8/12/2019 5_3516_832

1/18

18

Structure and properties of polymers

Copolymers

Copolymers: A polymer that consist of two or more dissimilar repeat units in

combination along its molecular chains.

- Copolymers, are polymers which has at least two different types of mers.

- They can differ in the way the mers are arranged so it can be classified into random

copolymer, alternating copolymer, blok copolymer, and graft copolymer.

- All types of copolymer depending on the polymerization process and the relative

fraction of these repeat unit types.

These different arrangement are

!andom copolymer A polymer in which two different repeat units are randomly

distributed along the molecular chain.

"#"#"#"$"#"$"#"#"$"#"$"$"

Alternating copolymer a copolymer in which two different repeat units alternate position

along the molecular chain.

-X-W-X-W-X-W-X-W-X-W-X-W-

%lock copolymer a linear copolymer

in which identical repeat units are clustered in blocks along the molecular chain.

"#"#"#"#"#"$"$"$"$"$"

8/12/2019 5_3516_832

2/18

18

&'ample (mpact modified polystyrene is a block copolymer that consisting of alternating

blocks of styrene and butadiene.

)raft copolymer A copolymer wherein homopolymer side branches of one monomer

type are grafted to homopolymer main chains of different monomer type.

The mer molecular weight for a copolymer can be determined by

$here

f*" mole fraction of repeat unit * in the polymer chain.

m*" molecular weight of repeat unit * in the polymer chain.

Polymer crystallinity

o Atomic arrangement in polymer crystals is more comple' than in metals or

ceramics.

o The unit cells are typically very large and comple' as molecules or chains

replace ions and or atoms in these structures.

8/12/2019 5_3516_832

3/18

18

o Think of it as packing of molecular chains in a geometrical array so the

polymer crystallinity can be defined as the backing of molecular chains to

produce an ordered atomic array.

o +ome parts of structure align during cooling to form crystalline region not

like -CC%CC metals/ chains align alongside each other.

&'ample fig.10.1/shows the unit cell of polyethylene and it2s relation to the molecular

chain structure, this unit cell has orthorhombic geometry 3a4b4c5.

6olecular substance having small molecular e.g.

water and methane / are normally either totally

crystallineas solids/ or totally amorphous as

li7uid/.

As a conse7uence of their size and often

comple'ity, polymer molecules are often partially

crystalline semicrystalline/, with crystalline

8/12/2019 5_3516_832

4/18

18

regions dispersed within amorphous material. %ecause, any disorder, kink in the long

chains induce an amorphous region.

The density of a crystalline polymer will be greater than an amorphous one of the same

material and molecular weight, since the chain are more closely packed together for the

crystalline structure.

The degree of crystallinity by weight may be determined from accurate density

measurement according to

$here

s" density of a specimen for which the present crystallinity is to be determined.

a" density of the completely amorphous polymer.

c" density of the completely crystalline polymer.

Hint egree of crystallinity ranges from 9":9/;

Factors effecting crystallinity:

!ate of cooling during solidification time is necessary for chains to move and align into

a crystal structure.

6er comple'ity crystallization less likely in comple' structures, simple polymers, such

as polyethylene, crystallize relatively easily.

Chain configuration linear polymers crystallize relatively easily, branches inhibit

crystallization, network polymers almost completely amorphous, crosslinked polymers

can be both crystalline and amorphous.

(somerism isotactic, syndiotactic polymers crystallize relatively easily " geometrical

regularity allows chains to fit together, atactic difficult to crystallize

Copolymerism easier to crystallize if mer arrangements are more regular " alternating,

block can crystallize more easily as compared to random and graft

6ore cryallinity higher density, more strength, higher resistance to dissolution and

softening by heating.

8/12/2019 5_3516_832

5/18

18

Hint:

linear polymers, crystallization is easily accomplished because there are few restrictions

to prevent chain alignment.

H.Wwhy the crosslinked are almost amorphoussing the answer to part a/, calculate the percent crystallinity of a

branched polyethylene that has a density of .:?9 g@cm. The density for the totally

amorphous material is .8B g@cm.

