New molecular materials for nonlinear optics: Preparation and detailed characterisation

Masters ThesisN E W M O L E C U L A R M AT E R I A L S F O R N O N L I N E A R O P T I C S

Preparation and detailed characterisation

(Nov molekulrn materily pro nelinern optiku: Prprava a detailncharakterizace)

J A N A M AT H A U S E R O V

S T U D Y P R O G R A M: ChemistryS T U D Y C O U R S E: Inorganic Chemistry

Department of Inorganic ChemistryFaculty of Science

Charles University in Prague

S U P E RV I S O R S:Doc. Ivan Nemec, PhDRNDr. Filip Uhlk, PhD

P R A G U E , 2 0 1 3

D E C L A R AT I O N : Hereby I declare that this thesis is a result of myown work and effort and that all my intellectual debts are acknowl-edged with due reference to literature or otherwise. Neither this thesisnor its significant part has been previously submitted for any degree.This master thesis was written in relation to the long term researchplan MSM0021620857 of the Ministry of Education of the Czech Re-public.

P R O H L E N : Prohlauji, e jsem zverecnou prci zpracovalasamostatne a e jsem uvedla vechny pouit informacn zdroje a lit-eraturu. Tato prce ani jej podstatn cst nebyla predloena k zsknjinho nebo stejnho akademickho titulu.Tato diplomov prce vznikla v souvislosti s reenm vzkumnhozmeru MSM0021620857.

Prague, 2/9/2013

Jana Mathauserov

ii

A B S T R A C T

In this thesis both an experimental and computational study of selected prospective materialsfor nonlinear optics is presented. The main focus was put on 2aminopyrimidinium hydrogen-phosphite (AMPPO3), a salt with high SHG efficiency and Type I. phase matching properties.AMPPO3 had been studied by our research group and its favourable qualities such as easypreparation, colourlessness, great water solubility and good crystallinity motivated further in-terest and the attempts for the growth of bulk single crystals from thermostated aqueous solu-tions as presented here. The products are meant to be probed by laser in a specific direction todetermine the maximum SHG efficiency, so far the polarized microscopy study revealing theextreme dispersion effects is provided.

AMPPO3 is an example of H-bond stabilized framework with both organic and inorganicbuilding blocks. The computational study was led as to understand the structure which is typicalfor one of the classes of compounds exhibiting nonlinear optical properties. DFT implementedin parallel CP2K was used and for the comparison three other optically active materials of sim-ilar composition were studied - urea, guanylurea hydrogenphosphite and 2aminopyrimidineboric acid 3/2 cocrystal. After the constrained geometry optimization and cell relaxation normalmode analysis was performed and well-fitted IR spectra were obtained and interpreted. Duringthe process of input tuning it was found that the setting for the optimization and vibrationalanalysis cannot be generalized and differs from once case to another as the type of stabilizinginteractions differs slightly as well.

Prospectively, another modelling approach based on ab initio molecular dynamics whichresembles better the experimental properties was proposed for future work on the model IRspectra or other properties.

A B S T R A K T

V predkldan prci je shrnuto studium vybranch novch materilu pro nelinern optikujak pomoc experimentlnch metod, tak s vyuitm vpocetn chemie. Prce je zamerena ze-jmna na hydrogenfosforitan aminopyrimidinia (AMPPO3), co je ltka studovan ji drve vna skupine a vykazujc vhodn vlastnosti jako snadn dostupnost, bezbarvost, dobr roz-pustnost ve vode a ochota krystalizovat. To bylo motivac k jejmu dalmu studiu a snaze vypes-tovat velk monokrystal z nasycenho vodnho roztoku udrovanho pri konstantn teplote.Studium prvnch krystalu polarizacnm mikroskopem odhalilo extrmn disperzi AMPPO3,dal generace krystalu bude pouita ke zmeren SHG na orientovanm krystalu.

AMPPO3 byl jako typick semiorganick materil s NLO vlastnostmi studovn tak teoret-icky pomoc DFT implementace v paralelizovanm CP2K. Pro porovnn byly jete studovnydal ltky typove podobnho sloen, vechny NLO aktivn: mocovina, hydrogenfosforitanguanylmocoviny a 2aminopyrimidin-borit kyselina 3/2 kokrystal. Experimentln strukturybyly zoptimalizovny tak, aby byl zachovn tvar bunky, kter byla pouze zrelaxovna. Po vi-bracn analze v rmci CP2K jsme obdreli IR spektra vborne odpovdajc experimentu, kterposlouily k detailn interpretaci normlnch mdu. Behem procesu optimalizace inputu bylozjiteno, e neexistuje zobecniteln postup pro vechny ltky s podobnm typem stabilizacestruktury, vpocetn studie mus bt vyvinuta specilne pro kadou ltku.

Jako vhled do budoucna byl navren postup pro zskn modelovch vibracnch spekter adalch vlastnost tak, aby vpocet lpe simuloval skutecn experimentln podmnky toho lzedoshnout za pomoci ab initio molekulov dynamiky.

iii

First of all,there is no luck.

There only iswork.

Max Black, Two broke girls

A C K N O W L E D G M E N T S

The access to computing and storage facilities owned by partiesand projects contributing to the National Grid Infrastructure MetaCen-trum, provided under the programme "Projects of Large Infrastructurefor Research, Development, and Innovations" (LM2010005) is highlyappreciated.

The prompt and thorough e-mail support from the MetaCentrumHelpdesk is highly appreciated also.

I would like to express my special thanks of gratitude to my adviserDoc. Ivan Nemec, PhD for the kind suggestions for the experimentalpart of my thesis, RNDr. Ivana Csarov, CSc. for having solved thenumerous single crystal X-ray analyses for this thesis and VlastimilaPitterov for measuring the powder X-ray diffraction data. I am alsovery grateful to Prof. Dr. Ladislav Bohat for providing me the chance tomake use not only of the unique facilities at the University of Colognebut also of his experienced suggestions while attempting to grow sin-gle bulk crystals of my compounds.

Secondly, I would also like to thank to RNDr. Filip Uhlk, PhD for thepriceless idea to use CP2K and for patient help at the beginning.

Last but not least I indeed owe my thanks to many other peoplewho were either helpful or full of ideas or both cases at once, but Ishall rather thank them in person.

iv

C O N T E N T S

i I N T R O D U C T I O N 11 M O T I VAT I O N 32 I N T R O D U C T I O N T O N O N L I N E A R O P T I C S 5

2.1 Nonlinear response in crystalline materials 52.2 Second harmonic generation (SHG) 6

3 M AT E R I A L S S E L E C T I O N F O R T H E S T U D Y 73.1 2-aminopyrimidine 8

3.1.1 AMPPO3 83.2 2,4-diaminopyrimidine 8

4 O P T I C A L M I C R O S C O P Y 94.1 Transport of light in media 9

ii E X P E R I M E N TA L PA R T 135 E X P E R I M E N TA L 15

5.1 Crystal growth of AMPPO3 from aqueous solutions 155.2 Microscopy 175.3 Synthesis of DAP salts 175.4 FTIR spectroscopy 185.5 Raman spectroscopy 185.6 Single crystal X-ray analysis 195.7 Powder X-ray diffraction 195.8 General postprocessing 19

6 S T R U C T U R A L A N A LY S I S O F dap B R A N C H O F F I N A LP R O D U C T S 216.1 Methods of structural studies 216.2 Products 21

6.2.1 DAP1 = (2,4-DAP)2,4-DAP(1+) perchlorate 226.2.2 DAP2 = 2,4-DAP(1+) perchlorate 246.2.3 DAP3 = 2,4-DAP(1+) nitrate 266.2.4 DAP4 = bis[2,4-DAP(1+)] hydrogenphosphate

tetrahydrate 296.2.5 DAP5 = bis[2,4-DAP(1+) hydrogenoxalate] mono-

hydrate 316.2.6 DAP6 = bis[2,4-DAP(1+)] malonate monohy-

drate 347 V I B R AT I O N A L S P E C T R A A S S I G N M E N T O F D A P B R A N C H

O F F I N A L P R O D U C T S 377.1 Theoretical spectra of DAP 377.2 Assignment 44

7.2.1 DAP1 = (2,4-DAP)2,4-DAP(1+) perchlorate 447.2.2 DAP2 = 2,4-DAP(1+) perchlorate 46

v

vi C O N T E N T S

7.2.3 DAP3 = 2,4-DAP(1+) nitrate 477.2.4 DAP4 = bis[2,4-DAP(1+)] hydrogenphosphate

tetrahydrate 497.2.5 DAP5 = bis[2,4-DAP(1+) hydrogenoxalate] mono-

hydrate 507.2.6 DAP6 = bis[2,4-DAP(1+)] malonate monohy-

drate 538 C R Y S TA L G R O W T H R E S U LT S 55

8.1 Purification 558.2 Purified AMPPO3 crystal growth 578.3 Microscope study 58

9 S U M M A RY O F T H E E X P E R I M E N TA L PA R T 61

iii C O M P U TAT I O N A L C H E M I S T R Y 6310 M O D E L L I N G O F E X P E R I M E N TA L P R O P E RT I E S 6511 P L A N E WAV E S V S . AT O M I C B A S I S S E T 6712 G A U S S I A N S I M U L AT I O N S R E S U LT S 71

12.1 Simple calculations of Urea 7112.2 Periodic boundary conditions in Gaussian 7212.3 Gaussian simulations: Conclusions 73

