Nile Red Jerry

7/29/2019 Nile Red Jerry

http://slidepdf.com/reader/full/nile-red-jerry 1/8

Anal..Chem. 1900, 62, 615-622 615

(4 ) Haddad. P., unpublished work, 1989.(5) Jandik, P.; Li, J. , unpublished work, 1989.(6 ) Small, H.; Stevens, T. S.; Bauman, W. C. Anal. Chem. 1976, 4 7 ,

LITERATURE CITED

Reaction Detection in Liqukl Chromatography; Krull, I. S., Ed.; Chro-matographic Science Series; Marcel Dekker: New York, 1986; Vol. 1601.34.

Nile Red as a Solvatochromic Dye for Measuring SolventStrength in Normal Liquids and Mixtures of Normal Liquids with

Supercritical and Near Critical Fluids

Jerry F. Deye an d T . A. Berger*

Hewlett-Packard, P.O. Bo x 900, Route 41 an d S ta rr Road, Auondale, Pennsylvania 19311 0900

Albert G. Anderson

Central Research an d Development Departm ent, E . I. du Pon t de Nemours an d Co., W ilmington, Delaware 19880-0328

Nlle Red was used as a solvatochromk probe to measure thesolvent strength of 85 pure llqulds and the resutts wer e com -pared to solvent strength scales from the better know n Rel-chardt’s hN)nd Et(Bo)dyes. Nlle Red Is also soluble In m epure sup ercrltkai flulds and In polar mixtures of supercrtticalfluids and m odlflers. The solvent strength of mlxtures of or-

ganlc mod ner s In CO,, Freon-13, and Fr-23 was measuredas a function of modlfler concentratlon wlth Nile Red andother so lvatochro mlc dyes . Nlle Red exhlblts large shlfts inthe wavelength of Its absorption maxlmum allowing subtlechanges In solvent strength wlth fluid composition to bequantlfled, and It Is stable In very strong acids. Relatlvescales of solvent strength com parlng different modtflers added

to C 02 are presented and co mpared to a wlde range of nor-mal liquids.

INTRODUCTION

The polarity range of solutes that can be separated bysupercritical fluid chromatography (SFC) can be greatly ex-tended by the addition of polar modifiers to supercriticalfluids. It has been w idely repqrted (1-4) that the initial smalladditio ns (less than 1% ) of polar m odifiers (like methanol)to superc ritical fluids (like carbon dioxide) significantly de-crease the chromatographic retention of solutes on packedcolumns including those with a bonded stationary phase. Suchlarge changes in retention caused by small changes in modifier

concentration appea r to be inconsistent with bonded phasechromatography (BPC). It is often assumed that bondedphases do not adequately cover silica surfaces and residualsilanol groups produce long retention and very poor peakshapes. Th is can be viewed as a competition between abonded phase retention mechanism and an adsorption re-tention mechanism which, in the absence of a modifier, favorsth e adsorption m echanism. Polar modifiers are assumed todecrease retention an d improve peak shapes by covering thesilanol sites (Le. blocking the adsorption mechan ism). It isuncertain whether, upon site deactivation, retention followsa bonded phase or adsorption mechanism.

Retention in BPC is normally a function of a mobile phasepolarity parameter, P’, which in turn is usually a linear

function of composition according to the equation (5)

p‘ = 4 a p a + 4 h p b (1)

where 4, and 4 b are the volume fractions of mobile phasecomponents a and b , and Paand P b are the solvent polarityparameters for the two pure components. These polarityparameters a re based on experimental solubility da ta likethose reported by Rohrschneider (6). Partition ratios (k’J renot necessarily linearly related to P’but retention changesare generally limited to a 2.2-fold decrease in k’for a changein P’by 1.0 unit (7). If the P’values for carbon dioxide andmethanol were 0 and 5.1, then th e addition of 1%methanolto carbon dioxide should decrease retention by only a fewpercent compared to retention using pure carbon dioxide.Instead, much larger changes in retention are common. Werecently observed tha t the addition of 1%methanol to carbondioxide decreased the reten tion of chrysene on a SA columnby 2.4 times ( k ’ = 17.9 in pu re COz to 7.6 in 1%methanol).Such large retention changes suggest th at eq 1 s inadequatefor predicting retention in packed column SFC.

Adsorption chromatography is characterized by largechanges in k’with small additions of polar modifiers and th us,at least superficially, appears t o offer a better explanation ofretention in packed column SFC than BPC. In the standa rdtheoretical treatment (8 ) , etention of solutes is related to th eadsorption energy of the mobile phase solvent on the sta-tionary phase through the equation

log KO = log V, + a(SO A , € ” )

where KO is the distrib ution co efficient for the adsorption ofthe solute, V , is th e adsorbent surface volume per weight, ais a measure of the activity of the adsorb ent, So the sampleadsorption energy on the standard-activity adso rbent, A, th esurface area covered by a molecule of th e m obile phase solven t,an d t o the adsorption energy of th e mobile phase solvent perunit area of the standard adsorbent. In binary fluids, themobile phase solvent strength, tab, tends to be dominated bythe more polar component, b

0003-2700/90/0362-0815$02.50/0 0 1990 American Chemical Society



616 A NA L Y TICA L CHE MIS TRY , VOL. 62 , NO . 6, M A R C H 15 , 1990

and pure com ponents, respectively, Nb is the mole fractionof b in the m ixture, and nb is related to molecular size. A plotof cab vs the mole fraction of b yields an ad sorption isothermfor b. Retention is, thus, related to surface pheno mena.Mobile phase interactions with the solute are generally ig-nored. Attempts to rationalize the solvent behavior of binarymixtures with c hromato graphic behavior (9,101usually con-tain an implicit assumption that the mixtures behave asregular solutions.

I t i s alsowell-known th at mobile phase intera ctions become

more important as he difference in the polarity of the solventcomponents increases (8). We interpret this as increasingdeviations from regular solution behavior. In the sta nda rdtheory, there ,is no framew ork for measuring these deviationsother th an through t he general equation for the net energyof adsorption (AE)

AE = E,, + mEal- E,1- E,,

where each term represents dim ensionless partial mo lar freeenergies and the subscripts x and s refer to sample and solvent,and a and 1refer to adso rbed a nd liquid phases, respectively.Ed- Ed epresents mobile phase interactions indepen dent ofthe stationary phase.

