Structure, Dynamics and Association of Thermosensitive Core-Shell Particles

Thermosensitive Core-Shell Partiles

DISSERTATION

zur Erlangung des akademishen Grades ein es

Doktors der Naturwissenshaften

- Dr. rer. nat. -

der Fakultät Biologie, Chemie und Geowissenshaften

der Universität Bayreuth

vorgelegt von

Jérme Crassous

geboren in Toulouse / Frankreih

Bayreuth, 2009

Prof. Dr. Matthias Ballau angefertigt.

VollständigerAbdruk der vonder Fakultät fürBiologie,Chemieund Geowissenshaften

der Universität Bayreuth zur Erlangung des akademishen Grades Eines doktors der

Naturwissenshaften genehmigtenDissertation.

Dissertationeingereihtam: 04.05.2009

Zulassung durhdie Promotionskommission: 06.05.2009

Wissenshaftlihes Kolloquium: 20.07.2009

Amtierender Dekan: Prof. Dr. Axel H.E. Müller

Prüfungsausshuss:

Prof. Dr. Matthias Ballau (Erstgutahter)

Prof. Dr. Nuri Aksel (Zweitgutahter)

Prof. Dr. Andreas Fery

Prof. Dr. Jürgen Senker ( Vorsitzender)

estdélivréde larainteet du

respet,vousle voyez selever,

dessinerl'idéeà grandsgestes,et

soudainrire de toutsonoeur,

ommeau plusbeau desjeux.

(Emile-AugusteChartier,dit

Alain)

A mes parents,

Withthe supportof the following people I ould nishthis thesis suessfully and would

liketo express my gratitude to allof them.

It's my great pleasure to thank Prof. Matthias Ballau for giving the opportunity and

freedom to do researh in this interesting eld under his supervision. His guidane and

support were a great help to always improve the quality of my work. Thanks to him I

realize whatit is tobe and behave asa researher.

IthankProf. NuriAkselandDr. LutzHeymannforthefruitfulommentsanddisussions

onerning my work and for sharingtheir experienein the eld of the rheology.

I thank Prof. Matthias Fuhs and his oworkers Oliver Henrih and David Hajnal for

the very fruitful ollaboration in the eld of the dynami and the development of the

theoritial model biased on the mode oupling theory. It was a great pleasure to work

with suh outstanding sientists. I would alsolike to thank Miriam Siebenbüger for her

tehnial support and for the measurements she performed duringher Diplomarbeit.

I amalsothankfultoDr. AlexanderWittemann andJohenRingforthe synthesisof the

system I used during my Phd. I also want to thank Dr. Alexander Wittemann for his

onstrutive suggestions and ollaborationin the eld of the olloidal assoiation.

Dr. Markus Drehsler,Prof. IshiTalmonandMarShrinneraregratefullyaknowledged

fortheir invaluableontributionandtireless enumerableattempts toget TEM andCryo-

TEM imagesas wellas BenjaminGöÿler for the SEM pitures.

I thank Prof. Norbert Willenbaher and Dr. Raphael Régisser for the ommonwork on

thepiezoeletriaxialvibrator,andinthesamesensethegroupofDr. WolfgangPehhold,

Theresia Groÿ and Dr. Ludwig Kirshenmann for the oneptionof this instrument.

Florian Shwaiger and prof. Werner Köhler are aknowledged for the experiments their

performed and their ontribution in the laser ontrolled miro-aggregation presented

herein.

I would liketo thank Dr. John Boso, Dr. Antonis Keralakis for the fruitful disussions

and ollaborations. I ampartiularly thankful to Pierre Millardfor the number of ideas

we developed together and for hisexpertise inthe polymer hemistry, to AdrianaMihut

for the onstant support and the time she spent to provide AFM pitures from various

of my projets, and to Christophe Rohette and to Sergiu Mihut for developing of the

programs for the SAXS and CryoTEM analysis.

IthankElisabethDüngfelderforherbureauratiworkwithalotofpatieneandkindness,

and Karlheinz Lautenbah for hisavailabilityand tehnial support.

Last butnot the least;Iexpress my gratitudetomyfamily members and friendsfortheir

strong supports, and simplyfor havingalways believed in me.

1 Introdution 1

2 Charaterization 4

2.1 Inuene ofthe degreeofrosslinkingonthe strutureand swellingbehav-

iorof thermosensitive ore-shell olloidallatexes. . . 4

2.1.1 Introdution . . . 4

2.1.2 Experimental . . . 5

2.1.3 Theoretialbakground. . . 9

2.1.4 Results and disussion . . . 11

2.1.5 Summary . . . 21

2.2 Quantitativeanalysis of polymer olloidsby normaland ryo-transmission eletron mirosopy.. . . 22

2.2.1 Introdution . . . 22

2.2.2 Experimental . . . 23

2.2.3 Theory . . . 24

2.2.4 Results and Disussion . . . 32

2.2.5 Summary . . . 41

2.3 Crystallization. . . 42

2.3.1 Introdution . . . 42

2.3.2 Experimental . . . 43

2.3.3 Eetive volume fration and rystallization . . . 43

2.3.4 Linear visoelasti behavior . . . 48

2.3.5 Flowurves and shear melting . . . 51

2.3.6 Summary . . . 52

3 Dynamis 53 3.1 Charaterization of the visoelasti behavior of omplex uids using the piezoeletri axialvibrator . . . 53

3.1.1 Introdution . . . 53

3.1.2 Theory . . . 54

3.1.3 Instruments . . . 56

3.1.4 Calibrationof the instrument and auray. . . 57

3.1.5 Visoelasti uids . . . 58

3.1.6 Summary . . . 64

3.2 Shearstresses ofolloidaldispersionsat theglass transitioninequilibrium and inow. . . 65

3.2.1 Introdution . . . 65

3.2.2 Experimentalsystem and methods . . . 66

3.2.3 Linear and non linear rheology. . . 67

3.2.4 Theory . . . 70

3.2.5 Comparisonof theory and experiment. . . 82

3.2.6 Summary . . . 96

4 Assoiation 97 4.1 Reversible self-assembly of omposite mirogels. . . 97

4.1.1 Introdution . . . 97

4.1.2 Coagulationkinetis . . . 98

4.1.3 Experimental . . . 100

4.1.4 Results and Disussion . . . 101

4.1.5 Summary . . . 110

4.2 Eletrostati Dipole Formationby Assoiation between Composite Miro- gels and Gold Nanopartiles . . . 112

4.2.1 Introdution . . . 112

4.2.2 Experimental . . . 113

4.2.3 Results and Disussion . . . 114

4.2.4 Summary . . . 121

5 Synopsis 123

6 Zusammenfassung 126

7 Abbreviations 129

8 Publiations 143

9 Erklärung 145

Gels omposed of ross-linked poly(

N

-isopropylarylamide) (PNIPAM) hains an un- dergo a phase transition as funtion of temperature in whih the network shrinks in a

ontinuous ordisontinuous fashion. The volume transitionin marosopinetworks has

been studiedextensively by T.Tanaka andothers [1,2℄. Areview onwork doneonthese

marosopisystemswasgivensometimeagoby Shibayama[3℄. Startingwithearlywork

byTanaka[4℄, manygroupshavedevelopedsynthesesofolloidalthermosensitivenetwork

by e.g. emulsionpolymerization. Two typesof partilesanbeprepared: Eitherthe par-

tilesonsisttotallyofaPNIPAM-network[512℄orthePNIPAM-networkispolymerized

onto a solid ore [1320℄. A great number of possible appliations have been disussed

for these systems that inlude widely separate elds ase. g. proteinadsorption [19, 20℄.

