~
Table V.
Analysis of Various Nonyl Phenol-Ethylene Oxide Condensates
Weight Konyl phenol 2 8 0 8 Trace Trace 0 5
Compound 3 mole S P E O 4 5 mole UPEO 6 0 mole TPEO 9 0 mole ?;€’EO I O 5 mole S P E O P = present a Ultraviolet procedure Table VI.
1 mole
52 0 0 8 Trace Trace 0 8
2 mole 34 6 10 2 2 1 0 2
1 2
$c
Composition
3 mole 7 9 22 3 6 9
oa
1 4
4 mole 0 8 19 8 12 7 1 4
0 7
5 mole P 15 9 18 5 4 5 11
6 mole
7 mole
8 mole
P P P P
P P P P
P P P P
Detnd mol w t 2b2 363 501 669 729
Comparison of Experimental Data with Theoretical Data Based on Poisson Distribution Function
Detnd. mol.
Compound 1vt.a Igepal 210 282 363 Polytergent B-150 501 Poll-tergent B-200 669 Polytergent B-300 729 Polytergent B-350 Vltraviolet procedure.
KPEO1 Found Theory 52.0 55.99 0.8 3.10 Trace 0.37 ruckerei, 1961. RECEIVED for review March 11, 1964. Accepted June 17, 1964.
Radiometric Determination of Adsorption Isotherms from Solution on Organic Substrate JEROME HABERMAN (and THOMAS C. CASTORINA Explosives laboratory, Feltman Research laboratories, Picatinny Arsenal, Dover, N . J .
b The adsorption by 6-octahydro1,3,5,7-tetranitro-s-tettazine (HMX) of the carbon-1 4-labeled quaternary salt, stearyltrimethyl lommonium bromide (STAB), from solution was studied. A solvent system consisting of 90% water and 10% ethanol gave acceptable isotherms on 10-micron HMX. The data are shown to b e quantitative and reproducible with a standard deviation of 0.01 mg. of STAB per gram of HMX. The knee portion of
the curve is below the critical micelle concentration of the STAB solution, indicating that the plateau represents true saturated adsorption.
P
on the adsorption of surfactants (adsorbates) from solution onto inorganic adsorbents (substrates) using radiotracer techniques have been reported ( 9 ) . The surface areas of these substrates were about REVIOUS STUDIES
5 to 50 sq. meters per gram. I n addition, the interactions of these adsorbents with the surfictants were sufficiently large to give q)ecific adsorbawes that were readily measured. However, no work has been reported on the al)plication of radiotracers for the determination of adsorption from solution by an organic substrate. Organic solids have relatively inacti1.e surface5 because of their small surface areas (nonporous) and low degree of VOL. 36. NO. IO, SEPTEMBER 1964
1917
interaction with surfactants. For this reason adsorption of surfactants from soliltion hy organic substrates was frrqiwntly considered to be unmeasurable. T h e quantity of surfactant adsorbed on tlie glassware used in a dctc,rmin:itioii ninde it extremely difficwlt to Attain :in absolute adsorption v:ilne. Iiw:iuse of the low surface xntivily I J I IIlIX (cdculated surface a r w t)ciiig lrss than 0.3 sq. meter per gram), a radiometric method of analysis furnished the only possible means of ~ ~ ~ ( ~ a s usinal1 r i r i g decreases in adsorbate roncentrations solely attributable t'o :ttiiorl~tionI)y H l I X The development of adsoi.l,tion isotherms for organic sul>st,ratcsby adsorption from solution will n ~ d wit l)ossit)le to measure induced chsngcs in the surface activity of organic exl)losi.iw, The chcmical and r)hysical pro1)vrticIs of tlie rst)losives would then be stiltlid as a functiori of t>hesesurface changes. This I)al)er describes a reproducible and qiiantitative radiomet,ric method for obtaining adsorption isothernts froin solution for the difficult 6 - oc*t:thydro - 1,3,5,7 - tetranitros-trtrazine -stearyltrimethyl ammonium broniidc. (11 MX-STAB) aqueous solution systtni . EXPERIMENTAL
