defined.The fact that there are no distinct spectral fea- tures in the low-wavenumber region suggests that the whole molecule is not fully ordered. These considerations indicate for solid 3 that the alkyl chain is ordered by taking the extended conformation while the oxyethylene chain is mostly disordered. cl 6E4C1 Of the three different spectra of this compound, the spectrum of solid 1 resembles the spectra of solids 1 and 2 of C16E3C,, except for a crystal-field split band at 1414 cm-' for solid 1 of C16E,Cl. It is shown for this solid that the molecular chain is extended except for the terminal oxyethylene segment that takes the gauche conformation [Fig. 2(d)]. The spectrum of solid 2 exhibits several distinctive features which are not observed in other spectra of C16E,C,.The bands at 294 and 1232 cm-are obvious indications of the presence of the trans-gauche-trans helical conformation in the oxyethy- lene chain. These key bands, together with others, establish the molecular conformation for solid 2; the oxyethylene chain adopts the helical conformation, while the alkyl chain is extended except for the oxyethylene-adjoining segment in the gauche conformation [Fig. 2(e)]. The overall spectral feature of solid 3 of C16E4C, coincides essentially with that of solid 3 of C,,E,C,. This shows that the alkyl chain takes the ordered extended conformation but the oxyethylene chain is mostly disordered. J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 Table 3 Conformations" of C8E,Cl and C8E, (rn = 1-4) in the solid state alkyl chain: oxyethylene-adjoining compound (C)CH,-CH ,-CH ,-O(C) C8E1ClC rapid cooling (solid 1) slow cooling (solid 2) C8E2C1C rapid cooling (solid 1) slow cooling (solid 2) C8E3C1c rapid cooling (solid 1) slow cooling (solid 2) C8E4ClC rapid cooling (solid 1) slow cooling (solid 2) ttt ttt ttt ttt ttt ttt tgt ttt alkyl chain: oxyethylene chain inner terminal (C)O -CH ,-CH, -O(C) (C)O-CH,-CH,-OCH, t t ttt tg t ttt tt t ttt t ttt t t t oxyethylene chain (C)CH, -CH ,-CH,-0( C) oxyethylene-adjoining inner (C)O-CH,-CH,-O(C) terminal (C)O-CH,-CH,OH C8E Id rapid cooling slow cooling t t t t t t t g+t t g+t C8E2d rapid cooling t t t t t t slow cooling t t t t t t rapid cooling C8E,d t t t t t t slow cooling t g t t g t rapid cooling C8E,d slow cooling " t and g denote trans and gauche conformations, respectively.Conformation of the alkyl chain is successive trans (t), with the conformation of the oxyethylene-adjoining segment as indicated. Ref. 6. Ref. 3. Conformational Behaviour of C,E,C, and C,E,: Effect of the End Group The important confdrmational features of C,,E,C, (m= 1-4) in the solid state are given in Table 2, in comparison with those of C16Em.3--5The conformations of C,,E,C, are summarized as follows: The molecules of C,,E,C, and C,,E,C, are in the all-trans extended conformation, while the molecules of C,,E3C, in the rapidly cooled and slowly cooled solids and of C,,E,C, in the rapidly cooled solid adopt the extended conformation except for the methoxy group terminated oxyethylene group.The C,,E,C, mol-ecules in the slowly cooled solid, on the other hand, take the helical conformation for the oxyethylene chain and the extended conformation for the alkyl chain, the oxyethylene- adjoining alkyl part being transformed into the helix. When the waxy solid of C,,E,C, and C16E4C1, obtained by cooling the liquid just below the melting point, is refrigerated, only the alkyl chain is crystallized by taking the extended conformation, but the oxyethylene chain is disordered. These conformational features of C16E,C, are compared with those of the hydroxy group terminated compounds C16Em.3--5The effect of the end group and the chain length on the molecular conformation is noted mostly in the oxyethylene-chain part.The conformational change of the inner oxyethylene segments from the extended to the helical structure takes place at rn = 4 for C16E,C1, while it takes place at m = 3 for C,&, (Table 2). The boundary com-pounds C,,E,C1 and C,,E, in fact adopt two conformations depending upon the solidification conditions. This indicates, under the notion of the conformational competition,20 that four or more oxyethylene units are necessary for C,,E,C, to establish the stable helical structure of the oxyethylene chain in competition