Solvatochromic dye measurem ents offer th e possibility ofdirectly measuring some function related t o P’or alternately

to mo bile phase free energies (i.e. E,, - E d ) independent ofchromatographic retention. The re is still considerable debateas to the value or general applicability of solvatochromicsolvent st ren gth scales (Le., refs 11and 12), but a t the leastthey provide an app roximate correlation to solvent polarity.The y also point out som e unexpected solution characteristicsof mixtures of polar modifiers and supercritical fluids. Rei-chardt (13) developed a solvent stren gth scale, Et(,,), usinga pyridinium betaine dy e, th e adsorption maximum of whichshifts depending on th e polarity of th e solvent in which it isdissolved. Note th at in this context “solvent strength” isstrictly related to solute (dye)-solvent interactions w hereasin adsorp tion chromatography “solvent streng th“ is really anelution strength an d normally has nothing to do with inter-actions in the mobile phase. We use the term “solvent

strength” as a measure of the interaction between solutes andsolvent in a fluid phase, independent of any station ary phaseinteractions.

Strong interactions in the mobile phase tend to producedeviations in retention in adsorption chromatography whichcannot be explained by mobile phase adsorption (8). In spiteof this, Reichardt’s Et(%) cale ha s been compared to tab valuesfrom adsorption chromatography (6) with “good” correlation,even though acidic and alcoholic solvents appear to p roducea substantially different relationship between Et(30) and cab

than other solvents. Kam lett and T aft and co-workers (14-16)have developed separate scales, using a variety of carefullychosen dyes, to separately account for polarizability, on-dipoleinteractions, acid-base, andhydrogen bonding. U nfortunately,th e dyes employed often produce only minor wavelength shifts

which make small changes in solvent strength difficult orimpossible to observe. For example, the absorbance maximumof 2-nitroan isol, a widely used dye in measurements on Kamletand Taft’s A* scale, shifts only 3.4 nm when the solvent ischanged from tetrahydrofuran to methanol (17).

In th e chemical engineering literature , clustering of polarmodifier molecules aro und polar solu te molecules (18-28) ina large excess of a supercritical fluid is widely accepted. Evenin mixtures con taining low bulk c oncentration s of modifier,solutes may be surround ed by a solvation sphere containinga high local concentratio n of the modifier. Such clusteringbecomes more pronounced as th e difference in polarity be-tween th e supercritical fluid and the modifier is increased.This suggests tha t a substantial part of the nonlinear behavior

of chromatographic systems using supercritical fluids mixedwith polar modifiers may be due to mobile phase, not sta-tionary phase, effects. If true, this ha s imp ortant implicationson retention mechanisms in both packed column SFC ndliquid chromatography.

Several solvatochromic studies have been performed (17,29-31) on pure supercritical fluids. Early unpublished rep ortsestablished th at pure carbon dioxide ( C 0 2 )was in the samepolarity range as hexane (32,33), and more detailed studiescorroborate and expand on this finding. Yonker et al. (17,

31) reported th e solvent strength of NH3, C 0 2 , nd N 20withchanging pressure, using solvatochromic dyes. In additio n,they measured th e solvent strength of C 0 2 ontaini ng severalconcentratio ns of methanol. Th ey concluded th at mobilephase solvent stren gth did not change dramatically wth smalladditions of polar modifiers. Instead, they attrib uted dramaticshifts in retention accompanying small additions of modifierto a modification of packed column statio nary phases throug hadsorption of the modifier (15)onto “active sites”.

Carbon dioxide and methanol possess vastly differentsolvent strengths on all solvent strength scales and shouldproduce extreme differences in retention in either BCP oradsorption chromatography. At th e same time, the nonlinearsolute-solvent interactions (clustering) in th e mobile phase,which are generally accepted in chem ical engineering, have

been virtually ignored in discussions of retention in packedcolumn SFC. Thi s raises the possibility th at retention mayfolow a BPC type mechanism b ut with a no nlinear relationshipbetween solvent strength and composition. Alternately,solvent strength measurements obtained by using solvato-chromic dyes may provide th e mean s of correcting for devi-ations between retention an d the adsorption isotherm for thesolvent on the station ary phase, once adsorption measure-men ts are made available. We believe th at polar modifiersstrongly interact with column packing materials, bu t we doub tsuch interactions represent a form of classic adsorptionchromatography. Th e present work employs Nile Red toprovide detailed measurements of the mobile phase solventstren gth of mixtu res of polar modifiers in C02,Freon-13, andFreon-23.

EXPERIMENTAL SECTION

Chemicals and Instrumentation. The adsorption maximaand the resulting calculated values of tr ansitio n energies for NileRed reported in Table I were collected by A. Anderson in 1983with an IBM 9430 spectrophotometer. Extrem e care was takento exclude water from the solvents. Th e rest of the work reportedhere was recently done with a Hewlett-Packard (HP) Model 8450Aphotodiode array spectrophotometer. A high-pressu re flow cellwas made of 316 stainless steel. Th e flow path was 1.5 mmdiameter with a 1cm long path length. Fused silica windows were6 mm thick and were transparent to UV to below 190 nm. Th ewindows were sealed, leak tight, on the ends of the flow path withKel-F washen compressed between the windows and the cellM y .A Hewlett-Packard Model 1084 liquid chromatograph w ith anadaptor kit for SFCwas used to pum p pure supercritical fluids.

A Brownlee Labs microsyringe pump was used to pump the puremodifiers. Flow from the two pumps was dynam ically mixedinside the HP 1084.

Dyes were purchased commercially, except for Re ichardt’sE,(,),structure I, which was synthesized according to procedures in theliterature (13). Nile Red, structure 11, was purchased fromEastm an K odak, and the other dyes were purchased from Aldrich.For th e most rece nt work involving modified su percriti cal fluids,solvents were of chromatographic grade or better; no specialprecautions were taken to exclude water. Th e effect of water waschecked by adding water to the modifiers. Only very smallconcentrations of water could be added to ternary mixtures ofwater, modifier, and carbon dioxide. This added water (eitherpurposely added or absorbed from room air) wa s most likelyresponsible for small differences between the 1983and some ofthe 1988 data. SFC grade carbon dioxide in aluminum cylinders



ANALYTICAL CHEMISTRY, VOL. 62, NO. 6, MARCH 15, 1990 61 7

Table 1. Comparison of Nile Red Tra nsition Energies to Reichardt's E,(,,) nd E,(,,) in Normal Liquids"

solvent

1 pentafluorophenol (40 "C )2 formic acid3 trifluoroacetic acid4 pentafluoropropionic acid5 heptafluorobutyric acid6 2,2,3,3,3-pentafluoropropanol7 pentadecafluorooctanoic acid

