phase stability and martensitic transitions in niti from ...€¦ · next generation resins for...

Phenolic Polymer Interactions with Water and Ethylene Glycol Solvents

Justin B. Haskins,1 Eric W. Bucholz,2 Charles W. Bauschlicher,3 Joshua D. Monk,1 John W. Lawson3

1AMA, Inc., Thermal Protection Materials Branch, NASA Ames Research Center2EAP, Thermal Protection Materials Branch, NASA Ames Research Center3Thermal Protection Materials Branch, NASA Ames Research Center

256th ACS National Meeting | August 19-23, 2018 | Boston, MA

Phase Stability and Martensitic Transitions in NiTi from First Principles Simulations

Justin Haskins1 and John Lawson2

1AMA Inc., Thermal Protection Materials Branch, NASA Ames Research Center2Thermal Protection Materials Branch, NASA Ames Research Center

147th Annual Meeting of The Minerals, Metals, and Materials Society | March 11-15, 2018 | Phoenix, AZ

https://ntrs.nasa.gov/search.jsp?R=20180006646 2020-01-17T22:41:55+00:00Z

Ablative Composites for Re-entry(carbon fiber/phenolic matrix)

Stackpoole, et al. AIAA (2008)

Ablative Heat Shields

2





American Institute of Aeronautics and Astronautics

5

Figure 3. Photograph of SRC after landing and prior to recovery (courtesy Scott Sandford).

Figure 4. Photograph of SRC transit from Wendover Airfield, UT with star map overlay (courtesy Bruce

Fischer).

V. Analysis

Several analysis activities have been initiated towards achieving the objectives listed previously and leveraging

the aforementioned data sources,. These activities are described herein along with a description of results. Detailed

reports describing the various analyses are referenced throughout.

A. Trajectory

The essential first step in the post-flight analysis is an estimate of the as-flown trajectory. The trajectory

information provides flight conditions (altitude, velocity, and orientation) from which the aero/thermal environment

is then assessed. Key data sources are the in-space navigation, airborne observation, ground photos, and terminal

descent radar. As described in Ref. 10, the program POST11

is used to propagate trajectories from the entry interface

condition through the beginning of radar acquisition to the point of drogue chute deployment. Nominal atmospheric

or aerodynamic properties are adjusted to reproduce the known trajectory state conditions.

In addition to the above, there are long-duration exposures of the SRC as it streaked across the sky. Using the

starfield as reference and knowing the location of the viewpoint, two ground images are used to determine the flight

Stardust

Mars Science Lander

Next Generation Resins for Heat Shields

2





P a g e | 24

Figure 1. Structures of phenolic chains considered in this study illustrating (a) ortho-ortho repeating novolac-type (OON) phenolic chain indicating ortho and para linking sites on a phenol ring, (b) ortho-para repeating novolac-type (OPN) phenolic chain, and (c) ortho-ortho repeating resole-type (RES) phenolic chain. Images on the right are the corresponding 9-phenol ring chains, and the chains are colored as: carbon backbone (red), non-backbone carbon (cyan), oxygen (yellow), and hydroxyl hydrogen (white). Translucent bonds indicate intramolecular h-bonding. Aromatic and methylene hydrogens are removed for clarity.

Phenolic (SOA)

Polyimides

Cyanate Esters

New resin chemistries for heat shields require different solvents for processing

4

P a g e | 24


Phenolic Polymers




147th Annual Meeting of The Minerals, Metals, and Materials Society | March 11-15, 2018 | Phoenix, AZortho-ortho novolac

ortho-para novolac

ortho-ortho resole

Design rules for SOA polymer and solvents

Polymers Solvents

ethylene glycol

water

H2O

FIG. 9: A nine segment phenolic model surrounded by a shell of water molecules. The

phenolic molecule has folded so the head and tail are adjacent.

36

Phenolic in Water

Quantum Chemical Calibration: understand basic polymer-solvent interactions and benchmark MD models

� combination of DFT, MP2, CCSD(T)

� water and ethylene glycol dimers

� solvent-monomer dimers

Molecular Dynamics Simulation: characterize polymer solubility in solvents

� OPLS-AA-SEI force field

� single polymers in large solvent boxes

� 500-100,000 solvent molecules

� polymers with 3-27 units

Outline





FIG. 5: Local minima for R2eg B species for di↵erent levels of theory. The relative energetics

are given for each level of theory in kcal/mol.

32

Ethylene glycol-phenolic dimer

7

Water and Ethylene Glycol Interactions





FIG. 2: The water and ethylene glycol dimers. For water, the hydrogen bond is shown. For

ethylene glycol, the O-H distance is only slightly longer than for water, 1.96 vs 1.92 A, but

the HOH angles are less than 90 degrees; therefore they are not consistent with classical

hydrogen bonds, and the O-H interactions are noted with a smaller line.