They havealsobeen reently usedas atemplatefor the redutionof metal nanopartiles

[2123℄for appliations inatalysis [22, 23℄. A omprehensive review onthe appliations

was given by Lyon[24℄.

They present a versatile phase behavior: on one hand they an behave like hard sphere

with aliquid-rystaltransitionbelowtheritialtemperature[7,9,25,26℄. On theother

hand in the absene of eletrostati stabilization or in saturated salt onentrations the

partiles beome attrative after the low ritial solution temperature. This leads to

a partially or totally reversible aggregation of the system in the dilute regime and to

the gelation of the system for higher onentrations [2730℄. The ne tuning of their

interpartiularpotentialan alsobeused forthe assoiationwith otherpartiles[31,32℄.

Reently,thethermosensitiveore-shellpartileshaveattratedrenewedinterestasmodel

olloids,inpartiular for aomprehensive study of the struture, dynamis,and owbe-

havior of onentrated suspensions [3341℄. Fig. 1.1 displays the overall struture and

the volume transition of these partiles in a shemati fashion: Immersed in water the

PNIPAM-shell of the partiles willswell if the temperature is low. However, raising the

temperatureinthesystem beyond32

o

Cleadstoavolumetransitioninwhihthenetwork

intheshell shrinksby expellingwater. Thus, theeetivevolumefration

φ ef f

^determin-

ing the hydrodynami volume of the partiles an be adjusted through the temperature

in the system. Hene, dense suspensions an be ahieved out of a rather dilute state by

lowering the temperature.

Sen et al. were the rst to present investigations of the rheology of suh ore-shell

partiles [34℄. The advantages of these thermosensitive partiles over the lassial hard

sphere partilesused in previous investigations [42, 43℄ of the ow behaviorare athand:

Thedensesuspensionisgeneratedinsitu thusavoidingshearandmehanialdeformation

during preparationand lling into a rheometridevie. Also, allprevious history aused

by shearing the suspension an simply be erased by raising the temperature and thus

lowering thevolumefration again. The highvolumefrationan then beadjustedagain

andapristinesamplebeinginfullequilibriumatalllengthsales anbegenerated. Sen

et al [34℄. showed that these "reversibly inatable spheres" an be used to study the

Figure1.1: Shemati representation of the volume transition in ore-shell mironetworks: The

polymer hains are axed to the surfae of the ore.

dependene of the visosity

η

^{on the shear rate}

γ ˙

^{. If the eetive volume fration of}

the partiles is not too high, a rst Newtonian region is observed if the shear rate

γ ˙

^is

small. Here the visosity

η 0

^{of the suspension measured in this rst Newtonian regime}

an be signiantly largerthan

η s

^{the one of the pure solvent. At highershear rates, the}

perturbationofthemirostrutureofthesuspensionbytheadvetiveforesannolonger

be restored by the Brownian motion of the partiles. Hene, signiant shear thinning

will result in whih the redued visosity

η/η s

^{is more and more lowered until one may}

speulate that aseond Newtonian region is reahed.

Reently, Sen's data [34℄ have been used to hek the preditions of mode-oupling

theory (MCT) [44, 45℄ for the ow behavior of onentrated suspensions [39, 40℄. Good

agreementwasreahedinthisomparisonemployingshematiMCT models[39℄. Hene,

this omparison suggests that the thermosensitive partilesshown in Fig. 1.1 present an

exellent model system for the study of the dynamis of suspensions in the viinity of

the glass transition. However, no fully quantitative omparison of theory [44, 45℄ and

experiment inluding a disussion of the t parameters ould not have been done before

this work.

This work is dediated to the study of this omposite partiles. The rst part desribes

the haraterization of the partiles. Here we present the rst study of thermosensitive

ore-shell partilesandtheir volumetransition(f. Figure1.1) by ryogenitransmission

eletronmirosopy(ryo-TEM).Thedependeneonthedegreeofrosslinkingandonthe

temperaturehas been rst investigated. A new method was developed toquantitatively

analyzed the TEM and CryoTEM images in order to aess to the internal struture

of olloids. The dierent observation are ompared to data obtained by dynami light

sattering and small-angle X-ray sattering. The last part of the hapter fous on the

olloidalrystallization of the partiles.

Theseondpartexploresthedynamisofthissystemintheviinityoftheglasstransition.

For this purpose a new rheologial set up ispresented in the rst setion to measure the

linearvisoelastiityofomplexuids onabrightfrequeny range. In theseondhapter

of this part, an interpretation of the dynamis of the olloidal ore-shell dispersion in

equilibriumand in owbiased onthe mode ouplingtheory is developed.

The lastpart of the thesis investigates the eld ofthe assoiation. Firstthe temperature

ontrolled self-assoiation of the system is put under srutiny. Then the assoiation of

2.1 Inuene of the degree of rosslinking on the

struture and swelling behavior of thermosensitive

ore-shell olloidal latexes.

2.1.1 Introdution

Environmentallysensitivemirogelshaveattratedonsiderableinterestduetotheirabil-

ity to swell and de-swell in response to external stimuli suh as temperature, pH or

light radiation [4648℄. A great number of possible appliations have been disussed for

these systems. A omprehensive review on the appliations was given by Nayak and

Lyon [24℄. Mirogels of poly(

N

-isopropylarylamide) (PNIPAM) rosslinked by

N, N ^′

^-

methylenebisarylamide (BIS) have been of partiular interest. The temperature of the

volumetransitionis loatedat 32

o

Cin aqueoussolution whih makesthem suitablean-

didatesforpossiblemedialappliationssuhasontrolledreleasesystems[19,20℄. Other

appliations inlude e.g. the use of suh systems as arriersfor metallinanopartiles in

atalysis [22, 23℄.

ThevolumetransitioninmarosopinetworkshasbeenstudiedextensivelybyT.Tanaka

andothers [14℄. A thermodynamianalysisof the transitionan bedoneintermsof the

lassial Flory-Rehner theory [4953℄. Hene, the volume transtition inmarosopi gels

seems to be rather well understood. For details the reader is deferred to the review of

Shibayama [3℄. Mirogels with dimensions inthe olloidaldomain have been the subjet

of a large number of experimental studies inreent years. The investigations range from

measurements of the marosopi properties, suh as turbidity [54, 55℄, high sensitive

sanning miroalorimetry [5557℄, rheology [7, 34℄, to experiments probing moleular

interations suh as nulear magneti resonane [56, 58℄, light sattering [7, 13, 34, 49,

55, 56, 58, 59℄, small-angle X-ray and neutron sattering [17, 49, 5964℄. Compared to

marosopigels,thedegreeofunderstandingofmirogelsismuhlessadvaned,however.

This hapter is devoted to a omprehensive study of thermosensitive ore-shell partiles

in aqueous solution. These partiles have been synthesized by us [17, 6365℄ and by

others [13, 59℄. They have been used as modelsystem for the study of the ow behavior

ofonentrated suspensions[39,40℄. Theresultsobtained sofarprovideanexellenttest

for the mode-ouplingtheory of the dynamis of dense olloidalsystems [39, 40, 44, 45℄.