8 hexafluoro-2-propanol9 m-cresol

10 phenol (45 "C)11 water12 hydrogen cyanide (15 "C)13 2,2,2-trichloroethanol14 2,2,2-trifluoroethanol15 heptafluorobutanol16 water/methanol 1:l (v/v)17 benzyl alcohol18 formamide19 ethylene glycol20 aniline21 pentadecafluorooctanol22 ethanol/water 8020 (v/v)23 acetic acid24 ethanolamine

25 ammonia (-34 "C)26 N-methylformamide27 methanol28 dimethyl sulfoxide29 quinoline30 ethanol31 I-butanol32 nitrobenzene33 1-propanol34 sulfolane35 benzonitrile36 butyrolactone37 1-octanol38 iodobenzene39 acetophenone

~ N R )

44.8545.0945.4745.4745.5045.5545.56

46.9347.4947.6548.2148.4648.8248.8649.3649.4350.2650.5150.5850.6551.0551.1651.3151.35

51.5151.9852.0252.0652.1352.1552.2152.2552.4052.4452.4452.5152.5652.6352.63

40 propionic acid 52.7141 1,3-dimethyl-2-imidazolidinone 2.71

42 nitromethane 52.7543 pyridine 52.79

wavelength.nm

637.5634.0628.8628.8628.4627.6627.5

609.2602.0600.0593.2590.0585.6585.2579.2578.4568.8566.0565.2564.5560.0558.8557.2556.8

555.0550.0549.6549.2548.4548.3547.6547.2545.6545.2545.2544.5544.0543.2543.2542.4542.4

542.0541.6

Et(30) solvent

44 N,N-dimethylformamide45 N,N-dimethylacetamide46 1-methyl-2-pyrrolidinone47 2-propanol48 carbon disulfide49 chloroform50 bromobenzene

57.2 51 2-methyl-2-propanol52 1,2-dichloroethane53 N,N-dichloroethane

55 methylene chloride63.1 54 chlorobenzene

53.9 56 ethyl acetoacetate57.2 57 phenylacetylene

58 acetonitrile59 anisol

50.8 46.5 60 fluorobenzene56.6 61 l,l,l -trifluoroacetone56.3 51.1 62 thiophene44.3 39.7 63 acetone

64 tetrahydrofuran53.6 65 benzene51.9 66 ethylene glycol dimethyl ether

67 toluene

68 ethyl acetate54.1 48.8 69 isobutylamine55.5 50.8 70 furan45.0 71 perfluorobenzene34.9 72 1,4-dioxane51.9 46.9 73 tert-butylamine50.2 45.0 74 C0,methanol 9O:lO (v/v)42.0 75 tetramethylethylenediamine50.7 45.6 76 N,N-diethylamine44.0 77 tert-butyl methyl ether42.0 78 m-xylene

79 diethyl ether80 C02/methanol 95:5 (v/v)81 carbon tetrachloride

41.3 82 triethylamine83 tributylamine84 cyclohexane

46.3 85 n-hexane40.2 37.3 86 pentane87 COz, 24 "C, 1270 psig

OTransition energy, E, of the solvatochromic band calculated by E = 28591.44/(nm) in kcal/mol.

52.8352.8752.8853.0253.1453.1853.26

53.2953.2953.3453.3453.4253.4253.5853.7853.7853.8953.9053.9053.9954.1954.4254.6454.81

54.9455.1555.1955.3555.3655.4555.6255.9355.9356.2456.4156.6856.9557.5057.8758.1658.63

59.0259.1259.54

nm

541.2540.8540.7539.2538.0537.6536.8

536.5536.5536.0536.0535.2535.2533.6531.6531.6530.5530.4530.4529.5527.6525.4523.2521.6

520.4518.4518.0516.5516.4515.6514.0511.2511.2508.4506.8504.4502.0497.2494.0491.6487.6

484.4483.6480.2

wavelength,

Et(30)

43.8 39.4

48.6 40.432.639.1 36.137.5

43.9 37.441.9

37.5 35.041.1 37.3

46.0 41.437.2

42.2 37.937.4 35.134.5 33.238.2 35.533.9 33.0

38.1 35.4

36.0 34.2

34.6

32.5

30.9

was obtained from S cott Specialty Gases, Plumsteadville, PA.Freons were obtained from Du Pont through Melchior h s t r o n g ,Edgemont, PA.Choosing Dyes. We were primarily interested in measuring

relatively subtle changes in solvent strength caused by smallchanges in solvent composition. Finding an appropriate dye orfamily of dyes became a bigger issue than expected. The dyeshad to be reasonably soluble in a wide range of pure solvents andsupercritical fluids. In addition, available instrumentation con-sisted of a spectrophotometer with limited wavelength resolution

(1nm spectral bandwidth between 200 and 400 nm, 2 nm spectralbandwidth between 400 and 800 nm). This required the use ofdyes exhibiting large solvatochromic wavelength shifts.

Kamlet and Tuft's x* Scale. The systematic approach ofKamlet and Taft (14-16) appears to offer the best means ofcomparing the solvent str engt h of modifier-supercritical fluidmixtures to normal liquids and deconvoluting the contributionsfrom dispersive forces, per man ent dipoles, acid-base, and hy-drogen bonding interactions. Unfortunately, our initial mea-surements with several of these dyes produced inadequate reso-lution of solvent strength differences and further work wasdropped.

Reichardt's Dyes. Since the system of Kamlet a nd Taft couldnot provide the desired sensitivity, other alternatives wereevaluated. Kosower's dye (34)was not soluble in a wide enoughrange of solvents. Reichardt's ( 1 3 , s )amily of pyridinium betaine

60

cc'4\T(,pia thyl athrr

57.5 t ' '.55I- '.

5 2 . 5 1

50 -

47.5-

45I

30 35 40 45 50 55 60

E t mo)

Figure 1. Comparison of transftion energies calculated from thewavelengths of maximum absorption of Nlle Red and EN301 solvato-chromic dyes in a number of common liquids. (a) is EX3,,)value cal-culated in ref 27 .