+olution a/

b/

8/12/2019 5_3516_832

6/18

18

Stability of Amorphous Pharmaceutical Solids: Crystal Growth

Mechanisms and Effect of Polymer Additives

Ye Sun, Lei Zhu,Tian Wu, Ting Cai,Erica M. Gunn, and Lian Yu

Author information Article notes Co!right and License information

Abstract

Go to"

IN!"#$CI"N

Amorphous solids can be produced by cooling li7uids, evaporating solutions, and

condensing vapors while avoiding crystallization. ther routes are known that lead to

amorphous solids= for e'ample, mechanically damaging crystals 1/ and removing waterfrom hydrated crystals ?/. Amorphous solids produced by cooling li7uids are commonly

called glasses. (n this process of glass formation, molecular motions become increasingly

slower with cooling until finally, at the so"called glass transition temperature Tg, the

system can no longer reach internal e7uilibrium with each decrease of temperature and

becomes kinetically frozen. $ith respect to molecular packing, amorphous solids are

usually envisioned as having significant local order e.g., each molecule having similar

number of nearest neighbors/, but lacking long"range order that characterizes molecular

packing in crystals.

Amorphous solids are generally more soluble and faster dissolving than their crystallinecounterparts, which makes them potentially useful for delivering poorly soluble drugs

whose bioavailability is limited by their low solubility. -or indomethacin, the solution

concentration reached by dissolving an amorphous solid has been found to be 9D1B times

higher than by dissolving a crystalline solid D9/. Amorphous ritonavir was found to

dissolve ca. ten times faster than crystalline ritonavir E/.

Amorphous drugs must resist their thermodynamic tendency to crystallize, for

crystallization negates their solubility advantages. The past two decades have seen active

research on amorphous pharmaceutical solids and their stability against crystallization,

and several reviews have appeared BD11/. +ome 7uestions studied in this conte't are

1. Can the crystallization rate of an organic glass be predicted by

extrapolating that of the corresponding liquid? This question is studied

to learn whether the crystallization of organic glasses can be treated as

the low-teperature !ersion of liquid-state crystallization. "ecause

http://www.ncbi.nlm.nih.gov/pubmed/?term=Sun%20Y%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Sun%20Y%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Zhu%20L%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Zhu%20L%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Wu%20T%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Cai%20T%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Gunn%20EM%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Gunn%20EM%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Yu%20L%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR1http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR2http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR3http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR5http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR6http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR7http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR11http://www.ncbi.nlm.nih.gov/pubmed/?term=Sun%20Y%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Zhu%20L%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Wu%20T%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Cai%20T%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Gunn%20EM%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pubmed/?term=Yu%20L%5Bauth%5Dhttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR1http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR2http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR3http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR5http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR6http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR7http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR11

8/12/2019 5_3516_832

7/18

18

crystallization in!ol!es nucleation and crystal growth# this question is

posed for each step.

$. What olecular otions in aorphous solids are associated with

crystallization?

%. &ow do free surfaces a'ect crystallization in organic glasses?

(. &ow does the crystallization of an aorphous solid depend on the

ethod of preparation )elt-cooling# solution-drying# !apor-

condensation# cryo-illing# and others* and its theral history?

+. &ow do excipients )especially polyers* a'ect the crystallization of

aorphous drugs# both during storage and dissolution? What attributes

of a polyer a,e it a good crystallization inhibitor?

. &ow soluble are crystalline drugs in polyers? )1$1(*# This question isstudied to learn the axial drug loading in a polyer atrix without

ris, of crystallization.

/. What is the e'ect of oisture on the stability of aorphous drugs and

forulations? )1+*

$e will focus this review on crystal growth in organic glasses. $e do so because of

recent progress in this area and because crystal growth in organic glasses has properties

unknown for and thus unpredictable from the behaviors of/ other materials. $e also

discuss recent work to understand the role of polymer additives in stabilizing amorphousdrugs against crystallization. The materials covered are relevant to Fuestions 1, ?, , and

9.

