Search

X-ray measurements have been made of changes in the lattice spacings of Rochelle salt between —90 and +40° C. Rough visual estimates have also been made of changes in the intensities of X-ray reflexions with temperature. The thermal properties of Rochelle salt have been correlated with its crystal structure. As with other crystals containing hydrogen bonds, thermal expansion is largest in the direction of these bonds, but above the upper Curie point this predominance largely disappears. Between the two Curie points the thermal expansion shows anomalies and is subject to hysteresis. Two main theoretical problems are discussed in the light of the observations. In the first place, the thermodynamic changes at the lower and upper Curie points have hitherto appeared to be transitions ‘of the second kind’, occurring within single crystals without change of phase. Fresh X-ray evidence shows that over the transition range of temperatures measurable differences can be detected between subcrystalline domains within a ‘single’ crystal. These differences reach a maximum value in certain directions in the crystal, but appear to be insufficient to cause it to break up into powder, with change of phase, at the two Curie points, so that most of the properties of a single crystal are retained over the whole range of temperatures. The experimental conclusions throw light on the general theory of phase transitions of the second kind. Secondly, the onset of anomalous dielectric properties at the lower Curie point is ascribed to the stretching of short hydrogen bonds owing to the thermal expansion of the crystal. This changes the bonds from a non-polar to a polar character and is correlated with thermal effects in other crystals containing hydrogen bonds. In extension of apparatus previously described a new controlled temperature chamber has been devised, which has been used with various X-ray cameras for single crystal work at temperatures down to — 90° C.

Considerable frequency dispersion of electrical conductivity has been ob­served in the liquid crystal phases of organic ionic melts, depending on their molecular structure and domain size. Dispersion began to be detectable above about 10000 Hz for sodium isovalerate (in which there are large domains), and above about 2000 Hz for sodium n -butyrate (domains about 0.1 mm across). Neither molten salt showed dispersion in a middle range of frequencies (between 1000 Hz and 100Hz). Previous researches had shown a marked diamagnetic anomaly of molten organic salts when these are cooled through their mesophases, while applying static magnetic fields. The kind of response observed can be attributed to interaction between ‘liquid crystal’ domains and the applied fields. A related effect has now been observed when a static magnetic field (approx. 10kG) was applied in the course of the above measurements of electrical conductivities. With sodium isovalerate melts, a magnetic field applied perpendicular to the electric field reduced by a factor of two the drop in electrical conductivity when the isotropic melt passed into its mesophase. No effect (≯ 2%) was detected with sodium n -butyrate under the same conditions.

A simple but sensitive horizon test is described to determine whether the mesophases of molten organic salts exhibit any detectable elastic-plastic threshold, before they begin to flow under even very small applied shear stresses. By using this test with sodium isovalerate (mesophase range 188–280 °C) no detectable threshold is found above about 260 °C; above this temperature the melt thus behaves as completely fluid. However, between this temperature and the first melting point (188 °C) a definite limit stress appears to be required to start flow. Other properties of the mesophase also show changes around 260 °C with some analogies to fluid–glass transi­tions in other systems. Molten sodium n -butyrate, when tested in a similar way, fails to exhibit any distinctive elastic-plastic threshold. So far as can be detected by the horizon test, it remains completely fluid over the whole of its meso­phase range (252–324 °C).

An improved method for making measurements normal to the layers in lamellar compounds is described. This preserves good contact under moderate pressures, while permitting unobstructed expansion of the specimen during intercalation, or changes of temperature. As a particular example detailed measurements are reported on the electrical resistance of graphite nitrates for temperatures traversing the thermal transformation at about 252 K. Magneto-resistance data have also been obtained. Mechanisms of charge transport perpendicular to the layers are discussed in relation to present findings for graphite nitrates. From the data available, extrinsic conduction paths would appear to make substantial contributions to the total c-axis conductivity.

The interleaved texture of human growth can sometimes be followed back to more than one kind of early experience, and often stems from more than one genetic element. Subtle and continuous elements affecting the life of Dudley Newitt came from his youthful experience and from his parentage. These had superimposed upon them uncommon events which helped to segment his life into strongly marked periods. In his origins Newitt was basically a Fen-man. His father’s great grandfather was a moderately prosperous farmer in Chatteris on the Isle of Ely, of which he was at one time High Constable. Dudley’s tenacity and toughness were characteristic of the Fens, though he was not without a romantic streak. When we visited Ely and its cathedral together in the 1950s it was not fanciful in me to observe the quickening of his countenance. But his own grandfather had left the Fens in search of his fortune when barely grown up—walking to London, eating vegetables from the fields. He married the daughter of an ironmonger in Ware in whose shop he found his first job on leaving the Fens. In typical Victorian entrepreneurial pattern for the industrious apprentice, he subsequently built racecourses, managed hotels, speculated in land in England and in South Africa, made a small fortune and lost two. Newitt’s maternal grandparents were a craftsman in Windsor who is reputed to have made the leather upholstery for Queen Victoria’s state coaches, and the daughter of a small farmer from the Welsh border country.

Various physical properties of alkali carboxylates can be used to study precursor effects immediately below or above their melting points. The present paper describes unusual pre-transition behaviour observed in the magnetic susceptibility of molten n -butyrates and isovalerates. This takes the form of a transitory enhancement of diamagnetic susceptibility, by nearly 10% in some cases, when the isotropic melt is cooled through the clearing point and into the mesophase region while a magnetic field (up to 1 T(10 4 G)) is applied. No corresponding changes are observed on taking the salt through the same temperature range but in the inverse direction. Again, no corresponding change is observed with an alkali carboxylate which shows no mesophase melting transition. Present findings can be accounted for on the assumption of cluster formation in the molten salts as these approach their upper melting points, which is also indicated by other evidence. The (transient) enhancement of molecular diamagnetism in this region of temperature may also indicate transient perturbations in the actual cation-anion binding near the transition temperature.