13 P L A N E WAV E B A S E D S I M U L AT I O N S 7713.1 Parallel computing and CP2K 7713.2 CP2K simulations 79

13.2.1 The theoretical finite difference IR spectrum 7913.2.2 The theoretical IR spectra from molecular dy-

namics 9314 S U M M A RY O F T H E C O M P U TAT I O N A L PA R T 9715 C O N C L U S I O N 99

iv A P P E N D I X 101A A P P E N D I X : S T R U C T U R E D E T E R M I N AT I O N D E TA I L S A N D

C R Y S TA L D ATA 103B A P P E N D I X : S U P P L E M E N TA RY M AT E R I A L 109C A P P E N D I X : X - R AY P O W D E R D I F F R A C T I O N D ATA 115D A P P E N D I X : C O M P L E T E I N P U T S 119

D.1 Inputs for Gaussian 119D.2 Inputs for finite difference spectra (CP2K) 120D.3 Inputs for Molecular dynamics (CP2K) 125

E A P P E N D I X : C O M PA R I S O N O F E X P E R I M E N TA L A N D C A L -C U L AT E D G E O M E T R Y PA R A M E T E R S 129

B I B L I O G R A P H Y 134

Part I

I N T R O D U C T I O N

1M O T I VAT I O N

Chemistry of crystalline materials with nonlinear properties is an inter-esting scientific field comprising methods of classical chemistry, suchas synthesis of materials with desired properties (based on previouslygained knowledge about compounds consisting of similar buildingblocks), crystal growth or characterisation via spectroscopy and X-raydiffraction techniques.Moreover, computational chemistry may be employed to predict or sim-ulate the behaviour of the studied compounds under various condi-tions; it should be, however, a bit more tedious task than running acalculation for a single-molecule model in vacuum. Such a study isgetting closer to experimental studies in the sense one needs to workwith the system of the interest for some time to get hands on and getthe feeling for how the system would behave under certain circum-stances.Materials with nonlinear response upon being exposed to externalelectrical field are certainly worth the study as they have importantrecent industrial applications in the areas of fast information process-ing and data storage, provided the selected material has convenientmechanical properties. [54] For instance, the bulk must be clear so therisk of overheating when irradiated is minimal, must form large good-quality single crystals as easily as possible and should be mechanicallyresistent. From the list it is obvious that the potentially vast groupof nonlinear materials shrinks to few compounds matching the cri-teria and many of them are so-called semiorganic materials, for thenonlinearity arises from the organic part while the inorganic buldingblock allegedly introduces enhanced crystallinity mechanical stability.[32, 28, 29]Hence the presented thesis deals mainly with semiorganic materialsbased on nitrogen-containing bases and organic or inorganic acids.

3

2I N T R O D U C T I O N T O N O N L I N E A R O P T I C S

Pi = 0j

(1)ij Ej (1)

By the look at Equation 1 one can recognize the well-known for-mula for the linear relation between polarization and applied electricfield. However, this is not entirely accurate formula when consideringa laser light beam of a great intensity. In that case, Equation 1 expandsin terms of applied field, Equation 2. This can be also done with otheroptical properties, such as refractive index.

Pi = 0j

(1)ij Ej +

(2)ijkEjEk +

(3)ijklEjEkEl + ... (2)

The higher-order parts of Equation 2 are non-negligible only for highelectric field intensities, but the dominant effect will always be the firstorder. The higher order effects may be considered a perturbation.

2.1 N O N L I N E A R R E S P O N S E I N C R Y S TA L L I N E M AT E R I A L S

Material gets polarized on a molecular level as it has been irradiatedby a light of electric field E. For the value of P the properties of materialare essential, these are in Equation 1 represented by , a frequencydependent matrix of susceptibilities. [48]

Thereby, a displacement is created as a function of the dielectric con-stant matrix ij and Ej while both of these are frequency-dependent.

Di = 0Ei + Pj = 0Ei + 0j

(1)ij Ej = 0

j

ijEj (3)

In the case of sum frequency generation two incident beams are com-bined together and the generated beam frequency equals to the sumof the incident beams frequencies. Therefore in Equation 3, for the ma-terial properties influencing the frequency dependant polarization weget

Pi = 1/2 0 jk

[ijk(3;2,1)Ej(1)Ek(2)

+ijk(3;2,1)Ej(1)Ek(2)](4)

The coefficients in Equation 4 are crucial for the nonlinear propertiesdescription as the nonlinear phenomena occurs when the polarization

5

6 I N T R O D U C T I O N T O N O N L I N E A R O P T I C S

of the material responds nonlinearly to the interaction with the elec-tric component of light. [19]

2.2 S E C O N D H A R M O N I C G E N E R AT I O N ( S H G )

SHG is a coherent optical process that can be classified as a specialcase of sum frequency generation by which an active medium havingbeen irradiated by a light beam of frequency produces a beam ofdouble the frequency (2). This phenomenon has been exploited inthe field of laser technique since long ago but newly also in the area ofoptical information processing, especially regarding organic materialswith commutable properties. [19]The direction of the output light is given by E21E2. For this to be asefficient as possible there should be a proper phase interaction of theincident light and the generated light. Specifically, the incident radi-ation wave and the generated wave should be in phase. This givesus the phase matching condition k2 = 2k1. So-called critical phase match-ing can be observed in single crystals of the highest optical quality. Itneeds the crystal to be oriented in such a way that the phase matchingreaches its maximum.For SHG to be observed the crystal has to be necessarily noncen-trosymmetric. Otherwise the polarization and electric field will exhibitan inversion symmetry, too. The SHG as well as any other second ordereffect will not be observable.

Polarization of the material is so largely dependant on the symme-try because of the origin of the nonlinear response. The principle lies inthe structure of the material, which is neutral as bulk but on a smallerscale shows non-uniformity with respect to the centre of symmetry ofeach unit cell. If such a non-uniformity is present, one can imagine thebulk as several polar chains of alternating positive and negative pointcharges. These charges are shifted when external electric field has beenapplied, but the shift is influenced by their mutual repulsive or attrac-tive interactions as well, which gives us larger displacement in certaindirections of the external field. Moreover, from this point of view it isalso clear, that the external field must be of similar strength as the in-teratomic forces. This is typically 105 108V/m which takes us to thenecessary use of lasers for nonlinear-effects probing. [49]

3M AT E R I A L S S E L E C T I O N F O R T H E S T U D Y

T H E B A S E S : Organic nitrogen bases are often used as buildingblocks for supramolecular structures as they are usually an ideal H-bond inducing building block. H-bonds are found between so-calledchemically complementary functional groups, for a N-H..N hydrogen tobe formed is therefore needed for example a ring nitrogen (the accep-tor of the hydrogen bond) and an aminogroup (the donor of the hy-drogen bond). The use of this knowledge may be exploited to roughlypredict the structure and properties of the compound being prepared.[23]Moreover, the hydrogen bond inducing compounds can be split intotwo groups considering the supramolecular scheme they usually in-duce - convergent and divergent donors. The convergent donors supportthe isolated clusters formation whrereas the divergent ones will mostlikely form a complex net framework. The latter group is the most im-portant one for crystal engineering and there are various examples the one-dimensional chains are formed by barbiturates and aminopy-ridines, melamine and cyanuric acid support rosetta-like net of H-bonds formation. [36]However, these typical bonding patterns may be very easily changedby either making a derivative of the organic base or inducing a new(preferably acidic) molecule into the framework to break the pattern.On the other hand there are some preferred interactions that hold oneven in the structure of the derivative, such as base-pair formation anal-ogous to the base pairs that stabilize DNA structure. The centrosym-metric pair formation leads to the compensation of dipole momentspresent on the ring, which has a largely stabilizing effect. However, asmentioned earlier, the pairs are centrosymmetric, hence this stabiliza-tion is not in favour of nonlinear optical material synthesis.Therefore the asymmetric electron distribution is enhanced by func-tionalization of the framework of the base by acids of various symme-try. (See for example [32, 51] or more recent [24]). In the contrary toexpectations, the use of organic chiral acids does not often lead to thedesired lack of the centre of symmetry as the C-H..O bonds are tooweak to compete with N-H..O or N-H..N. Therefore highly symmet-ric inorganic acid ions often serve better as they are willing to formN-H..O bonds with the base and help build a complex structure of H-bonds. [24]

7

8 M AT E R I A L S S E L E C T I O N F O R T H E S T U D Y

3.1 2 - A M I N O P Y R I M I D I N E

The presence of a ring nitrogen together with an aminogroup makesaminopyrimidines divergent donors with tendencies to form chains.Hence the derivatives of AMP have been widely used as buildingblocks that allegedly induce new H-bonding sites into the supramolec-ular structure. [39]

3.1.1 AMPPO3

Fot the advanced experimental and theoretical study aminopyrimi-dinium hydrogenphosphite was selected. It is an SHG active materialthat had been previously prepared [65] and studied in our group. [45]The SHG efficiency was previously measured on a powder sample,evaluated as 118% of the efficiency of the common standard potassiumdihydrogen phosphate (KDP). The compound was also found to exhibitphase-matching. However a major drawback of the previous studies isthat there was no measurement on an oriented single crystal, regard-ing not only the SHG efficiency and phase-matching properties, butalso the ability to form well-developed colourless crystals. [45]

3.2 2 , 4 - D I A M I N O P Y R I M I D I N E

Diaminopyrimidine is another organic base that has so far been ex-ploited mainly for its potential in medicine as a bactericide. [50, 52]In the field of supramolecular chemistry it is also attractive as it in-cludes two ring nitrogens and two aminogroups which offers plentyof variations of possible hydrogen bonding.In this thesis the aim was to prepare a series of salts/co-crystals ofDAP in the attempt to obtain a noncentrosymmetrical material to ac-company the study of AMPPO3.Various acidic compounds were chosen for reaction with DAP to in-vestigate its bonding properties. The inorganic acids with correspond-ing anions of Td and D3h symmetry were selected to propose a differ-ent hydrogen bond arrangement in each case. Apart from that, organicacids were chosen for their enhanced ability to form hydrogen bonds,mainly with the amino functional group on DAP.

