Download Solubility of Gases, Liquids and Solids in Liquids: An Introduction and more Study notes Chemistry in PDF only on Docsity!
SOLUBILITY E XPLAINED
B y A. G. S H ARPE , l\1. A., P H. D. , F.R.I .C.
L ec t~t1·e1· in I n01·ganic Gh emist1·y, Gamb1··idge Univ e;rsity
A solution is a homogeneous m ixture of t wo (or mor e) componen ts. In t his cont ext, homogeneity denotes n ot merely uniformi ty under observ ation by the eye, but in ca pability of separ ation into con st itu ents by any m echanical m ean s: in a true so lution th e particles of both con stit ue nts are separ ate molecules or ions, though ther e is so m etimes q uite st rong in te r acti on between one molecul e or i on and a sma ll n umb er of others in its i mm ediate environm en t. It is common p racti ce to discuss solub ility in terms of a solvent and one (or more) so lu tes; t he solvent is the compone nt th at dissolves the s ol~de , i .e. reta ins its own ph ysical sta te af ter addit ion of the so lu te. In some instances, however, it is not cl ear which component is the solvent and which the solute. Su ch a case exi sts for two liquids such as ethanol and water, which are miscible in a ll propor - t ions; th e distinction is of no r eal impor ta n ce in these circ um stances, but it is conventional to r efer t o the comp on ent pr esent in greater amoun t as the solven t. Solut ions are often classified on the ba sis of t heir physi cal state (gase ous, liquid, or sol id ) a nd t he physical st at e of the pur e so lut e at the ord inary te mp er atur e. Th ere are then nine types of solu tion, and these are listed, w ith an example of each , in the table.
Solution Gas in g as L iq ui d in gas .. So l id in gas .. Ga s in li qu id ..
Liqu id in liq u id So lid in liqu id Gas in sol id.. L iq ui d in solid Sol id in solid..
E xample Air ·w ater (vP. po u r) in nit r ogen Iod ine (vap our) in ni t.-ogen Car bon d i oxi d e in water (sod a - wate r ) Et h ano l in wate r Sod i um c hlor i de in wate r H ych·ogen in pa ll ad ium Morcury in sil ve 1.· (amal gam) Copper in ni ckel (co i nage a ll oy)
Thi s a rticle is concern ed mainly with so lut ion s of gases, li quids an d different t ypes of solid in liquids, s in ce these ar e the solutions
75
with which the chemi st, especia lly in t he early st ages of his training, is most concerned. Th e general pr inciples on wh ich solu b ility is discussed apply, however, to all types of solut ion, and the aim of this acco unt is to prov ide a brief intr oduc tion which , although r ea so nably self-con tained, will serve as an in tro du ct io n to a more advanced und er- standin g of the whole fi eld of solubility. No par t of this, it is i nteresting to note, is more di ffic ult to treat qua nti tatively t han t he sol ub ili ty of salts in water, the subj ect with which the st ud ent usua lly makes h is fu·s t acqua in ta nce with solubili ty ph enom ena. Most pa ir s of s ub sta nces (o th er th an pairs of gases) a re only partially miscibl e. When sucrose (ordina l'y su gar ) is stirred wi th wa ter at 25°, for examp le, the concentrat io n of t he solution event ua ll y reaches a consta,nt valu e; t he solut ion is th en said to be saturated, and i ts con cen tration is the solubility of sucro se in water at this te m per ature. Th is so lu b ility may be ex pr essed in sever al ways. From a physicochemi cal vie wpoint , the m ost f un da - m ental of these is in ter ms of t he ratio of the n umber of gra m-m olecul es, or moles, of sucrose to t he t otal nu mber of gram-molecules (of sucrose and water) in t he liqu id pha se- the mole j?-action of sucrose. Al so in co mmon use ar e molm·ity (the numb er of moles per litr e of solut ion), molality (numb er of moles per 1,000 g of sol ve nt ) and weight peT cent (number of grams of solu te per 100 g of so lution) as un its of solu bi lity; the seco nd of t hese is especia lly widely u sed in dealing with solutions of electrolytes, wher e the term 'gram -fo rmula-weigh t' r eplaces 'gr a m-m ole- cul e.' F or so lutions of gases in li quids t he absor p tion coefficient (the volume of gas, re du ce d to 0°0 at l at m, dissolved by one vol um e of solvent at the t e mp era tur e of th e experin1ent und er a par tial pr essur e of the gas of l atm ) is the unit m ost fr equently employed. Th e equili brium which exi sts in a saturated solu tion is, lilce a ll other equilibria, a d ynami c
76 EDUCATION IN CHEMISTRY