29

OPLS energetics within 2 kcal/mol of CCSD (T)

Water and Ethylene Glycol Dimers

FIG

.8:

Twoview

sof

the10

conform

ationsof

ethy

leneglycol.

35

FIG

.8:

Twoview

sof

the10

conform

ationsof

ethy

leneglycol.

35

Conformers of Ethylene Glycol

H = black; O = red; C = gray

8

Phenolic-Water Interactions





OPLS interactions within 2 kcal/mol of MP2/CBS

FIG. 3: Water bonded to the 2 segment phenolic models.

30

FIG. 3: Water bonded to the 2 segment phenolic models.

30

2 Unit Resole2 Unit Novolac 3 Unit Novolac

FIG. 4: Water bonded to the 3 segment phenolic model.

31


9

Phenolic-Ethylene Glycol Interactions





FIG. 6: Ethylene glycol bonded to the 2 segment phenolic models.

33

2 Unit Novolac

OPLS interactions within 3 kcal/mol of MP2/CBS

FIG. 6: Ethylene glycol bonded to the 2 segment phenolic models.

33

FIG. 7: Ethylene glycol bonded to the 3 segment phenolic model.

34

2 Unit Resole 3 Unit Novolac


FIG. 9: A nine segment phenolic model surrounded by a shell of water molecules. The

phenolic molecule has folded so the head and tail are adjacent.

36

Phenolic in Water

Quantum Chemical Calibration: understand basic polymer-solvent interactions and benchmark MD models

� combination of DFT, MP2, CCSD(T)

� water and ethylene glycol dimers

� solvent-monomer dimers

Molecular Dynamics Simulation: characterize polymer solubility in solvents

� OPLS-AA-SEI force field

� single polymers in large solvent boxes

� 500-100,000 solvent molecules

� polymers with 3-27 units

Outline





FIG. 5: Local minima for R2eg B species for di↵erent levels of theory. The relative energetics

are given for each level of theory in kcal/mol.

32

Ethylene glycol-phenolic dimer

Phenolic in Water Phenolic in Ethylene Glycol

10

Dynamics of Solvated Phenolic





Diffusion and viscosity of solvent strongly affect polymer dynamics

4. DISCUSSION

618 From a conformational perspective, the three types of low619 molecular weight phenolic chains considered here do not620 qualitatively differ when solvated in either EGL or H2O, as is621 evidenced by the collapsed oligomer structure that is illustrated622 in Figure 1 and from the Rg provided in Figure 2. This indicates623 that the solvation differences between EGL and H2O must624 result from the differences in the local H-bonding environment625 as well as the intermolecular interactions present within each626 phenolic chain’s solvation shell. Through structural analysis of627 the solvation shell, differences between the OON, OPN, and628 RES type phenolic chains are observed, as demonstrated in

629Figure 5. These observations suggest that the OON chain630interacts with EGL primarily through [010] and [001] type631hydrophobic interactions, whereas OPN and RES preferentially632interact through [100] type hydrophilic interactions. However,633OPN and RES also yield [010] and [001] type hydrophobic634interactions similar to those observed for OON but to a lesser635degree. In addition, the analysis of the H2O solvated systems636yields the same observations, with the numbers of hydrophilic637[100] type interactions increasing from OON to OPN to RES638type phenolic chains. The analysis of the H-bonding environ-639ment within each solvated phenolic system, detailed in Tables 6640and 7, supports the solvation shell observations, showing that

Figure 7. MSDs of the nine-ring phenolic chains in solution. Solvents are EGL (a, c, e) and H2O (b, d, f). Solutes are OON (a, b), OPN (c, d), andRES (e, f).

The Journal of Physical Chemistry B Article

DOI: 10.1021/acs.jpcb.7b00326J. Phys. Chem. B XXXX, XXX, XXX−XXX

K

Polymer Diffusion





4. DISCUSSION

618 From a conformational perspective, the three types of low619 molecular weight phenolic chains considered here do not620 qualitatively differ when solvated in either EGL or H2O, as is621 evidenced by the collapsed oligomer structure that is illustrated622 in Figure 1 and from the Rg provided in Figure 2. This indicates623 that the solvation differences between EGL and H2O must624 result from the differences in the local H-bonding environment625 as well as the intermolecular interactions present within each626 phenolic chain’s solvation shell. Through structural analysis of627 the solvation shell, differences between the OON, OPN, and628 RES type phenolic chains are observed, as demonstrated in

629Figure 5. These observations suggest that the OON chain630interacts with EGL primarily through [010] and [001] type631hydrophobic interactions, whereas OPN and RES preferentially632interact through [100] type hydrophilic interactions. However,633OPN and RES also yield [010] and [001] type hydrophobic634interactions similar to those observed for OON but to a lesser635degree. In addition, the analysis of the H2O solvated systems636yields the same observations, with the numbers of hydrophilic637[100] type interactions increasing from OON to OPN to RES638type phenolic chains. The analysis of the H-bonding environ-639ment within each solvated phenolic system, detailed in Tables 6640and 7, supports the solvation shell observations, showing that

Figure 7. MSDs of the nine-ring phenolic chains in solution. Solvents are EGL (a, c, e) and H2O (b, d, f). Solutes are OON (a, b), OPN (c, d), andRES (e, f).