4O P T I C A L M I C R O S C O P Y

4.1 T R A N S P O RT O F L I G H T I N M E D I A

The beam proceeding in a bulk solid changes its properties accordingto the properties of the material represented by dielectric constant .It is a constant dependent on the direction only in case the material isnot isotropic. The relation between displacement and electric field (5)shows that in isotropic materials the vectors D and E have the samedirection. The dielectric constant of anisotropic material has the formof a matrix hence the direction of D and E does not remain parallel.The dielectric constant reflects the susceptibility of the material, theability to be polarized. (6) In nonlinear materials may depend on theelectric field magnitude as well. (7)

D = E (5)

= 0 (1+ ) (6)

(E) = + 2E+ 3E2 (7)

One way to describe the direction-dependant properties of a crys-talline material is to introduce the indicatrix, an ellipsoid with thelength of its axes defined as the square of the refractive index in thecorresponding direction or equivalently as the value of the dielectricconstant in the corresponding direction. The refractive index is the ve-locity of the lightwave propagation in the media relative to the propa-gation in vacuum (e.g. the speed of light).

T R I C L I N I C A N D M O N O C L I N I C C R Y S TA L S : Regarding the stud-ied materials mainly two kind of symmetry setting are necessary to bementioned.Triclinic and monoclinic crystals are anisotropic as they have differ-ent refractive indices for every direction. On a molecular level this iscaused by the bonds being not uniformly strong in the whole bulk,they are stronger in a preferred direction instead. Typical example islayer formation, f.e. Figure 1 where the bonds are covalent inside thelayer whereas the layers are connected by hydrogen bonds. This is

9

10 O P T I C A L M I C R O S C O P Y

Figure 1: 2-Aminopyrimidine single crystal structure with hydrogen bondlayer stacking visualisation (view along the axis b).

a key property for crystals having nonlinear optical properties alongwith having no center of symmetry Chapter 2.The described anisotropy leads to birefringence (double refraction). How-ever, both the monoclinic and triclinic crystals are biaxial, which is aphenomena of having two optical axes, i. e. two directions in whichthe lightbeam suffers from no birefringence (the index of refractionperpendicular to the light propagation has a uniform value). In therepresentation of indicatrix this can be seen when the ellipsoid is cutin the direction perpendicular to the lightwave propagation; the cut iscircular.The symmetry constrains are not distingtively strict and therefore al-low more optical effects to occur.

T R I C L I N I C S Y M M E T R Y: In triclinic symmetry setting the relativepositions of the three axes are not fixed, they shift towards each otherwith the change of temperature, pressure and wavelength of the pass-ing light. The latter case is known as length dispersion as it is thelength of the indicatrix that changes while the right-angle arrange-ment stays intact.

M O N O C L I N I C S Y M M E T R Y: Monoclinic symmetry setting has amain symmetry axis of symmetry which coincides with the main axisof the indicatrix while the two remaining axes lie in the plane perpen-dicular to it. The length dispersion is still possible because the onlysymmetry constrain is for the main axis to be perpendicular to the

4.1 T R A N S P O RT O F L I G H T I N M E D I A 11

plane containing the other two axes while their position relatively oneto another is changeable.

Figure 2: Principle of polarization microscopy. [40] (For illustration; only onewavelength and only electric field vector is depicted).

There are several possibilities to introduce dispersion without break-ing the symmetry requirements for those two settings. In a triclinic sys-tem the dispersion is not constrained by symmetry requirement at all,in monoclinic systems the shift of the relative positions of the side axesis possible as long as their plane stays perpendicular to the main axis.One of the possible reasons for this effect to take place is the changeof wavelength of the incident light. We can easily observe the effect byusing white light, which includes the full spectrum. In that case theshift of the indicatrix will be different for each wavelength included inthe spectrum. This is the principle of length dispersion.The dispersion effect can be observed under the polarization micro-scope using white light in a setup consisting of polarizer and analysergrid filters with the crystal of the studied material placed betweenthem, Figure 2. The grid filter lines are perpendicular to each other,hence with no dispersion effect taking place there should be no lightpassing through this set. Thanks to length dispersion, however, notall the colours are filtered out. For some wavelengths the indicatrixis shifted so that its axes are not parallel to the polarizer and analyser.Once this happens the colour corresponding to the wavelength shouldbe visible after passing.However, length dispersion usually does not have a clearly visible ef-fect as the shift of the indicatrix is only subtle. The change of colour ofthe passing light takes place every time the length dispersion occursbut usually is not visible unless the dispersion is anomalously large. [13]

Part II

E X P E R I M E N TA L PA RT

5E X P E R I M E N TA L

5.1 C R Y S TA L G R O W T H O F A M P P O 3 F R O M A Q U E O U S S O L U -T I O N S

A M P - P O 3 : At first, the temperature dependence of solubility in wa-ter was measured (Figure 3) and the appropriate temperature for crys-tallization was evaluated, mainly regarding the theoretical requiredamount of the solid to prepare enough of the saturated solution.The setup for single crystal growth is shown in Figure 4.

The crystal seeds cca 0.5 cm large with well-defined crystal faceswere prepared and either glued to a ceramic desk with 2-phase epox-ide glue (Figure 4 left) or tied to a tripod (Figure 4 right). The tripodshould be no higher than 1/3 height of the beaker, the seeds are tiedon a thin nylon thread with its ends cut the shortest possible (the looseends could form a new crystallization centre). The hanging crystalshould be hung slightly above 1/2 of the height of the tripod. The stir-rer of matching length must be chosen regarding the shape and size ofthe beaker and the amount of saturated solution. There must be 2 cma gap between the stirrer and the tripod as well as between the stirrerand the solution surface.The surplus amount of the purified solid was put inside a 600 mLbeaker and cca 400 mL distilled water was poured over it. The mix-ture was closed with a lid without sealing, placed in thermostat withstirrer and kept under constant temperature of 38C. The mixture waschecked regularly and the solid material was added in case all of ithad dissolved. The saturated solution was obtained after three daysof saturation while there was only a small amount of solid remainingat the bottom of the beaker.The saturated solution was decanted into an Erlenmeyer flask and

heated 10C above the crystal growth temperature (48C) on a waterbath. The heating is necessary to dissolve the microcrystalline solidon the walls of the flask and the thermometer that are colder thanthe solution taken out of thermostat. The flask was closed with a lidwhile only a hole for a thermometer was kept so the vapours are mini-mized and the concentration remains the same. The solution was thenslowly cooled down 2C above the growth temperature. Crystal seedson a plate or on a tripod were placed into a thoroughly cleaned anddried beaker. The saturated solution was poured on an inside wall ofthe beaker to fill it up without the stream damaging the highly solu-ble crystal seeds. A glass lid with a bent glass stirring stick in it was

15

16 E X P E R I M E N TA L

Figure 3: The solubility of 2-aminopyrimidinium(1+) hydrogenphosphite inwater. (Experimental data, average values from 3 succesive mea-surements).

Figure 4: The two set-ups for crystal growth from saturated aqueous solutionused for 2-aminopyrimidinium(1+) hydrogenphosphite.

5.2 M I C R O S C O P Y 17

used to seal the beaker as it was glued to it with a polish as fast aspossible to prevent evaporation and temperature changes. The sealedbeaker was placed into the thermostat and left first 24 hours withoutstirring to let the system equillibriate. The following day stirring wasstarted. The crystal growth takes 8 to 12 weeks in if it does not have tobe interrupted because of the seeds having dissolved.

P O S S I B L E C O M P L I C AT I O N S : In case the seeds have dissolved thesaturation might not have been completed.The beaker must be opened, the tripod taken out and more of thesolid material must be added. Then the solution is resaturated, 1 or2 days of saturation following the already described procedure is usu-ally enough. The dissolving of the seeds may as well happen later,even after large crystals have already grown, and it is mainly due totemperature fluctuation inside the thermostat. Therefore it is neces-sary to have an excessive amount of the purified solid prepared. Inthe case of 2-aminopyrimidinium(1+) hydrogenphosphite it was cca1.5 kg for 3 x 600 mL beaker.

5.2 M I C R O S C O P Y

Mechanically selected crystals of high quality were investigated underclassical stereomicroscope Zeiss Stemi SV 11 to determine the qual-ity and physical parameters of the single crystal. Chosen best qualitycrystals were further investigated with polarisation microscope ZeissAxioplan 2. For taking the photos of the studied crystals the cameraAxioCam Icc 3 connected to the microscope was used.

For the polarisation measurement the crystal was placed betweentwo polarisation filters with grid lines perpendicular to each other.The white light source was passed through the polarizer, crystal andanalyzer. The crystal was rotated 100 with the 10 step to observe thedispersion effects causing the change of colour of the crystal.