with the stability of the extended structure of the hexadecyl chain.For C,&,, on the other hand, three or more oxyethylene units suffice to stabilize the helical struc- ture of the oxyethylene chain. According to the previous work on the shorter homologues C,E,C, and C,E,,, a similar conformational transformation of the oxyethylene chain from the extended to the helical structure is observed between m = 3 and 4for the methoxy group terminated com- pounds and at m= 3 for those terminated by a hydroxy group (Table 3). On the basis of these conformational charac- teristics associated with the end group, it is apparently shown that the oxyethylene chain with a terminal methoxy group, -(OCH,CH,),OCH,, is less effective than that with a ter- minal hydroxy group, -(OCH,CH,),OH, in retaining the intrinsic helical structure against the conformational effect of the alkyl chain to propagate into the oxyethylene chain.Booth and co-worker~~.~ have investigated the crystallinity of n-alkyloligo(oxyethy1ene)s and have shown that the hydroxy group terminated C,E, molecules crystallize into bilayers while the methoxy group terminated C,E,C, mol-ecules crystallize into monolayers. Two C,E, molecules are linked to each other through strong hydrogen bonds across the end-group plane to form the bilayer ~tructure.~ The more effective retention of the helical structure of the oxyethylene chain in C,E, than in C,E,C, molecules can be explained by the linkage structure of the two oxyethylene chains of a hydrogen-bonded dimer in the bilayer; the stabilities of the helical structures of the two oxyethylene chains are mutually reinforced.The different conformational behaviour of the methoxy group terminated and hydroxy group terminated n-alkyloligo(oxyethy1ene)s is thus explained by the different layer structures of the crystals. The previous spectral observa- tion that the LAM-1 wavenumbers of the helical oxyethylene chain for C,E, molecules are about half those for the corre- sponding C,E,Cl molecules has also been explained by the doubled chain length in the bilayer ~tructure.~ Disordered Conformation of the Oxyethylene Chain In solid 3 of C16E,Cl and C16E4Cl,the alkyl chain takes the ordered extended conformation, while the oxyethylene chain is mostly disordered. For C16ElCl and C16E,Cl, however, no such structures containing the disordered portion are found.Our previous study' on the polymorphic C16E3com-pound showed the existence of the same structure in the solid phase obtained by the same method (solid B). These experi- mental findings suggest that the C,E,Cl and C,E, com-pounds that contain a long alkyl chain and a relatively short oxyethylene chain tend to yield the solid phase in which the oxyethylene moiety is non-crystalline. The crystallization of the oxyethylene block in n-alkyloligo(oxyethy1ene)s is thus greatly influenced by the length of each of the two blocks constituting the molecule and the process yielding the solid phase. Booth and co-workers21*22have in fact classified the structures of the tri- block compounds C,E,C, on the basis of the crystallinity of the alkyl and oxyethylene chains ; the crystallinity depends significantly upon the relative lengths of the three blocks.Conformational Variety of n-Alkyloligo(oxyethylene)s The molecular conformation of a number of n-alkyloligo(oxyethy1ene)s in the solid state has been studied extensively by Raman spectroscopy in the present and pre- vious ~ork.'-~,~~ These studies have revealed that C,E, and C,E,C, compounds exhibit a rich variety of conformations in the solid state. The chain segments that characterize the con- formational state of the molecule are the oxyethylene-adjoining alkyl segment (C)CH,-CH,-CH,-O(C), the oxy- ethylene segment in the inner part (C)O-CH,-CH,-O(C), the hydroxy group terminated oxyethylene segment (C)O-CH,-CH,OH and the methoxy group terminated oxyethylene segment (C)O-CH, -CH, -OCH, .The alkyl chain generally takes the extended structure, but in some cases the oxyethylene-adjoining segment adopts the trans-gauche-trans conformation (e.g. fi form of C,Em4).A much greater variety is noted for the conformation of the oxyethylene chain. While the intrinsically stable conformation of repeated trans-gauche-trans is observed for the successive inner oxyethylene segments in