T h e C - 14 - STA13 I ~ l ) r l e t ,l i n the one position was obtnincd frorii the S e w England Nu(.Ir3r Corp., Vaml)ridge, IIass., and t l i l u t i v l to :t spwific activity of 0.05 m v . lwr grain. Thc HXIX was purified by successive .tallisations from acetone. This I)rocvtiure was shown to remove all of the rna/or impurity, hexahydro1,3,5-ti.initi,o-s-trinzincl (RDX), with C:H'~)(JIl-l4--t:tg~td1IL>X ( 3 ) . The siintill:ttion solution was composctl of 4 grams of diphenyloxazole (111'0) and 100 mg. of 1,4-bis-2-(5~ ~ t ~ e r i y l o x ~ z o l ~ ~ ) t (POPOP) ~ e n z e n e (scintillation grade) dissolved in 1 liter of a (j0 niixture of reagent grade toluene arid :tl)sulute alcohol. Procedure. ST.113 ADSORPTION l I 1%:.isu R M I(; NTS. ;\p pro xi ma te 1y 1 grain of EIlIX was weighed and 5 ml. of the C-14-STA\13 solution were pipettcd in a 2-dram vial. T h e scriirely cai)i)ed vial was equilibrated, with grntlr, agitation in a constant trnil)clr:iture Ilath set a t 30" i . 0 5 O C . , usi~ally overnight. ;\fter equilibration Lint1 suhsequent centrifugation, 1)rw;tution was taken to rinse the 1)iI)ot w i t h the solution to be counted :tnd t o cliscard the rinse solution before h I-ml. niakilig a determination. aliquot was 1)il)ettect into a 20-ml. vial, t h t l >vintill:ition solution was added, and the> s:iini)le .iws coiintcd in a liquid srinrill:it ii)n c w i i i i t c r . .\ 1-nil. aliquot from i l w i i i i t i n l soliltion was treated in the
Reagents.
13:
\:llIl('
Ill:tlllleI'.
ratio of the specific activities ( c ~ . l ) . n il, ) ( ~nil.) of the equilibrium to 'I'lw
initi:tl solutions multiplied by the initial , ~ , ~ u w ration iit (milligrams of ST.113 191 8
o
ANALYTICAL CHEMISTRY
1)er liter) gives thc equilibrium conwntration of STAH in milligrams per liter. The difference of thew two concmtrations is multiplied hy the total s:iniple volume and corrected for the adsorption upon the vial. The resulting net milligrams of ST.\I3 adwrhed I)er sample weight of IIMX :IN: thrn converted to sllecific adsorption in milligrams of ST.113 adsorbed I)er gram of IllIX.
SOILJBILII~Y ~)ETEHMINATIOXS.
I h T E R M I NA'TI 0 N
0F
'l"F:
CIt IT I C A L
MICELLEC O N C K ~ I ' I I A ~(IC. Il O I CN) OF ST.\Ii il'.ith thc cwuluctivity nicthod, sollitions of varyiiig conccntr:ttioris, some below the exlxctcd C l I C and others above tlie C l I C , wore 1)i~rI)aiwl with 10/90 eth:tnol-w:itcsr iolvent. .\ conductivity t)ridgc>w:is i i i r d t o rnmsiire the resistance of thcw solution.;. dip-ty1)e c ~ l was l c-nlibr:itetl with 0.1.V KC1 solution and iisrd to obtain thr specific resist an (CS. .\ 11 wadi n gs \YCI'V taken a t 30" C. 'I'he sl)cdica c m t f u c - tivities were c.alculattd by taking the reciprocals of the sl)wifict resist:tncv.;. h plot of the sliccific cwdric*tivitiw against, the concentrations gives an intersect corres1)onding to the CLIC. The C l l C \vas c-lieckeci by using the dye titration niethotl (4). J h s u r e ments of the 1)articnlr size clistrilmtion were made by the corivrntional inirroscopic method. DESoRPrlON
I>k:TF:RMlN.4TlON.
>!
Chi,-
bon-14-labeled HMX (specific artivity, 0.2 me. per gram of HSIX) was pl:tced in a 2-th.arn vial :tnd 5 ml. of n I O '90 ethanol-water solution of cold ST.4 I3 (100 mg. per liter) were added. The capped vial was equililmtrti for 2.4 hours a t 30" C. with constant shttking:. After centrifugation, 1 ml. of the suI)ernatant liquid was tr:msfcrreti to a rouiiting vial and the solvent was eval)oixt(d in an oven a t 100" C. Ulmn coolinq;, 1 ml. of dinlrthylsulfoside was : d ( t r r l to dissolve the HMX in the vial and scintillation solution was adticd. .\ frw mi1ligr:iins of t h tagged HJIX, ~ w i g h r d accurately to within 0.01 mg., srrveti as the reference htandarcl which \vas prepared for counting in the inanner described for the unknown. The ,wlubility of t,he B l I X dissolved in the adsorption solution W : L S c:tlculatc~tl I)y multiplying the activity of the unknowu by the recilmcal specific activity of the reference stantlard. The solubility of ST.\I3 in 10190 ethanol-water solution was forind tising the same radionirt,ric terliniqw as used for HMX, with the excel)tion that dimethylsiilfosidr w:ts excluded from t he scintillation solution.