5.3 S Y N T H E S I S O F D A P S A LT S

D A P S A LT S : The products derived from DAP were obtained bycrystallization of reaction mixtures prepared by simply dissolvingDAP in 20 - 50 mL distilled water and adding the correspondingamount of 2 mol/L acid solution.The molar ratio of DAP/acid was chosen to follow the expected sto-ichiometry of the product, hence it was always 1/1 at first. After the1/1 series has been completed, 1/2 ratio was tried out too and also2/1 in case of phosphate salt (DAP4) while most of the products of

18 E X P E R I M E N TA L

1/2 ratio correspond to the 1/1 product of the same starting materialcombination. The latter products are not listed whereas all the success-fully prepared DAP salts are presented in Table 1.

5.4 F T I R S P E C T R O S C O P Y

For the basic characterisation the diffuse reflectance method was used(DRIFTS), for the final characterisation however, the transmissionspectra of nujol and fluorolube mull were measured. The presentedtransmission IR spectra are composites of the corresponding nujolmull and the fluorolube mull spectra of the investigated compoundmerged together.The spectra were recorded on a Thermo Nicolet 6700 FTIR spec-trometer with 4 cm1 resolution, Happ-Genzel apodization and 4004000 cm1 wavenumber range.

All the presented spectra were postprocessed in OMNIC 7.1 [63] andWINFIRST 3.57 [46] and plotted in Gnuplot. [62]

5.5 R A M A N S P E C T R O S C O P Y

F T I R R A M A N : FT Raman spectra were recorded on a Thermo Nico-let 6700 FTIR spectrometer equipped with a Nexus FT Raman mod-ule (1064 nm Nd:YVO4 laser excitation, Happ-Genzel apodization,4 cm1 resolution, 3700150 cm1 range). The laser power of 1 Wwas used.All the presented spectra were postprocessed in OMNIC 7.1 [63].

R A M A N M I C R O S C O P Y: Mechanicaly separated crystals of selectedproducts were studied on dispersion Raman microscope.The spectra were recorded on a Thermo Scientific DXR Raman Micro-scope (Peltier-cooled CCD detector) interfaced to an Olympus micro-scope (10x and 50x objective lens) in the 503700 cm1 spectral regionwith a nominal spectral resolution of 4 cm1. The spectrometer wascalibrated by software-controlled calibration procedure using multi-ple neon emission lines (wavelength calibration), multiple polystyreneRaman bands (laser frequency calibration) and standardized whitelight sources (intensity calibration). The spectra were collected usingdiode-pumped solid state laser (532 nm) and frequency-stabilized sin-gle mode diode laser (780 nm). Laser power of 5 mW for micro-Ramanspectrometer was used.

5.6 S I N G L E C R Y S TA L X - R AY A N A LY S I S 19

All the presented spectra were postprocessed in OMNIC 7.1 [63].

5.6 S I N G L E C R Y S TA L X - R AY A N A LY S I S

The single crystal analysis was performed on a four-circle diffractome-ter with CCD detector Apex II (4000x4000 px). The source of the radi-ation was a lamp with a molybdene anode, maximum power of thehigh voltage source was 3 kW, the source was set to 50 kV (20 mA).The radiation went through a graphite monochromatore for Kwave-lenght to be selected (0.71073 ). The temperature during the anal-ysis was set to 150 K. Hydrogen atoms were found on a differen-tial Fourier map and further isotropically refined. Programs COL-LECT and DENZO were used for data processing while SIR97 [17]a SHELXL97 [58] were used for structure analysis.For further postprocessing, supplementary information or images gen-erating programmes Platon [59] and Diamond [12] were used.

5.7 P O W D E R X - R AY D I F F R A C T I O N

The powder diffraction data were measured on XPert PRO (PANalyt-ical) diffractometer with CuK lamp with the radiation wavelenght1.54178 . The 2Theta position range was set from 5 to 60 with thesimulated step 0.013 (50 s for 1 step).The data were postprocessed using XPert HighScore [14].

5.8 G E N E R A L P O S T P R O C E S S I N G

All the plots presented here were made in Gnuplot [62]. The modelvibrational spectra were obtained in a form of impulses and postpro-cessed in Gabedit [10], the normalization to [0;1] and convolution toLorentzian peaks with halfwidth 20 cm1 was performed.

6S T R U C T U R A L A N A LY S I S O F D A P B R A N C H O FF I N A L P R O D U C T S

6.1 M E T H O D S O F S T R U C T U R A L S T U D I E S

Single crystals were studied by single crystal X-ray diffraction analysisto characterize the crystal structure and bonding. The measurementparameters are discussed at each product in the following sections,the supplementary material including the selected bond lengths and an-gles as well as hydrogen bond parameters is presented in Appendix B.

However, for the analysis only one crystal was chosen and it mayor may have not represented the composition of the whole preparedbulk material. For that reason powder X-ray diffraction pattern wasmeasured on a homogeneous sample to be compared with theoreticalpowder pattern available from cif file after the single crystal analysis.By this means byproducts, if present, can be revealed. The powderdiffraction data obtained from the single crystal X-ray diffraction datacorresponds to the pattern of a completely pure compound as it isbased on a single crystal data. The simulated powder diffraction pat-terns are plotted to be compared with the experimental ones directlyin the following section. Moreover the most intensive diffractions havebeen listed out of the experimental data and presented in a separatesection Appendix C.

6.2 P R O D U C T S

The procedure yielding the following series of DAP-based product hasalready been described in 5 while the products themselves are listedin Table 1. This section investigates the structure and bonding of theproducts by using X-ray diffraction methods and discussing the re-sults.

21

22 S T R U C T U R A L A N A LY S I S O F dap B R A N C H O F F I N A L P R O D U C T S

C O M P O U N D N A M E A B B R E V I AT I O N D A P / A C I D R AT I O

(2,4-DAP)2,4-DAP(1+) perchlorate DAP1 1/12,4-DAP(1+) perchlorate DAP2 1/22,4-DAP(1+) nitrate DAP3 1/1bis[2,4-DAP(1+)] phosphate(2-) tetrahydrate DAP4 2/1bis[2,4-DAP(1+) oxalate] monohydrate DAP5 1/1bis[2,4-DAP(1+)] malonate(2-) monohydrate DAP6 1/1

Table 1: Successfully prepared DAP salts.

6.2.1 DAP1 = (2,4-DAP)2,4-DAP(1+) perchlorate

White crystals of (2,4-diaminopyrimidine)2,4-diaminopyrimidinium per-chlorate were obtained by slow crystallisation from mixture of 2,4-diaminopyrimidine and perchloric acid in 1/1 molar ratio. As shownon Figure 5, the building blocks consist of perchlorate anion and twoforms of diaminopyrimidine species - neutral 2,4-diaminopyrimidineand protonized diaminopyrimidinium cation. This creates additionalanisotropy and variations in the hydrogen bonding framework. How-ever by looking at Figure 5 it is clear that the crystal packing is of arelatively high symmetry in spite of the tangled and complex systemof hydrogen bonds.

DAP - perchloricacid The crystal structure is centrosymmetric (point group P1). The de-

tails of structure and its determination are given in Table 19 and Ap-pendix B, Listing 1, where the selected bond parameters are also listed.The framework consists of pairs of DAP and DAP(1+) with aminogroupsdirected to the opposite sides connected via N2-H2A..N3 hydrogenbonds. The donor of the H-bond is the aminogroup and the acceptoris the nitrogen of the neighbouring ring. These pairs are mutually in-terconnected via N2-H2B..O4 hydrogen bonds with perchlorate anion.A chain-like structure of DAP pairs as seen in Figure 5b is formedwhile the chains are connected only weakly by one N8-H8..N7 hydro-gen bond per each pair.

Figure 6 and shows experimental powder diffraction pattern, the listof diffractions is presented in Table 27. The powder X-ray diffractionpattern was compared to the pattern obtained from single crystal anal-ysis data. The two compared patterns are matching with slight differ-encies. The discrepancy between theoretical and experimental data isonly subtle and may be caused by temperature. The single crystal anal-ysis was determined at low temperature unlike the powder diffraction(See Table 19 Temperature). Therefore it may be claimed that the solidproduct is of reasonable purity.

6.2 P R O D U C T S 23

(a) Asymmetric unit of DAP1.

(b) Crystal packing of DAP1, with the chainlike structureconnected via hydrogen bonds.

Figure 5: DAP1 - Asymmetric unit and crystal packing, view along the crys-tallographic axis a.


Figure 6: Powder pattern of DAP1 compared with the pattern obtained fromcif file.

6.2.2 DAP2 = 2,4-DAP(1+) perchlorate

White crystals of 2,4-diaminopyrimidinium(1+) perchlorate were ob-tained by slow crystallisation from mixture of 2,4-diaminopyrimidineand perchloric acid in 1/2 molar ratio. As shown in Figure 7a, theasymmetric unit consists only of protonized DAP(1+) and perchlorateanion in the 1/1 ratio. The stoichiometry therefore does not follow themolar ratio of starting material.

