SOLUBILITY EXPLAINED

B y A. G. S HARPE , l\1.A., PH .D., F.R.I.C.

Lect~t1·e1· in In01·ganic Ghemist1·y, Gamb1··idge Unive;rsity

A solution is a homogeneous mixture of t wo with which the chemist, especially in t he (or more) components. In t his context, early stages of his training, is most concerned. homogeneity denotes not merely uniformity The general principles on which solubility is under observation by the eye, but incapability discussed apply, however, to all types of of separation into constituents by any solution, and the aim of this account is to mechanical means: in a true solution the provide a brief introduction which , although particles of both constit uents are separate reasonably self-contained, will serve as an molecules or ions, though there is sometimes introduction to a more advanced under­ quite strong interaction between one molecule standing of the whole field of solubility. No or ion and a small number of others in its part of this, it is interesting to note, is more immediate environment. It is common difficult to treat quantitatively t han t he practice to discuss solubility in terms of a solubility of salts in water, the subject with solvent and one (or more) solutes; t he which the student usually makes his fu·st solvent is the component that dissolves the acquaintance with solubility phenomena. sol~de , i .e. retains its own physical state after Most pairs of substances (other than pairs addit ion of the solute. In some instances, of gases) are only partially miscible. When however, it is not clear which component is sucrose (ordinal'y sugar) is stirred with water the solvent and which the solute. Such a at 25°, for example, the concentration of t he case exists for two liquids such as ethanol solution event ually reaches a consta,nt value; and water, which are miscible in all propor­ t he solution is then said to be saturated, and tions; the distinction is of no real importance its concentration is the solubility of sucrose in these circumstances, but it is conventional in water at this temperature. This solubility to refer to the component present in greater may be expressed in several ways. From a amount as the solvent. physicochemical viewpoint , the most funda­ Solutions are often classified on the basis of mental of these is in terms of t he ratio of the t heir physical state (gaseous, liquid, or solid) number of gram-molecules, or moles, of and t he physical state of the pur e solute at the sucrose to t he total number of gram-molecules ordinary temperature. There are then nine (of sucrose and water) in the liquid phase­ types of solution, and these are listed, with the mole j?-action of sucrose. Also in common an example of each , in the table. use are molm·ity (the number of moles per litre of solution), molality (number of moles per 1,000 g of solvent) and weight peT cent (number of grams of solute per 100 g of Solution E xample solution) as units of solubility; the second of Gas in gas Air t hese is especially widely used in dealing with L iq uid in gas .. ·w ater (vP.pou r) in nitrogen solutions of electrolytes, where the term Solid in gas .. Iodine (vapour) in nit.-ogen Gas in liquid .. Carbon dioxide in water (sod a- 'gram-formula-weight' replaces 'gram-mole­ water ) cule.' F or solutions of gases in liquids t he Liquid in liq uid Ethanol in water absorption coefficient (the volume of gas, Solid in liqu id Sodium chloride in wate r Gas in solid . . H ych·ogen in palladium reduced to 0°0 at l atm, dissolved by one L iquid in solid Morcury in silve1.· (amalgam) volume of solvent at the temperature of the Sol id in solid . . Copper in nickel (coinage alloy) experin1ent under a partial pressure of the gas of l atm) is the unit most frequently This article is concerned mainly with employed. solutions of gases, liquids and different t ypes The equilibrium which exists in a saturated of solid in liquids, since these are the solutions solution is, lilce all other equilibria, a dynamic 75 76 EDUCATION IN CHEMISTRY

one, as may be shown by isotopic tracer to the division of the total heat content (H) at experiments: the rate of dissolution of the absolute temperature T of a system into two solute is equal to its rate of separation from parts: that part which can be converted into the solution. The position of equilibrium is other forms of energy without change of affected by heat and pressure in accordance temperature is ]mown as j?-ee ene1·gy (G), with Le Chatelier's principle, and the solu­ whilst the part which is not convertible bility of a non-electrolyte is, in fact, an (H - G) is defined as the product of the equilibrium constant. F or substances which absolute temperature and a quantity called yield ions in solution, such as salts of formula the ent1·opy (S) of the system. Thus l\fX, the equilibrium is of the form H = 0 + TS MX (solid) ~ :M+ (in soln) + X- (in soln) or, for a change taking place at constant and the equilibrium constant is the product temperature T , of the activities {approximately, of the con­ D.H = D.G + T 6S centrations) of M+ and X- in solution, t he activities of solids being taken arbitrarily as The quantity S may also be looked at in unity; this equilibrium constant is known as another way, as a measure of the degree of the solubility prod1tct of l\fX at the tem­ randomness of the system. This aspect may