K

Mea

n Sq

uare

Disp

lace

men

t (nm

2 )

Simulation Time (ns) Simulation Time (ns)

Larger diffusion coefficients in water

P a g e | 24


ortho-ortho novolac

P a g e | 24


ortho-ortho novolac

12

13

Solvation Free Energy





Polymers more soluble in ethylene glycolResole most soluble polymer

Ethylene Glycol -27.1 -53.1 -84.0Water -8.5 -23.8 -46.5

P a g e | 24


ortho-ortho novolac

P a g e | 24


ortho-para novolac

P a g e | 24


ortho-ortho resole

Solvation Free Energy (kcal/mol)

14

P a g e | 29

Figure 3. First solvation shell of the OPN chain in a) EGL and b) H2O solvents. Hydrogen bonds are denoted in red.

Solvation Shell of Polymer





P a g e | 29

Figure 3. First solvation shell of the OPN chain in a) EGL and b) H2O solvents. Hydrogen bonds are denoted in red.

Solvation structure governs properties

Ethylene Glycol Water

polymer = gray; O = yellow; C = green; H = white

15

Solvated Polymer Structure





Self hydrogen bonding in ortho-ortho systemsOrtho-para novolac and ortho-ortho resole have free –OH groups

Pa

ge

| 2

4

Fi

gure

1.

Stru

ctur

es o

f ph

enol

ic c

hain

s co

nsid

ered

in

this

stu

dy i

llust

ratin

g (a

) or

tho-

orth

o re

peat

ing

novo

lac-

type

(OON

) phe

nolic

cha

in in

dica

ting

orth

o an

d pa

ra li

nkin

g si

tes

on a

phe

nol

ring,

(b) o

rtho-

para

rep

eatin

g no

vola

c-ty

pe (O

PN) p

heno

lic c

hain

, and

(c)

orth

o-or

tho

repe

atin

g re

sole

-type

(RES

) phe

nolic

cha

in. I

mag

es o

n th

e rig

ht a

re th

e co

rres

pond

ing

9-ph

enol

ring

cha

ins,

and

the

chai

ns a

re c

olor

ed a

s: c

arbo

n ba

ckbo

ne (

red)

, no

n-ba

ckbo

ne c

arbo

n (c

yan)

, ox

ygen

(y

ello

w),

and

hydr

oxyl

hyd

roge

n (w

hite

). Tr

ansl

ucen

t bon

ds in

dica

te in

tram

olec

ular

h-b

ondi

ng.

Aro

mat

ic a

nd m

ethy

lene

hyd

roge

ns a

re re

mov

ed fo

r cla

rity.

ortho-ortho novolac ortho-para novolac ortho-ortho resole

16

Specific Types of Polymer-Solvent Bonding





P a g e | 30

Figure 4. Illustrations of different interaction types between solvent molecules and solute element groups using a binary descriptor (1 = present, 0 = not present); specifically a) phenolic hydroxyl and solvent, denoted [100], b) center of phenolic ring and solvent, denoted [010], and c) phenolic methylene linker and solvent, denoted [001].

Figure 5. Average number of solvent/phenolic interactions during simulations of phenolic solvated in EGL. Binary code employs naming scheme illustrated in Fig. 4. Inset shows phenolic solvated in H2O.

Three primary interactions found in polymer-ethylene glycol solvation; one type in polymer-water solvation

O = red; C = green; H = white

17

P a g e | 30

Figure 4. Illustrations of different interaction types between solvent molecules and solute element groups using a binary descriptor (1 = present, 0 = not present); specifically a) phenolic hydroxyl and solvent, denoted [100], b) center of phenolic ring and solvent, denoted [010], and c) phenolic methylene linker and solvent, denoted [001].

Figure 5. Average number of solvent/phenolic interactions during simulations of phenolic solvated in EGL. Binary code employs naming scheme illustrated in Fig. 4. Inset shows phenolic solvated in H2O.