DAP - perchloricacid Regarding the simple stoichiometry the crystal packing is expected

to show a reasonable degree of uniformity also because DAP(1+) is ap-proximately planar and perchlorate anion has tetrahedral symmetry.The crystal structure of DAP2 is centrosymmetric with space groupP1. By looking at Figure 7b it is clear that the crystal packing is of arelatively high symmetry and the presence of centre of symmetry isclearly obvious. Also the framework structural pattern is similar tothe one of DAP1 (see Figure 5) there are pairs of DAP(1+) connectedvia N1-H12..N3 H-bond forming a chain-like structure via perchlorateanions, though the DAP/perchlorate ratio is different regarding thestoichiometry.

Further details of structure and its determination are provided inTable 20, the bond parameters are listed in Appendix B, Listing 2.

6.2 P R O D U C T S 25


(b) Crystal packing of DAP2.



Figure 8 shows experimental powder diffraction data of the sam-ple compared to the theoretical data. The listing of the experimen-tal diffraction positions and intensities is given in Table 27. The twocompared patterns are not entirely matching with slight discrepanciesmainly in the area around 20[2]. The discrepancy between theoret-ical and experimental data does not match powder diffraction peakspresent in starting material or in the 1/1 product DAP1, hence it ismost likely caused by the presence of an insignificant amount of an-other byproduct. However, the alleged byproduct is of minor amountand has worse crystallinity, therefore DAP2 can be easily mechanicallyseparated.

Figure 8: Powder pattern of DAP2 compared with the pattern obtained fromcif file.

6.2.3 DAP3 = 2,4-DAP(1+) nitrate

White crystals of 2,4-diaminopyrimidinium(1+) nitrate were obtainedfrom mixture of 2,4-diaminopyrimidine and nitric acid in 1/1 molarratio.As shown in Figure 9a, the asymmetric unit consists only of pro-tonized DAP and nitrate anion in the 1/1 ratio. The stoichiometrytherefore follows the molar ratio of starting material. Regarding thestarting compounds the crystal packing was expected to show stackedlayers because DAP(1+) is approximately planar as well as the nitrateanion. By looking at the crystal packing in Figure 9b it is clear that theframework is built from layers. Each layer consists of the network of

6.2 P R O D U C T S 27

DAP(1+) and nitrate anions connected via N4-H4..N3 and N4-H4..O3hydrogen bonds in such a fashion that the whole is planar. The ar-rangement has a relatively high symmetry and the presence of centreof symmetry is obvious from Figure 9. The space group is P1, furtherdetails of the structure and its determination are provided in Table 21.The list of bond parameters is to be found in Appendix B, Listing 3.

DAP - nitric acid

Figure 10 shows experimental powder diffraction pattern of thesample compared to the pattern obtained from single crystal analysisdata. The listing of experimental diffractions is provided in Table 26.Both the theoretical and experimental patterns show the most signifi-cant peaks at cca 18[2] and 28[2]. The match of the two patterns isnot entire but the discrepancies can be considered minor, therefore wecan conclude that DAP3 crystalline product is reasonably pure.

6.2 P R O D U C T S 29

Figure 10: Powder pattern of DAP3 compared with the pattern obtainedfrom cif file.

6.2.4 DAP4 = bis[2,4-DAP(1+)] hydrogenphosphate tetrahydrate

White crystals of bis[2,4-diaminopyrimidinium(1+)] hydrogenphosphatetetrahydrate were obtained by slow crystallisation from mixture of 2,4-diaminopyrimidine and phosphoric acid in 2/1 molar ratio. The 1/1ratio product is not mentioned as this system had already been stud-ied and its product published. [25]The reaction and analysis from [25] were reproduced successfully. The1/2 ratio product was found to be identical to the 1/1 product hencehas not been mentioned either.

As shown in Figure 11a, the asymmetric unit of DAP4 consists oftwo protonized DAP(1+) cations, a hydrogenphosphate anion andfour water molecules. The stoichiometry therefore follows the molarratio of starting material. Thanks to the presence of water moleculesnew possibilities of hydrogen bonding are induced. Therefore the crys-tal packing is expected to be complex with a network of hydrogenbonds to stabilize the structure. By looking at Figure 11b it is clear thatthe arrangement has a relatively high symmetry in spite of the com-plexity of the structure, the crystal structure centrosymmetric and haspoint group P1.DAP(1+) form pairs stacked plane-to-plane together in this structure.The pairs are not directly connected one to another, the connection isarranged via HPO4 and water molecules cluster forming a joint be-tween four pairs of DAP(1+). DAP(1+) is connected to hydrogenphos-


phate anion via N3-H2..O3 while phosphate anions are interconnectedvia water molecules which holds the structure together. Further detailsof structure and its determination are provided in Table 22, for the list-ing of the bonding parameters see Appendix B, Listing 5.




6.2 P R O D U C T S 31


DAP - phosphoricacidFigure 12 shows the comparison of experimental and theoretical

powder X-ray diffraction data, the listing of the diffractions is to befound in ??. It can be seen that the patterns match to a large extent,the main inconsistency lies in the intensities of the diffractions. How-ever, as long as the positions 2 match we can still conclude that theproduct is pure.

6.2.5 DAP5 = bis[2,4-DAP(1+) hydrogenoxalate] monohydrate

White polycrystalline material was obtained by slow crystallisationfrom mixture of 2,4-diaminopyrimidine and oxalic acid in 1/1 molarratio. The solid product was crystalline with small amount of a sec-ond phase of poorer crystallinity. The two phases were separated me-chanically for the crystalline product bis[2,4-diaminopyrimidinium(1+)hydrogenoxalate] monohydrate to be further analysed.As seen on Figure 13a, the building blocks are protonized DAP(1+),

a water molecule and a hydrogenoxalate counteranion. The presenceof water induces new structural patterns for hydrogen bonding in theframework and it is obvious from Figure 13b that the water is essen-tial for the crystal packing as the N4-H41..O1W and O1W-H1W..O4hydrogen bonds stabilize the structure via interconnecting DAP(1+)with oxalate anion. Figure 13c shows the planar DAP(1+) ions forminga ladder-like pattern with oxalates and water molecules on the sides.




(c) Ladderlike structure of DAP5.


6.2 P R O D U C T S 33

(a) Comparison of DAP5 with DAP5a.

Figure 14: Powder pattern of DAP5 compared with the pattern obtainedfrom cif file and with the pattern of the byproduct.


However, the typical paired DAP pattern remained which is obviousfor example from Figure 13b.Despite the complicated bonding patterns, the space group of DAP5 iscentrosymmetric C2/c. The details of structure and its determinationare provided in Table 23.

DAP - oxalic acid

Figure 14 shows experimental powder diffraction pattern of thesample compared to the pattern obtained from single crystal analysisdata. The ?? lists the diffractions as well as their intensities. The theo-retical and experimental compared patterns are matching reasonably.The discrepancy between theoretical and experimental data is onlysubtle and may be caused by the temperature effects, as single crystalanalysis was determined at low temperature (See Table 23 Tempera-ture). Therefore it may be claimed that the solid crystalline product isof a reasonable purity.

Due to the poor crystallinity of the rest of the solid product was notprobed via X-ray single crystal diffraction techniques. However, thepowder diffraction pattern was measured. By comparing it with DAP5powder pattern we can see they differ significantly. Therefore we canconclude that the reaction of DAP with oxalic acid most likely yieldedtwo different products (DAP5 and DAP5a). As there is no chance toobtain structure data directly from the single crystal X-ray diffraction,powder X-ray data were obtained (see Figure 14 and ??) and the el-ementary analysis was measured. The resulting stoichiometry is al-legedly C12H16N8O8 which resembles bis[2,4-DAP(1+) hydrogenox-alate], that is, dehydrated DAP5. Another possibility is for the stoi-chiometry to be simply a co-crystal 2,4-DAP-hydrogenoxalic acid, asthose two cannot be distinguished by the means of elemental analysis.

6.2.6 DAP6 = bis[2,4-DAP(1+)] malonate monohydrate

White crystalline material of bis[2,4-diaminopyrimidine]malonate mono-hydrate was obtained by slow crystallisation from mixture of 2,4-diaminopyrimidine and malonic acid in 1/1 molar ratio.The building blocks of the product consist of protonized DAP cation,a water molecule and a fully-deprotonized malonate counteranion, aspresented in Figure 15a. The crystal water is an important part ofthe structural patterns for hydrogen bonding in the framework, Fig-ure 15b shows that the water-induced hydrogen bonds stabilize theframework.The structural pattern consists of two alternating ladder-like chainsof stacked planar DAP(1+) rings, regarding the DAP plane orianta-tion these two ladders are almost perpendicular. Malonate and watermolecules are placed inbetween the DAP belts. Water mediates the

6.2 P R O D U C T S 35

connection between malonate and DAP via O1W-H1W..O2 and N1-H11..O1W bonds.The crystal structure is centrosymmetric P2/c. Further details of struc-ture and its determination are provided in Table 24, for the listing ofbond parameters see Appendix B, Listing 6.




DAP - malonicacidFigure 16 shows experimental powder diffraction pattern of the

sample compared to the pattern obtained from single crystal analy-sis data. The two compared patterns are matching with slight differ-encies. The discrepancy between theoretical and experimental data isonly subtle and may be caused by the temperature change, as singlecrystal analysis is determined at low temperatures unlike the powderdiffraction (See Table 24 Temperature). Therefore it may be claimedthat the solid product exhibits good purity.