perature chosen. For a salt of formula l\1X2, be illustrated by considering what happens if 2 the solubility product is [M+] [X-] , and so we m.i...-.,;: equal quantities of hot and cold on. Since [M+] and [X- ] are proportional to water : the resulting liquid all has the same the amount of dissolved salt, we see that the temperature, and instead of there being relation between solubility and solubility molecules describable as belonging to one product depends on the formula of the salt of two different categories, all are now mixed concerned. up-the randomness of the system has One feature of discussions of solubility increased. At the same time, although no which calls for particular comment is the use heat has been gained or lost, the possibility of the term 'insoluble.' )fost chemists tend of using the system to obtain electrical energy to describe as 'insoluble' any substance by putting thermocouples in the water at which dissolves in a liquid to the extent of different temperatw-es has disappeared. Gain less than about l g in 1,000 g of solvent. in entropy is therefore accompanied by a Silver chloride and cupric sulphide, for corresponding loss in free energy if there is example, are by this criterion insoluble in no change in the total heat content of the water. Their solubilities, l 0- 5 and 10- 22 system. Just as every substance has a g-formula-weight per litt·e, and solubility standard heat offormation, it has also a stan­ products, 10- 1 0 and I0- 44 (g-ionflitre)2, at dard free energy of formation and a 25°, however, are very different, and (as will standard entropy, and from tables of such be seen later) from the point of view of the quantities (which may be determined inde­ energy change which would be involved in pendently by a variety of methods) it is getting a g-formula-weight of solute into possible to compute changes in them for aqueous solution there is very much more different reactions. difference between silver chloride and copper The very existence of endothermic (heat­ sulphide than between silver chloride and absorbing) processes which take place spon­ sodium chloride, the solubility of which is taneously (e.g. the dissolution of most 0·62 g-formula-weight per litre at 25°. substances in water) is a proof that it is not ·we must now see why it is important to the change in heat content (t:JJ) which know something about such energy changes. determines how far a reaction goes, i .e. it s. It is well known that there are several forms equilibrium constant. This quantity is the of energy, such a,'3 mechanical, electrical, free-energy change, which is therefore a chemical and thermal energy. All forms of quantity of the greatest importance through­ energy can be converted into heat, but the out chemistry. It is not essential in obtainij1g converse is not always t rue, and this has led some understanding of the significance of the SOLUBILITY EXPLAINED 77

concept of free energy in chemistry to master tion or liberation of heat (if 6.H = 0, MJ. = the numerical form of all the relationships -T6.S). Usually, however, it will be involved, but one such relationship is of out­ necessary to consider heat terms as well as standing importance: if a reaction is carried entropy terms, and to what extent a solute A out on a very small scale (say, one-millionth will dissolve in a solvent B will then be deter­ gram-formula quantities) in the presence of mined la.rgely by the difference between the all the reactants and products at unit energy of interaction of A with B and the activity, and the free-energy change is energies of interaction of A with A, and B with multiplied (in t his case by a million) to give B. Looking at the problem in this way, we the so-called standard free-energy change, see that whether A dissolves in B depends 0 6.G , the latter is related to the equilibrium just as much on the properties of the solvent constant for t he reaction by an equation of as on the properties of the solute : hydro­ the form carbons, for example, dissolve only very - t:.G 0 = R TlnK slightly in water, not because of strong attrac­ where]( is the equilibrium constant expressed tion between hydrocarbon molecules in pure in suitable units, R is the gas constant, T the hydrocarbons but because of the hydrogen absolute temperature, and ln denotes loga­ bonding between molecules in liquid water; the G-H