618 A NA L Y TICA L CHE MIS TRY , V O L . 62 , N O . 6, MARCH 15 , 1990

dyes including Et(30), tructure 111, has been the most popular

0

i, R = t Bu, Et(26)I li, R = Phenyl, Et(30)

solvatochromic dye due, in part, to the fact tha t i t exhibits th elargest known shi fts in absorption maxima with changes in solventpolarity. Unfortunately, considerable concem has been expressedover this dye’s susceptibility to solvent hydrogen bonding power(16,26,36) nd it produces a hypsochromic shift with increasingsolvent polarity in di rect contrast to the family of dyes suggestedby Kamlet and Taft (14-16). is not very soluble in pure C 02and tends to p recipitate from solution as modifier concentrationbecomes small. The pen ta(t-butyl) (29, 35) and Et(2s) 37) de-rivatives were purp orted to be more soluble in COz than Et(30),but we found little difference in the so lubili tiesof and Et(,in carbon dioxide. However, we did observe a major decrease in

the molar absorptivity of both and Et(30)when COz was addedto many modifiers. Bubbling COz hrough solutions of eithe r ofthese dyes in polar solvents, such as methanol, resulted in therapid loss of visible color. Sub sequen t purging of the solutionswith either helium or nitrogen resulted in the rapid return of thecolor. Th e addition of acids also caused the color to disappear,while bases made color reappear. We concluded that these dyesare very susceptible to the presence of traces of water, solutionacidity, and solvent polarity. Thi s behavior led to uncertaintyin interp reting the results . #en water was excluded, significantconcentrations of carbonic acid could not form or dissociate, andspectra with usable levels of absorption could be obtained.

Nile Red. Nile Red has found sub stantial use as a lipophylicdye for staining fatt y tissues (38).However, although its solva-tochromic behavior has been noted, it has not been examinedextensively. Th e solvatochromism and fluorescence of oxazinedyes, especially those related to N ile Blue A (39), tructure IV ,

have also been noted (4043). Nile Red was previously used with

11 , R - 0, ile Re dIV, R NH, Nile BlueA

a limited number of solvents as a solvatochromic probe of solventbasicity in polar aprotic media (42, 43). Because of it s photo-chemical stability and strong fluorescence peak (which is alsosolvatochromic), Nile Red has been suggested as a laser dye (46).

Itslow basicity (pK,

= 1.00i .05) (47)ermits observation ofits solvatochromic peak in acidic media (48). Although not ideal,

Nile Red provides the best available combination of characteristicsfor our requirements. Nile Red is soluble in solvents at theextremes of polarity of intere st. Th is dye exhibits large shif tsin the wavelength of maximum absorbance with solvent polarity(>160nm s hift between carbon dioxide and trifluoroaceticacid).Nile Red is uniquely stable in extremely acidic media and is notsusceptible to a loss of molar absorptivity in the presence of acidslike Reichardt’s dye. The absorbance maximum of Nile Red doesnot appear to shift significantly when subjected to hydrogenbonding solvents. It also produces a bathochromic wavelengthshift consistent with stabilization of the excited state in p -r* orT-T* electronic transition s and comparable with the T* scale ofKamlet and Taft (16). This contrasts with the hypsochromiocshift of Reichardt’s E,(,,,.

Nile Red is fluorescent in some solvents. In very acidic, polarsolve nts, absorption bands of the protona ted form of Nile Redappear. Fluorescence emission band ($ overlap the solvatochromicabsorption bands of the protonated form of Nile Red to createwhat appears to be a triplet in the absorption sp ectra. Thesolvatochromic absorption bands of the regular form of Nile Redare still present but are sometimes not very obvious. To avoidconfusion and to deconvolute the overlapping bands, compositionwas varied while spe ctra were continuously monitored. Com-position was varied by making mixtures of a low polarity solventand the high polarity solvent of inte rest. A t the beginning of the

experiment, single spectral bands were observed. As the exper-iment progressed and increasing concentrations of th e more polarsolute was added , the band due to the normal form of the dyewas observed to shi ft. New bands appeared b ut could be dif-ferentiated from the band of interest.

RESULTS AND DISCUSSION

Measurements in Normal Liquids. Th e wavelength ofmaximum abso rption of Nile Red was measured in a num berof common liquids used as solvents or modifiers in both LCand packed column SFC. Wavelength measurements wereconverted t o transition energies in kilocalories per mole bydividing 28592 by wavelength in nanometers (34,49). NileRed transition energies are designated E ( y ) .Th e Nile Redabsorption maxima an d transition energies in a large number

of solvents are presented in Table I. Values for Reichardt’sEt(30) and Et(26) re also presented for comparison. Neitherof the latter dyes is stable in acidic media. From these d atait appears t ha t Nile Red, because it is more stable in acidicmedia, is somewhat more useful than Et(30) r Et(zo n meas-uring th e solvent streng th of acids an d othe r very polar sol-vents. Solution s of Nile Red in neat, strong acids, such astrifluoroacetic acid, maintained nearly constant color formonths. In solutions of bases, Nile Red tends to lose colorover several days.

Some of the data in Table I are also presented graphically,in Figure 1,where values of E(m)or Nile Red and Reichardt’sEt(,,) are plotted again st each other. Th e figure shows th atEt(,,) is extremely sensitive to h ydrogen bon ding solvents, aspreviously observed by oth ers (36), ut, in contrast, tha t Nile

Red is much less sensitive to hydrogen bonding an d protondonor/acce ptor solvents. Neither scale, by itself, provides aclear picture of the relative streng ths of th e solvents. Not etha t the da ta on the left and right extremes of Figure 1allowparallel lines to be drawn with an offset occurring in the regionwhere hydrogen bonding a nd do nor/accepto r characteristicsare important.

Results in Fluids Used in SFC. Methanol Modifier inCarbon Dioxide and Freons. Values of E ( N R ) vs Et(30) formixtures of methanol in carbon dioxide, Freon-13, andFreon-23 are presented in Figure 2. Th e lower dashed lineis the normal solvent E) vs Et(3o)urve from Figure 1. Thenumb ers next to th e curves are volume percents of methanolmixed int o the various fluids.

T h e C 0 2dat a were obtained by using very dry m ethanol,

but the intensity of absorption for Et(m)was low, making i tsomew hat difficult to be certain of the wavelength of maxi-mu m absorp tion. Tw o values of Et(30) for metha nol arepresented as half sha ded circles in Figure 2. Th e less polarvalue was obtained by using methanol from a freshly openedbottle, and t he more polar value was obtained from an older,used bottle of methan ol. Th e two values represent t he ex-tremes observed over several months using methanol with nospecial preparation. Freshly opened bottles tended to produceslightly less polar readings tha n older bottles t ha t had beenopened many times. Freshly opened bottles gave th e sameresults as scrupulously dried modifier. Small concentrationsof water purposely added to t he modifier gave small spectralshifts similar to those observed w ith “old” bottles of modifier.