i\
typical adsorption was run with an initial concentration of ST.iI% corresponding to the equilibrium roncentration in the platcau region of the isotherm added to a weighed .saml)le of HXIX. The su1)rrnatant solution was counted in thc u s u d mnnner--i.r., discarding 1 nil. and counting thc srcond milliliter. l'hv tltlilvtrd sul)(~rnatant liquitl w:is iiic.i,cxsrtl with f i w h solvent l)y an :iniount oqiial to t h v volume removed. The I ) ~ O C T * S of reequilibration and sucwssive tlilritions was relieated until the ralculatrd ST.\Ii
IIII I
I
b 100
EQUILIBRIUM CONCN mg /LITER Figure 1. Adsorption of STAB on HMX from xylene-saturated water and pure water at 30" f 0.05" C. 0 0
Xylene-saturated water Pure water
cwnwntration remaining I'rior to equiIi1)r:ition cwrwponded to the equilihration value ne:tr the \tart of the awcwlirig /)ortion of the (wrve. RESULTS A N D DiSCUSSlON
The choice of a solvent for solution s r degree ndwi,l)tion is critical I ) ( ~ ~ t i ithe of atlsorl)tion of surfactants is strongly affected by t h r solvent. This relntionship may I)c :tttributed to the cornpctition of the wlvcnt nnd wrf:icLtant for t h v sitl)+tr:itc~surface arid t h o cwnlwtition of' thr solvcnt and hiihstratr surf:we for the surfactant. .\ usrful solvrnt, for tho H l I X hul)strate is onc in whivh the adsorbent has less than 1 0 P molar solubility, the surfactant has a soliil)ility grrntcr than molar, and the affinity of the solvent for the surfartant is less than that of the siirfartant for the :itisoi.limt. Initid eslwiments using ST.\13 in :tclueoiis solution showed that, while :~(lsoq)tiondid occIir, there was a lack of rcyducibility in the data, and that isotherms could not be developed for the average 10-micron particle size H l I X i i s ~ din this study. I t was not I)ossiLle to t%st:iblisha plateau, because the highcst cffective concentration of ST.\I%tv:is in the neighborhood of 30 nix. of S'l'.\l% 1)cr liter ('3 X iO--bJ[) a t t,he equili1)riwn concentration, as Figure 1 shows. .\ solvent mixture consisting of sy l rnc-sa t 11rat (id w a tcr was invest i-
I
I
EQUILIBRIUM CONCEFITRATION,mO I LITER
Figure 2. Adsorption isotherm for STAB on silica in ai 90/10 waterethanol solvent
gated but the preliminary result,s were no better than those obtained from the experiment> with liiire water, as may be s w n fmni Figurr I . 1;'ailui~to ctstal)lish a plateau in rach c a w is attributed to the low solutiility of S'l',115 in water ( 10-4.\1) :ind i n sylrne-saturated water 12 x I O - ~ M ) , Rather than have additional (pantitic's of HXIX ronsumed in a cwnt i r i r i c d mircah for LL suitahle solvent, silica u a i i i w l a,< the standard in detmiiining c+fwtivt, :itlsorl)tion of STXI3 from various other solvents. Ethanol dissolvcd o v ~ r1yo of ST.113. Ho\\-c,vc~,no :rtlsorl)tion on the silica took 1)lac.e. Succcssi.ic dilutions with water l q m d a 50% mixture showed adsorption to be greatest in a mixture conhisting of 9O/lO water-ethanol. The rel)rothic*ilileresults obtained from this S'I'.