Go to"

%AS M"#ES "% C!&SA' G!"() IN "!GANIC G'ASSES

6any studies have observed that the linear velocity uof crystal growth in a one"

component li7uid typically increases and then decreases with supercooling.

-igure1illustrates this pattern for crystal growth in li7uid o"terphenyl TG/, a well"studied small"molecule organic li7uid. The u vs. temperature plot is a bell"shaped curve

between the melting point Tmand the glass transition temperature Tgtriangles/. This

pattern e'ists because at small supercooling, the growth rate is limited by thermodynamic

driving force, and at larger supercooling, the growth rate is limited by molecular mobility

in the li7uid. -or TG, uclosely tracks the self"diffusion coefficientDopen symbols/ at

large enough supercooling TH ?89 I/, over several orders of magnitude of change 1ED

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR12http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR14http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR15http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR16http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR12http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR14http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR15http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR16

8/12/2019 5_3516_832

8/18

18

18/. This relation *ustifies the description of the crystal growth process as diffusion"

controlled, and is consistent with the common view that molecular diffusion in the li7uid

defines the kinetic barrier for crystal growth 1:,?/.

0ig. 1

aCrystal growth rate uand self-di'usion coecient Dof liquid and glassyo-

terphenyl. Tg 2 $( 3. b4hotoicrographs of 5C growth at $(6 3

iffusion"controlled growth serves as a reference point for fast modes of crystal growth

that can emerge as organic li7uids are cooled to become glasses. These growth modes

lead to crystal growth rates orders of magnitude faster than e'pected for diffusion"

controlled growth. ne such growth mode happens in the interior of a glass, and another

occurs at the free surface. These phenomena are unknown or uncommon for other classes

of glass"forming li7uids. $e review below key observations concerning these growth

modes and their current e'planations.

Go to"

G'ASS*C!&SA' G!"() M"#E

-igure1shows that while the crystal growth rate utracks the diffusion

coefficientDabove Tg, ubecomes orders of magnitude faster with a temperature drop of

a few I filled circles/ ?1,??/. This growth mode, termed )C glass"to"crystal/, is so fast

that it is not limited by molecular diffusion in the bulk li7uid. This phenomenon was

apparently first noted by )reet and Turnbull in 1:EB ?1/ and then studied systematically

by guni and coworkers since 1::9 ??,?/. The phenomenon is remarkable because on

cooling, the loss of li7uid"like mobility activates fast crystal growth, and on heating, thegain of li7uid"like mobility disrupts the fast crystal growth that occurs in the glass state.

To our knowledge, the abrupt activation of fast crystal growth has not been reported for

non"organic li7uids. +ince there is no corresponding increase in the diffusivity at the

onset of )C growth, this growth mode has been called JdiffusionlessK, in contrast to

diffusion"controlled growth.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR18http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR19http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR20http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR21http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR22http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR22http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR21http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR22http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR23http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR18http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR19http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR20http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig1/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR21http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR22http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR21http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR22http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR23

8/12/2019 5_3516_832

9/18

18

To date, )C growth has been reported for more than ten organic li7uids ?D?E/,

including nifedipine L(-/, a poorly water soluble calcium channel blocker drug ?0/. -or

L(-, the rate of crystal growth shows an abrupt tenfold increase as the temperature is

decreased from above to below Tg19 I/.

