Geometries and Excited States MUCCC34 - the 17th MU3C ...discus/muccc/muccc34/Bradley-MUCCC34pd… · Entry Isomer Starting Geometry Final Geometry Energy (kJ/mol) 1 Cis Hughes’

Geometries and Excited States of Long Wavelength BF

2-Azo Dyes

to Guide Target Selection

Colin D. BradleyFreshman Double Major in Biochemistry & Molecular Biology and Spanish

Gillmore Research Group, Department of Chemistry, Hope CollegeHolland, MI

[email protected], [email protected]

MUCCC34 - the 17th MU3C Online Virtual Winter Conference - February 4-6, 2020

Azo Dyes

● Conventional azo dyes are generally Ar-N=N-Ar’ where Ar is some sort of aryl group.

● Robust, easy to synthesize, very stable.

● Dyes change SHAPE substantially when trans/cis isomerization occurs with light and/or heat.

● Absorbance changes generally much more modest (e.g., not sufficient for photochromic lenses).

● Known for a very long time (1863), with uses in a range of dyes, including food dyes since 1906.

● Renewed interest since 1960s, and especially since late 1990s, for their shape change within

structured polymer or liquid crystalline materials (e.g., photomechanical or photonastic materials

that bend in response to light.)

● Limited by the relatively narrow short wavelength absorbance window of the dyes.

https://www.fda.gov/about-fda/fdas-evolving-regulatory-powers/part-i-1906-food-and-drugs-act-and-its-enforcementhttps://en.wikipedia.org/wiki/Azo_dye

Azo Dyes

Energy

UV IR

Conventional Azo Dyes

Higher energy light:• Competitive

Absorption• Photodegradation• Biologically

Incompatible

Longer Wavelength Azo Dyes

Energy

UV IR

Conventional Azo Dyes Aprahamian’s BF2-Azo Dyes

Higher energy light:• Competitive

Absorption• Photodegradation• Biologically

Incompatible

1. Yang, Y.; Hughes, P.; Aprahamian, I. J. Am. Chem. Soc. 2012, 134, 15221-15224.

2. Yang, Y.; Hughes, P.; Aprahamian, I. J. Am. Chem. Soc. 2014, 136, 13190-13193.

We can repeat Aprahamian’s synthesis...

and Aprahamian’s photochemistry & spectroscopy.

Trans to Cis Isomerization - 350 W Hg Arc Lamp with RG610 Long-Pass Filter

cis transcis trans

We can even improve upon the synthesis!● First two steps shortened to one!● Higher yields!● Lower temperature, shorter time, cyanation with NaI catalysis● And we can use substituted quinaldines

Aprahamian’s Dyes and Ours

Entry R1

R2

trans lambdamax

(nm) cis lmax

(nm)

1 H H 530 4752 H OMe 594 5253 H NMe

2680 ~630

4 H pyrrolidinyl 690 ~6505 H piperidinyl 681 ~6306 H methylpiperazinyl 661 ~6007 H morpholinyl 651 ~590

8 H Et 545 487

9 H C≡CH 542 482

10 H Br 537 479

11 Br Br 547 492

12 Br MeO 607 531

13 C≡CH C≡CH tbd tbd

14 C≡CH Br tbd tbd

15 C≡CH MeO tbd tbd

We were (perhaps obviously) trying to add synthetic handles (alkynes, aryl halides) to incorporate the dyes into polymers via Sonogashira, Thiol-yne, and Azide Click coupling reactions.

That was part of a collaboration toward longer-wavelength photomechanical materials that has stalled a bit both on our collaborators side (polymer substrate and testbed) and ours.

But along the way we discovered ways to put substituents on the previously unstudied quinoline ring, and saw they had similar effects as substituents on the phenyl ring (e.g., compare entries 1 vs 10 vs 11, and 2 vs 12, on the previous slide).

So now we’ve got an ACS PRF grant to...

Prepare & Study a Diverse Library of Dyes!

● Preliminary evidence shows that EDG’s have similar effects on both rings, and that the effects are additive.