\I~-water-ethanol isotherm bhown in Figure 2. The solubility of STAl3 in the IO; 90 ethanol water solvent, determined by the radiometric technique, was 3.3 x 10-3x as compared to I O - ~ M in pure water. -1 cursory drtermiriat,ion on HMX using this solvent mixture gave every indication that an isotherm similar to the one obtained for sil.ica is also possible for HhIX. However, for a quantitative determination, a detailed study of all possible parameters affecting adsorption had to be conducted and several related problems had to be solved. At first, when relatively high concentrations of STAB (about 500 m g per liter) were used, some anomalous results were obtained. An example of these results was the rapparent adsorption of 0.050 mg. of STAB per gram of HMX as compared to 0.150 mg. of STXI3 per gram of HMX obtained previously for the same sample of the HhIX and ST-XB solution. These erratic result!; were produced because solutions of surfactants form micelles at certain critical concentrations. Micelles ( I ) arc: aggregations of long chain molecules into large charged units which are in dyn,zmic equilibrium with the monomer species. These micelles may contain up to 100 or more mononier molecules. ;Striking changes in ~ a r i o u sphysical prmoperties of solu-
tions timy oc(*iir in the wgioti of t h r critical niicelle c o n c ~ ~ ~ t r a t i o(CAIC). ii 800 For example, there is evitlenw of the solution becoming colloitlal in this 700 c region. lIicelle forrnatiori may bti tiriic -0 x 600 dependent and , tl t c m b f c w , sol i i t i c )i is 1ii:i.v be subject, on long to 1)rok 500 gressive decreases in monomer con2 c centration (6). This effect m:ty':iccwrint 3 400 0 for the anomalous results ol)stsrvcd in $ 300 adsorption st'udies with solutions of V long chain molecules. 2 200 I n addition, :is the i,otti,(,tiir:tiiofi LL V of the monomer' species is increased, L 100 v) there is an increase in the number of ' " ' " micelles, which in turn c ~ c i r r c ~ i ~ ~ i r i ~ l i r i ~ l ~O 0 l 100 200 300 400 S O 600 decreases t'he initial monotiier conSTAB, mg/LITER centration ( 2 ) . .Ilthouph the substrate Figure 3. Determination of critical is exposed to an apparently increasinq micelle concentration of STAB in 90/10 equilihium coric,c~ntr:ttioti diiriiig t t w water-ethanol ljrocess of o1)t:iining isotticwi., 1 extent of adsorlltion may decrease I)(:cause the concentration of thc ftw monomer is ~ c t l i i c ~hy ~ l t l i r f ( J I ' l n 3 ~i ( l i i of micelles. .1(lsorl ition i i o t h i t is have hetm relmtctl in w1iic.h thc. :i(lwrl)tion rraclietl a x ~ 1ir:itioti t Ivvc~l at 1 tic' equilibrium conc.c~ntratiori wliicti allproximately coitic,icletl wit,h tlic: ( ' l l ( : of t h r long chsiri c~lcctrolytcs used (,s). The criticd niic~vlle cnonrcxntration of the ST.113 in tlw >olvctnt by,t('tn NXS determined by ci )ntluc,tivi t y i i w : L w t x s ments ( 7 ) . The sliecific c,oiitluc.tiviti(,s of a series of solutions werc d ( ~ t c ~ r n i i n c ~ t 1 a t 30" C. and plotted against the concentration. The CSZC was found at the intersection of the two curves as may be seen in Figure 3. 'I'he CMC f o r S'I'.il3 in this solvent nictlium is 5.1 x 1 ( ) - 4 , \ [ (200 nig. per liter). The ( : \ I ( ' ol)t:iincil was checked by using the dye titrntion method ( 4 ) . Xs a result of stritlicss on the CllIC of ST.113, adsorlltiun of thc stable monomeric form of SrI'.Ll% in 9OjlO water-c.thano1 was confined by an upper conccntration of a h i t 240 ma. per liter. l i ( 8
0
0
m300 0
-X
-U
U
Y
X
5
I
$ a ti
200
C
0
E LOO
ADSORPTION ISOTHERM ON 2-DRAM VIALS
IO0
200
300
400
500
EQUILIBRIUM CONCENTRATION m g /LITER
30" + 0.05"
C. 90/1 0
VOL. 36, NO. 10, SEPTEMBER 1964
1919
Figure 4. Adsorption isotherm of C-14 STAB on HMX at water-ethanol solvent
Table I.
Desorption Experiments
.ht1(iunt
S‘l’.ll3
adsorbed, nig. per gr:trn
€INX 0 134 0 130 0 1:30 0 123
0.12(5 0 129 0 100 0 058 0 . 0x5
C:ilcd. eq. 17q. (‘(inc’n., concm., m g . rtig. per liter per liter 44 1 178
93 80
69 19 33
1T6 108 66 39 21 15
20
9
15
6
1)ilution of blanks
467 278 167
280 167
HMX.