one, as may be shown by isotopic tracer experiments: the rate of dissolution of the so lute is equal to its rate of separ ation from the sol ution. The position of equilibrium is affected by heat and pre ss ure in accordance wi th Le Chatelier's principle, and the solu- bility of a non-el ectrolyte is, in fa ct, an equilibrium constant. F or substances which yield ions in solution, such as sal ts of formula l\fX, the equilibrium is of the form MX (solid) ~ :M+ ( in so ln ) + X- (in soln)
and th e equilibrium constant is the produ ct of the activities {approximatel y, of the con- centrations) of M+ and X- in solution , t he activities of solids being taken arbitrarily as unity; this equilibrium co nstant is known as t he so lubility prod1tct of l\fX at the tem- perature chosen. For a salt of formula l\1X 2 , the solubility produ ct is [M +] [X-] 2 , an d so on. Since [M +] and [X - ] are propor tional to the amount of dissol ve d salt , we see that the relation be tween solubility and solubili ty product depends on the formul a of the salt concerned. One f eature of discussions of solubility which calls for par ticu lar co mme nt is the use of the te rm ' inso luble.' )fost chemists tend to describe as 'insolubl e' any sub sta nce which dissolves in a liquid to the extent of less t h an about l g in 1,000 g of sol ve nt. Silver chloride and cupric sulphide, for exampl e, are by this criterion insoluble in water. Their so lubilities, l 0- 5 and 10- 22 g-form ula-weight per lit t·e, a nd solubili ty products, 10- 1 0^ and I 0- 44 (g -ionflit r e) 2 , at 25°, however, are very different , and (as will be seen lat er ) from t he point of view of the ener gy change which would be involved in getting a g-formul a-weight of so lut e into aqueous sol ution there is very much more difference between silver chloride and copper sulphide than between silver chlori de and sodium chlo ride, the solubili ty of which is 0·62 g-formula-weig ht p er litre at 25°. ·we mu st now see why it is importa nt to know s omething about such ener gy ch an ges. It is well known that t here are several forms of energy, such a,'3 me chani cal, electrical, chemical and the rmal ener gy. A ll forms of energy can be converted into heat, but t he converse is not always t ru e, and th is ha s led
to the divisi on of the total heat content (H) at
abso lute temp e rature T of a system into two parts: that part which can be conv erted into other forms of energy with out change of t e mperatur e is ]mown as j?-ee ene1·gy (G), wh il st the part which is not conv ertible (H - G) is defin ed as the product of the absolute tempe rature a nd a quant ity called the ent1·op y (S) of the system. Thus
H = 0 + TS
or, for a chan ge taking pla ce at co ns tant te mpera ture T , D.H = D.G + T 6S
Th e quantity S m ay a lso be looked at in another way, as a measure of the d eg ree of randomness of the system. This aspect m ay be illu strated by considering wh at happens if we m.i...-.,;: eq ual quantities of hot a nd co ld water : the resulting liquid a ll has the same t e mperature, a nd in stead of there being molecules describable as belonging to one of two different categori es, all are now mix ed up-the randomness of the system h as increased. At the same time, a lthough no h eat has been gained or lo st, the possibility of using the system to obtain electrical energy by putting thermocouples in the water at different temperatw-es has di sapp ear ed. Gain in entropy is therefore accompanied by a corr esponding loss in free e nergy if there is no change in the total h eat cont ent of the system. Ju st as every substance has a standard h eat offor mation, it h as also a stan - dard free energy of formation and a s tandard en tropy, and from tables of such quantities (which may be de termin ed inde- pendently by a vari ety of methods) it is possible to com pute changes in them for different reactions. The very existence of end othermi c (h eat- absorbing) processes which take place spon - taneously (e.g. the dissolution of mo st sub stan ces in water) is a proof that it is not t he change in h eat content (t:JJ) which determ ines how far a reaction goes, i .e. it s. equilibri um constant. This quantity is the free-energy chan ge, which is therefore a quantity of the greatest importan ce through- out chemistry. I t is not essential in obta inij1g some under standing of the significanc e of th e
78 EDUCATION^ IN^ CHEMISTRY