a relatively long oxyethylene chain (a and fi forms of C,Em4; C8E4Cl and C16E4Cl),the extended trans-trans-trans conformation is usually attained in a short oxyethylene chain except for the terminal part, when a long alkyl chain is linked to the oxyethylene (y and yt forms of C,Em4).The conformation of the hydroxy group ter- minated oxyethylene segment in C,E, is either trans-gauche-trans (y form) or trans-trans-trans (yr form).For some of the C,E,Cl compounds with a short oxyethylene chain, the methoxy group terminated oxyethylene segment adopts the trans-gauche-trans conformation, whereas the inner oxyethy- lene segments are in the trans-trans-trans conformation (C,E2Cl, C8E,C, and CI6E3C,).Other C,E,Cl compounds with a short oxyethylene chain prefer the trans-trans-trans J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 conformation for the methoxy group terminated oxyethylene segment together with the inner oxyethylene segments (C16E1C1 and C16E2C1).The conformational variety as revealed for a large number of C,E, and C,E,C, compounds is associated primarily with the high flexibility of the oxyethylene chain which assumes the most appropriate conformation by adapting itself to its environment.The molecular conformation of the block com- pounds consisting of the oxyethylene and alkyl chains is practically determined by the relative lengths of the constitu- ent blocks and by the end group of the oxyethylene chain. Conclusions Raman spectroscopic studies of n-alkyloligo(oxyethy1ene)sin the solid state have shown that the end group of the oxyethy- lene chain affects the conformation of the molecule. The dif- ference in the conformational behaviour between the compounds terminated by methoxy and hydroxy groups results from the difference of the layer structures of the crys- tals.The strong hydrogen bonds between the hydroxy group terminated oxyethylene chains in the bilayer are responsible for the more effective retention of the helical structure of the oxyethylene chain in the C,E, molecules in competition with the stability of the extended structure of the alkyl chain. Conformational variety is one of the most prominent struc- tural features of n-alkyloligo(oxyethy1ene)s. This property stems from the high flexibility of the oxyethylene chain. The peculiar conformational behaviour plays an important role in various functional substances that contain the oxyethylene chain.We thank Dr. Koichi Fukuhara of Hiroshima University for valuable discussions. References 1 H. Matsuura and K. Fukuhara, Chem. Lett., 1984,933. 2 H. Matsuura and K. Fukuhara, J. Phys. Chem., 1986,90,3057. 3 H. Matsuura and K. Fukuhara, J. Phys. Chem., 1987,91,6139. 4 H. Matsuura, K. Fukuhara, S. Masatoki and M. Sakakibara, J. Am. Chem. SOC., 1991,113,1193. 5 S. Masatoki, K. Fukuhara and H. Matsuura, J. Chem. SOC., Faraday Trans., 1993,89,4079. 6 S. Masatoki, H.Matsuura and K. Fukuhara, J. Raman Spectro-sc., 1994,25, in the press. 7 K. Viras, F. Viras, C. Campbell, T. A. King and C. Booth, J. Chem. SOC., Faraday Trans. 2,1987,83,917. 8 J. R. Craven, 2.Hao and C. Booth, J. Chem. SOC., Furaday Trans., 1991,87, 1183. 9 J. R. Craven, K. Viras, A. J. Masters and C. Booth, J. Chem. SOC., Faraday Trans., 1991,87,3677. 10 H. H. Freedman and R. A. Dubois, Tetrahedron Lett., 1975, 3251. 11 T. Gibson, J. Org. Chem., 1980,45, 1095. 12 H. Matsuura and K. Fukuhara, J. Polym. Sci., Part B: Polym. Phys., 1986,24,1383. 13 S. Mizushima and T. Shimanouchi, J. Am. Chem. SOC., 1949,71, 1320. 14 R. F. Schaufele and T. Shimanouchi, J. Chem. Phys., 1967, 47, 3605. 15 T. Shimanouchi, H. Matsuura, Y. Ogawa and I. Harada, J. Phys. Chem. Ref: Data, 1978,7,1323. 16 H. Matsuura, K. Fukuhara and H. Tamaoki, J. Mol. Struct., 1987,156,293. 17 H. Matsuura, Comput. Chem., 1990,14,59. 18 H. G.Olf and B. Fanconi, J. Chem. Phys., 1973,59,534. 19 F. J. Boerio and J. L. Koenig, J. Chem. Phys., 1970,52, 3425. 20 K. Fukuhara and H. Matsuura, Chem. Lett., 1987, 1549. 21 R.C. Domszy and C. Booth, Makromol. Chem., 1982,183,1051. 22 H. H. Teo, T. G. E. Swales, R.C. Domszy, F. Heatley and C. Booth, Makromol. Chem., 1983,184,861. 23 H. Matsuura, K. Fukuhara and 0. Hiraoka, J. Mol. Struct., 1988,189,249. Paper 4/02904F; Received 16th May, 1994
ISSN:0956-5000
DOI:10.1039/FT9949002769
出版商:RSC
年代:1994
数据来源: RSC
|
|