X-Ray measurements have been made on single crystals of caesium chloride from 20 to 469° C 495° C traversing the transition CsCl (b. c. c.) ⇌ 469°C CsCl (f. c. c.). Contrary to previously published statements which were all based on CsCl powder, the lattice expansion of the b. c. c. crystal remains approximately a linear function of temperature to within 1° of the transition point. Above about 380° C the intensity of X-ray reflexions decreases more rapidly than the Debye-Waller theory predicts. Tentative explanations are discussed. As the temperature rises intensity measurements show that the mean amplitude of vibration of the Cs + ion increases faster than that of the Cl - ion. Little, if any, correlation is found between the orientation of a single crystal of b. c. c. CsCl and the assemblage of crystallites into which it breaks up above the transition point. The number of such crystallites formed is small, indicating that the rate of nucleus formation is slow compared with their rate of growth. No ‘interaction’ between the two phases can be inferred from the measurements of lattice spacings. A simple hypothesis is suggested to explain the tendency for CsCl to form distorted single crystals.

Hydrogen electrolyzed into palladium changes the internal friction. In torsional oscillations at 0$\cdot $54 and 0$\cdot $22 c/s an immediate large increase of internal friction is observed after charging, or after removing some of the hydrogen by electrolysis. On ageing, the internal friction in such specimens falls to a limiting value, without appreciable loss of hydrogen. This decrease of internal friction obeys approximately a unimolecular rate law, with velocity constant markedly affected by the grain size of the metal. Decay is fastest in annealed specimens with large grain sizes, where in the limit the internal friction returns almost to the value for pure palladium, notwithstanding the presence of about 0$\cdot $6 H/Pd. In fine-grained charged specimens the internal friction decays to limiting values rather larger than for the pure metal. The observations on palladium/hydrogen suggest that there is a diversity of relaxation processes in the alloy. Electrolysis produces large but temporary 'cold-working' which is accompanied by high internal friction. Rested alloys show permanent relaxation effects related to the initial grain size. Brief discussion is given of measured effects of hydrogen on Young's modulus and the rigidity modulus of palladium, both of which are somewhat decreased on charging with hydrogen.

A microthermostat design for single crystal studies is described which gives control of specimen temperature in the range 30 to 100°C to an accuracy of ± 0.1 degC. For X-ray diffraction experiments the furnace comprises a beryllium tube chamber. For optical studies, the control system is combined with a chamber suitable for mounting on the stage of a polarizing microscope. Temperature cycles with single crystals of potassium nitrite indicate that the thermal transformation around 40°C from a monoclinic structure to one which is probably cubic, is continuous and reversible. Hysteresis is found in the transformation; high- and low-temperature structures co-exist in a hybrid single crystal over a temperature range.

Observations are recorded on X-ray spectra diffracted from cathodes of iron and palladium during actual discharge and diffusion of hydrogen. For iron, no expansion of the crystal lattice greater than 0$\cdot $0002 angstrom could be observed during such diffusion, and the broadening observed of X-ray diffraction spectra indicated no preferred crystallographic planes for the evolution of molecular hydrogen. Blisters were formed across the faces of single crystals of iron, when made a cathode, indicating that diffusion of hydrogen occurs through the crystal grains rather than along crystal boundaries. For palladium, an expansion was observed during actual electrolysis to an extent of 0$\cdot $0170 $\pm $ 0$\cdot $0002 angstrom. On ceasing electrolysis, there was reversion to a less-expanded lattice of palladium hydride. Broadening of the reflexions observed during electrolysis indicated no preferred planes for evolution of hydrogen. The significance of these findings is discussed for mechanisms of overvoltage and of diffusion of hydrogen through metals.

Extensive randomization of molecular orientations or ‘rotation’ is generally thought to occur on melting of certain fairly simple substances. The onset of ‘rotation’ on melting of more complex molecules is reviewed. Special attention has been directed to the behaviour of ‘stiff’ molecules such as benzene derivatives and fused aromatic rings, for which intra­molecular crumpling in the melt is likely to be unimportant. The fractional increase in volume ∆Vf/Vson melting has been evaluated for a range of compounds from existing density measurements. ∆Vf/Vshas also been determined experimentally for the molecules 2, 6-dimethyl naphthalene, 2, 3-dimethyl naphthalene, acenaphthene, anthracene, phenanthrene and chrysene. Using bond-distances and Van der Waals repulsion envelopes established by X-ray studies, the increase in volume required to permit ‘rotation’ about different mole­cular axes in the melt has been calculated. Comparison with the volume increase on melting observed experimentally shows that in many melts containing ‘stiff molecules’ there can be no ‘rotation’ of the molecules. In the present experiments no anomalies occur in such melts even 50° above the melting-point, but it is not yet possible to say whether ‘rotational’ transitions would be observed at higher temperatures in the liquids, similar to the well- known transitions in crystals. On the other hand, ‘rotation’ about at least one of the mole­cular-axes probably does accompany melting for some of the molecules studied, including benzene, bromobenzene,p-xylene, acenaphthene, naphthalene and 2, 6- but not 2, 3-dimethyl naphthalene.

Graphite as deposited by cracking hydrocarbons around 2000 °C shows marked preferred orientation with consequent anisotropy, but still contains numerous structural defects. Procedures have been investigated for removing these by recrystallization around 3000 °C under a shearing stress, resulting from applying uniaxial pressures of about 400 atm to cylindrical specimens. Subsequent annealing at around 3400 °C under light confinement yields blocks of graphite whose crystal properties are comparable with those of small natural single crystals, as indicated by various tests.