bond is much less polar than the rithm to base e, or 2·3 log10. Substitution of the values of R and T leads (for T = 25°C) 0- H bond, and hence bonds of the latter to the equation type interact with one another preferentially. We shall now discuss some different types -t:.G0 (in kcal) = 1·36 log K 10 of solution to illustrate the relative impor­ The special feature of this relationship which tance of heat and entropy terms in deter­ must be noted is that it is an exponential mining the position of equilibrium. one. Since it is the logarithm of the equili­ 0 brium constant that is related to -6.G , a SOLUTIONS IN GA.SES small change in the latter quantity makes a Since the molecules of an ideal gas exert no very large difference to how far a process force of attraction on one another, a gas as a goes; a change of 4 kcal in 6.G, in fact, alters solvent offers no resistance to dilution with ]( by a factor of a thousand. In the case of another substance; there is always a positive an ionic solid dissolving in water, the energies entropy of mixing, and the position of of interaction of the ions with one another in equilibrium is determined largely by the the solid and with water molecules in solution amount of energy needed to separate the are both very large, and it is the small units present in the structure of the solute. difference between these two large quantities If this is another gas, a negligible amow1t of (neither of which may be accurately known) energy is needed, and hence all gases are which determines the magnitude of the completely miscible. If it is a liquid or a solubility product (the equilibrium constant) solid, the latent heat of Yaporization or and hence of the solubility; this is why the sublimation must be absorbed in order to calculation of solubilities of salts from thermal transform it into the vapour phase; this then data presents such a formidable task. mixes freely with the gas, which is considered This competitive picture of the dissolution to be the solvent. The influence of tem­ of salts may now be generalized to include all perature is thus confined to its effect on the processes of solution. There will always be an vapour pressure of the solute, and solubilities increase in the entropy of the system when of liquids and solids in gases always increase two constituents mi..~ to form a solution, with rise in temperatm·e. because of the increase in randomness, and Occasionally there is evidence for specific in cases where there is no heat of solution this interaction between a gaseous solvent and a increase in entropy provides the driving force solid solute; some metal oxides (e.g. BeO) are for t he dissolution: such a case is illustrated significantly soluble in compressed steam at by many pairs of organic liquids, which are high temperatures owing to hydroxide freely soluble in one another without ahsorp- formation. 78 EDUCATION IN CHEMISTRY

SOLUTIONS OF GASES lK LIQUIDS ideal solubility of any gas if the vapour Solubilities of gases in liquids are usually pressure above its liquid is known at that cited as absorption coefficients, as has been temperature (i.e. if it is below its critical mentioned earlier; since, however, by Avo­ temperature). At 25°0, for example, the gadro's law the number of moles of a gas is vapour pressure of pure liquid ethane is 42 proportional to its volume at standard tem­ atm; if JJ2 = l atm, x2 = 1(42 = 0·024 mole­ perature and pressure, the absorption co­ fraction. For gases above their critical efficient is evidently proportional to the con­ points the hypothetical vapour pressure of the centration in moles per litre. One of the liquid can be estimated (from the Clausius­ difficulties which arises in the interpretation Clapeyron equation) on the assumption that of absorption coefficients (and, indeed, of the heat of vaporization remains constant; many other solubility data) is that the in this way it can be shown that the ideal analytical methods employed in their deter­ solubility of nitrogen should be about I0-3 mination do not reveal t he chemical form in mole-fraction. which the solute has dissolved. It may These calculated values, it should be noted, reasonably be supposed that nitrogen and apply to any solvent so long as the solutions oxygen dissolve unchanged, for they are show ideal behaviour; the actual values for sparingly soluble in all solvents; but hydrogen ethane and nitrogen in solvents such as chloride in water is largely converted into hydrocarbons and carbon tetrachloride are about 0·018 and 6 X I0- 4 mole-fractions, hydrated H 30 + and Cl- ions ; the fact that 2M ammonia solution has quite a strong respectively, showing that there is some odour, whereas 2M hydrochloric acid is departure from ideality. It is, however, odourless, indicates that for ammonia the very striking that so simple a treatment leads degree of