● We want to study positions other than just the 6-position we’ve studied so far. (We’ll forgo the 3- and 8-positions to avoid steric and/or neighboring group participation effects.)

● 6 phenyl substituents x 8 quinoline substituents x 4 quinoline positions = 192 possibilities!

● Sometimes, push-pull systems can absorb at even longer wavelengths.● This may also weaken the N=N bond slightly and improve quantum yield

and thermal reversion rate!● 120 more possibilities!!!

Prepare & Study a Diverse Library of Dyes!

● Aprahamian & Hughes actually showed by TD-DFT that the cyano group, necessary for the synthesis, is a liability in blue-shifting the dye, relative to a hypothetical H or OH.

● We wonder if we can convert the CN to something else after the synthetic step in which it is necessary.

● Many more possible targets!!!!!

Prepare & Study a Diverse Library of Dyes!

So what am *I* doing???

● We need to use computation to select a dozen or two most attractive and diverse targets.

● Aprahamian’s computational collaborator, Russell Hughes, computed geometries using

B3LYP/6-311G**++ in Jaguar, and excited states and UV/Vis absorbance using

TDDFT(B3LYP/TZ2P) or (B3LYP-D3/TZ2P) in ADF, all in vacuo.

● I am working to adapt his methods to use WebMO (18.1) and Gaussian (16 rev B.01) on

the MU3C cluster.

● I’m just getting started, but I’m working to:

○ ensure that I can get as good DFT geometries as possible (so far varying input

geometry and basis set), or at least as are neccessary, and to

○ conduct TDDFT computations to predict UV-Vis spectra (especially the longest

wavelength bands of both cis & trans) with as good agreement with Hughes, and

more importantly with experiment, as possible!

● Secondarily I hope to minimize compute times required and see if less

computationally expensive methods are feasible, particularly for geometries.

Geometries#N B3LYP/6-31G(d) OPT Geom=Connectivity, #N B3LYP/6-311+G(2d,p) OPT Geom=Connectivity

Running geometry optimization on a trans isomer, with one job using 6-311+G(2d,p) and the other using 6-31G(d) both from the same starting geometry (entries 2 and 6), produced two molecules with the phenyl moiety and BF

2 group slightly off from each other.

Running geometry optimization in 6-311+G(2d,p) on two cis starting points (entries 7 and 8), one with the phenyl moiety nearly flat with the rest of the molecule and the other with the phenyl moiety nearly bisecting, produced a pair of enantiomers, but essentially identical geometries/energies.

Geometries#N B3LYP/6-31G(d) OPT Geom=Connectivity, #N B3LYP/6-311+G(2d,p) OPT Geom=Connectivity

Entry Isomer Starting Geometry Final Geometry Energy (kJ/mol)

1 Cis Hughes’ and Aprahamian’s A -2883068.1644720722 Trans Hughes’ and Aprahamian’s B -2883087.545074393 Cis Everything “Wiggled” C -2883068.1643145424 Trans Everything “Wiggled” D -2883087.545126898

5 Cis Hughes’ and Aprahamian’s E -2883911.10956166

6 Trans Hughes’ and Aprahamian’s F -2883926.21049363

7 Cis Phenyl Flat G -2883911.10958792

8 Cis Phenyl Bisecting H -2883911.10956166

A C E G

B D F H

Computed Absorbance Spectra#N TD(NStates=50) B3LYP/cc-pVTZ Geom=Connectivity

Cis/Trans R1

R2

Hughes’ Calc (nm) Our Calc (nm) Experimental (nm)

Cis* H H 482 481 475

Trans* H H 510 512 530Cis** H H (482) 479 475

Trans** H H (510) 509 530Cis H H (482) 484 475

Trans H H (510) 515 530

* These calculations were performed directly on Hughes’ and Aprahamian’s reported geometries. ** These calculations were performed on geometries started from Hughes’ but reoptimized in Gaussian B3LYP/6-311+G(2d,p).