GC Growth Favors Certain Crystal Structures

+un et al. studied )C growth using the polymorphs of 9"methyl"?"3?"

nitrophenyl/amino5""thiophenecarbonitrile, named !M for its numerous red, orange,

and yellow polymorphs and the top system for the number of coe'isting polymorphs of

known structures in the Cambridge +tructure atabase ?B,?8/. %ecause polymorphs

share the same li7uid and glass, which polymorph grows and which polymorph shows

)C growth can reveal the dependence of the phenomenon on crystal structure. f seven

polymorphs of !M whose crystal growth can be studied near TgTg N ?E I/, four

polymorphs show )C mode, while the other three do not ?9/. The polymorphs that showthe )C growth have more isotropic molecular packing and greater densities than those

that do not. -or a molecule in a polymorph showing )C growth, the closest neighbors are

at appro'imately the same distance, as one e'pects for the packing in the li7uid state,

whereas a molecule in a polymorph not showing )C growth has the closest neighbors at

very different distances. +imilar isotropic packing characterizes other crystal structures

showing )C growth, including TG, toluene, and salol. This finding suggests that fast

crystal growth may occur if sufficient similarity e'ists between molecular packing in the

li7uid and the crystalline state.

GC Growth has Precursor in the Equilibrium Liquid

+un et al. reported that )C growth is not truly a growth mode suddenly emerging

near Tgbut already e'isting in the form of fast"growing fibers in the e7uilibrium li7uid up

to about 1.19 Tg?:/. (f the growth rates of these fibers are plotted against temperature,

they fall smoothly in line with rates of the fully activated )C growth near and below Tg.

(t is also observed that the actively growing tips of the fibers are the preferred site for

activation of the compact, spherulitic )C growth upon cooling below the onset

temperature for )C growth. -igure ?shows such an e'ample. The spherulitic )C growth

of MT0 a polymorph of !M/ was interrupted by a "I temperature increase from

?EB I to ?B I. uring holding at ?B I, the compact )C growth appeared to cease, but

close e'amination revealed fiber"like crystals e'tending into the li7uid. After the

temperature is returned to ?EB I, new )C growth was initiated predominately on the tips

of the actively growing, far"reaching fibers. #i et al. reported a similar observation for

)C growth in TG /.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR23http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR26http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR24http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR27http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR28http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR28http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR25http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR29http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig2/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig2/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR30http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR23http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR26http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR24http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR27http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR28http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR25http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR29http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig2/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR30

8/12/2019 5_3516_832

10/18

18

0ig. $

a0ibers of 7T8( )a 9:7 polyorph* eerging at $/8 3 in $+8 in fro a

spherulite pre!iously grown at $/ 3. Crossed polarizers were used to re!eal

the ;bers and resulted in dar, bac,ground. b

8/12/2019 5_3516_832

11/18

18

in the bulk -ig./. 0/ This phenomenon results in a thin surface layer of crystals

around a slower"crystallizing interior. The fast surface crystal growth can be halted by an

ultra"thin coating e.g., 1 nm of gold and D? nm of polymer/ 9/, which suggests that

surface molecular mobility is an enabler for the phenomenon. The finding also suggests a

way to stabilize amorphous drugs against crystallization see later/. (t was established byother studies that the linear velocity of crystal growth in the interior of an (6C li7uid is

appro'imately proportional to the diffusion coefficient as temperature

approaches TgE,B/, a relation e'pected for diffusion"controlled growth. Thus, fast

surface crystal growth is another mechanism by which crystal growth rate e'ceeds that

e'pected for diffusion"controlled growth.

0ig. %

aCrystal growth rates of = >C in the bul, and at the free

surface. b4hotographs of = >C growing at the free surface. The saple is on

a circular co!er glass. The fast surface crystal growth can be inhibited by a

thin coating of gold ...

The significant difference between surface and bulk crystallization rates leads to anunusual crystallization kinetics of amorphous (6C 0/. The degree of crystallinity rises

initially and then stabilizes at levels well below 1 ; crystallinity. The initial rise is due

to surface crystallization, and the subse7uent leveling off to slower bulk crystallization.

This phenomenon also causes a particle"size dependence of crystallization kinetics the

crystallinity JplateauK increases with decreasing particle size increasing surface@volume

ratio/. -ast surface crystallization is consistent with the observation that mechanical

damages of organic glasses accelerate their crystallization :/.