A N A L Y T I C A L C H E M I ST R Y , VOL. 62, NO. 6, MARCH 15, 1990 61 9

Table 11. Wavelengths of A bsorption Maxima and Calculated Transition Energies for M ixtures of Common Modifiers withCarbon Dioxide ( at 25 " C)

~ N R )

MeOH CHBCN CH2Cl2 TH F

% kcall kcall kcall kcallmodifier nm mol nm mol nm mol nm mol

1" 488 58.62" 494 57.93a 498 57 .4

4" 500 57.25 504 56.7 496 57.6 492 58.1 490 58.310 512 55.8 504 56.7 496 57.6 494 57.920 522 54.8 512 55.8 504 56.7 498 57.440 532 53.7 522 54.8 516 55.4 508 56.360 540 52.9 526 54.4 524 54.6 518 55.280 548 52.2 532 53.7 530 53.9 54.5100 554 51.6 538 53.1 538 53.1 528 54.1

Collected at 29 " C.

50 t

\'.

\

. \\

\

\

\

30 35 4 0 45 50 55 60

Ef (30)

Figure 2. Compar lso n of energ ies, as in F igure 1, but for mix tures ofmethan ol w i th three supercr i t i ca l f lu ids. Lower dash ed l ine is sameas line in Fig. 1, f o r r e fe r ence .

The results in Figure 2 are very informative. On eithersolvent strength scale (E(*) r Et(3o))he polarity of the purefluids follows the order Freon-13 < C 0 2 < Freon-23. Un-published chromatographic results have demonstrated to usth at these p ure fluids follow the sa me order in their elutionstrength of many solutes from a wide range of stationaryphases. Th e addition of methanol as modifier appears to shiftthe solvent strength of the mixtures dramatically. Th e larger

th e difference in solvent strength between the two pure fluids(supercritical fluid and modifier), the more dramatically thesolvent streng th of the mixture shifts. Th us, methanol-Freon-13 mixtures appear more polar than methanol-COzmixtures of similar composition (on either solvent stren gthscale). This can be explained by using the concept of clus-tering (17-28). Polar m olecules are only sparingly soluble inFreon-13 and tend t o be rejected by it. At low concentrationsthey do not, however, form a second phase. Instea d, polarmolecules tend to cluster together forming a local environm enttha t is much more polar than the bulk composition wouldsuggest. Th e greater the difference in polarity between thecomponents, the more intense the clustering and the morenonlinear the relationship between solvent strength andcomposition.

The solvent strength measurements summ arized in Figure2 do not precisely parallel the elution strengths of thesemixtures in chromatography. Mixtures of meth anol inFreon-13 are always less able to elute polar solutes thanmixtures of methanol in carbon dioxide at the same concen-

trat ion of modifier. We believe this is due to intense inter-actions between the modifier and the relatively polar sta-tionary phases we usually use in packed column SFC (i.e.cyanopropyl or diol bonded phases on silica). If the super -critical fluid essentially rejects the m odifier, causing clusteri ngin the mobile phase, then i t is probable t ha t the re will alsobe a strong association between the polar modifier and thepolar station ary phase. Pooling of the modifier on the sta-tionary phase causes a change in the polarity and/ or volumeof the stationary phase.

None of the dy es is soluble in pure Freon-13, so the positionof the unmodified fluid is not shown in Figure 2. We couldnot produce mixtures of over 12% m ethanol in Freon-13 attemperatures between room tempe rature (25 "C) and 80 "Cand over a wide range of pressures. Wi th greater tha n 12%methanol adde d to Freon-13, the base-line noise increaseddramatically, suggesting the fluid broke down into two phases.

2-Propanol in Hexane. Mixtures of 2-propanol in hexanewere also evaluated with Nile Red to determine whether thebehavior of methanol in nonpo lar supe rcritical fluids is per-ceptibly different in comparison tomix tures of norma l liquids.Hexane is very nonpolar while 2-propanol is moderately polar.Hexane and 2-propanol were chosen because they have oneof the greatest differences in polarity possible among binarypairs of miscible, normal liquids. Th e results, show n in Figure3, indicate similar but less dramatic behavior compared to th ebehavior observed for mixtures in supercritical fluids (seeFigure 2). Th e similarity between the curves in Figures 2 and3gave some reassurance that the e ffed s observed were realistic

mea sure men ts of solvent strength. If plots like Figures 2 or3 are linear, local composition and bulk composition areprobably t he same. Th e relatively more linear behavior of2-propanol in hexane, compared to alcohols in carbon d ioxidesuggests th at th ere is less clustering of the 2-propanol. Th edifference in behavior between the mix tures containing hexaneand carbon dioxide can be attributed to the more limitedcompress ibility of hexane comp ared to carbon dioxide (1419).

Other Modifiers i n Carbon Dioxide. The N ile Red tran-sition energies of mixtu res of m ethanol, aceton itrile, methylen echloride, and tetrah ydrofu ran in carbon dioxide are listed inTable I1and shown in Figure 4. In Figure 4, the vertical lineson the extreme left and right represent th e solvent strengthsof pure carbon dioxide and pure methanol, respectively.Horizontal lines connect these two extremes. Th e top three



620 ANALYT I CAL CHEMISTRY, VOL. 62 , NO. 6, MARCH 15, 1990

-

-

-<- _

-

60 I " " " '56.62

57.18

E ( ~ ~ )

57 .76

58.35

\

\

\

\

\

\

I? 20 4060 807'HF

5 13 20 40 60 80

CH3CN

lp 20 5080

5 IO 20 40 60 80I I / I ,

I 2 3 4 5 IO 20 40 60 80

I I I / I I I I I I .

4 5 I

30 35 40 45 50 55 60

E t(30)

Figure 3. Comparison as in Figures 1 and 2 but for mixtures of twonormal liquids, 2-propanol and hexane.

FREON- 23

25

I

2o MeOH

I 1

60 5 9 58 57 5 6 55 54 53 52 51

E(NR) , K C A L / MOLE

Figure 4. Transition energies obtained by using Nile R ed for mixturesof carbon dioxide and common modifiers.

horizontal lines are broken by intermediate vertical linesrepresenting the solvent strength of pure Freon-23, tetra-hydrofuran (TH F), and acetonitr i le (ACN). The num bersabove the tick mark s on th e horizontal lines are the concen-tration of th e more polar modifier in the less polar solvent,in volume percent. Thu s, on the bottom line the number 10appea rs about half way between the vertical lines represe ntingpure carbon dioxide and pure methanol. Thismeans that 10%methanol in carbon dioxide has a solvent strength one mightnormally expect from 50% methan ol in carbon dioxide (as-suming ideal solutions). Apparently, the dye molecules aresurrounded by a solvation sph ere th at ha s a higher concen-tration of -modifier tha n is present in th e bulk.