Single crystals of pure potassium cyanide, and of potassium cyanide containing various amounts of sodium cyanide in solid solution, have been subjected to cyclic temperature changes traversing the thermal transformation at around - 110 degrees C. By means of Laue and Bragg photographs, persistence of the crystal axes has been observed in a 'normal cycle' high $\rightarrow $ low $\rightarrow $ high, but a metastable crystal form was also produced by a particular sequence of changes. There was considerable hysteresis in the transformation of single crystals. Introducing lattice disturbances by incorporating sodium cyanide in solid solution broadened the transformation. Co-existence was observed of various domains of the low-temperature form in 'hybrid' single crystals; there was also evidence for the co-existence of domains of the low- and high-temperature forms over a range of temperatures. These results are discussed in relation to theories of continuous thermodynamic transformations in solids.

Using synthetic near-ideal graphite obtained by hot-pressing and annealing pyrolytic material, we have prepared graphite nitrates from the first to the fourth sequence, under controlled conditions. For the first sequence compound, present determinations on form I (room temperature) agree with previous publications and confirm the stacking sequence A | A | A | A . Form I sequences 2, 3 and 4 show well defined stacking of the graphite networks, characterized by X-ray methods as follows: sequence network stacking* system unit cell (Å) space group 2 A | AB | BC | CA | A rhombohedral a H = 2⋅46 R 3 ¯ m c H = 33⋅45 3 A | ABA | ACA | A orthorhombic a = 2⋅46 Cmc 2 1 b = 4⋅26 c = 28⋅96 4 A | ABAB | BCBC | CACA | A rhombohedral a H = 2⋅46 R 3 ¯ m c H = 53⋅5 * The vertical lines imply layers of intercalate. In all these compounds, the stacking sequences found can be generated from the stacking in normal graphite, ABAB , if the change AB → A | A on entry of successive intercalate layers involves the movement of a boundary dislocation through the structure. Since it is observed that the end-product (sequence N ) under any given conditions is formed by way of a systematic numerical succession of higher sequence compounds ( N +2, N +1) as intermediates, this fundamental stacking change must reverse so that AB ⇌ A | A . When any of these sequences of graphite nitrate are cooled through a λ transformation around —20°C, increased ordering appears in the low temperature structures (form II). Each intercalate layer now exhibits two-dimensional long range order. In addition, successive intercalated layers within any crystallite are stacked in a non-random way, which in the case of the first sequence compound implies a structure that is periodic in three dimensions. Within any intercalated layer in form I, the molecular packing resembles that in liquids. In a given crystal compound the numerical sequence of such layers is regular, and successive intercalate layers occur after N carbon hexagon planes. In form II, molecular packing in the intercalated layers is crystalline. Some correlation also appears between the occupied sites in different filled layers. Implications of this order-disorder transformation are briefly discussed.

Halides that are active Friedel Craft catalysts with aromatic compounds have been used to prepare synthetic metals. Aluminium bromide AlBr 3 intercalates readily into well oriented graphite between 70 and 80 °C. Second stage intercalation compounds are formed with approximate composition (C 33 AlBr 3.6 ). Aluminium chloride AlCl 3 does not intercalate spontaneously to any appreciable extent, unless free chlorine is also present. In this case, intercalation begins at about 60 °C and reaches a limit after several hours between 70 and 80 °C to produce a first stage intercalate with approximate composition (C 18 + AlCl 4 - .AlCl 3 ). In the direction of the a -axes both halide intercalates show good metallic conductivity with p characteristics. Unlike the parent graphite, temperature coefficients of resistivity of these synthetic metals are comparable with those of natural metals, indicating that thermal vibrations of charged atoms or molecules in the crystal compound contribute a dominant scattering mechanism controlling the electrical conductivity.

Melting mechanisms have been studied for diphenyl, and o -, m - and p -terphenyl. The volume increase on melting is largest for the linear polyphenyls and smallest for branched molecules which can interlock in the melt. In no case will the volume increase permit the ‘onset of rotation’ on melting. Viscosity measurements give marked indications of the growth of cybotactic groups in the melts on approaching the freezing-points, especially in the case of o -terphenyl, where the , ratio H vap / E n reaches exceptionally low values near the freezing-point. Plots of fluidity against specific volume and the marked tendency of the melt of o- terphenyl to supercool give further indications of the presence of interlocked groups of molecules in the melt.

The heat of formation of 1 g. atom of metal from the atoms has been calcu­lated to differ by about 1·6 kcal. for hydrogen and deuterium (Wigner and Huntingdon 1935). Allowing for the difference in the zero point energies of molecular hydrogen and deuterium (1·8 kcal. /mol. according to Rideal and Melville (1935)), the heat of formation of the metal from molecular hydrogen is smaller by 0·7 kcal./g. atom than from molecular deuterium. The difference in zero point energy between H and D in theα-phase is 0·5 kcal. /g. atom (Rideal and Melville). No data are available for theβ-phase. This difference is about 6 % of the heat of solution in forming theβ-phase, and since the expansion of the palladium lattice is sensitive to small energy changes (cf. Part I), an appreciable difference in expansions may be expected in the two alloys (cf. the discussion of other types of chemical linkage, Ubbelohde (1936). Measurements of the lattice expansion, and change of resistance on alloying with the two metals were undertaken to test this.

One of the most important thermodynamic properties of a crystal in relation to its structure is the thermal expansion, which provides information on the lattice vibrations, and thus indirectly on the binding forces. The experiments described in this paper were carried out with a view to obtaining fresh information on hydrogen and hydroxyl bonds in crystals from measurements of the thermal expansion. A comparison of the thermal expansion of corresponding- hydrogen and deuterium compounds from about 90° K to room temperature was made in order to distinguish between lattice deformations due to zero point energy and thermal energy respectively. The marked anisotropy of thermal expansion in crystals containing hydrogen bonds throws interesting light on the structure.