SOLUTIONS OF GASES lK LIQUIDS Solubilities of gases in liquids are u sually cited as absorption coefficients, as has been mentioned earlier; sinc e, however, by Avo- gadro's law the number of moles of a gas is proportional to its volume at standard tem - perature and pre ssure, the absorption co- efficient is evidently propor tional to the con- cent ra tion in moles per litre. One of the difficulties which arises in the interpretation of absorption coefficients (and, indeed, of many other solub ility data) is that the anal ytical methods employed in thei r deter- minati on do not reveal t he chemical form in which the solute has dissolved. It may reasonably be supposed that nitrogen a nd oxygen dissolve unchanged, for they are sparingly soluble in all solvents; but h ydrogen chloride in water is largely converted into hydrated H 3 0 + and Cl- ions ; the fa ct that 2M ammonia soluti on has quite a str ong odour, whereas 2M hydrochloric acid is odourless, indicates that for ammonia the degree of interaction with the solvent is small er. All very soluble gases do, in fact, undergo chemical change to an appreciable extent in the so lvents in which they are very so luble. The solubil ity of a gas at a given tem- perature is proportional to the pressure (H enry's Law). If n 1 and n 2 are the numb ers of moles of so lvent and gas in the sat ura.ted soluti on at pressure p 2 of the gas,
n2 f nl = kp If the gas is not very so luble, n 2 is much small er than nv and n 1 is then nearly equal to ~ + n 2, so that n 2f (n 1 + n 2 ) = x2 = kp2, where x 2 is the mole-fr-action of sol ute in the saturat ed solution. Now if we think of the sol vent as non-volatile, p 2 becomes the vapo ur pressure of t he volatile- solut e present in mole-fraction x 2 in the solution. If H enry's Law applies over the whole con- ce ntration range fi-om pure solvent to pure liquefied so lut e, then when x 2 is unity p (^2)
becomes p 2 °, the vapour pressure of the pure
solute, so th at k = 1 fp 2 °. Hence
X2 = P2 /P2 ° or P2 = X2P2 ° This is, of course, Raoult's law for a liq u id mixture. With its aid we can calculate the
ideal solubility of any gas if the vapour pressure above its liquid is known at that temperature (i.e. if it is be low it s criti cal temperature). At 25 °0, for examp le, the vapour pressure of pure l iquid ethane is 42 atm; if JJ 2 = l atm, x 2 = 1(42 = 0·024 mole- fraction. For gases above their critical points the h ypothetical vapour pressure of the liquid can be estimated (from the Clausius- Clapeyron equation) on the assu mpt ion that the heat of vaporization remains constant; in t hi s way it can be sho wn that the ideal so lubility of nitrogen should be about I0-^3 mole-fraction. These calc ulated values, it sho uld be noted, app ly to any solvent so long as the solutions show ideal behaviour; the actua l values for ethane and ni trogen in solvents such as hydrocarbons and carbon tetrachloride are about 0·018 and 6 X I0 - 4 mole -fractions, respectively, showing that there is some departure from ideality. It is, however, very strik ing that so s imple a treatment l eads to a result which is in err or by less than a factor of two. When water is the solve nt this is not so, for the intermolecular attra ction in liquid water comp licates the pictur e b~ malung all gases which dissolve in molecular form much less sol uble than they are calcu- lated to be; the gain in entropy of mixing is offset by the work which has to be done against this intermolecular attracti on in forcing molecules of sol vent apart. Water is the outstanding example of a polar solvent with strong int ermolecular attraction, bu t nitrobenzene, aniline and ethanol also show this behaviour to som e extent; another manifestation of this in the case of ethanol is the fact that its boiling p oint is 100° higher than that of dimethyl ether, which has the same molecular weight.
Th e inverse dependence of x 2 on p 2 ° at a
particular te mp eratur e also explains two other features of the so lubility of gases. Since the pressure above the liquid is less for gases with relatively high boiling points, gases such as butane and carbon dioxide are much more soluble in all solven ts than are hydro- gen, o:s.."y gen, nitrogen and carbon monoxide. Furthermo re, since for any gas p 2 o increases with in cr ease in temperature, the solubilities of a ll gases decrease with rise in temperatur e.
SOLUBILI'fY EXPLAINED 79
The latter conclusion may al so be r eached by Le Chate lier's princip l e: diss olution of a gas in a liquid may be considered as in volving first the condensation of the gas, followed by mixing of the two liquids ; since the con- densation stage is accompanied by lib e ration of heat, lowering the temperature will cause mor e gas t{) di ssolve.