Lifedipine L(-/ and griseofulvin )+-, an antifungal drug/ also e'hibit fast surface

crystal growth in the glassy state 8,:/. At the same temperature relative to Tg, thesurface crystal growth rate usof L(- is ca. 1 times faster than that of (6C, and the usof

)+- is ca. 1 times faster than that of (6C :/. +urface"enhanced crystal growth of

organic glasses contrasts the comparable rates of crystal growth at the free surface and in

the interior of metallic and silicate glasses 0D0/.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig3/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig3/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR34http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR36http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR37http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR37http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig3/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR34http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR9http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR38http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR39http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR39http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR40http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR43http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig3/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig3/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR34http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR36http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR37http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig3/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR34http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR9http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR38http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR39http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR39http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR40http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR43

8/12/2019 5_3516_832

12/18

18

Surface Crystals Grow Upward !oward Free Space"

+un et al. reported that surface crystals rise substantially above the glass surface while

growing laterally, without penetrating deep into the bulk. 00/ -or the two polymorphs of

(6C Q and R/ studied, the growth front can be hundreds of nanometers above the glass

surface. -igure 0shows typical images collected by light microscopy S6/ and atomicforce microscopy A-6/ of Q (6C grown at the surface of an (6C glass film at TgD ? I.

The A-6 height image shows that surface crystals at the growth front rise significantly

above the flat glass surface.

0ig. (

@ight icroscopy )@* aand band atoic force icroscopy )A0* ciages of

B >C crystals grown at the surface of a 1+ thic, glass at (8DC. The A0

scan in cco!ered the square in b.Arrowindicates ad!ance direction ...

+un et al. also studied surface crystal growth in films with different thicknesses, from

9 nm to 19 m, to assess how the phenomenon depends on the amount of bulk material

underneath. -ilms thinner than 9 nm were prepared by spin"coating. -or Q (6C, the

crystal growth rate near Tgchanges little with film thickness until it decreases below ca.

nm= the surface growth of R (6C shows no dependence on film thickness down to

18 nm, the thinnest film in which growth of R (6C could be observed. These results

argue that surface crystal growth on (6C glasses is not perturbed by reducing the glass

thickness to a few hundred nanometers, and that the surface crystal layer is appro'imately

a few hundred nanometers thick.

Models for Fast Surface Crystal Growth

Current views differ on how crystal growth rate should change on going from the interior

of a glass to the free surface. +chmelzer and coworkers hypothesize that growing high"

density crystals in low"density glass causes elastic strain and lowers the thermodynamic

driving force 09/, and that on going from the bulk to the surface, the elastic strain

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR44http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig4/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig4/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR45http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig4/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR44http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig4/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig4/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR45

8/12/2019 5_3516_832

13/18

18

diminishes and crystallization becomes faster. Analyzing the same process of growing

denser crystals in less dense glasses, however, Tanaka concludes that the stress around a

crystal growing in a glass Jshould provide the free volume to the particles surrounding

the crystal, increase their mobility, and help further crystallization.K 1,?/ %ased on this

model, crystal growth rate at the free surface would be slower than that in the bulk. (nanother view, the different packing of molecules at the surface is thought to be

responsible for the faster crystallization 0E/. This scenario seems unlikely considering

the fact that the surface crystal layer can be hundreds of nanometers thick, much thicker

than the mobile surface layer typically envisioned a few nanometers/. Another type of

model emphasizes the greater molecular mobility at the surface 9,00/, reasoning that if

crystal growth rate is limited by molecular mobility, the enhanced mobility of surface

molecules can accelerate crystal growth. This model is consistent with the upward"lateral

growth of surface crystals. (n this scenario, crystallizing molecules would be drawn to the

crystal, climb up, and deposit at the growth sites. This model is also consistent with the

inhibitory effect of surface crystal growth by nanocoating 9/, which presumably

reduces the high surface mobility to bulk level.