Changes in ultra-violet absorption spectra have been measured for nitrates and nitrites of the alkali metals and for ammonium, silver and thallous nitrates, in relation to thermal transformations in these solids. A gradual shift of absorption bands is found to accompany thermal expansion of all the crystals. More abrupt changes in absorption spectra are observed at transformation points from one crystal structure to another; in many instances these two effects of temperature can be correlated. In cases where crystal structures are known, shifts in absorption maxima due to changes in temperature have been interpreted. in relation to changes in the distances between cations and anions. By means of the techniques described, ultra-violet absorption spectroscopy provides a sensitive means for studying various thermodynamic effects in phase transformations. Thus with ammonium nitrate, hysteresis with both superheating and supercooling is observed for the transitions IV ^ III ( T c ~ 32 °C) and III ^ II ( T c ~ 88 °C), but not II ^ I ( T c ~ 125 °C). With potassium nitrate during cyclic changes of temperature a (metastable) crystal form III appears. With potassium nitrite, a new crystal transformation showing hysteresis ( T c ~ 40 °C) is readily detected by ultra-violet spectroscopy.

Experiments on the abnormal conductance of caesium chloride (Harpur, Moss & Ubbelohde 1955) show that this can be altered in various ways. Methods described include heating the salt at 400° C in atmospheres with controlled partial pressures of oxygen, or in nitrogen containing gaseous hydrogen chloride. The conductances of samples of caesium chloride treated with hydrogen chloride show a pre-transition increase of lattice defects on approaching the transitionbcc→fccat 469° C. Micro-analyses show a ratio ([Cs+] — [Cl–])/[CsCl] in the salt as ordinarily obtained, ranging from about 0·000018 to 0·00013. After heating in oxygen this can rise to as much as 0·00066. It is suggested that some of the vacant chloride-ion sites are occupied by the O2-ion. After prolonged heating in oxygen caesium chloride exhibits oxidizing properties which are tentatatively attributed to a proportion of divalent caesium ions of the order of magnitude [Cs2+]/[Cs+] = 0·0007. After treatment with a mixture of hydrogen chloride and nitrogen, this oxidizing power and the excess caesium ratio fall below the limit of detection. Possible conductance mechanisms are discussed in the light of these experiments. The results also give additional information about the nature of the thermal transition. This shows marked hysteresis, suggesting some smearing of the free-energy curves of the two phases due to co-existence, within one crystal, of subcrystalline regions of the two crystal phases over a narrow range of temperatures, with contributions to the free energy from strain and internal surface energy.

Specimens of well oriented stress-annealed pyrolytic graphite mounted as electrical conductors have been oxidized under controlled conditions to give the acid salts, graphite nitrate and graphite bisulphate. Progressive intercalation in forming these compounds has been followed by measurements of the electrical resistivity, the magneto-resistance, and the Hall effect. Various electronic properties confirm a model for the acid salts in which layers of carbon hexagon networks act as macro-cations, separated by layers in which acid anions are linked to additional molecules of acid by hydrogen bonds. In conformity with this model, as the concentration rises, the resistance drops, at first steeply, finally asymptoting to a limiting value characteristic of a synthetic metal. The Hall effect which is negative for the near-ideal graphites used as starting materials passes through zero and finally becomes positive, confirming the general nature of the charge transfer hitherto assumed. The magneto-resistance measured a t 6 kG decreases, to become practically zero as intercalation proceeds.

An investigation of the properties of metallic hydrogen is of particular interest, on account of the simple structure of a metal composed of protons and electrons. A calculation of the energy change on forming a body centred lattice of metallic hydrogen from hydrogen atoms has recently been made by Wigner and Huntingdon (1935). They find that the energy of formation of metallic hydrogen from hydrogen atoms would be 10 kcal., and of metallic deuterium from deuterium atoms 11·6 kcal., but that metallic hydrogen (Graham’s hydrogenium) would only be stable relatively to covalent hydrogen at pressures not less than 2·5 × 10 5 atmospheres. Experiments on hydrogen at these pressures have not yet been realized. It is however possible to study the properties of metallic hydrogen in a number of alloys. At high temperatures, hydrogen dissolves to an appreci­able extent in a number of metals (Sieverts and Gotta 1928), but at low temperatures the only examples where hydrogen dissolves appreciably without forming covalent links (as GeH 4 ) or negative ions (as LiH) are the “metallic” hydrides of transitional elements such as palladium, tantalum, titanium, etc. The exceptional position of these low temperature alloys of metallic hydrogen is referred to again below. The evidence that hydrogen is in the metallic state has been reviewed (Ubbelohde 1931) and may be summarized by the statements that the hydrogen dissolves as atoms, and that at least a portion of these atoms is ionized to give electrons and protons.

Velocity/composition relationships and detonation limits have been determined for the system S c =C 2 N 2 +O 2 , diluted with C 2 N 2 , O 2 , He or Ar. Limit values found are, for 1 in. diameter stainless steel tubes system moles% diluent at limit S c =C 2 N 2 51·9 ±0·1 + O 2 72·0 ±0·1 + He 82·20 ±0·1 + Ar 90·4 ±0·2 Systematic though minor fluctuations of velocity are found with the cyanogen systems, similar to those previously detected with S H =2H 2 +O 2 and S D )=2D 2 +O 2 . The system S c =C 2 N 2 shows a rise in Mach product γ M 2 on approaching the cyanogen-rich limit. This appears to be characteristic of the formation of solid detonation products, whose energy of condensation can feed back into the detonation wave under suitable conditions. Argon as diluent extends the detonation limits of cyanogen/oxygen mixtures more than does helium . Similar mass effects were previously found with fuel molecules containing hydrogen. This indicates that Soret gradients of hydrogen atoms at the detonation front are not a unique or even a very likely explanation of the mass effects observed. More probably, argon differs from helium in its role as an energy transfer catalyst.

Ultra-violet absorption spectra of ions in crystals are sensitive to their environment of neighbouring ions. This fact is utilized to study melting processes in nitrates and nitrites, with particular reference to changes in the average shell of nearest neighbours around any ion, on passing from crystal to melt. For the nitrates of sodium, potassium, rubidium, caesium, thallium and silver, for ammonium nitrate and for the nitrites of sodium and potassium this shell appears to contract on melting, which is attributed to the formation of association complexes in the melt. For the two crystal hydrates HNO 3 . H 2 O and HNO 3 . 3H 2 O, corresponding changes are observed, suggesting that the structure and absorption spectra correspond with the ions (OH 3 )+ and NO - 3 in the crystals, likewise giving rise to tighter association complexes on melting. Pure nitric acid on the other hand behaves in a manner more closely similar to isopropyl nitrate.