SO LUTION S OF LIQU ID S IN LTQUTDS Mixtures of two volatile liquids wh ich are miscibl e in all pr oport ions are in many ways like solu t ions of gases in liquid s. If in t h e
pr eceding solution p 2 ° is less t han l atm at
25° and the ph ase de si gnated the solvent ha s an appr eci abl e vap our pressure, we hav e a typ i cal binary liquid system. Ea ch li q uid in s uch a mix t ur e exerts a vapour pre ss ur e equal t o its mole-fraction t im es its vapour pressure in a pur e s tate at the same temperatu re (Raoult 's Law) ; sol ution s for which this law hold s are call ed ideal solutions, a nd typical examples are u sually mixtures of very similar s ubstances such as ethy l bro mide and ethyl iodide, or benzene and toluene. It would be rather s urprising if there were many pairs of li q uid s which obey ed R aoul t's law exactly, since t he ex iste nce of a s ub stan ce in the liq uid phase at all is a proof of the pre sence of qu i te s trong int ermolecular attr actions which can be o ver come only by s upply ing the heat of vap orizati o n. Wh en a n id ea l solut ion is form ed from t wo liquids th ere is no change in he at content; the driving force t owards mi x ing is provide d by the increase in en tropy. There are , of course, many p air s of li qu ids which are miscible in all prop or tion s to yield mi xt ur es which show d ev iations from Raou lt 's law, e.g. diethyl e ther and acetone, h eptane a nd e tha nol, ethanol .and water , perchlori c acid a nd water. Th ese devi ations arise from t h e differences in pr opert ies of the con sti tue nts of the pa irs, a nd in so me cases there ar e. appreciable h eats of mi x ing. The natur e of the se dev ia tion s fr om Raoult 's l aw is, however , of minor int er est in conn ec tion with solubility, a nd will not be discussed furt her h ere. When two liqui ds are less simil ar than the members of the pair s ci ted above, partial miscibility usua lly occurs; if a lit tle ph en ol,
aniline or ether, for example, is added to water, i t dissolves, but furth er additions soon re s ult in the form ation of two layers, one a saturated solution of the organic com pound in water , and the other a saturated solution of water in the organic compoun d. Similar b ehav iour also r es ults with aniline and n- hexan e, or m ethanol and carbon disulphide. Gre at er difference in nature (e .g. water and carbon disulphide, eth a nol and mercury) results in immis cibility. The mi scibility of liquids is often di scu ssed in term s of their internal p 1·essu1·es, whi ch may be reg arded as meas ures of the dependen ce of internal energy on volume, i .e. of the s tr en gth of the cohesive forces. The se may be estimated roughl y by sever al methods, typical va lues at ordinary temperature bein g : hexan e, 2,000; ca,r bon tetr achloride, 3,500 ; phen ol, 5,000; water, 16,000 atm. In creasing difference in internal pre ssures l eads to de cr eas ing so lubility; increasing temperature o vercom es the co- hesive forces, and many pairs of partially miscible liquid s become complet e ly miscib le at hig her temperatures. Th e observant reade r will have no ticed that no que stion of absorpt ion or lib eration of a heat of fu sion or vapori zation arises in the diss olution of one liquid in another ; the overall heat of s olution is t herefore near ly always s mall and arises from differ ent degrees of non-id eality in the li q uids con- cern ed. Clearly, entro py consid erations p lay an imp or tant part; but it can not be said t hat at t he presen t t im e there is any simple t heory wh ich deals satisfactorily with all of the great variety of behaviour found in liquid- li quid systems.