7V I B R AT I O N A L S P E C T R A A S S I G N M E N T O F D A PB R A N C H O F F I N A L P R O D U C T S

S P E C T R O S C O P Y: Infrared and Raman spectra of the investigatedDAP-based compounds have been recorded and assigned.The assignment of the spectra is based on theoretical vibrational spec-tra for a single DAP molecule. DAP in the investigated materials isnot an isolated molecule and even pure DAP in bulk consists of themolecules interconnected via hydrogen bonds. Therefore the isolatedmolecule is not an exact representation of the electronic structure andpredicted vibrational modes must be shifted with respect to the realones.Another reason causing inaccuracies in the normal mode positionis the band broadening and anharmonic effect. These cannot be de-scribed by a harmonic oscillator approximation used in the calcula-tions presented in the following section.To assign the spectra, atomic basis set and Gaussian 09 [30] calculationas well as plane wave basis set and CP2K [60] non-periodic approachhave been tried out and compared.

7.1 T H E O R E T I C A L S P E C T R A O F D A P

DAP related molecules were not intended as models for more thor-ough computational study that is presented further in Chapter 10,that was especially for the presence of two aminogroups that resultsin more complex H-bond structures which might be tedious to workwith as there usually is not one universal approach. The model of sin-gle molecule should be highly innacurate in terms of intermolecularinteractions that influence vibrational spectra because, as can be seenthrough the following section,NH2related modes make up the mainpart of the vibrational activity of DAP salts. However, there are meth-ods designated for approximative calculations like this one and themode assignment may as well resemble reality reasonably. [11] Hencefor the purpose of interpretation of IR and Raman spectra widely usedsingle molecule approach was applied as follows.

M E T H O D S : At first the infrared spectra of a single molecule in gasphase were calculated from the most widely spread method: DFT withhybrid functionals in Gaussian. The hybrid B3LYP functional and 6-311G(d,p) atomic basis set were used to obtain a reasonable agreement

37

38 V I B R AT I O N A L S P E C T R A A S S I G N M E N T O F D A P B R A N C H O F F I N A L P R O D U C T S

of the normal mode frequencies as shown in Figure 17a.

Apart from that, the non-periodic single molecule normal mode cal-culation in CP2K was performed. The results are provided in Figure 18while the specifics of CP2K use are thoroughly described in Chap-ter 13. Single DAP molecule was placed into the centre of a cubic box,the box edge length was set to 10 to make sure the whole moleculewas inside and to ensure the correct truncation of plane waves. TheMartyna-Tuckermann Poisson solver for plane wave non-periodic cal-culation was used and the normal mode analysis was performed aftera geometry optimization on the same level of theory. To make sureof a good convergence and high precision, tight convergence criteriawere set (convergence criterion for maximum geometry change 104

Angstrom, SCF target gradient 1012) and DFT functionals (PBE [33]and BLYP [11]) with large QZV2P-GTH basis set and correspondingGTH pseudopotentials [55] were used.

R E S U LT S : The results and assignment of the set of all three calcula-tions for neutral DAP are presented in Table 2 with their relative inten-sities in Table 3. The Lorentzian simulation of all the spectra comparedto the experimental ones are provided in Figure 17a, Figure 17b andFigure 18. The experimental spectra of DAP(1+) are represented bythe spectra for DAP(1+)Cl(-) prepared by a procedure analogous toSection 5.3. The counteranion is atomic, therefore his presence resultsin no bands in vibrational spectra and only the vibrational modes ofthe cation are left.It is obvious that B3LYP Gaussian calculation provides frequencies rea-sonably matching the experimental ones. The positions of valence CHor NH vibrations in both IR and Raman spectra (from 3000 to 3500cm1) are shifted and the intensities do not match the experiment atall. For N-H groups both these errors might have been caused by insuf-ficient treatment of hydrogen bonds, which is not possible in principlewhen using a single molecule in vacuum model. The C-H group is notaffected by H-bonds in such a large extent as O-H or N-H hence theshift in vibrational modes must be caused by the inevitable inaccuracyin the computational method. In spite of these imprecisions the calcu-lated spectrum is reliable enough to be possibly used for normal modeassignment.Figure 19a and Figure 19b provide the comparison between cationicand neutral DAP spectra calculated at the same level of theory. It isobvious that the discrepancy is only minor for IR spectrum and al-most negligible for Raman spectrum. Moreover it affects mostly theN-H deformation vibrations (region 1600 - 1700 cm1 ). These are,by closer look at the forementioned region in Table 2, not very welldistinguishable as the region contains (NH2) and (ring) modes of

7.1 T H E O R E T I C A L S P E C T R A O F D A P 39

(a) Predicted and experimental IR spectra of DAP(1+).

(b) Predicted and experimental Raman spectra of DAP(1+).

Figure 17: Calculated (green) and experimental (blue) IR and Raman spectraof DAP(1+). Green Lorentzian peaks correspond to the theoreti-cal normal modes on the B3LYP/6-311g(d,p) level of theory calcu-lated for a single molecule in Gaussian 09.


Figure 18: Predicted and experimental vibrational spectra of DAP ob-tained having used various computational approaches. GreenLorentzians correspond to the Gaussian model spectra while bluespectrum presents the experimental result. Pink Lorentzian peakscorrespond to cp2k non-periodic plane-wave approach havingused PBE and BLYP functionals, respectively.


variable intensity. Therefore for the interpretation of experimental vi-brational modes there is no significant difference in the use of DAPor DAP(1+) model spectra in spite of the fact that most of the DAP-related products contain DAP(1+).

The plane wave approach is usually best to be used for periodicalcalculation as the periodicity is inherently included in the plane waveformalism. To perform an isolated molecule calculation the moleculehas to be formally placed inside a unit cell twice as large as themolecule, as the electron density should vanish to zero at the edgesof the unit cell. It is apparent from Figure 18 that BLYP describes thevalence CH and NH modes slightly better but overall the PBE andBLYP results are not significantly different. Also, it is obvious that theintensities of the bands match well the experiment, but unfortunatelycp2k failed to predict the nitrogen-related modes around 1600 cm1,those are shifted towards much higher frequencies. Apart from thisshift the assignment of the harmonic modes is similar to the Gaussianresult.Considering the cost of the calculations in terms of cpu time Gaussianshould be put forward as the simple isolated molecule B3LYP calcula-tion was done in 11 minutes using a desk computer whereas cp2k sim-ulations were performed in a parallel fashion on 24 processors. Walltime for BLYP optimization and successive normal mode analysis was1 hour 48 min, but cpu time was 43 hours 22 minutes.Therefore Gaussian can be highly recommended in case a quick nor-mal mode analysis is needed. The assignment of the vibrational spec-tra in the following sections will be based mainly on the Gaussian re-sults; that is for the cationic part on Table 2 and for anions on literature[47].


(a) Predicted IR spectra of DAP (black) and DAP(1+) (blue).

(b) Predicted Raman spectra of DAP (black) and DAP(1+) (blue).

Figure 19: Comparison of model spectra (Lorentzian peaks simulation) ofneutral and protonized DAP on the B3LYP/6-311G(d,p) level oftheory calculated for a single molecule in Gaussian 09.


Gaussian/B3LYP cp2k/BLYP cp2k/PBEcm1 Assign. cm1 Assign. cm1 Assign.

182 (NH2) 138 (NH2)218 (NH2) 258 (NH2) 218 (NH2)

291 (ring), (NH2) 265 (NH2)329 (NH2) 306 (NH2)358 (NH2) 341 (NH2) 314 (NH2)375 (NH2)448 (NH2) 440 (NH2), (CH) 430 (NH2), (CH)460 (NH2) 464 (ring) 454 (ring), (NH2)497 (NH2) 490 (NH2), pi(NCC) 475 (NH2), (NH2)522 (NH2) 523 (NH2)

531 (NH2) 533 (NH2), (NH2)563 (ring) 573 (NH2) 543 (NH2), (NH2)608 (ring) 558 (ring)718 (CH) 700 (ring), (NH2) 703 (ring)785 (CH) 730 (ring), (NH2) 706 (ring)798 (ring), (CH) 799 (ring) 775 (ring)817 (ring) 813 (ring), (NH2) 819 (CH)

824 (CH) 842 (ring), (NH2)898 (ring), (NH2) 920 (ring)

983 (NH2), (CH) 947 (ring), (NH2) 949 (ring), (NH2)993 (CH) 960 (CH) 955 (CH)1002 (ring), (NH2) 1061 (CH), (NH2)1048 (NH2) 1101 (CH), (NH2) 1071 (NH2)1128 (CH), (ring,CC) 1126 (ring), (NH2) 1123 (CH), (NH2)1178 (NH2), (CH) 1167 (NH2), (CH) 1161 (ring), (NH2)

1164 (NH2), (CH)1296 (ring),(NH2) 1273 (CH), (NH2) 1283 (CH), (NH2)1342 (CH), (ring,CN) 1336 (ring) 1316 (ring)1393 (CH), (ring) 1371 (ring)

1409 (CH), (NH2) 1417 (CH), (NH2)1478 s(NH2), (CN)1507 (CH), (NH) 1503 (CH) 1526 (CH), (ring)

1519 (ring), (CH) 1537 (ring), (CH)1605 (ring),(NH2)1620 s(NH2)1640 (ring), s(NH2)1653 s(NH2), (ring)

1885 s(NH2) 1871 s(NH2)2041 s(NH2) 2035 s(NH2)

3140 (CH) 3039 (CH) 3060 (CH)3195 (CH) 3128 (CH) 3134 (CH)3589 s(NH2) 3279 s(NH2) 3331 s(NH2)3605 s(NH2) 3344 s(NH2) 3400 s(NH2)3710 as(NH2) 3519 as(NH2) 3570 as(NH2)3733 as(NH2) 3562 as(NH2) 3605 as(NH2)

Table 2: Results and comparison of the set of DFT calculations of IR spectrafor a single molecule of DAP: Gaussian B3LYP/6-311g(d,p); cp2kBLYP/QZV2P and cp2k PBE/QZV2P respectively.