Measurements have been made on the changes in lattice spacing of various phthalocyanines, over a range of temperatures from 90 to 600° K. Rough visual estimates have also been obtained of changes in the intensities of reflexions with rise in temperature. The molecular movements in the crystals have been calculated from the experimental data. Attention is drawn to the difference in behaviour of hydrogen and platinum phthalocyanines, which is probably due to the increased space required by the platinum atoms as the temperature rises, compared with the hydrogen in hydrogen phthalocyanine. Apparatus described includes a new design of furnace, suitable for use with various X-ray cameras for single crystal work, and a moving film oscillation camera, which uses the multiple exposure principle to determine small changes in lattice spacing with high relative accuracy.

Flames consist of maintained chemical reactions, and the concentrations of labile chemical species such as hydrogen atoms, may greatly exceed the concentrations for chemical equilibrium at the flame temperature. This should influence the attack on metals by flames. To test this, apparatus has been constructed for measuring the rate of permeation of hydrogen and other gases from flames through metal septa. Relative rates of permeation have been compared for a hydrogen/oxygen/nitrogen flame: (1) with metal septa consisting of decarburized chrome steel, chrome steel, decarburized chromium/nickel alloy steel, or nickel. (2) with metal septa consisting of decarburized steel or nickel, when small quantities of sulphur dioxide or hydrogen sulphide are mixed with the hydrogen/oxygen/nitrogen flame. For the different metals and the same reference flame (free from sulphur) relative rates of permeation at around 820° C are in the sequence: nickel > chrome steel ≃ decarburized chromium/nickel alloy steel > decarburized chrome steel. Approximate relative rates are nickel, 1000; chrome steel, 300; decarburized chromium/ nickel alloy steel 300; decarburized chrome steel 10. The addition of small quantities of sulphur dioxide to the flame at once reduces the concentration of hydrogen atoms, as observed with a molybdenum trioxide test. Comparable amounts of hydrogen sulphide added to the flame have little or no effect. Both types of sulphur compound rapidly change the rate of permeation of hydrogen through the metal septa from the flame. An increased permeation rate is observed for decarburized chrome steel, and a decreased permeation rate for a decarburized chromium/nickel/iron alloy. An initial drop in permeation rate through chrome steel with sulphur dioxide in the flame is probably due to the decreased concentration of hydrogen atoms. The subsequent effects for both sulphur dioxide and hydrogen sulphide are attributed to the formation of sulphur compounds either on the surface or within the crystal lattice of the metals.

Complex theories of transfer from translational to internal energy in gaseous collisions can be tested in an accurate and fairly simple way, by comparing catalytic efficiencies of transfer by the isotopic molecules H2O and D2O. Results are given for experiments with the vibrators N2O, CO2, SO2, CH3Cl, C2H4and cyclopropane. The ratio of efficiencies conforms with the ratio of the masses of H2O and D2O for C2H4, and within the limits of error for CH3Cl, but not for the other vibrators. With each of these molecules, in different ways, more than one catalytic mechanism appears to be operative for the transfer of internal energy.

Evidence has been collected which shows that knock in an internal combustion engine cylinder is intimately connected with the presence of peroxidic substances of characteristic structure formed in the last part of the charge to burn. Such peroxides might be formed either in the primary oxidation of the hydrocarbon, or from aldehydes which are known to be an intermediate product, The experiments described, in the paper cited, left somewhat undecided the question whether or not peroxides arising from aldehydes formed during combustion are respon­sible for engine knock, although there was evidence against such a con­clusion. It was also left uncertain to what extent aldehydes other than formaldehyde were present in the gases sampled from the engine cylinder, The purpose of this note is to describe some further experiments on these questions, which make it very unlikely that aldehyde peroxides are those that are mainly responsible for knock; this conclusion strengthens the evidence for the primary formation of peroxides in the oxidation of the higher hydrocarbons. First, experiments on various fuels show that aldehydes other than formaldehyde are formed in amounts inadequate to account for the knock observed; and formaldehyde is shown not to cause knock even when its concentration is much greater than ever arises from the com­bustion of fuels which knock. Furthermore, there is no correlation between the amount of higher aldehydes formed by any fuel and its tendency to knock. Secondly, experiments were made to test whether fuels which do not readily knock fail to do so on account of their not forming sufficient primary peroxide of suitable structure, or because of a radically different mechanism of combustion, The results suggest that the former is the reason; for instance, amylene containing appreciable amounts of olefine peroxide (p. 112) did not knock, whereas an addition to it of small quantities of amyl nitrite had a pro-knock effect even more pronounced than its effect on a fuel such as petrol.

The conclusion from previous experiments on the isotope effect in crystals (Robertson and Ubbelohde 1939) was that on substituting D for H in hydroxyl bonds (length 2.75-285 A) there was only a small change in the lattice spacings, whereas for the short hydrogen bonds (length ~2.55 A) there was a considerable expansion. From this conclusion it was suggested that resonance effects must contribute an appreciable proportion of the binding energy in short hydrogen bonds. One difficulty in this suggestion arose from the complicated crystal structures used, for which inferences about changes in bond lengths cannot be very direct. The aim of the present experiments was to substantiate the former results by investigating further examples of the isotope effect. In order to have a structurally simple example of the hydroxyl bond, pentaerythritol C(CH 2 OH) 4 was compared with C(CH 2 OD) 4 . The salt hydrates SrCl 2 6H 2 O and CuSO 4 5H 2 O were also compared with the corresponding deuterium compounds. For the short hydrogen bond a comparison was made between KH 2 PO 4 and KD 2 PO 4 . The (unknown) structures of NaHSO 4 and fumaric acid were also investigated. Finally, in order to obtain information on possible “hydrogen bonds” between atoms other than oxygen, urea was compared with CO(ND 2 ) 2 and KHF 2 with KDF 2 . In addition to confirming the previous conclusions, certain fresh facts have emerged about the behaviour of hydrogen bonds in crystals, as is recorded below.