)unn et al. tested the models of +chmelzer and coworkers and of Ionishi and Tanaka

using the polymorphs of carbamazepine C%U/, an anticonvulsant drug 0B/. C%U has

four known polymorphs with different densities, three polymorphs of which were

observed to grow at the surface and in the bulk of C%U glasses. The model of +chmelzer

and coworkers predicts that us@ubthe ratio of surface and bulk crystal growth rates/

increases with crystal density, whereas the model of Ionishi and Tanaka predicts the

opposite. )unn et al. found that there is no consistent increase or decrease of us@ubwithcrystal density -ig. 9/, indicating that crystal density has no controlling effect on the

difference between surface and bulk crystal growth rates.

0ig. +

usEubvs. crystal density for three C"F polyorphs at %8% and %1% 3)openand closed symbols# respecti!ely*

To test whether surface diffusion can support surface crystal growth, Uhu et al.

determined the surface self"diffusion coefficient of (6C glasses 08/. +urface diffusion

has been well studied for metals and semiconductors 0:/, but no data e'isted on organic

solids before Uhu et al.2s work. To determine the self"diffusion of (6C glass, the classic

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR31http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR32http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR46http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR44http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR47http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig5/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig5/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR48http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR49http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig5/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR31http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR32http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR46http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR44http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR47http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig5/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig5/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR48http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR49

8/12/2019 5_3516_832

14/18

18

method of surface smoothing 9/ was used. riven by surface tension, an initially

corrugated surface flattens over time by various mechanisms, among which surface

diffusion dominates the kinetics at short length scales and low temperatures. The

smoothing of surface gratings was followed at a constant temperature in dry nitrogen

with an atomic force microscope, which measured the grating2s amplitude, or an opticalmicroscope, which measured the grating2s diffraction intensity. (t was found that surface

diffusion on (6C glasses is at least one million times faster than bulk diffusion,

indicating the e'istence of a highly mobile surface -ig. E/. This finding is consistent

with recent reports of surface mobility for polymer glasses 91D90/, and a small"molecule

glass"forming li7uid 99/. The finding that e'ceptionally stable organic glasses can be

prepared by vapor deposition is also linked to enhanced surface mobility 9E,9B/. At Tg V

? I, the surface crystal growth front of (6C advances 1 nm or one molecular layer per

second= during this time, an average molecule in the bulk diffuses ?Dvt/.9 N .1 nm and a

surface molecule could diffuse ?Dst/.9 N 1 nm. This analysis suggests that surface

diffusion is fast enough to sustain observed surface crystal growth while bulk diffusion is

not.

0ig.

8/12/2019 5_3516_832

15/18

18

Although amorphous pharmaceutical formulations may contain polymers as the ma*or

component, recent studies have e'amined the use of low"concentration polymers as

crystallization inhibitors ?0,E8,E:/. +uch studies are a necessary first step for

understanding more comple' formulations, and could discover effective polymer

additives that significantly improve the properties of amorphous drugs. (n these studies,the light doping of polymers a few wt ;/ does not significantly change the

thermodynamic driving force of crystallization and the dynamics of the glasses, allowing

a close e'amination of other factors affecting crystallization.

Polymer #dditives can $ave %ifferent Effects on &ul' and Surface Crystal Growth

(shida et al. observed that doping an L(- glass with 1 wt ; of Golyvinylpyrrolidone

GWG/ I19 can slow crystal growth in the bulk by a factor of 1 at 1 I Tg V 1? I/

?0/. This inhibitory effect is remarkable and suggests that low"concentration polymer

additives can substantially stabilize amorphous drugs against crystallization. Theirfinding was substantiated by Cai et al.E:/, who showed that the logarithm of the bulk

crystal growth rate decreases linearly with the concentration of GWG in weight percent

-ig. B/. At ? wt; GWG I19, the bulk crystal growth rate is slowed from .? mm@week to

.1 mm@year. Iestur et al. observed a similar linear relation between log uand w@w ;

GWG for crystal growth in li7uid felodipine containing GWG at temperatures substantially

above TgE8/. They reported a weaker inhibitory effect of GWG on crystal growth in li7uid

felodipine than the effect observed by Cai et al. for crystal growth in GWG"doped L(-

glasses -ig. B, the line labeled JbulkK/. This difference probably reflects the greater

power of a polymer dopant to inhibit crystal growth in a glass than in a low"viscosity

li7uid.