Wavenumber RamanIntensity

IR Intensity Wavenumber RamanIntensity

IR Intensity

cm1 cm1

182 0.003 0.172 1048 0.013 0.128218 0.018 0.001 1128 0.018 0.230329 0.006 0.338 1178 0.003 0.028358 0.589 0.186 1296 0.023 0.127375 0.147 0.365 1342 0.020 0.006448 0.029 0.015 1393 0.063 0.146460 0.003 0.127 1478 0.339 0.001497 0.083 0.146 1507 0.040 0.006522 0.001 0.184 1605 0.190 0.020563 0.009 0.500 1620 0.021 0.027608 0.002 0.569 1640 1.000 0.395718 0.003 0.071 1653 0.361 0.085785 0.018 0.177 3140 0.037 0.519798 0.007 1.000 3195 0.013 0.502817 0.063 0.000 3589 0.061 0.435983 0.010 0.338 3605 0.081 0.439993 0.001 0.050 3710 0.040 0.1521002 0.000 0.341 3733 0.055 0.139

Table 3: Comparison of the relative intensities of normal modes in DFT calcu-lations of IR and Raman spectra for a single molecule of DAP Gaus-sian B3LYP/6-311G(d,p). The Raman Intensity was obtained fromRaInt [16] for solid state spectra with 1064 nm excitation laser.

7.2 A S S I G N M E N T

7.2.1 DAP1 = (2,4-DAP)2,4-DAP(1+) perchlorate

DAP1The strongest Raman-active modes in Figure 20 are those correspond-ing to perchlorate valence vibrations. On the other hand in the IR spec-trum the NH modes and perchlorate modes are equally intense. TheIR spectrum also shows reasonably separated and distinguishable re-gion of N-H and C-H vibrations. The broader band around 2000 cm1

is supposedly an overtone band of the ring vibrations (an anharmoniceffect). Apart from that the complete assignment of normal modes ispresented in Table 4.

7.2 A S S I G N M E N T 45

Figure 20: IR (in transmittance) and Raman spectra of DAP1.

I R R A M A N A S S I G N M E N T

221 w (NH2)350 w (NH2)

432 w 458 w (NH2)509 w 504 w (NH2)564 w 588 m (ring)634 m 626 w d(ClO4)690 w (ring)781 w 787 s (CH)815 w (ring)

940 s s(ClO4)1051 sh (NH2), (CH)

993 w (NH2)1100 vs 1115 w d(ClO4)1260 m 1261 w (ring), (NH2)1385 sh 1337 w (CH), (ring,CN)

1391 w (CH), (ring)1462 s1504 s (CH), (NH)1553 s s(NH2)1629 s 1628 w s(NH2), (ring)1671 s (CH)1900 w overtone2130 w overtone3150 bs 3096 w (CH)3350 s (NH2)3440 s (NH2)

Table 4: IR and Raman spectra assignment of DAP1.


7.2.2 DAP2 = 2,4-DAP(1+) perchlorate

DAP2The perchlorate vibrational modes are very significant just as in thespectra of DAP1, Figure 20. However, the bands corresponding to N-H vibrations and the overall shape of spectra is different, which makesclear that the bonding parameters, especially hydrogen bonds, differa lot from those of DAP1. The complete assignment of normal modesis presented in Table 5.


7.2 A S S I G N M E N T 47


366 w (NH2)467 w 473 w (NH2)481 w 496 w (NH2)543 w 550 w (NH2)585 m 588 m (ring)627 s d(ClO4)669 m (ring)766 m 779 s (CH)822 m (ring)

924 s s(ClO4)919 m 978 w (NH2), (CH)1051 vs 1039 w (NH2)1100 vs 1123 w d(ClO4)1128 vs 1170 w (CH), (ring,CC)1232 m 1238 w (ring), (NH2)1372 m 1353 w (CH), (ring,CN)

1383 w (CH), (ring)1460 w s(NH2), (CN), (CH)

1511 s 1506 w (CH), (NH)1622 s 1614 w s(NH2)1671 s 1658 w s(NH2), (ring)1952 w overtone3090 s (CH)3170 bs 3104 w (CH)3300 bs (NH2)3445 s (NH2)


7.2.3 DAP3 = 2,4-DAP(1+) nitrate

DAP3Vibrational modes of nitrate group usually have great intensitieswhich also applies for DAP3 as seen from Figure 22. The symmetricbend and stretching modes of nitrate are the strongest peaks in Ra-man spectrum, the asymmetric stretching mode is the strongest bandin IR. The complete assignment is presented in Table 6.




183 w (NH2)389 m (NH2)402 w (NH2)

481 w 504 m (NH2)516 w 550 m (NH2)557 w 595 s (NH2)669 w (NO3)704 w 725 w (CH)759 w 787 s (CH)787 m (CH)794 m (ring), (CH)945 w 985 m (NH2), (CH)1072 m 1047 s s(NO3)1128 m 1108 w (NH2)1184 m 1184 w (NH2), (CH)1232 m 1245 w (ring), (NH2)1337 vs 1345 m (CH), (ring)1378 vs 1398 m d(NO3)1455 m 1444 w s(NH2), (CN), (CH)1525 m 1528 w (CH), (NH)1650 vs 1666 w s(NH2), (ring)1685 vs s(NH2), (ring)1952 w overtone2860 sh (CH)2945 m (CH)3120 bs 3096 w (NH2)

3196 w (NH2)3306 m (NH2)3430 s (NH2)


7.2 A S S I G N M E N T 49

7.2.4 DAP4 = bis[2,4-DAP(1+)] hydrogenphosphate tetrahydrate

DAP4The normal modes of hydrogenphosphate are present as typical strongbands in IR around 970 cm1 and 1060 cm1 as seen from Figure 23.Most of the remaining modes belong to DAP, including the large re-gion of C-H and N-H vibrations, which follows the stoichiometry ratioof DAP/acid (2/1) present in this product. The complete assignmentis presented in Table 7.




535 s 517 m (NH2), (PO4)668 bs 552 m (ring)758 s 587 m (CH)786 s 783 s (CH)807 s (ring), (CH)870 m 853 w974 s 972 m s(PO4)1065 s d(PO4)1092 s 1136 w (NH2)1190 m (NH2), (CH)1245 m 1232 w (ring), (NH2)1259 m 1280 w (ring), (NH2)1329 m 1334 m (CH), (ring,CN)1405 m 1407 m (CH), (ring)1461 m 1456 w s(NH2), (CN), (CH)1524 s 1526 m (CH), (NH)1649 s 1666 w (ring), s(NH2)1698 s 1701 vw s(NH2), (ring)1882 m overtone1933 m overtone1996 m overtone3145 bs 3073 w (CH)3445 sh 3122 w (NH2)


7.2.5 DAP5 = bis[2,4-DAP(1+) hydrogenoxalate] monohydrate

DAP5The reaction of DAP and oxalic acid yields DAP5 with a minor mi-crocrystalline byproduct of unknown crystal structure. IR and Ramanspectra of DAP5 as well as of the byproduct were measured and arediscussed in this section.By looking at Figure 24 and Figure 25 one can see the spectra are differ-ent, especially in the region of carboxylic acid modes (cca 1300 cm1).There is one main peak in DAP5 IR whereas the byproduct shows sev-eral split bands in that area. Also the region of C-H and N-H valencestretches of the byproduct is split into a number of narrow bands.This shows the two phases probably differ in the ionization of the car-boxylic acid and in the way the building blocks are connected via hy-drogen bonds.DAP5 definitely includes oxalate(1-) ion as confirmed by the X-rayanalysis. (Chapter 6) Carboxylate anion modes should be found in theIR spectra as two bands about 1500 cm1 and 1600 cm1 representing(respectively) the symmetric and asymmetric stretch of COO- groupwhile all the CO bonds have equal length due to delocalisation. It isobvious from Figure 24 that these bands are present but partially cov-ered by other bands, mainly nitrogen related ones. The C=O valencestretch at 1685 cm1 referring to the protonized carboxylate group ispresent as well and is equally intensive as the COO- modes, whichfollows the fact, that half of the carboxylate groups present is ionizedwhile the other half is not.DAP5a has a very strong band close to 1700 cm1 Figure 25 that refers

7.2 A S S I G N M E N T 51

to valence stretch of C=O band hence the product most likely containsnon-ionized oxalic acid. The complete assignment is presented in Ta-ble 8 and Table 9, where the assignment of oxalate vibrational modesis based on literature [18].


Figure 25: IR (in transmittance) and Raman spectra of DAP5a.