Previously available studies of isotope effects in catalysts for the transfer of vibrational energy have been extended by comparing catalytic effects of NH3and ND3. For each of the three vibrators, ethylene, methyl chloride, andcyclopropane, vibration-vibration transfer appears to be only of subsidiary importance in the overall catalysis observed. With these vibrators, good agreement is found between observed relative efficiencies of H2O and D2O, and H3N and D3N as energy transfer catalysts. With ethylene, relative efficiencies of H3N and D3N can probably be attributed to mass differences. But even after allowing for mass differences of the isotopic molecules, collisions with some of the other vibrators show residual predominance of the hydrogen over the deuterium derivative as catalyst. This can possibly be attributed to effects of rotation during a collision, on the steric factor. To advance our knowledge about energy transfer further, more accurate information is needed about the origin and magnitude of inter-molecular repulsions.

Detonation limits have been studied for binary mixtures of tetramethylsilane ( TMS ) with oxygen and also for ternary mixtures including up to 92% of helium by volume. The limits for the binary system are 1⋅8% and 48 to 50% TMS by volume. The object of these researches was to study factors controlling transition from detonation to deflagration, particularly when a time lag is introduced in the liberation of chemical energy by the need to condense and crystallize or to polymerize non-volatile reaction products before the large heat changes involved can contribute to the propagation of detonation. Solids formed in the detonating mixtures were examined to determine their composition, density, structure, surface area and electrical conductivity. As the composition of the mixture is progressively changed, a systematic trend is observed in properties of the solid products. This appears to correspond with different modes of propagation of detonation. Starting with the oxygen rich limit where carbon-free silica is formed, the solids change colour rapidly from white to black in a transition region between 28 and 32% by volume of TMS , remaining black for richer mixtures. Black solids incorporate appreciable proportions of carbon. Two peaks in ‘detonation strength’ characterize different regions, in which the trends of fluctuations in the velocity of propagation of detonation correspond with changes in the solid products. Calculations make it clear that the heat of formation of various solids contributes towards the hydrodynamic propagation of the detonation wave. However, the condensation of silica appears to act in a manner different from the formation of carbon. Extremely dilute mixtures of TMS will still detonate. For binary mixtures, the lower limit is found at 1⋅8% TMS by volume. For ternary mixtures, detonation is still found with 0⋅9% TMS , 7⋅8% oxygen and 91⋅2% helium. Some implications of this finding are discussed.

The isotope effect has been measured for (NH 4 )H 2 PO 4 . Expansion of the crystal lattice takes place on substituting deuterium for hydrogen. The expansion is in the direction of the hydrogen bonds, and is of the same magnitude as in the isomorphous KH 2 PO 4 . A contraction of the lattice at right angles to the plane containing the hydrogen bonds is considerably larger for (NH 4 )H 2 PO 4 than for KH 2 PO 4 . From an application of the calculus of errors, the accuracy of the results is estimated to be α 11 = + 0.00269 ± 0.00006, α 33 = - 0.00123 ± 0.00009. Further investigations on the monoclinic form of KD 2 PO 4 has shown the presence of curious diffuse reflexions in moving-film photographs. Attempts to obtain (NH 4 )D 2 PO 4 or KH 2 PO 4 in the monoclinic structure were unsuccessful.

Melting parameters are reviewed for the nitrates of Li, Na, K, Rb, Cs and Ag, and are com­pared with values for the corresponding halides. Previously published information has been supplemented by experiments on the viscosity of molten AgNO 3 , and on the dilatometry and electrical conductivity of AgNO 3 in the solid and melt. It is suggested that the low melting points of group I nitrates, compared with the corre­sponding halides, is due to the formation of association complexes in the nitrate melts. This process could contribute a term to the overall entropy of fusion S f without greatly increasing the heat of fusion H f . Since T f = H f / S f , this would explain the unusually low melting points. The small volume changes of fusion of nitrates, where measured, and the abnormal activation energies for the viscosities of molten nitrates, support the suggested mechanism of melting.

Some comparisons were made between the lamellar crystalline compounds of graphite and intercalary compounds formed of a certain number of other crysthls with shell structure. Although there is a certain number of resemblances, the graphite is almost unique with respect to the amphoteric character of its hexagonal lattices. They can be considered as aromatic macromolecules. In graphite, the interaction between these macromolecules is only weak. They preserve the essence of their aromatic structure in numerous crystalline compounds. Because of the aromatic character there is produced, in a distinct fashion, in the crystulline compounds a charge transfer between these macromolecules and various intercalary groups. This causes profound modifications in the electronic properties of solids. By interposing electron donor groups such as the atoms of an alkali metal or electron acceptor groups such as the halogens or acid radicals, a new class of very conductive solids is obtained. Anisotropic studies of the electrical properties of these solids parallel and perpendicular to the hexagonal layers of the carbon are analyzed. They give important information on the charge transfer to or from aromatic macromolecules. Impurities or crystalline faults of the original graphite modify the formation and the behavior of its crystalline compounds. Some impuritymore » atoms at the edges can facilitate the intercalation of groups in the crystalline compounds and can also affect the electron levels in the macromolecules of carbon, modifying as a result the free bonding energy. Lattice faults can seriously perturb the planeity of aromatic lattices because of the impurity atoms bound to a hold or a prominence of the lattice. Spiral dislocations of the lattice were detected in some graphites. It is unknown to what extent these faults or other faults can affect the conclusions, based on crystalline compounds. (tr-auth)« less

When amalgams are formed by discharge of the ions Li + , Na + , K + , Ca ++ , Mg ++ , NH + 4 at a mercury cathode in a glass vessel, a wedge of electrolyte penetrates between the amalgam and the glass. The effect is not observed for the discharge of Ag + , Zn ++ , Cd ++ or OH + 3 ions. To elucidate the origin of the forces maintaining the wedge in equilibrium, the wedge thickness and the curvature of the mercury/electrolyte interface have been investigated in various ways. The apex of the wedge was studied by interferometry. Measurements have been made during electrolysis of the interfacial tensions mercury/electrolyte and mercury/gaseous nitrogen. Possible theoretical interpretations are discussed. A marked reduction of the surface tension which is observed at certain amalgam/electrolyte interfaces is attributed to strong surface adsorption of the metal especially when the atom is more readily polarized than mercury. The electro-positive character of the amalgams and this reduction of surface tension explain some of the phenomena of the wedge effect.