A further point ommanding attention is the rystallization of these partiles. Given

the fat that the shell of these partiles onsist of a ompressible network, this point is

ertainly inneed of further eluidation and willbe disussed in the hapter 2.3.

In this hapter we demonstrate that ryogeni transmission eletron mirosopy (Cryo-

Table2.1: Synthesis of the ore partiles.

Styrol[

g

^{℄ 193.2}

NiPAM [

g

^{℄ 10.5}

SDS[

g

^{℄ 1.79}

KPS [

g

^{℄ 0.39}

H

2

^O[

g

^{℄ 706}

TEM) was highly suited to study these ore-shell partiles in situ [22, 23, 65℄. Cryo-

TEM allows usto visualize the partilesdiretly inthe aqueous phase by shok-freezing

of a suspension of the partiles. The volume transition was for the rst time diretly

made visible at dierent temperatures, inluding temperatures below and above room

temperature, and anbeompared todata obtained by small-angleX-ray sattering and

dynamilightsattering. Moreover,the swellingofthenetworkismodeledintermsofthe

Flory-Rehnertheory. Speialattentionispaidtotheinterplayofthedegreeofrosslinking

of the partiles and the phase behaviorat high volumefrations and willbedisussed in

the setion rystallization.

2.1.2 Experimental

Synthesis and puriation

The ore-shell partiles were synthesized in a two-step reationas desribed in ref. [17℄.

The ore partiles were obtained by emulsion polymerization and used as seed for the

radial polymerizationof the ross-linked shell.

Chemial

N

-isopropylarylamide (NIPAM; Aldrih),

N, N ^′

-methylenebisarylamide (BIS; Fluka), sodium dodeyl sulfate (SDS; Fluka), and potassiumperoxodisulfate (KPS; Fluka) were

used as reeived. Styrene (BASF) was washed with KOH solution and distilled prior

to use. Water was puried using reverse osmosis (MilliRO; Millipore) and ion exhange

(MilliQ;Millipore).

Core latex

Emulsionpolymerizationhas been doneusing a1-Laskequipped withastirrer,areux

ondenser,andathermometer. Thereipefortheorelatexisgiveninthefollowing: SDS

and NIPAM were dissolved in pure water with stirring and the solution is degassed by

repeated evauation under nitrogen atmosphere. After addition of styrene, the mixture

is heated to 80

o C

^{under an atmosphere of nitrogen. The initiator KPS dissolved in}

15

mL

^{of water is added while the mixture is stirred with 300}

rpm

^{. After 8}

h

^the

latex is ooled to room temperature and ltered through glass wool to remove traes of

oagulum. Puriationwasdonebydialysisofthe latexagainst2.5

· 10 ⁻³ M

^KClsolution

for approximately 3 weeks (Mediell, 12000-14000

Dalton

^{). The masses of the dierent}

reatants are summarizedin the table 2.1.

Core-shell latex

Table 2.2:Synthesis of the ore-shell Laties.

Core-shell Latex KS1 KS2 KS3 KS4

(ross-linking[

mol.%

^{℄) 1.25 2.5 5 2.5}

CoreLatex [

wt.%

^{℄ 20.1 18.9 21 19.5}

CoreLatex [

g

^{℄ 199.0 211.5 190.5 205.1}

NiPA [

g

^{℄ 38.0 38.0 38.0 19.0}

BIS [

g

^{℄ 0.6480 1.2959 2.5885 0.6470}

KPS in10

ml

^H

2

^O[

g

^{℄ 0.3834 0.3814 0.3812 0.3838}

H

2

^O[

g

^{℄ 542.4 535.8 568.2 363.5}

m P S /m shell

^{1.06 1.03 - 1.05}

The seeded emulsion polymerization for the ore-shell system under onsideration here

wasdoneusing100

g

^{oftheorelatexdilutedwith 320}

g

^{ofdeionizedwater togetherwith}

20

g

^{of NIPAM and 1.43}

g

^{of BIS. No additionalSDS wasadded in this step. After this}

stirred mixturehas been heated to 80

o C

^{,the reationis startedby the additionof 0.201}

g

^{of KPS (dissolved in 15}

mL

^{of water) and the entire mixture is allowed tostir for 4}

h

at this temperature. After ooling to room temperature the latex has been puried by

exhaustiveserumreplaementagainstpuriedwater (membrane: ellulosenitratewith a

0.10-

µm

^{pore width supplied by Shleiher and Shuell). The ells ontain 750}

ml

^{. The}

puriationwas performedonira 10

wt.%

^{solutionunder1, 2}

bar

^{nitrogenand used to}

onentrate the initialsolutions and to adjust the salt onentration. The masses of the

dierentreatantsusedforthesynthesisofthedierentore-shellsystemsaresummarized

in the table 2.2.

Methods

Transmissioneletron mirosopy

Samplesfor TEMwere prepared by plainga drop ofthe 0.2

wt.%

^{solution onaarbon-}

oatedoppergrid. Afterfewseonds,exess solutionwasremoved by blottingwithlter

paper. Theryo-TEMpreparationwasdoneondilutesamples(0.2

wt.%

^{). Thesamplewas}

kept at roomtemperature and vitried rapidly by the methoddesribed previously [66℄.

AfewmirolitersofdilutedemulsionwereplaedonabareopperTEMgrid (Plano,600

mesh) held by the tweezers of the ControlledEnvironmentVitriationSystem (CEVS).

The dimensions of the holes where the sample is absorbed and vitried are

35 × 35 × 10 µm

^{. The exess liquidwas removed with lterpaper. Typially the lmthikness where}

the partiles are investigated ranges between 1

µm

^{and the diameter of the partiles}

(

∼

¹⁰⁰

nm

^{). This sample was ryo-xed by rapid immersingintoliquid ethaneooled to}

-180

o C

^{in a ryo-box (Carl Zeiss NTS GmbH). The speimen was inserted into a ryo-}

transfer holder (CT3500, Gatan, Munih, Germany) and transferred to a Zeiss EM922

EFTEM (Zeiss NTS GmbH, Oberkohen, Germany). Examinations were arried out at

temperatures around -180

o C

^{. The TEM was operated at an aeleration voltage of 200}

kV

^{. Zero-losslteredimagesweretaken underredueddoseonditions(}

<

²¹⁰⁰⁰

e ⁻ /nm ²

⁾

with an aperture

α ₀ = 10 mrad

^{at amagniation of}

16000X

^{. All imageswere reorded}

Table 2.3:Summary of the dierent parameters used for the normalization of the sattering in-

tensity prole (see textfor further details).

Systems

c [g/cm ³ ] crosslinking [mol.%] _m ^{m core}

shell N/V [nm ⁻³ ]

Core 0.060 - - 9.62.10

−8

KS1 0.032 1.25 1.19 1.99.10

−8

KS2 0.023 2.50 1.15 2.79.10

−8

KS3 0.035 5.00 1.04

1

2.87.10

−8

digitallyby a bottom-mounted 16bit CCDamera system (UltraSan 1000, Gatan). To

avoid any saturation of the gray values all the measurements were taken with intensity

below 15000, onsidering that the maximum value for a 16 bit amera is

2 ¹⁶

^{. Images}

havebeen proessedwithadigitalimagingproessingsystem (DigitalMirograph 3.9for

GMS1.4, Gatan). The experiment at45

o C

^{were performedinan OxfordCT-3500 (now:}

Gatan, Pleasanton, CA) ryo-holder, and were examined in an FEI (The Netherlands)

T12 G

2

dediated ryogeni-temperature transmissioneletron mirosope.