interaction with the solvent is to a result which is in error by less than a smaller. All very soluble gases do, in fact, factor of two. When water is the solvent this undergo chemical change to an appreciable is not so, for the intermolecular attraction in extent in the solvents in which they are very liquid water complicates the picture b~­ soluble. malung all gases which dissolve in molecular The solubility of a gas at a given tem­ form much less soluble than they are calcu­ perature is proportional to the pressure lated to be; the gain in entropy of mixing is offset by the work which has to be done (H enry's Law). If n1 and n 2 are the numbers of moles of solvent and gas in the satura.ted against this intermolecular attraction in solution at pressure p of the gas, forcing molecules of solvent apart. Water is 2 the outstanding example of a polar solvent n2fnl = kp2 with strong intermolecular attraction, but

If the gas is not very soluble, n 2 is much nitrobenzene, aniline and ethanol also show smaller than nv and n1 is then nearly equal this behaviour to some extent; another manifestation of this in the case of ethanol is to ~ + n 2, so that n 2f (n1 + n 2) = x2 = kp2, where x2 is the mole-fr-action of solute in the the fact that its boiling point is 100° higher saturated solution. Now if we think of the than that of dimethyl ether, which has the solvent as non-volatile, p 2 becomes the same molecular weight. vapour pressure of t he volatile- solute The inverse dependence of x 2 on p 2 ° at a present in mole-fraction x2 in the solution. particular temperature also explains two If H enry's Law applies over the whole con­ other features of the solubility of gases. centration range fi-om pure solvent to pure Since the pressure above the liquid is less for liquefied solute, then when x 2 is unity p 2 gases with relatively high boiling points, gases becomes p 2°, the vapour pressure of the pure such as butane and carbon dioxide are much solute, so that k = 1 fp2 °. Hence more soluble in all solvents than are hydro­ gen, o:s.."ygen, nitrogen and carbon monoxide. X2 = P2/P2 ° or P2 = X2P2 ° Furthermore, since for any gas p 2o increases This is, of course, Raoult's law for a liquid with increase in temperature, the solubilities mixture. With its aid we can calculate the of all gases decrease with rise in temperature. SOLUBILI'fY EXPLAINED 79

The latter conclusion may also be reached by aniline or ether, for example, is added to Le Chatelier's principle: dissolution of a gas water, it dissolves, but further additions soon in a liquid may be considered as involving result in the formation of two layers, one a first the condensation of the gas, followed by saturated solution of the organic compound mixing of the two liquids ; since the con­ in water, and the other a saturated solution densation stage is accompanied by liberation of water in the organic compound. Similar of heat, lowering the temperature will cause behaviour also results with aniline and n­ more gas t{) dissolve. hexane, or methanol and carbon disulphide. Greater difference in nature (e .g. water and SOLUTIONS OF LIQUIDS IN LTQUTDS carbon disulphide, ethanol and mercury) Mixtures of two volatile liquids which are results in immiscibility. The miscibility of miscible in all proportions are in many ways liquids is often discussed in terms of their like solutions of gases in liquids. If in the internal p1·essu1·es, which may be regarded as

preceding solution p 2 ° is less than l atm at measures of the dependence of internal energy 25° and the phase designated the solvent has on volume, i .e. of the strength of the cohesive an appreciable vapour pressure, we have a forces. These may be estimated roughly by typical binary liquid system. Each liquid in several methods, typical values at ordinary such a mixture exerts a vapour pressure equal temperature being : hexane, 2,000; ca,rbon t o its mole-fraction times its vapour pressure tetrachloride, 3,500 ; phenol, 5,000; water, in a pure state at the same temperature 16,000 atm. Increasing difference in internal (Raoult's Law) ; solutions for which this law pressures leads to decreasing solubility; holds are called ideal solutions, and typical increasing temperature overcomes the co­ examples are usually mixtures of very similar hesive forces, and many pairs of partially substances such as ethyl bromide and ethyl miscible liquids become complet ely miscible iodide, or benzene and toluene. It would be at higher temperatures. rather surprising if there were many pairs of The observant reader will have noticed liquids which obeyed R aoult's law exactly, that no question of absorption or liberation since t he existence of a substance in the of a heat of fusion or vaporization arises in liquid phase at all is a proof of the presence the dissolution of one liquid in another; the of quite strong