It is generally accepted that conduction in graphites is predominantly by way of phonon transport; heat transfer due to the migration of charge carriers in a temperature gradient is unlikely to correspond with more than 1% of the total in each principal crystallographic direction. With the crystal compounds for which the charge carrier concentration is increased from about 1 per 104 C atoms to 1 per 24 C atoms this situation may not apply. As is well known from earlier theoretical and experimental studies on diamond, sapphire, and other materials, the scattering of phonons by static imperfections such as grain boundaries, dislocations, clusters and point defects makes the thermal conductance at low values of TIOD a very sensitive measure of crystalline perfection, where OD is the Debye characteristic temperature. Measurement of thermal conductance in relation to systematic changes of disorder yields information not at present obtainable in other ways for well-oriented graphites. Phonon migration in this highly anisotropic material is also of special interest because the assignment of lattice vibrations to different groups of frequencies corresponding with different vibrational modes of the carbonhexagonnetworkrequires the adoption of characteristic temperatures OD which cover a very wide range. Although it is probable that the electronic contribution to the thermal conductance is negligible, phonon-electron interactions [ 83 ] 6-2

Vibrations of sulphur dioxide show two separate relaxation times. Values of r 1 = 6.0 × 10 -8 s for the vibrations of 519 cm -1 and r 2 = 1.2 × 10 -6 s for the vibrations of 1151 and 1361 cm -1 have been derived from new ultrasonic measurements, in fair agreement with earlier work. Molecules studied as possible vibration-translation catalysts included ethane, ethylene, water, and n -hexane. No enhanced efficiency of energy transfer was observed with ethylene. Ethane and water molecules were found to be only moderately efficient catalysts; proportionally, they exert a greater effect for processes associated with r 1 . Deuterium oxide is found to be somewhat more efficient than water. n -Hexane is highly efficient for both the r 1 and r 2 relaxation processes of sulphur dioxide. These observations are discussed in relation to various mechanisms for the catalysis of energy transfer.

Velocity data have been determined as a detailed function of composition for the systems: S H = 2H 2 + O 2 diluted with H 2 , O 2 or CO 2 and S D = 2D 2 + O 2 diluted with D 2 or O 2 . Detonation limits have been established. Each mixture received standard initiation from a detonation in {2H 2 + O 2 }. Velocities were determined 8 ft. downstream from the initiating impact. Within this distance, either a constant velocity of detonation front was reached, or the mixture under test had failed to propagate detonation. Near the limit concentrations, under certain conditions an alternative constant velocity régime established itself. These detonation velocities were still greater than the velocity of sound in the unreacted mixture but about 30% below those of the ‘normal' régime, which alone extends throughout the entire composition range between the limits. Evidence is discussed that this alternative régime is associated with a different reaction mechanism in the detonation front, with hydrogen (deuterium) peroxide as end-product. Subsequent decomposition of this hydrogen peroxide apparently leads to the formation of a second detonation wave. This follows the first with a higher velocity and eventually overtakes it, thereby setting up the normal velocity régime, for which water is the end-product. Possible factors leading to such dual detonation régimes are discussed.

Minor fluctuations of velocity are frequently observed in gaseous mixtures whose compositions lie near to the limiting values of sustained propagation of detonation. Various possible origins of these fluctuations are discussed. In cylindrical tubes, they commonly take the form of a spinning detonation front. In order to study these effects in detail, previously described methods for making accurate observations near the limits have been extended in various ways. Use of multiple probes set radially in a standard detonation assembly permitted determination of the profiles of spinning ionization fronts to the detonation wave. In some cases, the existence of multiple spin heads detected by these means could be confirmed by soot-trace techniques. For any given mixture, fluctuations in the form of spinning ionization fronts are often remarkably reproducible in a particular tube.

Alkali thiocyanates are remarkable for their low melting-points (KSCN 175°C; RbSCN 195°C). Mechanisms of melting of the sodium and potassium salts have been investigated by determining ionic conductivities of solid and melt, viscosities of the melts and, in the case of the potassium salt, entropy and volume changes on melting. It seems likely that the low melting-points of alkali thiocyanates arise from the contribution of association complexes in the melt to the entropy of fusion. These increase the entropy of fusion without requiring large heat or volume changes. In such cases melting-points tend to be low in accordance with the equation, T f = H f / S f . Marked homophase premelting is observed in solid KSCN; the experiments show hardly any discontinuity in volume at the melting-point.

The dispersion of sound in pure ethylene has been reinvestigated at the frequencies 151 and 303 kc/s, at pressures ranging from 20 to 76 cm Hg and at temperatures from —3.4 to 25° C. On the basis that this dispersion is due to hindrances in the transfer translational ⇌vibrational energy, the ethylene molecule has been used as a test system for comparing the efficiencies of pick-up of vibrational energy on collision with other molecules. Particular attention has been paid to paraffin hydrocarbons. From the observed dispersions in mixtures of ethylene with small proportions of other gases, the approximate number of collisions required for transfer of one quantum of vibrational energy from ethylene (l→0) is calculated. The high effectiveness in transmitting vibrational energy by the higher paraffin hydrocarbons investigated is discussed in terms of molecular flexibility and other aspects of the hydrocarbon structure.