All other calculations were performed on geometries started from Hughes’ but reoptimized in Gaussian B3LYP/6-31G(d).

Computed Absorbance Spectra#N TD(NStates=50) B3LYP/cc-pVTZ Geom=Connectivity

Cis/Trans R1

R2

Hughes’ Calc (nm) Our Calc (nm) Experimental (nm)

Cis H H (482) 484 475Trans H H (510) 515 530

Cis H OMe 517 525Trans H OMe 536 594

Cis H NMe2 556 630

Trans H NMe2 565 680

Cis H Br 495 479Trans H Br 526 537

Cis Br OMe 529 531Trans Br OMe 548 607

Cis OMe OMe 531 ?Trans OMe OMe 558 ?

Cis Acetyl OMe 502 ?Trans Acetyl OMe 541 ?

Calculations were performed on geometries started from Hughes’ but reoptimized in Gaussian B3LYP/6-31G(d).

Future WorkPretty much all of it is “Future Work” - we’ve really only been at this for about 2 weeks!

● Clearly our values are not yet predictive / useful!

● Perhaps Hughes is right that higher level geometry calculations are required?

● We will also explore the use of solvent models on both geometry and excited state

calculations, which Hughes said were not necessary. (However we find it interesting

that while Aprahamian & Hughes state that DFT calculations of absorbance guided

their choice of substrates, they only report geometries not wavelengths in their

supporting info for the paper where they varied the substituents!)

● We can also consider varying the basis set and/or functional for geometry and/or

excited state calculations, or more nuanced parameters.

● Eventually, once we have computed absorbance values that agree well with

experiment for compounds prepared to date, we will use computations to help

narrow the several hundred potential targets to a more manageable dozen or two to

synthesize and study experimentally over the next 2-3 years.

Questions for the MU3C community…Remember… I’m a freshman, and my mentor is an experimentalist and total hack at computation!

● How do folks determine if they are getting the “same” (or “same enough”) geometry from OPT calcs at different basis sets or even functionals (so energies are different)? Visually? Run energy only calcs on both OPT outputs with same functional/basis set? Manual graphical overlays? Some set of bond angles, dihedrals, bond lengths, etc? Some global least squares method? What tools do you recommend?

● What should we try first to improve our calcs agreement with experiment?○ Is it more likely a geometry issue or excited states issue?○ Higher level geometries? Different functional? Solvent in geometries? ○ Different level TDDFT? Different functional? # of states? Solvent in excited

state calculations?

● What other questions SHOULD we be asking?

Acknowledgments• Mentor: Prof. Jason G. Gillmore• Current JGG Grp labmates (synthesis on this project): Thomas Cygan & Ethan Cramer• Current Funding: ACS Petroleum Research Fund undergraduate research grant #60174-UR1, and MU3C (NSF

MRIs OAC-1919571, CHE-1039925, CHE-0520704)• Helpful computational conversations with: Professors Hughes (Dartmouth), Alvarado (UWRF), D’Acchioli

(UWSP), Lord (GVSU), Polik & Krueger (Hope) - MU3C folk rock!

• Former work on BF2-Azo Incorporation into Polymers:

○ Collaborator: Prof. Matthew L. Smith, Hope College Engineering Department○ Gillmore / Smith Undergraduate Research Team:

Brandon Derstine, Sean Gitter, Marcus Brinks, Addison Duda, Brian Simonich,Emily Hofmeyer, Connor Kuhlmann, Alyssa VanZanten, Jessica KorteAleksandra Masiak, Jessica Scott, Drew Bennett, Nick Olen CHEM 256B projects: 8 of the above, plus Jonah Kaveh, Dan Clark

○ Funded by: Hope College Chemistry Department Undergraduate Research Fund & Schaap Research Fellows program, Beckman Scholars Program, Michigan Space Grant Consortium Seed Grant, Henry Dreyfus Teacher-Scholar Award, and Prof. Matt Smith’s Start-up funds (Hope College Engineering Department, and Dean of Natural & Applied Sciences)