Specific Types of Polymer-Solvent Bonding





Hydrogen bonding and carbon linker interactions dominate

P a g e | 24


linker

Hydrogen Bonding





Hydrogen bonding common to both solvents and most prevalent bonding18O = red; C = green; H = white

19

517 intermittent autocorrelation function. Unlike the continuous518 definition, with the intermittent definition, Sij(t + t0) is519 permitted to transition freely between 1 and 0 for all times t520 + t0.521 Using the intermittent autocorrelation function, the stabilities522 of the three types of H-bonds at 298 K within the solvated

f6 523 systems are illustrated for OON in Figure 6a,b, OPN in Figure524 6c,d, and RES in Figure 6e,f. In addition, solvated EGL and525 H2O systems are represented in Figure 6a,c,e and b,d,f,526 respectively. As can be seen, intramolecular “P to P” H-bonds527 within OON and RES are mostly constant and long lived,528 whereas they decay quickly within OPN. The two intermo-

529lecular H-bond types, “P to S” and “S to P”, decay much faster530in H2O than in EGL. It is also observed that “P to S” type H-531bonds generally decay slowly in both H2O and EGL than “S to532P” type H-bonds. The observed asymmetry between “P to S”533and “S to P” type H-bonds is likely due to the differences in the534partial charges of each hydroxyl type in the solvated systems.535For intermolecular H-bonds involving OPN and OON chains,536the partial charge difference is greater when the phenol537hydroxyl is the H-bond donor than when it is the H-bond538acceptor, which leads to an increased stability of “P to S” type539H-bonds. For RES, on the other hand, the partial charge540differences are less pronounced when the methylol hydroxyls

Figure 6. Hydrogen bond autocorrelations using intermittent definition for nine-ring solvated phenolic simulations at 298 K. Solvents are EGL (a, c,e) and H2O (b, d, f). Solutes are OON (a, b), OPN (c, d), and RES (e, f). Legend notes: P and S refer to phenolic and solvent OHs, respectively.



I

Polymer-Solvent Hydrogen Bonds





517 intermittent autocorrelation function. Unlike the continuous518 definition, with the intermittent definition, Sij(t + t0) is519 permitted to transition freely between 1 and 0 for all times t520 + t0.521 Using the intermittent autocorrelation function, the stabilities522 of the three types of H-bonds at 298 K within the solvated

f6 523 systems are illustrated for OON in Figure 6a,b, OPN in Figure524 6c,d, and RES in Figure 6e,f. In addition, solvated EGL and525 H2O systems are represented in Figure 6a,c,e and b,d,f,526 respectively. As can be seen, intramolecular “P to P” H-bonds527 within OON and RES are mostly constant and long lived,528 whereas they decay quickly within OPN. The two intermo-

529lecular H-bond types, “P to S” and “S to P”, decay much faster530in H2O than in EGL. It is also observed that “P to S” type H-531bonds generally decay slowly in both H2O and EGL than “S to532P” type H-bonds. The observed asymmetry between “P to S”533and “S to P” type H-bonds is likely due to the differences in the534partial charges of each hydroxyl type in the solvated systems.535For intermolecular H-bonds involving OPN and OON chains,536the partial charge difference is greater when the phenol537hydroxyl is the H-bond donor than when it is the H-bond538acceptor, which leads to an increased stability of “P to S” type539H-bonds. For RES, on the other hand, the partial charge540differences are less pronounced when the methylol hydroxyls

Figure 6. Hydrogen bond autocorrelations using intermittent definition for nine-ring solvated phenolic simulations at 298 K. Solvents are EGL (a, c,e) and H2O (b, d, f). Solutes are OON (a, b), OPN (c, d), and RES (e, f). Legend notes: P and S refer to phenolic and solvent OHs, respectively.



I

Hydr

ogen

Bon

d Au

toco

rrel

atio

n

Simulation Time (ns) Simulation Time (ns)

P a g e | 24


ortho-ortho novolac

Hydrogen bonding more persistent in ethylene glycol

P a g e | 24


ortho-ortho novolac

20

•OPLS-AA-SEI energetics agree with high accuracy CCSD(T) solvent computations

•OPLS-AA-SEI polymer-solvent interactions within a few kcal/mol of MP2/CBS computations

•Ethylene glycol more readily solvates the polymers because of more, longer-lived hydrogen bonding than water

•Resole is more soluble than the novolac polymers: more hydrogen bonding and hydrophobic-hydrophobic interactions

Conclusions





Bucholz, et al. JPCB 121, 2839 (2017)Bauschlicher, et al. JPCB 121, 2852 (2017)

Questions?





Bucholz, et al. JPCB 121, 2839 (2017)Bauschlicher, et al. JPCB 121, 2852 (2017)

phase stability and martensitic transitions in niti from ...€¦ · next generation resins for...

Documents