Crystallization of Organic Glasses

Ye Sun, Lei Zhu, Ting Cai, Hanmi Xi, Mariko Hasebe, Tian Wu, Hajime Ishida, Melgardt de

Villliers, Mark Ediger, and Lian Yu

University of Wisconsin – MadisonSchool of Pharmacy & Dept. of Chemistry

With thanks to: NSF, PRF, AstraZeneca

Crystalline and amorphous SiO2

Crystalline SiO2 (quartz)Density 2.65 g/cm3

Amorphous SiO2 (glass)Density 2.20 g/cm3

Crystalline and amorphous indomethacin

Amorphous solid (glass). Tg = 42 ºC More soluble than crystals (5 – 17 x)

Cl

N

CO2H

O

O

Figure from Chen, …, Stowell. JACS 2002. Solubilities from Hancock & Parks 2000, Murdande et al.. 2010, Alonzo et al. 2010

P21 , m.p. 153 ºC

P-1, m.p. 160 ºC

The glass transition region

Standard way to make a glass: Cooling a liquid without crystallization

Other ways: Grinding Freeze or spray dryingDrying crystalline

hydratesVapor deposition

Is there a problem?Can glasses really crystallize?

Obsidian: Natural silicate glass stable for millions of years – but eventually crystallizes

Ishida, Wu, and Yu. J. Pharm. Sci. 2007

A spherulite growing in amorphous nifedipine at 30 ºC (Tg – 12 ºC)u = 40 m/day

20 m20 m

HN

O

O

O

ON+

-O

O

(1)A bulk mode (GC) activated near Tg

(2)A surface mode

This talkFast crystal growth

in one-component organic glasses

-16

-15

-14

-13

-12

-11

-10

-9

-8

-7

-13

-12

-11

-10

-9

-8

-7

-6

-5

-4

-3

220 240 260 280 300 320 340

Log

u,

(m/s

)

T, K

Tg

Diffusion controlled

growth

Tm

Log

D,

(cm

2 /s)

▲ log u Magill & Li (1973)□ log D Mapes et al. (2006)○ log D Fujara et al. (1992)

u D

Crystal growth rate in a one-component liquid

OTP

Thermo-dynamic control

glass liquid

L

C

u

Wilson, Frenkel, Turnbull: Diffusion defines kinetic barrier of crystal growth

▲ log u Magill & Li● log u Hikima et al.□ log D Mapes et al.○ log D Fujara et al.

A new, fast growth mode (“GC”) is activated near Tg

OTP

• First observed for OTP by Greet and Turnbull in 1967

• Studied by Oguni and coworkers since 1995

• Unknown for inorganic and polymeric glass formers

-13

-12

-11

-10

-9

-8

-7

-6

-5

-4

-3

220 240 260 280 300 320 340

-16

-15

-14

-13

-12

-11

-10

-9

-8

-7

Log

u,

(m/s

)

T, K

Tg

“GC growth”Not controlled by

bulk diffusion

Diffusion controlled

growth

Tm

Log

D,

(cm

2 /s)

What does GC growth look like?

t0 + 40 min t0 + 60 mint0

100 m

Normal growth at 298 K

GC growth at 248 K

GC crystals are “real” crystals!• Same unit cell as normal crystals• Nearly the same m.p. and Hm

• Growth is also mainly along b

b

-11

-10

-9

-8

-7

-6

-5

-4

200 220 240 260 280 300 320 340

Scherer et al.Scherer et al. (protuberance)Magill & LiHikima et al.Greet &TurnbullGreetKonishi & TanakaThis work

Log

u, m

/s

T, K

Tg = 246 Kcompact

spherulites fibers

249 K

GC growth has fast-growing fibers as precursors

Compact growth

Fiber growth

100 m

Sun et al. J. Phys. Chem. B 2008, 112, 661; Xi et al. J. Chem. Phys. 2009, 130, 094508

Studying GC growth with polymorphs:From the same liquid/glass, which polymorph

shows GC growth, and which does not?

A B

liquid/glass

C

ROY polymorphs

(1) R P-1mp 106.2 oC= 21.7° ROY

(6) ORP Pbcamp 97 ºC= 39.4°

(4) OP P21/cmp 112.7 oC= 46.1°

(2) ON P21/cmp 114.8oC= 52.6°

(5) YN P-1, mp 99 ºC= 104.1°

(3) Y P21/cmp 109.8 oC= 104.7°

N

S

HNO O

C

N

CH3

JACS.2000

100 m

Y04

R

RL

JACS 2005a

(9) YT04 P21/cmp 106.9oC = 112.8°

(7) RPL

JACS 2001

(8) Y04

R05

Y04

L

(10) R05

JACS 2005b

Some ROY polymorphs show GC growth;some do not

Polymorphs of ROY

(1) R P-1mp 106.2 oC= 21.7° ROY

(6) ORP Pbcamp 97 ºC, = 39.4°

(4) OP P21/cmp 112.7 oC= 46.1°

(2) ON P21/cmp 114.8oC= 52.6°

(5) YN P-1, mp 99 ºC= 104.1°

(3) Y P21/cmp 109.8 oC= 104.7°

N

S

HNO O

C

N

CH3

J. Am. Chem. Soc.2000, 122, 585

100 m

Y04

R

RL

JACS 2005a

(9) YT04 P21/cmp 106.9oC = 112.8°

(7) RPL

JACS 2001

(8) Y04

R05

Y04

L

(10) R05

JACS 2005b

Sun, … J. Phys. Chem. B 2008, 112, 661 and 5594

-10

-8

-6

-4

-2

0

2

4

6

220 240 260 280 300 320 340 360 380

-11

-10

-9

-8

-7

-6

-5

-4

-3

220 240 260 280 300 320 340 360 380

T, K

Tg

No GC growth

YT04 GC growth

ON (No GC growth)