In any process of adsorption, dynamic equilibrium is reached when the number of molecules condensing in unit time on the adsorbing surface is equal to the molecules evaporating. Langmuir obtained the well-known relation between the mass adsorbed and the pressure m = k 1 p /( k 2 p + b ) by writing the number of molecules evaporating as proportional to the number already adsorbed, and the number condensing as equal to na 0 kp , where p is the pressure, n the fraction of the adsorbing space unoccupied, k a constant derived from the kinetic theory of gases, and a 0 an accommodation coefficient, which was supposed to differ very little from unity. Using these assumptions, adsorption equilibria should be reached so rapidly as to be practically instantaneous. If we regard the process of adsorption as analogous to a chemical reaction, involving changes in behaviour of the valency electrons of the adsorbed molecules and the adsorbent, the assumption that a 0 differs little from unity, corresponds with the assumption that every collision is fruitful in a gas reaction. H. S. Taylor has pointed out in a recent paper that for some adsorption processes it is apparently necessary to assume quite large energies of activation. Instead of all the molecules colliding with “unoccupied” spaces being adsorbed, only a fraction e -E/RT of the impinging molecules change over into adsorbed molecules.

Measurements are reported of the electrical resistance of graphite/bromine of various compositions. Some of the experiments refer to natural graphite, but the majority deal with specimens of pyrolytic graphite which show good orientation with respect to both a- and c-axes. Changes of resistance in these two directions have been measured as a function of bromine up take to limiting concentrations; cyclic sorption— desorption studies have been included where appropriate. Results are discussed in the light of current theories about the electron band structure in graphite and in relation to the nature of the bonding between bromine and graphite.

Extended families of synthetic p metals can be prepared by controlled intercalation of electron acceptor molecules between the carbon hexagon networks of well oriented graphite. Resistivities and Hall coefficients have been determined down to 1.8 K for graphite bisulphates and (to a more limited extent) for graphite perchlorates. For many of these dilute synthetic metals, peak anomalies are found in Hall effects at around 20 K . These are discussed in relation to possible changes in the structure of the intercalate layers with temperature.

A low-temperature micro-thermostat is described, suitable for single-crystal X-ray work. Thermal expansions of KH 2 PO 4 (and in less detail of KH 2 AsO 4 ) have been measured by X-ray methods from room temperatures down to about 80° K. Both these crystals differ from some previously investigated organic compounds with short hydrogen bonds, in that maximum thermal contraction on cooling occurs along the c axis, which is approximately perpendicular to the planes containing the hydrogen bonds in these two crystals. This axis becomes electrically polarized below a thermal ‘transition point’. Remarkable effects are observed in the neighbourhood of the transition point, similar to those previously described for Rochelle salt. When the tetragonal single crystals are cooled below the respective Curie points (123° K for KH 2 PO 4 and 96° K for KH 2 AsO 4 ) they change into crystal ‘hybrids’, in which subcrystalline units of lower symmetry pack so as to preserve the c axis and either the (100) or (010) planes of the original tetragonal structure, and to maintain electrical neutrality on the average. When a ‘hybrid ’ crystal is warmed above the Curie point, the subcrystalline units merge after a varying lapse of time into the original tetragonal lattice. This change from single crystal to hybrid and back can be repeated at will by cooling and heating. The structure of these crystal hybrids has been investigated by both Bragg and Laue methods. On the basis of the new experimental evidence and of previous work, an explanation is offered for phase transitions ‘of the second kind’, preferably termed ‘continuous’, in terms of crystal structure. It is known that ‘discontinuous’ phase changes at a transition ‘point’ involve a rearrangement of molecules within the single crystal of one structure, to give a powder of crystals with a different structure. But if the new crystals can pack as sub­crystalline units into the original crystal lattice, as is possible when the difference in structure is small and not too much strain is involved, the phase change will occur continuously over a range of temperatures instead of discontinuously at a transition point. The X-ray studies described give new information about the thermodynamics of crystals, and promise to throw light on certain problems of imperfect crystals. Possible biological analogues of hybrid crystals are briefly referred to.

Measurements have been made of the densities of liquid potassium nitrate containing up to 50 equivalents per cent of barium nitrate. Within the temperature range from the liquidus to 440 °C, isotherms of equivalent volume as a function of equivalent composition were found to be straight lines. The excess volume of mixing for any pair of melts in this composition range is thus zero. Partial equivalent volumes of both components are independent of composition at a given temperature, and are given by V̄ m (KNO 3 ) = 53.17 + 0.0229( t - 300), V̄ m (Ba(NO 3 ) 2 ) = 41.75 + 0.0107( t - 300), where t is the temperature in degrees centigrade. This somewhat unusual mixing behaviour is discussed in relation to possible structures for nitrate melts, and in relation to available information about excess volumes of mixing for other ionic melts and liquids generally.

Procedures are described for maintaining good chemical stability in molten alkali-metal carboxylates, up to about 350 °C. Valid physical measurements can be made and the fluids can be used up to about this temperature, above which spontaneous decomposition of the anions is difficult to repress. Molten sodium propionate has a useful liquid range of about 60 °C and sodium isobutyrate of about 90 °C. Sodium n -butyrate transforms into a ‘liquid crystal’ at about 250 °0 and into the isotropic liquid at 324 °C. For sodium isovalerate corresponding transition points are respectively 188 and 280 °C. Thermodynamic measurements are reported of volume and enthalpy changes at transition and melting points. Transport parameters measured include the viscosity and the electrical conductivity. The viscosity of these ionic melts undergoes a steep decrease at the transition from mesomorphic to isotropic liquids. Jumps in ionic conductivity are found both at the melting and clearing points. Even for the isotropic liquids, the ratio of viscosity to electrical conductivity is exceptionally high, compared with the other ionic melts. Mechanisms of melting for these ionic solids are discussed.