94
100
V i t h the establishinent of the parameters for solution adsorption using ST.115 in a water-ethanol solvent mixture, it becanie possible to construct an isotherm for HMX. .\ typical isotherm on H l I X is plotted in Figure 4. The apparent adsorption isotherm of ST.il%on H l I X may be characterized as H-type according to Giles et al. (5). H-type isothtrnis are tylical of which a low affinity surface interacts with a high affinity adsorbate, The plateau, or beginning of the lincar lmrtion of the curve above the knee of the isotherm. undoubtedly represents a saturation of the surface because the knee corresponds to an equilibrium concentration below the C l I C of 200 ma. per liter. This amount may be a complete monolayer, but not necessarily a close-packed one. ’The monolayer at the HJlX surface Irobably includes both Polvent and adsorbate niolerules. The reversibility of the adsorption of STAl3 on HAIX was established bl- successive dilutions of the equilibriiini concentration starting a t the right end of the plateau and continuing to the left of the liner portion of the isotherm. Table I shows the aniounts of ST.il%
...
esperirnental conditions was detei,niined rridionirtricdly to be (1.36 X 10-5 niolar) iri 10;’90 ethanol-water containing 500 ing. of S‘Y.113 lwr liter. .\ny solul)ilit>-, however slight as observed for HJIX, could lead to (-ontinuc,cl mchange with the si1rfac.c and thus ..snioothingJ’ or “riJughing” of Imticlw. I n addition. becmise the particle rlistribut ion follows a Gaussian-type ciirve, it is conc.eivablc that a I)rocess of growth of the larg,er liarticks could take place a t the expense of the snialler O I I C . ~ . The ,stability of thr H l I X surface and Ixirticlt zize distriliution during the of ~ ~ ~ u i l i l ~ rwas, a t i ~therrfore, ~n gated. This was donc by measririiig the' saturation atl.sor.l)tion value ( ~ S‘I’.iH f ~ ) e unit r wiJight of HLIX as a func8tic:n (if time. The rrhults obtained for a -erie.s of four wi)licate espwimmts g ~ v c:in avcrage of 0.238 nig. of W.\H a(l>iji.bed lit’r g r a ~ i i of IIAlX with iisndard deviation of 0.01 ing. over a Ii(,riid varying from ont-half hour to rncnty-four hourz of eqiiilibi~ation. I3ecmsi~ tlics ainount of STAH ad-
’1 920
sorbed is Iroportional to the surface area in a given initial concentration, it can be concluded from the results that the particle size antilor surface area of HMX is ap1)arently stable in the 10/90 ethanol-water solvent ovpr a period of 24 hours. Consistent with this observation the time factor was not taken into consideration in dwe1oi)ing adsori)tion isotherms and equilibrations were usually run overnight. The rei)roducibility of this radiometric rnethod of adwrption from solution is shown by the standard deviation of 0.01 obtained for sis replicate esperinients having an average value of 0.243 mg. of ST.113 per gram of
ANALYTICAL CHEMISTRY
adsorbed lier pram of H l l X for several equilibrium concentrations as rompared to corres1)onding calculated values based on dilution. In the plateau region, these sets of values are comparable. Below the knee, the esperimental values are greater than the calculated value*, indicative of desorbed SI.il3. L-sing the sanie technique, the dilution of blanks shows that the data are reproducible within experimental error. 1hese tlesoq)tion experiments were repeated using solvent saturated with H l I X and the results obtained were identical. The reversibility of aclsorption was thus clearly established and it was, therefore, concluded that ST.\B is physically adsorbed onto HMX from the water-ethanol solvent mixture. 7
,
ACKNOWLEDGMENT
The author5 express their gratitude to -1.C‘. Zettlenioyer and Samuel Helf f o i their interezt and helpful qugpe\tion\. LITERATURE CITED
( 1 ) Adanison, ;ti itr.> “Physicd Chemistry of Surfaces, p. 373, Interscience,
S e n Tork, 1960. ( 2 ) Ibid., p . 377. ( 3 ) Castorina, T. C., Holahm, F. S., (;raybush, R. J., Kriufninn, J. V.R., Helf, s., J . .4m. Chem. SOC.82, 1617 (1960). (1)Corrin, 11. L., Harkins, iV. I)., I h i d . , 69, G79 (1947). ( 3 ) (;iles, C. H., hIacEwan, T. H., S a k h n a , S. S . ,Smith, I)., J . C‘hem. Sac. 1960, 3973. (6) S a s h , T., J . Colloid Sci. 14, 59 (1959). ( 7 ) Osipow, L. I., “Surface Chemistry,“ p. 165, Reinhold, S e x York, 1962. 18) ?‘arnainushi, R., Tarnaki, I