0ig. /

Gi'erent dependences of bul, and surface crystal growth rates in an H>0

glass on 4I4-31+ concentration

(t is noteworthy that the strong inhibitory effect of GWG is lost if its molecular weight is

reduced to that of a dimer. $hereas GWGs of different molecular weights have

comparable performance as crystallization inhibitors, the WG dimer has virtually no

inhibitory effect. This observation indicates the importance of high molecular weight for

an effective crystallization inhibitor. %ecause the WG dimer and the GWGs have similar

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR24http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR68http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR24http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR68http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR24http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR68http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR24http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR68http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/

8/12/2019 5_3516_832

16/18

18

interactions with L(- molecules, the analysis of JdirectK intermolecular interactions alone

is insufficient for predicting their effectiveness as crystallization inhibitors.

Although GWG additives can strongly inhibit bulk crystal growth in L(- glasses, their

effect on surface crystal growth is much weaker E:/. -igureBshows that both log usand

log ubdecrease linearly with increasing concentration of GWG I19 in weight percent,

log ubdecreases appro'imately three times faster than log us. As a result of the stronger

inhibition of bulk crystal growth, the thickness of the propagating surface crystal layer is

substantially thinner in the presence of GWG additives -ig.8/.

0ig. 6

J'ect of 4I4 on crystal growth in an H>0 glass at %1% 3. a4ure H>0. bH>0

containing 1 K wEw4I4-31+. ubL bul, growth rateM usL surface growth

rate. t8is the tie to start trac,ing crystal growth. aand bshare the

sae scale bar...

Cai et al. E:/ considered several e'planations for the weaker inhibition of surface crystalgrowth by polymer additives than bulk crystal growth 1/ lower polymer concentration at

the surface than in the bulk= ?/ upward growth of surface crystals making the process

less sensitive to polymer impurities= and / surface molecular mobility making polymers

less effective as crystal growth inhibitors. -urther work is needed to determine which

e'planation accurately accounts for the effect observed. !egardless of the e'planation,

the effect observed argues that it might be profitable to complement bulk doping with

surface stabilization in developing technologies to stabilize amorphous solids with

polymer additives.

Polymer (anocoatin) for *nhibitin) Surface Crystal Growth

$u et al. demonstrated that surface crystal growth on organic glasses can be inhibited

with a coating only a few nm thick 0,9/. Coatings of very different materials and

thicknesses have been found effective= for e'ample, 1 nm of gold and D? nm of

polymer deposited layer"by"layer through electrostatic assembly. >nder a coating, the

rate of surface crystal growth is decreased to that of bulk crystal growth. &ven the growth

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig8/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig8/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig8/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR34http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig8/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig7/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig8/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/figure/Fig8/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR69http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR34http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35

8/12/2019 5_3516_832

17/18

18

of e'isting surface crystals is halted. Although multiple layers of polyelectrolytes can be

deposited, a single layer proves sufficiently effective for inhibiting crystallization on (6C

glasses. The ultra"thin polymer coating still permitted fast dissolution of amorphous (6C,

while improving its wetting and flow.

The finding of $u et al.9/ supports the view that fast surface crystal growth of

amorphous (6C is enabled by the mobility of a thin layer of surface molecules, and that

this mobility can be suppressed by a coating of only a few nanometers thick. The

effectiveness of ultra"thin polymer coatings for halting surface crystallization suggests a

general way to stabilize amorphous drugs with the addition of only a small amount of

polymers. The a7ueous coating solutions used in electrostatic deposition of polymers are

compatible with drugs of low a7ueous solubility. +uch hydrophobic amorphous drugs

may dissolve and crystallize during a coating process that uses organic solvents.