Dynami light sattering

Dynami light sattering (DLS) was done using a Peters ALV 5000 light sattering go-

niometerequipped with a He-Ne laser (

λ =

^632.8

nm

^{). The} temperature was ontrolled with an auray of 0.1

o C

^{. The sampleswere highlydiluted (}

c = 2.5.10 ⁻³ wt.%

^{)to pre-}

vent multiple sattering and ltered through a

1.2 µm

^{lter to remove dust. The salt}

onentration in KCl was set to 10

−4 mol.L ⁻¹

^{and 5.10}

⁻² mol.L ⁻¹

^{. The} measurements were performedat asattering angleof 90

o

for temperatures between 10 and 50

o C

^.

Small-angle X-Ray sattering

Small-angle X-Ray sattering experiments have been performed on both ore and ore-

shellsystems. MostoftheSAXSmeasurementsreportedherehavebeen performedatthe

ID2beamlineatthe EuropeanSynhrotronRadiationFaility(ESRF,Grenoble, Frane).

The diameterof the X-ray beam was 150

µm

^{and the inident wave length equals to 0.1}

nm

^{. SAXS pattern were reorded with a} two-dimensional amera loated at a distane of 5

m

^{fromthe sample withinan evauated ighttube. The bakground sattering has}

been subtrated from the data and orretions were made for spatialdistortions and for

the detetor eieny. The onentrations of the laties varies between 2 and 6

wt.%

(see Table 2.3). For the latex onentrations used here we assume that the inuene of

interpartiular interferenes an be dismissed without problems and that the struture

fator

S(q)

^{is equalto 1[17, 67℄.}

In order to hek the detetor the same ore solution has been measured on a modied

Kratkyamerafor

q

^{between 0.03 and4}

nm ⁻¹

^{. The desriptionof the ameraandof the}

evaluationof the sattering isgiven elsewhere [17℄.

The density of the shell has been alulated onsidering the value of the density of the

polystyrene ore (1.0525

g/cm ³

^{),the density ofthe ore-shell fortheKS2 at25}

^o C

^(1.098

g/cm ³

^{) and the mass ratio}

m P S /m shell

^determined gravimetrially (1.03) using the for- mula:

̺ shell = 1 − (m P S /m shell )/(1 + m P S /m shell )

̺ ⁻¹ _core−shell − ̺ ⁻¹ _core (m P S /m shell )/(1 + m P S /m shell )

^(2.1)

Theshell densityderivesfromthis alulationisequalto1.149

g/cm ³

^{. The samealula-}

tion performed this time onsidering the density of the ore partiles (1.059

g/cm ³

^{) and}

the mass ratio between the orepartiles and the shell polymerizedin the seondstep of

thepolymerization

m core /m shell

^{(1.15)givesavalueof1.147}

g/cm ³

^{. Thesamealulation}

performed on the KS1 onsidering the density of the ore-shell measured at 20

o C

^(1.098

g/cm ³

^{) and the dierent} mass-ratios (

m P S /m shell = 1.06

^,

m core /m shell = 1.19

^{) gives re-}

spetively adensity of 1.151 and 1.148

g/cm ³

^{. The dierent results for the two systems}

obtained fromthe two alulationsare in good agreement withinthe experimentalerror,

whihis mostly omingfrom the determinationof the mass ratioby gravimetry. Forthe

rest of the work the density for the PNIPAM and for the ross-linked shell will be on-

sidered equalto 1.149

g/cm ³

^{. In this way the density value of the shell is slightly higher}

than the density of pure PNIPAM in water as determined by Shibayama and al. (1.140

g/cm ³

^{)[1℄, whih is naturalonsidering the} ross-linkingof the system.

The eletroni density has been alulated in

electrons/nm ³

^{using the formula:}

̺ e = N A .̺.n e ⁻

M

^(2.2)

with

̺

^{the density of the system,}

M

^and

n e ⁻

^{the moleular weight and the number}

of eletrons per onstituting moleules. From the density values the exess eletroni

density

∆̺ e

^{of the} ross-linked shell follows as 45.5

e ⁻ /nm ³

^{. The respetive quantity of}

polystyrene is 7.5

e ⁻ /nm ³

^{at 25}

^o C

^{. Thesenumbers denethe ontrastinSAXS ofthese}

polymers in water.

Thesatteringdensityprolehavebeennormalizedbythenumberofpartilespervolume

N/V

⁽ⁱⁿ

nm ⁻³

^{) in order to obtain the sattering of one single partile}

I ₀

^{. The quantity}

N/V

^{derives from the mass} onentration of the dispersion

c

⁽ⁱⁿ

g/cm ³

^{), from the ratio}

ore/shell

m core /m shell

^{determinedby gravimetry,and fromthe radiusof theore}

R c

^and

itsdensity (1.059

g/cm ³

^{)as follows:}

N/V = c.(m _{P S} /m _shell )/(1 + m _{P S} /m _shell )

(4/3)π̺ c R ³ _c

^(2.3)

Tothispurposethevalueof

R c

^{wasonsideredequalto52}

nm

^{fromthegaussiantofthe}

size distribution determined from the ryoTEM analysis (see setion 2.2). The dierent

parameters for the normalization of the urves are indiated in the table 2.3. Note that

themassrationore/shelloftheKS3hasnotbeendeterminedgravimetriallybutderived

fromthe phase diagrampresent inthe setionrystallization (see setion 2.3).

2.1.3 Theoretial bakground

Flory-Rehner theory

ThemarosopistateofahomogeneousneutralgelisdesribedwithinthelassialFlory-

Rehner theory. Here we followthe exposition of this modelgiven inRef. [49℄. Hene, it

sues todelineate the mainsteps.

The net osmoti pressure within the gelis given by

Π = k b T a ³

(

− φ − ln(1 − φ) − χφ ² + φ 0

N Gel

"

1 2

φ φ 0

− φ

φ 0

1/3 #)

(2.4)

where

k B

^{is the Boltzmann onstant,}

a

^{is the monomer segment length,}

χ

^{is the Flory}

interation parameter,

φ

^{is thepolymervolumefration,}

φ 0

^{refers tothe polymer volume}

frationatareferenestateand

N gel

^{istheaveragedegreeof}polymerizationofthepolymer hain between two rosslinking points. For systems undergoing isotropi swelling, the

swelling of the mirogel an be desribed as the ratio of the average polymer volume

fration

φ

^{and the average polymer volume fration}

φ ref

^{in the ollapsed state}

φ φ ref

=

R H,ref 3

− R c 3

R H 3

− R c 3

(2.5)

with

R H

^the hydrodynami radius of the ore shell at the temperature

T

^and

R H,ref

^the

radius et the referene state after the omplete ollapse of the shell measured at 45

o C

^.

R c

^{denotes the radius of the orepartilesdetermined fromthe ryogenized} transmission eletron mirosopy. The Flory interation parameter

χ

^{isgiven by}

χ = ∆F

k _b T = ∆H − T ∆S k _b T = 1

2 − A

1 − Θ T

(2.6)

where

A = (2∆S + k B )/2k B

^and

Θ = 2∆H/(2∆S + k B

^).