intermolecular attractions overall heat of solution is t herefore nearly which can be overcome only by supplying always small and arises from different the heat of vaporization. When a n ideal degrees of non-ideality in the liquids con­ solution is formed from two liquids there is cerned. Clearly, entropy considerations play no change in heat content; the driving force an important part; but it cannot be said towards mixing is provided by the increase that at t he present t ime there is any simple in entropy . theory which deals satisfactorily with all of There are, of course, many pairs of liquids the great variety of behaviour found in which are miscible in all proportions to yield liquid- liquid systems. mixtures which show deviations from Raoult's law, e.g. diethyl ether and acetone, heptane and ethanol, ethanol .and water, SOLU'l'IONS OF SOLIDS IN LIQUIDS perchloric acid and water. These deviations Three types of solid will be considered in arise from the differences in properties of the the following discussion: those having mole­ constituents of the pairs, and in some cases cular, macromolecular and essentially ionic there are. appreciable heats of mixing. The structUI·es. nature of these deviations from Raoult's law In molecular solids t he shortest distances is, however, of minor interest in connection between atoms ·in different molecules are with solubility, and will not be discussed about twice t he interatomic distances within further here. m olecules; the intermolecular bonding is When two liquids are less similar than the weak, as is shown by the low heat s of sublima­ members of the pairs cited above, partial tion. Among substances which crystallize miscibility usually occurs; if a little phenol, with molecular structmes at the ordinary 80 EDUCATION IN CHEMl Sl'RY temperature are iodine, mercuric chloride and solubility of most organic compounds in it; nearly all organic compounds. the organic compounds which are soluble in Macromolecular structures, such as those water all contain groups such as hydroxyl, of diamond, silica and red phosphorus, con­ carboxyl or sulphonyl, which can interact tain no discrete molecules; strong covalent with water molecules as strongly as the latter bonds extend through the structure in two or interact with one another. Typical com­ more directions, and it is only with great pounds in these classes are low-molecular­ difficulty that the lattice can be broken up. weight alcohols and acids, polyhydric alcohols Ionic solids contain symmetrical arrange­ such as sugars, and soaps. ments of charged ions held together by For a macromolecular solid which is electrostatic bonding, the energy of the in1agined to form an ideal molecular solution, lattice relative to isolated ions at infinity similar considerations should apply; in this being proportional to 1/(1"+ + L), where case, however, the latent heat of fusion is (1·+ + r_) is the sum of the ionic radii, i .e. extremely high and hence the solubility in all the interionic distance. Although salts have solvents is extremely low. very high heats of fusion and vaporization, For ionic solids t he position of equilibrium the ions are strongly attracted by polar is determined mainly by the balance of lattice molecules (especially water) and solvation energies and solvation energies. In a more energies of ions are sometimes greater than refined treatment the latter would be split lattice energies of salts containing them. into heat and T x (entropy of solvation) If a molecular solid is melted (thereby terms, but for present purposes a discussion absorbing the latent heat of fusion), t he based on lattice energy and hydration heats resulting liquid may form an ideal solution will suffice; as has been stated earlier , both with a solvent; when this is so, the heat of are very large quantities, but it is the small fusion is equal to the heat of solution, and it difference between them which determines can be shown by a rather lengthy t hermo­ solubilities. In the case of sodium chloride, dynamic argument (which will not be repro­ for example, it requires absorption of 184 duced here) that the ideal solubility under kcalfg-formula-weight to separate the solid constant pressm·e is given by into ions at infinity; since the heat of solution of sodium chloride in water is nearly zero ln.x= _ LR,(~ 1' -.!..T ) (hence the independence of solubility on tem­ 0 perature) the sum of the hydration heats of where '1' is the temperature at which the Na+ and CI- must also be 184 kcal. Most mole-fraction of the solute in the solution is other salts absorb a few kcal of heat on dis­ x, L 1 is the molar heat of fusion of the solid, solution in water , and hence are more soluble and T 0 is its melting point. This equation in hot than in cold water ; the increase in may, for example, be used to calculate the solubility