Go to"

C"NC'$SI"NS

$e have reviewed some recent progress toward understanding the crystallization of

amorphous organic solids. -ast modes of crystal growth can emerge as organic li7uids are

cooled to become glasses. ne such growth mode, the glass"to"crystal or )C mode,

occurs in the bulk, and another fast growth mode e'ists at the free surface, both leading to

crystal growth rates orders of magnitude faster than predicted by theories that assume

diffusion defines the kinetic barrier of crystallization. These findings indicate that such

Jmolecular mobilityK measures as diffusivity, viscosity, and structural rela'ation time arepoor indicators of crystallization rates in organic glasses, and new theories are needed to

account for these phenomena. $ith the aid of polymorphs, recent studies have found that

)C growth favors more isotropically packed and denser crystal structures and is

kinetically similar to polymorphic conversion. Among the e'planations proposed for )C

growth, we favor the view that the process is solid"state transformation enabled by local

mobility in glasses.

(t is noteworthy that free surfaces of organic glasses can enhance not only crystal

nucleation a well anticipated effect/ but also crystal growth. +urface crystals on organic

glasses rise upward as they grow laterally, a growth mechanism that is unavailable tobulk crystals and that effectively utilizes higher surface molecular mobility. +tudies with

crystal polymorphs established that the degree to which crystal growth rate is enhanced

on going from the bulk to the surface is not controlled by the crystalDglass density

difference as predicted by the models of +chmelzer and coworkers and of Ionishi and

Tanaka. $e attribute fast surface crystal growth to surface molecular mobility.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/#CR35http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3385820/

8/12/2019 5_3516_832

18/18

18

The emergence of fast modes of crystal growth near the glass transition temperature

makes it invalid to predict the rates of crystallization in organic glasses by e'trapolating

the corresponding rates in the li7uid state. The importance of free surfaces in accelerating

the crystallization of amorphous drugs demonstrates that searches for molecular motions

responsible for crystallization must not be limited to bulk motions, and must includesurface mobility. There has been recent progress in measuring surface diffusion on

organic glasses, and the results indicate that surface diffusion can be orders of magnitude

faster than bulk diffusion and fast enough to support the surface crystal growth observed.

!ecent work has found that low"concentration polymer additives can be remarkably

effective in slowing bulk crystal growth in organic glasses, but their effect on surface

crystal growth is much weaker. (t was also discovered recently that ultra"thin polymer

coatings can inhibit surface crystallization, as well as improving the flow and wetting of a

hydrophobic drug. These results suggest the possibility of using low"concentration

polymer additives to stabilize amorphous drugs= for e'ample, a bulk additive to inhibitbulk crystallization and an ultra"thin surface coating to halt surface crystallization.

(n searches for effective polymers as crystallization inhibitors, attention has been paid to

JdirectK intermolecular interactions such as hydrogen bonding between drugs and

polymers. !ecent work, however, has highlighted the importance of the molecular weight

of the inhibitor. !elative to polyvinylpyrrolidone, the WG dimer has little effect on crystal

growth in nifedipine glasses. %ecause the dimer and the polymer have similar JdirectK

interactions with the drug, this finding argues that molecular weight is an important factor

for an effective crystallization inhibitor.

(mportant 7uestions remain concerning the stability of amorphous drugs against

crystallization. The mechanistic details are still lacking for fast crystal growth in the bulk

and at the surface of organic glasses, and for the emergence of fast modes of crystal

growth as organic li7uids are cooled to become glasses. (t is unclear what factors define

the degree to which crystal growth rate is enhanced on going from the interior to the

surface of an organic glass, and why fast surface crystal growth seems more prevalent for

organic glasses. The molecular motions responsible for crystallization in glasses remain

to be better understood. (t is unknown how different factors combine to define effective

crystallization inhibitors for amorphous drugs strength of JdirectK intermolecular

interactions, molecular weight, miscibility, and perhaps others. $e still do not know

whether the mechanism of crystal growth changes with increasing concentrations of

polymer additives. $ith better understanding of crystallization in organic glasses, more

accurate models may be formulated and more informative e'periments be conducted to

design amorphous pharmaceutical formulations with good physicochemical stability.