∆S

^and

∆H

^{are the hanges}

in entropy and enthalpy of the proess, respetively. It has been shown that

χ

^inreases

nonlinearly with inreasing onentration of polymer (see e.g. Ref. [68℄ and further

literature ited therein)

χ(T, φ) = χ ₁ (T ) + χ ₂ φ + χ ₃ φ ² + ...

^(2.7)

with

χ 1

orrespondingto equation(3). Following Ref. [49℄wewillonly onsider the rst order of the

φ

-expansion, whih leads tothe following expression for

χ

χ = ∆F k b T = 1

2 − A

1 − Θ T

+ χ 2 φ

^(2.8)

Thermodynamiequilibriumforthe gelisattainedwhen

Π = 0

^{,i.e., ifthe pressureinside}

and outside the gel is the same. Combining eq. 2.4 and eq. 2.8, the equilibrium line in

the

T − φ

^{phase diagramis given by}

T Π=0 = Aφ ² Θ

− φ − ln(1 − φ) + A − ¹ 2

φ ² − χ 2 φ ³ + _N ^{φ 0}

Gel

1 2

φ φ 0

−

φ φ 0

1/3

^(2.9)

Small-angle X-ray sattering

The sattering intensity

I(q)

^{measured for a suspension of partiles with spherial sym-}

metry may be rendered as the produt of

I 0 (q)

^{, the sattering intensity of an isolated}

partile,and

S(q)

^{,the struture fator that takes intoaount the mutual interation of}

the partiles:

I(q) = (N/V )I 0 (q)S(q)

^(2.10)

where

N/V

^{denotes thenumberdensityofthesatteringobjets. Apreviousdisussionof}

S(q)forsystemsofspherialpartileshasdemonstratedthattheinueneofthestruture

fatorisrestritedtotheregionofsmallest

q

^{valueswhenthe}onentrationofthepartiles issmall. Its inueneontothe measuredsattering intensity anthereforebedisregarded

in the present analysis. Hene,

S(q) = 1

^{willbe assumed in the following[67℄.}

The sattering intensity of one single partile an be deomposed in priniple in three

terms [17, 63,64, 67℄:

I 0 (q) = I part (q) + I f luc,P S (q) + I f luc,shell (q)

^(2.11)

I part (q)

^{is the part of}

I 0 (q)

^{due to the ore-shell struture of the partiles (i.e., the}

sattering intensity aused by omposite partiles having a homogeneous ore and

shell) [63, 64℄. The ore and the shell are haraterized by dierent eletron densities.

I f luc,P S (q)

^and

I f luc,shell (q)

^{refer to the thermal utuation of the PS ore and the}

PNIPAM shell respetively. The shell, however, does not onsist of a solid material but

of a polymeri network whih exhibits stati inhomogeneities and thermal utuations,

for this reason we negleted the ontribution of the utuation of the PS ore and we

only take into aount

I f luc,shell (q)

^{. For spherial symmetri partiles with radius}

R

^,

I part (q)

^{is equalto}

B ² (q)

^{where the sattering amplitude}

B(q)

^{is given by.}

B (q) = 4π Z R

0

φ(r)[̺ e,p (r) − ̺ e,w ]r ² sin(qr)

qr dr

^(2.12)

The sattering ontrast is the dierene of the sattering length density of the polymer

and the surrounding solvent

∆̺ e (r) = ̺ e,p (r) − ̺ e,w

^{. By} multiplying the polymer fration

φ(r)

^{prole by the sattering ontrast} respetively of the polystyrene for the ore (

∆̺ _{e,P S} = 7.5 e.u/nm ³

^{) and of the} ross-linked PNIPAM for the shell (

∆̺ e,P N IP AM = 45.5 e.u/nm ³

^{; see setion Methods), we obtained the eletron density}

prole neessary for the alulation of the sattering intensity.

Figure2.1: Cryo-TEM mirographs of a 0.2

wt.%

^{aqueous suspension of the PS/PNIPAMore-}

shellpartilesfordierentdegreesofrosslinking: (a)KS11.25M

%

^,(b)KS22.5M

%

and ()KS3 5 M

%

^{. The sampleswere kept at room} temperature before vitriation.

The ore onsists of polystyrene and the orona of PNIPAM ross-linked with BIS.

The full and dashed lines show the hydrodynami radii respetively of the ore and

ore-shell partiles as determined by DLS.

The polydispersity an be desribed by a normalized Gaussian number distribution [17,

67℄:

D(R, σ) = 1 σ √

2π exp

− (R − h R i ) ² 2σ ²

(2.13)

with

h R i

^{the average radius and}

σ

^{the standard deviation. Here, it sues to mention}

thatthepolydispersitysmearsout thedeepminimaof

I part (q)

^{toaertainextent[63,64℄.}

Fortheevaluationof thepartofthe satteringausedbythethermaldensity utuations

within the shell

I f luc (q)

^itis appropriateto use the empirialformula[63, 64℄:

I f luc = I f luct (0)

1 + ξ ² q ²

^(2.14)

where the average orrelation length in the network is desribed by

ξ

^.

I f luc

^ontributes

signiantly onlyin the high

q

^regime.

2.1.4 Results and disussion

Cryogeni eletron mirosopy

The synthesis of the ore-shell partiles proeeds in two steps [17℄: First a poly(styrene)

ore is synthesized by onventional emulsion polymerization. The ore partiles thus

obtained are pratiallymonodisperseand well-dened. A radius of 52.0

nm

^{and apoly-}

dispersity of 4

%

^{were derived from the ryoTEM mirographs (see setion 2.2), whereas}

the dynami lightsattering gives a value of 55.0

nm

^{between 8 and 45}

^o C

^{. As expeted}

theradiusoftheorepartilesasobservedbyDLShasnodependeneonthetemperature.

Figure2.2: Cryo-TEM mirographs of a 0.2

wt.%

^{aqueous suspension of the PS/PNIPAMore-}

shell partiles. The samplewasmaintained at23

o

C(left-hand side)and 45

o

C(right-

hand side) before vitriation. The ores onsists of polystyrene and the orona of

ross-linked PNIPAM with BIS. The irle around the ore marks the ore-radius

determined by dynami light sattering in solution. The irles around the entire

partile givesthehydrodynamiradius

R _H

^{oftheore-shellpartilesagaindetermined}

by dynami light satteringtaken from Fig. 2.7

It needs to be noted that the ore partiles bear a small number of hemially bound

harges on their surfae. This is due to the synthesis of the ores whih proeeds

through a onventional emulsion polymerization. These harges keep the solutionstable

even at high temperature. In a seond step the thermosensitive shell is polymerized at

highertemperatures(80

o C

^{)ontotheseorepartilesinaseededemulsion}polymerization.

Fig. 2.1 shows the mirographs obtained for dierent degrees of rosslinking by ryo-

TEM. Forthe analysisa suspension of the partilesisshok-frozen inliquidethane. The

water is superooled by this proedure to form a glass and the partiles an diretly be

studied upon in-situ. Fig. 2.1 shows that the ore-shell partiles are indeed narrowly

distributed. Moreover, the PNIPAM shell is learly visible in these pitures without

usingany ontrastagent. Allthe polystyrene oresobserved are overed by thePNIPAM

shell leading toa partially spherialshape. This is aompanied by parts of the network

of higher and lower transmission whih an be assigned to the density utuations

and the spatial inhomogeneities in the network. This orresponds to the additional

ontributionseeninSAXS measurementsof similarore-shell partiles. Asarguedinref.