S OLU'l'IONS OF S OLID S IN LIQUIDS Three t y pes of solid will be consid ered in the following discus sion: t hose having mole- cular, ma cromolecular and essentially ionic structUI·es. In molecul ar solids t he s hortest dista nces between atoms · in differ e nt mol ecu les are a b out twice t he int eratomi c di stan ces w ithin m olecules ; the int ermolecular bo nding is weak, as is s hown by the low h eat s of s ublim a- tion. Among s ub stanc es which cr ystallize with molecul ar s tructme s at th e ordinary
SOLUBILITY EXPLAINED 81
s um of t wo solv ati on energies approach the l attice energy of t he solid. H ence salts, if th ey dissol ve at a ll , will normally do so o nly in media of high E, such as water, ammonia , nitrom e thane and dimethy lformamide. Th e few sal ts t hat dissolve in or ganic solve nt s of low dielec tri c con stan t (e .g. e ther) n early alw ays contain very l arge ions a nd h ave very low l attice energies; tetr abutylam- monium pict·ate is an e xample of s uch a compou nd. Th e f act that the latti ce energy of an ionic solid depends inversely on r + + t·_ taken togeth er, whil st the s um of t he sol vation energies is proportional to 1jr·+ + 1jr-_ taken separ atel y, has other instru ctive conse- quences. For a set of co mp ou nd s h aving t he sa me st ru ctur e and r-_ ~ r·+ the latti ce ene rgies ch ange li ttle with increasing t"+ ; 1/L is constant and very sma ll , whil st 1jr·+ decreases rapidl y wi th increase in 1·+· H ence so lubilit ies decrease as r+ in creases; this is w ha t happ ens among, f or examp le, the alkal i- m etal chloropl at in ates or cobaltinitri tes or the a lkaline-ear th m etal s ulphate s in w ater. Inso lubil ity in water t hus need n ot den ote covalent chara cter ; there is, for examp le, no t t he sligh test evidence to suggest that t he bonding between th e m etal a nd the anion in calcium fluoride, bari um sulphat e or pota ss ium chloroplatinate is n ot ionic. · Silver ion ha s a lm ost t he sam e radiu s as s odium ion, and silver fluoride is fr· co ly soluble in water. Sil ver chloride, how ever , is very sparingl .v soluble, and it is in s tru ct ive to co mpar e the pau·s ~aF a nd NaC l, a nd AgF and AgCI. The observed lattice energies of these foue co mpounds (a ll of which h ave t he KaCl stru ctur e) arc 216, 184, 228 a nd 216 kcaljg -fo rmula-wcigh t respectively. Des pite the difference in size between F - and CI-, we n ot ice that the difference in lat tice ener gy is o nly 12 kcal for the silver sal ts; for tho so dium salts it has the mu ch lar ger value of 32 kcal. W e t herefor e infer th at there is ab out 20 kcal of non-elec trostat ic bond in g stabi lizing th e solid in t he case of silver chlorid e; this t er m ha s no co unterpart in the hyd ra tion energies of Ag+ a nd Cl- in w ater, wh ich ar e the sam e as in solutions of silve r fluoride and so dium chl or ide, a nd it is t here- fore t he ma in f acto r und e rly ing t he striking
differ ence in so lubility in water between sil ver .flu o rid e a nd s ilver chl orid e. The foregoing acco unt of t he solubili ty of salts h as been ba sed on an el ectrostatic model for the sol vated i on in so lu tion, the cha r ge on the ion in teracting with the uneven ch arge di stri bution in the water or othe r solvent molecule. The re is som et imes, however , ver_, - strong ev iden ce th at the solvation pr ocess is n ot a simple elec tro s tati c one. Thu s when a. param ag n etic Co3+ ion is h ydr a.ted, it becomes di amagn etic, showing that an elec tr onic r e- arran gemen t has taken place. 'Wh en C r- (H 2 0) 63 " ions from chro me alum dissol ve in (^1) 80 -lab c ll ed water , i t is fo und th at no exchange with ·wa ter from t he al um take s place during several hour s; ev id en tly the water attac h ed to the ca ti on is not feee to exchange, a fact which again s uggests coval ent bo ndin g. Fina lly, the hi ghly specific in ter - action between s alts of a few metals and or gani c liquids, e .g. the solubility of silver .fluoroborate (whi ch has the same stru ctur e as potass ium ft. uor oborate) in benz ene and to luene, ind icates covalent in teract ion of a rath er special kind , in t hi s case between th e aro mati c comp o und and t he silver ion. In a ll th ese cases, however , it s hould be remembered t h at the position of equil ibrium is determin ed by the difference be tw een the la ttice ener gy of the solid a nd th e magnil~tde (not the. nat~o·e) of the in ter acti on between ions a nd sol vent: it is the magni t u de of the free- ener gy change wh ich determines how f ar the process of dissolution goes.