Th e da ta are presented in Figure 4 in a form at similar tothe format used to express elution strength (cab) informationin adsorption chromatography (50).However, in Figure 4, henonlinearity is a mobile phase phenomenon involving thesolute and the modifier whereas in adsorption chromatographythe nonlinearity is due to a mobile phase-stationary phaseinteraction.

56 .06i,0I+?-- --- - $

' 5%

4 9 0

4 58.95

25 30 35 40 4 5 5 0

T E M P , 'C

Figure 5 . Effect of temperature on absorba nce maxima of Nile R edin mixtures of carbon dioxide and methand. The numbers next to thecurves represent the volume percent of methanol in carbon dioxide.

Effects of Temperature and Pressure. One operatingmode in SF C involves density program ming where partition

ratio's ( k ? can be changed by a factor of at least lo3 imes (i.e.51). Such large variations in k', achievable with changes indensity, have led many people to equate large changes indensity with large changes in solvent strength. Suc h obser-vations need to be put in perspective. Much larger changesin k ' re achievable in LC by changing fluid composition. Inadsorption chrom atography, a change in solvent from hexaneto methanol (through m ultistep subs titution s) can result ina change in solute retention by asmuch as lo 9 or 1Olo times.Thu s, the apparen tly large changes in k'in SFC , achievableusing only changes in the density of common supercriticalfluids, re small compared to changes ach ievable in LC. Theydo not necessarily represent large changes in solvent streng thand can only cover a small fraction of the full range of solven tstreng ths available with normal liquids. Experiments were

undertaken to determine the range of solvent strengthsachievable with density variations compared to compositionchanges using Nile Red tra nsitio n energies.

Temperature. Several autho rs have shown (52,531 ha t ,for a pure f luid at constant density, there are no abruptchanges in chromatographic retenti on as tempe rature is in-creased from sub- to supercritical. Hya tt (29)reported similarmobile phase solvent strengths at sub- and supercriticaltemperatures of a number of Kamlet and Taft (14-16) dyes.We also observed only modest shifts in th e wavelength of NileRed absorption with changing temperature. These wavelengthchanges are similar in ma gnitude to shifts caused by smallchanges in fluid composition as shown in Figure 5. It is alsointeresting to observe th at solvent streng th decreases whenthe fluid temp erature is changed from subcritical to super-critical.

Th e density vs temp erature an d pressure relationships ofmethano l-carbon dioxide mixtures were not knownwhen thesemeasu rements were taken. However, subsequent measure-ments of density using a "U" tube densitometer shows thatan increase in temperature from 24 to 50 "C (at 130 bar)decreases density from 0.87 to 0.69 g/cm3. This esult suggeststha t negative temperature programming of constant compo-sition binary mixtures should produce increasing solventstrength which agrees with gen eral experience in capillary SFC.

Pressure. Very little change in solvent streng th was ob-served when pressure was increased from 76 ba r (1100 si )to 241 bar (3500 psi) at tempe ratures from 25 to 40 OC. Thewavelength of maximum absorbance changes no more than



ANALYTICAL CHEMISTRY, VOL. 62, NO. 6, MARCH 15, 1990 821

been discussed (49). Binary m ixtures of a supercritical fluidand a polar modifier appear to allow much wider variationin mixture solvent strength through composition adjustmenttha n is possible by using binary pairs of normal liquids. Suchwide variations allow nonlinearities between solvent strengthand composition to be much more apparent. Solvatochromicshift measurem ents using Nile Red indicate a strong associ-ation between the dye and polar modifiers. Such “association”,or “local compo sition”,or “clustering” means that polar solutesare surrounded by a solvation sphere containing a substantially

higher modifier concentration, and experience a subsequ entlyhigher polarity, tha n expected from bulk composition. Suchmobile phase effects must have an mpact on chromatographicretention, but th e exact nature of such effects has not yet beendetermin ed. If packed column SFC follows a mechanismcharacteristic of BPC, the equation for solvent polarity, P’,mu st be modified to account for the difference between bulkand local composition. If adsorption of the mobile phasesolvent onto th e stationary phase is important in retention,the equations for adsorption chromatography must be ex-tended to quan tify specific solute-solvent effects in the mo bilephase. Th ere is a reasonable likelihood th at some aspects ofboth BPC and adsorption chromatography impact retentionin normal phase, packed column S FC.

In supercritica l fluids containing p olar modifiers, the pri-

mary mobile phase interactions experienced by polar solutesshould be with clusters of the modifier. Syst em densityprimarily determines the na ture of the modifier cluster andthu s has only a secondary effect on retention adjus tmen t (alimited range of solvent strength adjustm ent). Alternately,density program ming is likely to produce very subtle changesin solvent strength and will likely be most useful in differ-entiating very similar molecules.

ACKNOWLEDGMENT

We than k Professor Keith Joh nsto n of the Chemical En-gineering Department of the University of Texas and Dr.Clement Yonker of Battelle Northwest Lab s for helpful dis-cussions. We also wish to than k the Du Po nt Company forallowing us to publish th e Nile Red m easuremen ts in Table

I.LITERATURE CITED

(1) Blllie, A. L.; Grelbrokk, T. Anal. Chem. 1985, 57, 2239.(2) Gere, D. R. Hewlett-Packard Applkatlon Note AN800-2, 1983.(3 ) Taylor, L. T. Atlas of Chromatograms.J. Chromatogr. Scl . 1987, 25 ,

4-5.(4) Yonker, C. R.; Smith, R. D. J. Chromatogr. 1986, 367, 25-32.(5) Snyder, L. R. J. Chromatogr. 1974, 92 , 223-230.(6)Rohrschnelder,L. Anal. Chem. 1973, 45. 1241.(7) Snyder, L . R.; Kkkland, J. J. In tmdwtkm to MDdem UquM Chrometog-

raphy, 2nded.; ohn Wiley and Sons: New York, 1979; Chapter 7.3.(8) Snyder, L. R. princ@ks of Adsorption Chromatography; Marcel Dek-

ker: New York, 1968; Chapter 8.(9) Mourier, P.; Sassiat, P.; Caude, M.; Rosset, R. J. Chromatogr. 1986,

(10) Smith, P. L .; Cooper. W. T. Chromtographia 1988, 25(1), 55-60.(11) Sjostrom, M.; Wold, S. Acta Chem. Scand. 1981, 835. 537-554.(12) Kamlet. M. J.; Taft, R. W. Acta Chem. Scand. 1985, 835 , 611-628.(13) Dlmroth. K.; Reichardt, C.; Slepmann, T.; Bohimann, F. Justus LWigs

Ann. Chem. 1963, 667, 1.(14) Kamlet, M. J.; Taft. R. W. J. Am. Chem. SOC.1976, 98 , 377-383.(15) Taft, R . W.; Kamlet, M. J. J. Am . Chem. Soc. 1978, 98, 2886-2894.(16) Kamlet. M. J.; Abboud, J. L.; Tan, R. W. J . A m . C h em . SOC.1977.