[49, 63, 64℄, the sattering intensity ontains a term related to spatial inhomogeneities

of the network found for marosopi networks and predited by theory [3℄. Hene,

Figure2.1providesadiret visualproofof animportantonlusiondrawn fromprevious

sattering measurements. Moreover, the present mirographs suggest that these utu-

ations lead to a slightly irregular shape that may be also embodied in the ontribution

to the sattering intensity measured at higher sattering angles. The bukling of the

shell whih isdereasing withinreasing rosslinkingan be relatedtothe instabilitiesof

swellinggelsourringatthe surfaeofswollengelsaxed tosolidsubstrate. Areviewof

thestudiesofthiseetrelatedtomarosopisystemswasgivenbyBoudaoudetal. [69℄.

Figure 2.1 also demonstrates that the thermosensitive shell is in some ases not fully

attahed to the ore. This sheds new light on the seond step in the synthesis of the

ore-shell partiles: The analysis of the ore partilesby SAXS showed that the addition

of 5

%

^{NIPAM during the synthesis of the ore leads to a thin shell of PNIPAM at the}

surfae of the ore partiles [17℄. The shell will be bound to the ore most probably

by hain transfer of the growing PNIPAM network to the thin PNIPAM-shell overing

the ore. The mirographs demonstrate, however, that this binding is inomplete. At

high temperatures duringthe synthesis of theshell the growingnetwork isollapsed onto

the ore. Thus, the shell is expeted to be rather homogeneous at temperatures above

the volume transition. This was shown experimentally by SANS [63℄. It will also be

shown below that the volume fration

φ 0

^{whih follows from the} Flory-Rehner analysis willbehighanddemonstratethesmalldegreeofswellingofthenetworkduringsynthesis.

However, haintransfer does not leadto omplete attahmentof the shell tothe ores in

thisstep. Hene,thethree-dimensionalswellingoftheshellbelowthetransitionmustlead

toa partial detahment of the shells. Fig. 2.1demonstrates that this eet isdereasing

with inreasing degree of ross-linkingasexpeted.

The phase transition inthe shell an be diretly imaged by CryoTEM analysis. Fig. 2.2

is an example of the mirographs resulting from the system KS2 quenhed from 45

o C

^.

Here we hose ahigher magniation to display the details of the partilesmore learly.

Naturally,thisexperimentismorediultbeausevitriationmustbemuhfasterthan

the relaxation time haraterizing the shrinking kinetis of the partiles. However, Fig.

2.2b in omparison to 2.2a learly shows that the partiles have shrunken onsiderably.

Moreover, theshellhasbeenompatedbythisshrinkingproessandprovidesnowatight

envelope of the ores. This isexpeted given the fat that the shell has been attahed to

theore atevenhighertemperatures. Moreover, theompatnessofthe shellhadalready

been dedued fromSANS-measurements [63, 64℄.

Small-angle X-ray sattering

Core partiles

First the sattering intensity prole has been evaluated for the ore partiles. Fig. 2.3

presentstheexperimentalsatteringintensityofoneisolatedpartiles

I ₀ (q)

^obtainedfrom

the synhrotron and from the modied Kratky amera. Both measurements superpose

until

q = 0.6 nm ⁻¹

^{, even if the resolution of the synhrotron is better for the small}

q

^{. Above this}

q

^{value the signal is beoming to noisy to be evaluated in opposition}

to the Kratky amera, whih is more appropriate for the higher

q

^{ause of its shorter}

distane soure detetor. For this reason the following sattering intensities

I 0 (q)

^have

beenevaluatedonlyup to0.6

nm ⁻¹

^{. Asimpletonsideringan}homogeneouspolystyrene partileof 52

nm

^{sueedstodesribethe positionofthesidemaxima. Athigher}

q

^-values

themeasured satteringisonsiderablyhigherthan theonealulatedforahomogeneous

sphere. This fatonrms the presene of a thin PNIPAM layerat the interfae.

The best twas obtained fora ore-shell prolewith a dense polystyrene ore of 48

nm

^.

The SAXS analysis of the ore partiles demonstrates furthermore that a small fration

of PNIPAM is loated in a thin 2

nm

^{shell at the surfae of the partiles. The eletron}

density of this shell (23.0

e ⁻ nm ⁻³

^{) exhibits a} onsiderably higher density than the ore (7.5

e ⁻ nm ⁻³

^{)and ontributes} onsiderably to the sattering intensity. The sensitivity of

10 ^-4 10 ^-2 10 ⁰ 10 ²

0 0.2 0.4 0.6

q [nm ^-1 ] I/ (N /V ) [n m 2 ]

0 5 10 15 20 25

0 20 40 60

r [nm]

r e ,p (r )- r e ,w [e .u /n m 3 ]

Figure2.3: Satteringintensity of an isolated partile,

I ₀ (q) = I(q)/(N/V )

^{obtained for the ore}

partiles from the synhrotron (irles) and from the modied Kratky amera (tri-

angles). The dashed line presents the sattering intensity prole of a pure 52

nm

polystyrene partiles. The solid line is the best t obtained for a ore-shell system

with a 48

nm

polystyrene ore and a 2

nm

^{thin PNIPAM shell with an eletroni}

density of 23

e ⁻ /nm ³

^{(see inset).}

the SAXS to detet thin polymer layer at solid ore interfae has been already found in

former studies for similar systems [64℄ and also in the adsorption of surfatant on ore

laties [70℄. The t proedure also shows that the size distribution of the ore partiles

is rather small with a polydispersity of 5.0

%

^. Considering the eletron density of the PNIPAM alulatedformerlythis value orresponds toapolymervolume frationof 0.5.

The mass perentage of PNIPAM in the ore deriving fromthis analysis is 6.7

%

^{, whih}

remains rather lose to the 5

%

^{introdued at the beginning of the} opolymerization of the ore partiles. Moreover the average size of 50

nm

^{is in good agreement with the}

average value obtained by TEM and ryoTEM with a deviation of less than 4

%

^(see

setion 2.2). The deviation with the hydrodynami radius of 55

nm

^{determined by DLS}

ismuhhigher. Neverthelessthisvaluereferstoanintensityweightedaveragewhereasthe

two others methods refer to number weighted average. A number distribution obtained

fromtheCONTINanalysiswillthuslay around50

nm

^{ingoodagreementwiththeothers}

methods.

Core-shell partiles

The same analysis has been performed on dierent ore-shell systems. Fig. 2.4 presents

the dierent

I 0 (q)

^{obtained for dierent degrees of} rosslinking.

I 0 (q)

^{desribes a single}

maximumfor1.25

mol.%

rosslinking,thentwomaxima for2.5

mol.%

^{andthreemaxima}

for 5

mol.%

^{. Moreover the intensity in the low}

q

^{region is inreasing. This learly}

indiatesthatinreasingthedegreeofrosslinkingleadstomoredenedandmoreompat

partiles. Moreoverthe rstmaximum isslightlyshifted tothe leftwhihisanindiation

of aderease inthe sizeof the partilesinagreement withthe diretobservationdone by

ryoTEM. A paraboliprolehas been onsidered for the shell asproposedby Berndt et

al. in their investigation of PNIPAM mirogels [60℄, PNIPAM/PNIPMAM [61, 71℄ and

PNIPMAM/PNIPAM omposite mirogels[62℄.