SO LGTIO SS IN SOLIDS Th o ch aracte ri st ic feature of a solid is that it conta ins ato ms, molecul es or ions occupy- ing particular positions in a rigid structUl'e. When the solute ca n be accommodated in the sol vent l attice with o ut appr eciable change, tho format i on of a so lid so lu t i on takes pl ace r eadily: examples a t·e pr ovided by 1J-dichloro- bcnzcno and p-dibro mobenzene, potassi um chl oropl atinate and pota ssiu m bromoplati- nate , magnes ium oxide, MgO, and li thi um ferri c oxide, LiFe0 2 (Mg 2 +, Li + and F o^3 + are the same size, and the aver age char ge on Li+ and Fe3+ is the s ame as on Mg^2 +), and s ih·cr and gold. In such cases, there is a negligible r han gc in h eat content when a solid solution
2 }J )) UCAT ION I N CH E 31IS'l'RY
is formed. Gases and liquids, on the o th er hand, are seldom very solu ble in so lids: the fact that they are gases and liq uids implies a simple molecu lar stru ct ur e in the ir so lid phases, and solubili ty is t hen to be expect ed on ly in elemen ts or co mp ounds cr ystallizing wi th simil ar simple molecular stru ctures. For a gas or a liq uid to dissol ve in a solid requires work to be done in fo rc ing apar t the units in the so lid stru ct ure an d, since the interatomic, intermolecul ar or i nteri on ic fo rces in solids are rel ative ly str ong, this ener gy fac tor milita tes against solubili ty. Gases usua lly interact with solids only to tho extent of form ing a monolayer held by surface forces and a weakly he ld l ayer, only a few molecules t hick, on top of this: W here the solu bility is appreciable, as in the case of hydrogen in palladium and ot her m etals, str ong chemical bond ing must be invol ved to compensate not only for the work done in
expa ndin g t ho metal lat t ice, but also fo r the energy needed to br eak t he strong bo nds (103 k caljmole) in hydrogen , which is pr esent in the solid phase as single atomic uni ts.
CONCLU S IO N The posit ions of a ll equilibria are deter - mined by the rm odynamic factors. Wh en solu te and so lvent are in the same physical state, the increase in entr opy on mix ing ensmes so lu bili ty so long as (i) in the case of liquids, th ey do n ot differ to o wi dely in their p ro p erties a nd inte rn al pressur es; (ii ) in the case of so lids, th ey have t he same str u ctu re and, if ionic, con tain ions of similar sizes. Fo r oth er cases, heats of vaporization or fusion are involved as well as ent r·opies of mix ing. In a ll cases, however, t ho key to the inter pretation of solu bility lies in the evalu ation of the fr ee-ener gy change wh ich acco mpanies the pr ocess of dissolution.
CLE AJ.~ I NG OF APP ARATUS: T HE ALCOHOL -N ITRIC AC ID R EACTION"
At one t ime it was custom ary to cl ea nse dirty appar at u s-es pecially con taining ta rr y m atter-by m eans of ni t ric ac id and alcohol. IL seems th at in the early days t his pr ocedur e was not considered pa rti cula rly dangerous, but in r ecent years its h azard ous n ature has been incr easingly e mph asized a nd serious accidents h ave in f act occmred. The reacti on is well- know n to be auto - catalytic. W hen pure n itri c ac id and alcohol are mixed, no reacti on occm s at fi rst, but r eact ion st ar ts almost explosiv ely after a vary in g lapse of t ime and is accompanied by viol ent bump ing and t he form at ion of nitrogen oxides. It seems likely th at the man ufacturers of nitric acid have progr ess- ively supplied a pmer and p urer pr odu ct and th at, in the early days, a treache rous latent peri od was n ot a ma rked f eatu re of the react ion. T he d iffi culty can be av oided by t he deliberate ad dition of nitrous acid. If 5-10 per cent of sod ium nitrite is stirr ed into ordinary nitri c ac id (density 1·42 ), a red -colomed acid, containing dinitrogcn tetroxide, is immediately obtain ed. If to
this alcohol is added , r eaction s tarts at once. This mod ifi cation of the usual procedure has now been used for cleaning ap paratus over a numb er of years, and so far no indica- t ion that it migh t pr ove dangerous has been noted. The 'bumpin g' th at is so character - istic of the pure nitric acid r eacti on does not occm. The reaction starts sm oothly at tho smface of tho liquid, a nd i ts onset is so ra p id th at there would seem to be little ri sk of adding enou gh alcohol, in one lot, to gener ate danger ous pr essure in a flask. Taking elem enta ry preca ut ions (fume-cup - board, avoid na rr ow-necked vessels and so on) it is difficul t to see in wh at way it might pr ove danger ous. All th at is necessary is th at the nitric acid, before use, should have a rich reddish -brown colom. If this single po int is attended to, even the mistaken use of sodium ni trate fo r nitri te should at once give warning th at alc ohol mu st not be added. If desired t he nitri c acid m ixture m ay be k ept bottled r eady for use.
R. E. D. CLARK Cambridge Co llege of Ar ts & Technology.