(17) Frye, S. L.; Yonker, C. R.; Kalkwarf, D. R.; Smith, R. D. I n Supercrlt i-cal Fluids: Chemical and Engineering Applications; Squires, T. G.,Paulaltis, M. E.,Eds.;American Chemical Soclety: Washington, DC,1987; Chapter 3.

(18) Kim, S.; Johnston, K. P. A I C h E J . 1987, 33(10), 1603-1611.(19) Johnston, K. P., prhrate communication.(20) Johnston, K. P.; Kim, S.; Wong. J. M. Fluid Phase Equlllb. 1887, 36

(21) Deiters, U. K. Fluid Phase Equilib. 1982, 8 , 123-129; 1983, 73,

(22) Whltlng, W. 6.; Prausnttz, J. M. Fluidph ase Equillb. 1962, 9. 119-147.(23) Mathlas. P. M.; Copeman, T. W. Fluidphas e EquHlb. 1983, 7391-108.(24) Hu, Y.; Azevedo, E.G.; Prauznitz, J. M. FluidPhase Equillb. 1883, 13,

(25) Nitsche, K.-S.; Suppan, P. Chimia 1982, 36(9), 346-348.

353, 61-75.

99, 6027-6036.

39-82.

109- 120.

351-360.

1.2 -

1.0 -

0.8.8 -lo g h’

0.6 -

A B

X METHANOL

1.4 1.4 -

0 0 2 5 5 7 5 IO-6 METHANOL 59 56 57 66 55

E(NR), KCAL/MOLE

Flgure 8. Preliminary correlation of retention with b,.etention ofthe phenykhiohydantoin derivatives of proline (diamonds), isoleucine(squares), and a-am inobuty rlc acid (circles ) on a cyanopropyl columnas a function of (a) methanol modifier concentration and (b) solventstrength measured by using Nile Red.

2-4 nm, corresponding to a change in E ( N R ) of less tha n 0.5.

Th e spectrophotometer resolution (2 nm) is on the same orderof magn itude as the observed spectral shifts. Such findingsare consistent with the observation that density changes

produce much smaller changes in retention in SFC thancomposition changes produce in LC.

Results similar to those reported for pure carbon dioxide,were also observed with mixtures of modifiers in carbon di-oxide. Th e wavelength of maximu m absor bance was primarilydetermined by th e amou nt of modifier added an d was rela-tively insensitive to the pressure or density of the fluid, at leastat high densities (>0.7g/cm3). This is consistent with previousobservations (18) hat the concentration of modifier insideclusters increases as the density decreases. The intensity ofinteraction between solutes and m odifier in the clusters ac-tually increases, but the number of clusters decreases sub-stantially. Th is leads to a drop in solubility of polar solutesin the mixture (27). Clearly, solubility of such solutes iscontrolled by the identity and concentration of the modifier

and the natu re of the clustering phenomenon. Th e minimaleffect of pressure on the apparent solvent strength of themixtures suggests that chromatographic retention of polarsolutes will likely be more a function of compositon thandensity when p olar modifiers are added to supercritical fluids.

Preliminary Correlation with Chromatography. Theseparation of p henylthiohydantoin (P TH ) derivatives of aminoacids have been extensively studied in our laboratory (54) asa function of methano l concentration in carbon dioxide. Plotsof log k’vs percent methanol are nonlinear, as shown in Figure6a for three PT H am ino acids. Plots of the same retentiondata , but substi tuting E( NR ) for methanol concentration,produce substantially improved linearity, as shown in Figure6b. Residual nonlinearities in Figure 6b suggest th at oth erimp ortant interactions, likely related to modifier-stationaryphase interactions, are not being accounted for when log k’is plotted against E(NR).

CONCLUSIONS

The retention mechanism in many packed column SFCseparations is still a matter of debate. The clustering of polarmodifier molecules has been shown to produce local regionsof high solvent stren gth . Suc h clustering produces verynonlinear solven t strengt h vs composition profiies and suggestsa poasible alternativ e to or an enhancement of previous modelsof retention. Th e possibility th at solvent mixtures producenonlinear changes in solute-solvent interactions vs compo-sition has not been widely considered in chromatograph y eventhough discrepancies between retention a nd composition have



622 Anal. Chem. 1990, 62, 22-625

(26) Figueras, J. J . A m . Chem. Soc. 1871, 93, 3255-3263.(27) Dobbs, J. M. ; Wong, J. M ., Lahlere, R. J.; Johnston, K. P. Ind. Eng.

Chem. Res. 1887. 26, 56-65.(28) Walsh, J. M .; Ikonomou. G. D.; Donohue, M . D. Fluid Fhase Eguilib.

(29) Hyatt, J. A. J . Org. Chem. 1984, 49 , 5097-5101.(30) Slgman, M . E.; Lindby, S. M. ; Leffler, J. E. J . A m . C h em . S o c . 1985.

(31) Yonker, C. R.; Frye, S. L.; Kalkwarf, D. R.; Smith, R. D. J . M y s .Chem. 1986. 90 , 3022-3026.

(32) Bente, Paul: Weaver, Harry. HewlettPackard, private communication.(33) Deye, J. F.; Anderson, A. G., Central Research and Development De-

partment. E. I. du Pont de Nemows, unpublished work.(34) K osow er, E. M. J . A m . C h em . Soc. 1958. 6 0 , 3253, 3261, 3267.