10 ^-2 10 ⁰ 10 ² 10

0 0.2 0.4 0.6

q [nm ¹ ] I/ (N /V ) [n m 2 ]

Figure2.4: Sattering intensities

I ₀ (q) = I(q)/(N/V )

^{obtained for the dierent degrees of}

rosslinking: KS1 (1.25

mol.%

^{) (squares); KS2 (2.5}

mol.%

^{) (irles) and KS3 (5}

mol.%

)(triangles).

Table 2.4:Fit parameters used forthe alulationof thesatteringintensityprole

I ₀ (q)

^{(see g.}

2.5).

m core /m _shell

^,

ξ

^and

R

^{arethemassratioore/shell,theorrelationlengthandthe}

overall size of the systemsderived from this analysis.

R _H

^isthe hydrodynami radius derived fromthe dynami light sattering at23

o C

^{(see gure 2.7). The} orresponding polymer volume frationprole are given in the g. 2.6.

Systems

K R hw [nm] σ

^[nm℄

P DI [%] _m ^{m core}

shell ξ [nm] R [nm] R H

^[nm℄

KS1 0.160 85 20 8.0 1.38 10 105 124.6

KS2 0.284 77 17 8.0 1.14 7 94 112.4

KS3 0.435 72 12 6.0 1.03 7 84 107.0

The following relation has been used to desribe the polymer volume fration prole for

the rosslinked shell [61,62, 71℄:

Kφ(r) =



 



 



1 : r ≤ R c

1 − (R hw − r + σ) ² /(2σ ² ) : R c < r ≤ R hw

(r − R hw + σ) ² /(2σ ² ) : R hw < r ≤ R hw + σ 0 : R hw + σ < r

(2.15)

K

^{is a prefator,}

R c

^{is the radius of the ore and}

R hw

^{isthe half-widthradius and}

σ

^the

half-width.

The prole forthe ore has been kept idential to the one derived from the ore analysis

in the preedent setion. The prole of the ore-shell partileshas been then introdued

inequation2.12toalulate

I part (q)

^{. The} polydispersity whihissmearingthemaximum has been introdued in term of a Gaussian distribution. Fig. 2.5 displays the dierent

sattering intensity proles normalized by N/V and the best t obtained for eah

system. The dashed lines refer to

I part (q)

^{and the dotted lines refer to}

I f luc (q)

^{. The}

best t derives from the sum of these two ontributions and is displayed by the solid

line. The ts provide a good desription of the experimental set of data on the

q

^range

10 ^-3 10 ^-1 10 ¹ 10 ³

0 0.2 0.4 0.6

q [nm ^-1 ] I/ (N /V ) [n m 2 ]

10 ^-3 10 ^-1 10 ¹ 10 ³

0 0.2 0.4 0.6

q [nm ^-1 ] I/ (N /V ) [n m 2 ]

10 ^-3 10 ^-1 10 ¹ 10 ³

0 0.2 0.4 0.6

q [nm ^-1 ] I/ (N /V ) [n m 2 ]

Ks1

Ks2

Ks3

Figure2.5: Form fator

P (q) = I(q)/(N/V )

^{obtained for dierent degree of} rosslinking: KS1 (1.25

mol.%

^),KS2(2.5

mol.%

^),KS3(5

mol.%

^{). Theirlesdisplaythe}experimental measurements. The long dashed lines are the alulated

I _part (q)

^{, whereas the dashed}

lines represent the ontributionof the thermaldensityutuations

I _{f luc} (q)

^{. The sum}

of this two ontributions are given by the solid lines, whih orrespond of the best

0 0.25 0.50 0.75 1.00

0 25 50 75 100

r [nm]

f

Figure2.6: Radialpolymereetivevolumefration

φ(r)

^{obtainedfromtheSAXSanalysisforthe}

dierent degrees of rosslinking: KS1 (1.25

mol.%

^{) (dotted line); KS2 (2.5}

mol.%

⁾

(dashed line), KS3 (5

mol.%

^{)(solid line). The proles onsist on a dense 48}

nm

polystyrene ore witha 2

nm

^{thin PNIPAMshell onto whiha the} rossslinked shell has beenpolymerized. The analysisonsiders aparaboli prolefor theshell asgiven

in equation 2.15. The dierent tparameters are given in table2.4.

investigated. The dierent t parameters are summarized in the table 2.4. The polymer

volume fration prole an be extrated from the t of the sattering experiments and

is presented in Fig. 2.6. As already observed by ryoTEM, inreasing the degree of

rosslinking leads to a more ompat struture and to a more pronouned depletion at

the ore/shell interfae . The ore/shell mass ratio derived from the dierent proles is

in good agreement with the value obtained by gravitometry, exept for the lower degree

of rosslinking. This ould be attributed by a lak of ontrast for a too diuse shell.

Moreoverthesize isalsodereasingasalreadyobserved fromthe dynamilightsattering

experiments (see g. 2.7). Nevertheless the radius is about 16, 16 and 21

%

^smaller

ompared tothe hydrodynami radiusof the KS1, KS2and KS3 determinedby dynami

lightsattering. Thiswas attributedinaformer study tothe presene ofdanglinghains

whih ould not be deteted by SAXS or SANS but onlyby DLS [63℄. Nevertheless the

diretimagingofthepartilesby ryoTEMevidened thestrongbuklingofthepartiles.

The disrepany between the two methods an be explained as follows: Mirogels are

dynami strutures whih exhibit thermalutuations. Moreover, the synthesis leads to

the buklingup ofthe shell asdisussed already. Hene, the shape of the partilesis not

perfetlyspherial. A rotationalaveragehene willresult inalarger size. This pointwill

be disussed infurther detailsinthe next hapterdediating tothe quantitativeanalysis

of the ryo-TEM mirograhs (see hapter 2.2).

Thermodynamis of the phase transition

The volume transitionwithin the shell an easilybestudied by dynamilightsattering

(DLS).Figure2.7shows thedependeneofthehydrodynamiradius

R _H

^{ofthe omposite}

60 70 80 90 100 110 120 130 140 150

280 290 300 310 320

T [K]

R H [n m ]

Figure2.7: Hydrodynami radii of the ore-shell latexes versus temperature for dierent degrees

of rosslinking, as determined by DLS (irles: 1.25

mol.%

^{, squares: 2.5}

mol.%

^,

triangles: 5

mol.%

^{). Full symbols represent the} measurements without addition of salt, whereas hollow symbols display the measurements performed by adding 5.10

−2

mol.L ⁻¹

^KCl.

mirogelsdeterminedbyDLSasfuntionofthetemperature.

R H

^{dereasesgraduallywith}

temperatureuntilasharpvolumetransitionfromswollen tounswollen statestakesplae,

reahinganalollapsedsizeatatransitiontemperaturebetween34and38

o C

^depending

onthedegreeofrosslinking. Inreasingthedegreeofrosslinkingthetransitionbeomes

moreontinuousandtheollapsestateisshiftedtohighertemperatures. Withoutaddition

of saltthis proess isthermoreversible withoutany hysteresis.