Log

u, (

m/s

)

Log

u, (

m/s

)

ON

YT04

YN

Y

OP

R

+ = R05

y shift

+11

+8

+5

+2

a

b

T, K

Tt

GC growth (compact

spherulites)

fibers

YT04 fibers

Tt

Anisotropic packing: No GC growth

More isotropic packing: GC

growth

Radial distribution functions

Crystal structures showing GC growth are more “liquid like”

Num

ber

of m

olec

ules

YN

ON

YT04

Y

0

1

2

3

3 4 5 6 7 8 9 10

01234

3 4 5 6 7 8 9 10

01234

3 4 5 6 7 8 9 10

01234

3 4 5 6 7 8 9 10

01234

3 4 5 6 7 8 9 10

OP

R0

1

2

3

3 4 5 6 7 8 9 10

r, Ǻ

261 K

295 K

100μm

100 μm

R

YN

L/G R2

GC growth is similar kinetically to polymorphic conversion, but “feels” the glass transition

Tg

Log

u, (

m/s

)

T, K

-11

-10

-9

-8

-7

-6

220 240 260 280 300

YN R

L YN

G R(GC)

Origin of GC growth: Still an open question

• Bulk relaxation. But the process is absent in ROY and aged away in OTP

• Tension at crystal/glass interface. But fibers grow rapidly above Tg, and no expected “autocatalysis”

• Solid-state transition similar to polymorphic conversion. But no predictive power

• Molecular mobility at grain-boundaries. But no predictive power

• …

(1)A bulk mode (GC) activated near Tg

(2)A surface mode

This talkFast crystal growth

in one-component organic glasses

“Whereas in many metallic glasses nucleation has been observed to be enhanced at the surface, growth rates are usually quite comparable with those in the bulk.” U. Koster (Mat. Sci. & Eng. 1988, 97, 233)

[For silicate glasses,] “The crystal growth velocities of crystals in the volume and of the surface layer in the glass volume, as well as of isolated crystals on the glass surface are equal.” Diaz-Mora et al. (J. Non-Crystalline Solids 2000, 273, 81)

Crystal growth in the bulk

Crystal growth at the free surface

glasscrystal glass

crystal

-11

-10

-9

-8

-7

-6

280 300 320 340 360 380 400 420 440

For organic glasses, crystal growth can be much faster at the surface than in the bulk

Wu, T.; Yu, L. J. Phys. Chem. B 2006, 100, 15694; Pharm. Res. 2006, 23, 2350

Tg

Cover Glass

Cover GlassGlass Crystal

Cover GlassGlass Crystal

Log

u, m

/s

T, K

Crystal growth rate ofIMC ( polymorph)

bulk

surface

What do surface crystals look like?

a

c

1500nm

b

2µm

200µm 300nm

a

1µm

b

c

400µm

Growth front is above glass surface

Sun et al. PNAS, in press

IMC IMC

Growth of IMC surface crystals at 22 ºC

0 2 h 4 h

6 h 8 h

5 m

Mariko Hasebe

Surface crystallization of amorphous IMC can be inhibited with a nanocoating

Wu, Sun, Li, de Villiers, and Yu. Langmuir 2007, 23, 5148

200μm 200μm

Uncoated sample

7 days at 40 ºC

Coated with 10 nm gold or 2 – 30 nm polymer

7 days at 40 ºC

15 mm

t = 96 hr

400µm

194 hr 298 hr

a-IMC

air

-IMC

Surface crystal layer can be quite thin

250µm

d=480nm 410nm 220nmd=10µm 180nm 135nm 100nm

IMC glass

cover glass (22×22 mm2 )

cover glass (18×18 mm2 )

scratch

scra

tch

Glass Crystal

Si

d = 50 nm – 15 m

Expt. 1: Direct observation

Expt. 2: Vary glass thickness d Sun et al.

PNAS, in press

No effect New morphology, slower growth

Surface diffusion is fast on amorphous IMC

Glass

Mobile surface

Lei Zhu-21

-19

-17

-15

-13

290 300 310 320 330 340 350

log

D, m

2 /s

T, K

Tg

Ds (surface)

Dv (bulk)

b

Origin for surface-enhanced crystal growth: Still an open question

• Surface mobility

• Surface crystals can exploit opportunity to grow upward

• Tension from crystal growth is better released at the surface than in the bulk

• …

All these models imply generality of the phenomenon. But is it?

GC growth• A new bulk growth mode is activated near Tg.

It is not limited by bulk diffusion• Favors “liquid-like” structures• Similar to polymorphic transformation

Surface growth• Surface enhances growth, not just nucleation• Inhibited by nano-coating• Correlates with surface molecular mobility

Both modes are known only or mainly for organic glasses

Summary: Fast crystal growth in organic glasses