(35) Reichardt, C.; Harbusch-Gornert, E. Justus Liebigs Ann. Chem. 1983,721.(36) Fowler, F.W.; Katritzky, A . R.; Rutherford, R. J. D. J . Cbem. SOC.8

1971, 460.(37) Andersen, A. G., unpublished work.(38) H.J. Conn’s Biob@C~?/ talns, 9th ed.;Lillie, R. D., Ed.; The Williams

and Wlkins Co.: Baltimore, MD, 1977; pp 372-430.(39) Relssig. T. D.R.P. 45268, 1888.(40) Mohlau. R.; Uhlmann, K. Justus LieMg.9 Ann. Chem. 1898, 289, 90 .(41) Kehrmann, F. 6 e r . DTsch. Chem. Ges. 804, 37 , 3581.

1987, 33, 295-314.

107, 1471-1472.

(42) Thorpe, J. F. J . Chem. Soc . 1907, 91 , 324.(43) Kehrman, F.; Herzbaum, A. Ber. Dtsch. Chem. Qes. 1917, 50 , 873.(44) Kollina. 0.W. Anal. Chem. 1976. 48. 884.(45j Kolling; 0. . Anal. mtn.1968; 38; 424.(46) Bastlng, D.; Ouw, D.; Schafer, F. P. Opt . Commun. 1876, 18, 260.(47) Stuzka, V .; Simanek, V. Collect. Czech. Chem . Common. 1973, 38 ,

194.(48) Davis, M. M.; Hetzer, H. . Anal. Chem. 1966, 38 , 451.(49) Dorsey, J. G. Chromatugraphy 1987, 2(5), 34-41.(50) Saunders, D. L. Anal. Chem. 1974, 46, 472.(51) van Wassen, U.; Swaid, I.; Schneider, G. M . Angew. Chem., In t . Ed.

Engl. 1980, 19, 575-587.(52) Lauer, H. H.; Mcmanigill, D.; Board, R. D. Anal. Chem. 1983, 5 5 ,

1370.

(53) Doran, T. Gas Chromatugraphy 1972; Applied Science Publishers:Essex. 1973; pp 133-143.(54) Berger, T. A.; Deye, J. F.;Ashraf-Korassani, M. ; Taylor. L. T. J. Cbro-

matogf . Sci . 1989. 27 , 105.

RECEIVEDor review Jun e 27,1989. Accepted December 15,1989. Th is document is contribution num ber 4980 from DuPont’s Central Research Department.

Extraction of Tervalent Lanthanides as Hydroxide Complexes

with Tri-n -octylphosphine Oxide

Ted Cecconie and Henry Freiser*

Strategic M etals Recovery Research Facili ty, Departm ent of Chemistry, University of Arizona, Tucson, Arizona 85721

The extraction behavior of lanthanum( I II ) , praseodymium-( I I I ) , europkrm( I I I ) , te rb ium(I I I ) , hdmium(I I I ) , and yt -terbium(I I I ) from dilute chloride solutions into chloroformsolutions contalnlng trkn-oc tylpho sph lne oxide (TOPO) wasexam ined. The lanthanid es were found to extract as hy-droxide complexes of the form Ln(OH),.nTOPO. The ex-traction selectivity for thls system increases with increasingTOPO conc entration, rlvallng those for commo n acidic org a-nophosphorus extractants In the TOPO c onc entration range0.025-0.100 M and surpassing them at higher concentratkns.When TOPO concentrations of 0.25 M or greater are em-ployed, the extraction system is, overall, the most selectivefo r lanthanlde separations to date. Thls Is the first report of

the lanthanldes extracting as hydroxide co mplexes.

INTRODUCTION

The use of monodentate neutral phosphorus compoundsin metal ion extraction iswidely known (I). These compoundsgenerally function asadducta nts, auxiliary complexing agents,which enhance th e extractab ility of coordinatively unsatu ratedchelates by replacing water m olecules remaining coordinatedto the complexed metal ion. Th e report of extraction oflanthanides and actinides by sulfur ligands in the presenceof tri-n-octylphosphine oxide (TOPO ) ( 2 , 3 ) ed us o examinesystems such asquinolin-%thiol a nd dithizone in the presenceof T OP O for selected Ln3+. On observing the surprising ex tentof extraction in the absence of the S-containing ligand, wedecided to examine this unusual phenomenon further. Ourhypothesis of the formation and extraction of a lanthanidehydroxide complex was sounexpected th at we took particularcare to examine and reject alternate explanations.

EXPERIMENTAL SECTION

Apparatus. Infrared spectra were obtained with a Perkin-Elmer 1800Fourier transform infrared spectrometer. 31PNM Rspectra were obtained at 250 MHz with a Bruker WM-250spectrometer. Phosphorus and lanthan ide concentrations were

determined with a Perkin-Elmer 6500 ICP spectrometer.Reagents. Stock solutions containing lanthanum(III), pra-

seodymium(III), europium(III), terbiu m(III), holmium(III), andytterbium(II1) were prepared from chloride salts (Alfa Inorganics,99.9% ). Chloroform (AR grade) was washed 3 times withdeionized water prior to use.

TOPO (courtesy of A merican Cyanamid, >9 9% ) was purifiedto remove all traces of acidic impurities, such as di(n-octy1)-phosphoric acid, which are found in commercial sources of TOPO.This pu rification involved equilibrating a concentrated (0.5 M)TOPO solution in chloroform 8 times with an equal volume ofa 0.1M acetate so lution, pH = 5 , to remove the acidic impurities(4 ) . The phosphorus level of all except the first of the eight washeswas below the inductively coupled plasma limit of detection forP ( 5 X lo4 M). Th e acetate-washed solution was equilibrated12 times with a n equal volume of deionized water to remove any

acetic acid extracted by TOPO ( 5 , 6 ) .The final TOPO solutionwas placed in a Rotovap to remove the chloroform solvent, withthe resulting oily material placed in a vacuum desiccator tocrystallize overnight: mp, 50-52 “C; IR (KBr),qll& (P=O);31PNMR (CDCI,), b 48.7 ppm.

Distribution Studies. An unbuffered solution (p = 0.1,HCl/NaCl) containing the lanthanides a t - 0”’ M was equili-brated w ith an equal volume of a TOP O solution in chloroformfor at least 15 min, a time found ade quate for equilibrium to bereached. Th e equilibrium pH of the aqueous phase was measuredafter phase separation. Metal distribution results were obtainedby inductively coupled plasma atomic emission spectrometry(ICP-AES) nalysis. The lanthanide concentration remaining inthe aqueous phase was determ ined directly, employing manganeseas an intemal standard . T he concentration of lanthanide extractedinto the o rganic phase was determined by back extraction into

0003-2700/90/0362-0622$02.50/00 990 Am erican Chemical Society

Nile Red Jerry

Documents