The omparison between the overall size observed from the mirographs and the hy-

drodynami radius as determined by the DLS an be observed in the Fig. 2.7. The

hydrodynami radius

R H

^{as measured by DLS indiated in eah ase as a shed irle}

around one partile evidently provides an appropriate measure of the average radius of

the partiles. Moreover we found that the overall radius of the partiles from these mi-

rographs isingoodagreementwiththe hydrodynamiradius measuredat45

o C

^(dashed

irle). This indiates that the proess of quenhing is suiently fast to preserve the

high-temperature struture. This nding is quite important inasmuh it shows that the

method of preparation does not disturb the struture of the thermosensitive partiles.

This fat is of great importane when determining the eetive volume fration of the

partilesin a onentrated suspension. Addition of 5.10

−2 mol.L ⁻¹

^{KClleads to a slight}

shrinking of the partiles. This phenomenon has been already investigated in a reent

study [72℄. The additionof saltsreens the residualeletrostatiinteration ofthe parti-

les. Hene, at higher temperatures the dispersions beome unstable and aggregate [72℄.

Forthe systems underonsiderationaggregationtakes plaeabove32

o C

^{for theKS1 and}

about 33

o C

^{for the KS2 andKS3. Evidently,}experimentsaiming atanunderstanding of the ow behavior of stable suspensions must be done below these temperatures. On the

other hand, salt must be added for a suient sreening of the eletrostati interation

in orderto obtaina modeldispersions that interats solelythrough steri repulsion.

We an now disuss the modeling of the swelling data shown in Fig. 2.7 interms of the

Flory-Rehnertheory. Parameter of the dierent sets ofdata isthe degreeof rosslinking.

280 285 290 295 300 305 310 315 320

0.05 0.1 0.2 0.5 1

f

T [K ]

Figure2.8: Experimental phase diagram

T − φ

^{for dierent degrees of} rosslinkings (full irles:

1.25

mol.%

^{, full squares: 2.5}

mol.%

^{, hollow triangles: 5}

mol.%

^{). Lines present}

the ts obtained from eq.(2.9℄. The vertial dashed line marks the referene volume

fration

φ ₀ = 0.7

^{in the ollapsed state.}

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

290 295 300 305 310 315

T [K]

c

Figure2.9: Solvent parameter

χ

^{as determined from the ts of g. 2.8 for dierent degrees of}

rosslinking (full irles: 1.25

mol.%

^{, full squares: 2.5}

mol.%

^{, hollow triangles: 5}

mol.%

^{). Dereasing the degree of} rosslinking the PNIPAM network shrinks upon heating from a ontinuous to a disontinuous fashion to reah a ollapsed state for

χ =

^1.

χ =

^{0.5is indiated by the dashed line andlays} approximately at 32

o C

^,whih

orresponds tothe LCSTof pure PNIPAMin aqueous solution.

Table2.5: Parameters of theFlory-Rehner t. (eq. 2.9 andFig. 2.8).

KS1 KS2 KS3

n(BIS)/n(NiPAM) [mol.

%

^{℄ 1,25 2,50 5,00}

φ ₀

^{0,7 0,7 0,7}

A

^{-8,7 -8,7 -8,7}

χ 2

^{0,9 0,9 0,9}

Θ

^{[K℄ 312 314 316}

N Gel

^{80 45 22}

LCST [

o

C℄ 31.7 32.3 32.2

T (χ = 1)

^[

^o C

^{℄ 35.1 36.2 38.2}

Thetproedureusedtomodelthephasetransitionisthe sameasreportedreently[49℄.

Thets arepresented togetherwiththe experimentaldata asshown inthe

T − φ

^diagram

(Fig. 2.8). Theresultantttingparameters aresummarizedinthetable 2.5. Considering

that onlythe amount ofrosslinkeris hanging,we keep the same value for

φ 0

^,

A

^and

χ 2

for allthe systems and we onlyvary

θ

^and

N

^.

For the present system the best agreement for the referene polymer volume fration in

the ollapsed state has been found for

φ 0 = 0.7

^{. This value has already been expeted}

from the previous analysis of the partiles by SAXS and SANS [63℄. Indeed as reported

by reent nulear magneti resonane measurements water moleules are still present in

the shell above the LCST but they are strongly onned [58℄.

The

N

^{values (see table 2.4) found are} proportionalto the degree of rosslinkingbut are about two times larger than those orresponding to the rosslinking in a homogeneous

network. Aontentof2.5

mol.%

^{oftherosslinkerBISwouldorrespondto}

N gel =

^{20. This}

disrepanyanbetraedbaktothe inhomogeneitiesinthePNIPAMmirogels. Indeed,

Wu et al. [73℄ investigated the polymerization of NIPAM and BIS during the mirogel

synthesis. The rosslinker was found to be onsumed faster than the NIPAM indiating

that the partiles are unlikely to have a uniform omposition. This nding has been

onrmed by SAXSand SANS,revealing thatthe segment density inthe swollen state is

not homogeneous, but gradually deays at the surfae [60, 63℄. Moreover high-sensitive

alorimetristudy haveonrmedthisassumption[55℄. Given thevariousunertaintiesof

the Flory-Rehner analysis, the present ts seem to providea suient desription of the

data. Moreover, itshouldbekept inmindthatthe originaltheory hasbeen developedfor

marosopi, three-dimensional networks while it is applied here to mirosopi systems

whihan swell onlyalong the radial diretion.

Fig. 2.9 displays the evolution of the solvent parameter

χ

^{derived from the t from the}

g. 2.8asfuntionofthetemperatureforthethreesystems. TheLCSTthenorresponds

to the temperature where

χ

^{is equal to 0.5. We found that inreasing the} ross-linking slightly shifts the LCST to higher temperature between 1.25 and 2.5

mol.%

rosslinker, but the LCST found from this analysis is rather lose to 32

o C

^{whih orresponds to}

the LCST of linear PNIPAM hains [55℄. On the ontrary the temperature where the

shell is totallyollapsed obtained from

χ = 1

^{shifts from35.1 to38.2}

^o C

^{with inreasing}

rosslinking, whih an be attributed to a higher rubber elastiity of the network. This

illustratesthe transitionfromasharptoaontinuousvolumetransitionby inreasingthe

rosslinkingof the shell.

The present analysis thus demonstrates that the ore-shell mirogelsan be modelled in

the same way asmarosopi systems.

2.1.5 Summary

In this hapter, omposite PS/PNIPAM ore-shell mirogels with dierent degrees of

rosslinking have been synthesized and haraterized by ryogeni transmission eletron

mirosopy, small-angle X-rays sattering and dynami light sattering. The analysis

demonstratesthatthe shellformsawell-denednetwork aroundthe pratiallymonodis-

perseorepartiles. Inreasingthedegreeofrosslinkingwasalsofoundtoleadtosmaller

anddenserpartiles. Moreover,diretimagingofthepartilesbyCryo-TEMshowsthein-

homogeneitieswithinthenetwork. Cryo-TEM shows alsothe buklingof theshellaused

by the one-dimensionswelling ofthe shell. This bukling eet whih iswell-known from

marosopi systems leads to a slightly irregular shape partially explaining the disrep-

aniesbetween the SAXSand the DLS. Moreovera parabolidensity prole forthe shell

has been evidened by SAXS. The volume transition within the shell of these partiles

an bedesribed very well by the Flory-Rehnertheory. All results demonstrate that the

two-stepsynthesisofthepartilesleadstowell-denedpartilessuitableasmodelsystems

for studying the dynamisof onentrated suspensions.