is, however, usually relatively ideal solubility of naphthalene in any solvent small. In principle the hydration energy of as 0·31 mole-fraction at 25°C. The solubili­ an ion of radius 1" and charge z in a solvent of ties in benzene, toluene, chlorobenzene and dielectric constant E is given by the expression chloroform (all of which are non-polar liquids having internal pressures similar to z2e2 (l _.!:) 21· E t hat of liquid naphthalene) are all close to this value; t hose in aniline and acetone agree but the radius of an isolated ion in solution is less well; in water , naphthalene is very not necessarily the same as that of the same sparingly soluble, since the strong attraction ion in a crystal, and E need not be the same of water molecules for one another (made near an ion as it is in the pure solvent; in manifest in the high boiling point, entropy of practice, therefore, empirical methods play a vaporization, and internal pressure of water) large part in obtaining values for solvation serves to exclude molecules of the non-polar energies. Theimportanceofthe (1- 1/E) term, solute. It is, in fact, the abnormal character however, is very great, for it means that only of water which is responsible for the low in sol vents of high dielectric constant will the SOLUBILITY EXPLAINED 81 sum of t wo solvation energies approach the difference in solubility in water between silver lattice energy of t he solid. H ence salts, if .fluoride and silver chloride. they dissolve at all, will normally do so only The foregoing account of t he solubility of in media of high E, such as water, ammonia, salts has been based on an electrostatic model nitromethane and dimethylformamide. The for the solvated ion in solution, the charge on few salts t hat dissolve in organic solvents of the ion interacting with the uneven charge low dielectric constant (e .g. ether) nearly distribution in the water or other solvent always contain very large ions and have molecule. There is sometimes, however, ver_,­ very low lattice energies; tetrabutylam­ strong evidence that the solvation process is monium pict·ate is an example of such a not a simple electrostatic one. Thus when a. compound. paramagnetic Co3+ ion is hydra.ted, it becomes The fact that the lattice energy of an ionic diamagnetic, showing that an electronic re­ solid depends inversely on r + + t·_ taken arrangement has taken place. 'When 3 together, whilst the sum of t he solvation Cr- (H20)6 " ions from chrome alum dissolve in energies is proportional to 1jr·+ + 1jr-_ taken 180 -labclled water, it is found that no separately, has other instructive conse­ exchange with ·water from t he alum takes quences. For a set of compounds having the place during several hours; evidently the same structure and r-_ ~ r·+ the lattice water attached to the cation is not feee to energies change little with increasing t"+ ; exchange, a fact which again suggests covalent 1/L is constant and very small, whilst 1jr·+ bonding. Finally, the highly specific inter­ decreases rapidly with increase in 1·+· H ence action between salts of a few metals and solubilit ies decrease as r+ increases; this is organic liquids, e.g. the solubility of silver what happens among, for example, the alkali­ .fluoroborate (which has the same structure metal chloroplatinates or cobaltinitrites or the as potassium ft.uoroborate) in benzene and alkaline-earth metal sulphates in water. toluene, indicates covalent interaction of a Insolubility in water thus need not denote rather special kind, in this case between the covalent character ; there is, for example, not aromatic compound and t he silver ion. In all the slightest evidence to suggest that the these cases, however, it should be remembered bonding between the metal and the anion that the position of equilibrium is determined in calcium fluoride, barium sulphate or by the difference between the lattice energy potassium chloroplatinate is not ionic. of the solid and the magnil~tde (not the. · Silver ion has almost the same radius as nat~o·e) of the interaction between ions and sodium ion, and silver fluoride is fr·coly solvent: it is the magnitude of the free­ soluble in water. Silver chloride, however, is energy change which determines how far the very sparingl.v soluble, and it is instructive process of dissolution goes. to compare the pau·s ~aF and NaCl, and AgF and AgCI. The observed lattice energies SOLGTIOSS IN SOLIDS of these foue compounds (all of which have Tho characteristic feature of a solid is that the KaCl structure) arc 216, 184, 228 it contains atoms, molecules or ions occupy­ and 216 kcaljg-formula-wcight respectively. ing particular positions in a rigid structUl'e. Despite the difference in size between F - and When the solute can be accommodated in the CI-, we notice that the difference in lattice solvent lattice without appreciable change, energy is only 12 kcal for the silver salts; for tho formation of a solid solut ion takes place tho sodium salts it has the much larger value readily: examples at·e provided by 1J-dichloro­ of 32 kcal. We t herefore infer that there is bcnzcno and p-dibromobenzene, potassium about 20 kcal of non-electrostatic bonding chloroplatinate and potassium bromoplati­ stabilizing the solid in the case of silver nate, magnesium oxide, MgO, and lithium 2 3 chloride; this t erm has no counterpart in the ferric oxide, LiFe02 (Mg +, Li+ and Fo + are hydration energies of Ag+ and Cl- in water, the same size, and the average charge on Li+ which are the same as in solutions of silver and Fe3+ is the same as on Mg2+), and sih·cr fluoride and sodium chloride, and it is t here­ and gold. In such cases, there is a negligible fore the main factor underlying t he striking rhangc in heat content when a solid solution 2 }J))UCATION I N CHE 31IS'l'RY is formed. Gases and liquids, on the other expanding t ho metal lat tice, but also for the hand, are seldom very soluble in solids: the energy needed to break t he strong bonds fact that they are gases and liquids implies a (103 kcaljmole) in hydrogen, which is present simple molecular structure in their solid in the solid phase as single atomic units. phases, and solubility is t hen to be expected only in elements or compounds crystallizing CONCLUSION with similar simple molecular structures. The posit ions of all equilibria are deter­ For a gas or a liquid to dissolve in a solid mined by thermodynamic factors. When requires work to be done in forcing apart the solute and solvent are in the same physical units in the solid structure and, since the state, the increase in entropy on mixing interatomic, intermolecular or interionic ensmes solubility so long as (i) in the case of forces in solids are relatively strong, this liquids, they do not differ too widely in their energy factor militates against solubility. properties and internal pressures; (ii) in the Gases usually interact with solids only to tho case of solids, they have the same structure extent of forming a monolayer held by and, if ionic, contain ions of similar sizes. surface forces and a weakly held layer, only For other cases, heats of vaporization or a few molecules t hick, on top of this: Where fusion are involved as well as entr·opies of the solubility is appreciable, as in the case of mixing. In all cases, however, t ho key to hydrogen in palladium and ot her metals, the interpretation of solubility lies in the strong chemical bonding must be involved to evaluation of the free-energy change which compensate not only for the work done in accompanies the process of dissolution.

CLE AJ.~I NG OF APPARATUS: THE ALCOHOL -NITRIC ACID REACTION" At one time it was customary to cleanse this alcohol is added, reaction starts at once. dirty apparatus-especially containing tarry This modifi cation of the usual procedure matter-by means of nitric acid and alcohol. has now been used for cleaning apparatus IL seems that in the early days t his procedure over a number of years, and so far no indica­ was not considered particularly dangerous, tion that it might prove dangerous has been but in recent years its hazardous nature has noted. The 'bumping' that is so character ­ been increasingly emphasized and serious istic of the pure nitric acid reaction does not accidents have in fact occmred. occm. The reaction starts smoothly at tho The reaction is well-known to be auto­ smface of tho liquid, and its onset is so catalytic. When pure nitric acid and alcohol rapid that there would seem to be little are mixed, no reaction occm s at first, risk of adding enough alcohol, in one lot, but reaction starts almost explosively after to generate dangerous pressure in a flask. a varying lapse of time and is accompanied Taking elementary precautions (fume-cup­ by violent bumping and t he formation of board, avoid narrow-necked vessels and so on) nitrogen oxides. It seems likely that the it is difficult to see in what way it might manufacturers of nitric acid have progress­ prove dangerous. All that is necessary is ively supplied a pmer and purer product that the nitric acid, before use, should have and that, in the early days, a treacherous a rich reddish-brown colom . If this single latent period was not a marked feature of the point is attended to, even the mistaken use of reaction. sodium nitrate for nitrite should at once The difficulty can be avoided by t he give warning that alcohol must not be deliberate addition of nitrous acid. If added. If desired t he nitric acid mixture 5-10 per cent of sod ium nitrite is stirred may be kept bottled ready for use. into ordinary nitric acid (density 1·42), a red-colomed acid, containing dinitrogcn R. E. D. CLARK tetroxide, is immediately obtained. If to Cambridge College of Arts & Technology.