357 m (NH2)418 w 377 m (NH2)474 w (NH2)550 w (NH2)704 s 695 w (ring)

787 w (CH)829 w (CC,ox)877 w 894 m (CC,ox)

963 w (CH)1051 m 1110 w (NH2)1225 vs 1207 w (ring), (NH2)1330 m 1345 m (CH), (ring,CN)1392 m 1406 w (CH), (ring)

1437 w s(NH2), (CN), (CH)1525 m 1521 w (CH), (NH), (COO,ox)1594 s 1605 w s(NH2), (ring)1643 s 1666 w (COO,ox)1685 s (CO,ox), s(NH2), (ring)

1704 w s(NH2), (ring)1938 w overtone3145 bs 3104 w (CH)3306 m (NH2)3405 m (NH2)



374 w (NH2)453 w 443 w (NH2)

511 w (NH2)550 w 550 w (NH2)

588 m (ring)724 m (CH)759 w 787 s (CH)808 w (ring)824 m 855 w (CC,ox)864 m (CC,ox)933 w 970 w (CH)1051 w (NH2)1163 m 1131 w (CH), (ring,CC)1240 s 1245 w (ring), (NH2)1309 s 1337 w (CH), (ring,CN)1399 s 1406 m (CO,ox), (CH), (ring)

1460 w s(NH2), (CN), (CH)1511 m 1521 w (CH), (NH)1685 vs 1658 w (CO,ox), s(NH2), (ring)1768 m 1773 w (CO,ox)3135 bs 3127 w (NH2)3400 s (NH2)

Table 9: IR and Raman spectra assignment of DAP5a.

7.2 A S S I G N M E N T 53

7.2.6 DAP6 = bis[2,4-DAP(1+)] malonate monohydrate

DAP6The vibrational modes of malonate are relatively subtle when com-pared to the DAP modes. (Figure 26) In the Raman spectrum, thestrongest peak corresponds to DAP mode while malonate modes areonly of medium intensity. This property of the Raman spectrum fol-lows the stoichiometry of the compound and shows that Raman spec-tra may be used as an estimate of it.Malonic acid is fully deprotonized which, as explained in the previoussection, is confirmed by the IR spectrum region 1300 1600 cm1. Twoequally strong bands representing the stretching vibrations of COO-group with delocalized electrons have been found ??.The complete assignment is presented in Table 10, where the malonatevibrational modes assignment is based on literature [22].




366 w (NH2)446 w 435 w (NH2)

511 m (NH2)550 m 550 m (CCO,mal), (NH2)571 w (NH2)648 m 603 m (ring)696 m (ring)767 m 787 s (CH)794 w (CH)815 w (ring)926 m (CH,mal)968 m 970 m (CH,mal)1086 w 1039 w (NH2)

1115 w (CH), (ring,CC)1192 w (NH2), (CH)

1253 s 1268 w (CCH,mal), (ring), (NH2)1337 s 1345 m (CH), (ring,CN)1365 s 1398 m d(CH2,mal)1392 s (CH), (ring)1462 s (COO,mal)1504 s 1520 w (CH), (NH)1559 s 1551 w (COO,mal)1636 s 1635 w s(NH2)1657 vs (ring), s(NH2)1870 w overtone2280 m ?

2943 w ?3140 bs 3096 w (CH)3300 m (NH2)3487 m (NH2)


8C RY S TA L G R O W T H R E S U LT S

C H O S E N M E T H O D O F C R Y S TA L G R O W T H : For the optical prop-erties to be probed a large high-quality single crystal is required andthere is a lot of more or less complicated methods of crystal growth,dependant of the properties of the studied material.The compound of our main interest 2-aminopyrimidinium(1+) hydro-genphosphite (AMPPO3) happens to be fairly watersoluble with thesolubility increasing with rising temperature (partially Figure 3) un-til the point of decomposition which is roughly 170C for phosphitesalt (see Table 11). This was important for the preparation of the largeamount of starting material. An analogous approach as in Section 5.3was applied, but on a large scale, therefore the dissolving of the start-ing materials was accelerated by heating up the mixture. Also, it isclear that it is perfectly safe to set up a crystal growth at larger thanroom temperature.Thanks to the water solubility and stability of the studied material themethod of aqueous solution crystal growth was chosen as efficient.The experimental setup is thoroughly described in Section 5.1.

N A M E M E LT I N G P O I N T [C ] D E C O M P O S I T I O N [C ]

AMPSO4 160 158AMPHCl 193 186AMPPO3 171 173AMPH2PO4 172 175

Table 11: Experimental melting and decomposition points of selected saltsderived from AMP.

8.1 P U R I F I C AT I O N

The crystal growth of AMPPO3 was preceded by synthesis of kilo-gram amount of the material. However, complications arose due tolow purity of the commercially available starting material. Thereforethe syntheses of the compounds of interest had to be started by purifi-cation of 2-aminopyrimidine.The first attempted purification method was simple filtration and recrys-tallization. This approach purified the starting deep yellow materialpartially, to a lighter yellow colour. (Figure 27) However, the entirelypure powder AMPPO3 should be white. This approach was there-fore not ideal but the purification continues during the single crystalgrowth so not entirely pure starting material might be sufficient.

55

56 C R Y S TA L G R O W T H R E S U LT S

Figure 27: Crude AMPPO3 with visible double refraction.

Moreover, the seeds prepared from the pale yellow material were largeenough for the single crystal growth to be started, and pure enough fordouble refraction to be observed. As seen in Figure 27 the seeds hada regular shape with crystal phases reasonably well developed, whichis essential for the single crystal to grow.

However, the growing crystals either dissolved after 12 weeks orgot inclusions on them, Figure 28a. The first case shows that the re-maining impurities made it difficult for the system to stay equillibri-ate while in the latter case the impurities and irregularities on the seedsurface simply formed a new side-seed. It is obvious that in case ofthe studied compound more sophisticated purification methos needsto be applied.

(a) Crystal seeds with inclusions. (b) Twinned AMPPO3.

Figure 28: AMPPO3: (a) Crystal seeds with inclusions of AMPPO3 salt after5 days in saturated solution. The crystal as a whole did not startgrowing, the impurities formed the seeds for crystal growth in-stead. (b) Twinned seeds of AMPPO3 (purified).

To complete the purification, both the prepared product and thestarting material were dissolved at slightly higher temperature (cca60C) and purified on the highest purity activated charcoal. After thor-

8.2 P U R I F I E D A M P P O 3 C R Y S TA L G R O W T H 57

ough filtration the solutions were slowly evaporated to obtain crystalclear solid material. Small samples were taken aside and evaporatedeven slower to obtain crystal seeds large enough for the single crystalgrowth to be started.The charcoal purification was about equally efficient for purification ofeither 2-aminopyrimidine or its phosphite salt, unlike the simple re-crystallisation which is remarkably more efficient for 2-aminopyrimidine.

8.2 P U R I F I E D A M P P O 3 C R Y S TA L G R O W T H

The seeds of the charcoal purified material were left growing in theaqueous solution with stirring in a thermostat with the temperatureregulated and kept on 38C (Figure 4).After 9 weeks the first generation crystals (Figure 29) were obtainedand taken out of the mother solution. It was revealed that a group ofcrystals was obtained rather than a single crystal. It was caused bythe natural tendencies of the material to form twins, an effect often ob-served in crystals with large surface strain. These tendencies were con-firmed by studying the seeds under the microscope. From Figure 28bit is obvious that the seeds were twinned even after the charcoal pu-rification. Nevertheless, the single crystal areas can be cut off as thewhole bulk of a size 4,4 x 2,5 x 0,6 cm3 is large enough to be split intopieces. The cut is feasible and fast on a water-dipped wire cutter be-cause of the high water solubility of AMPPO3. The single crystal cutsmay be used as the best obtainable seeds for a new crystal growth.The method of severalgenerations crystal growth is the standard pro-cedure for obtaining the high optical quality large bulk single crystals.

Figure 29: Bulk AMPPO3 4,4x2,5x0,6 cm3 large after 9 weeks growth in satu-rated aqueous solution.


8.3 M I C R O S C O P E S T U D Y

Both the charcoal-purified and impure crystals were observed undernormal and polarization microscope. (See Chapter 4 for details.)However, only after the purification on charcoal clear enough and un-twinned crystals have been prepared so the characterization under thepolarization microscope could have been performed more properlyonly after that. It was revealed, that the studied material has enormousdispersion which is supposedly related to the nonlinear properties ofthe compound.

The dispersion was not observable before the purification, as seenin Figure 30.On the contrary, after the charcoal filter purification the length disper-sion is observed. (Chapter 4) Based on this microscopy study we canconclude that for observing all the optical effects crystals of the high-est possible purity are necessary.A mechanically selected small single crystal was placed under the po-larization microscope with white light source and shifted with the stepof 10relatively to the original position (and to the plane of polarizedlight passing through the polarizer at the same time). At each stepthe apparent change of colour of the crystal was observed. (Figure 31)That means the indicatrix is deformed and for certain wavelengths itoccurs in a larger extent. Moreover, the deformation of the indicatrixmust be significant as the change of the colour is unusually brightlyvisible. The reason for this to happen lies in the highly anisotropicstructure of the material with layer-like framework interconnected viahydrogen bonds.Among other things obvious from Figure 31 we can also notice thecrack through the crystal. It showed up while the microscopy studywas proceeding and it served to release the strain present in the ma-terial. The reason for the strain must also be the layered molecularstructure of AMPPO3 and is related to the already mentioned tenden-cies of the material to form twins.

8.3 M I C R O S C O P E S T U D Y 59

Figure 30: The twinned crystal of AMPPO3 salt under polarization micro-scope. With the gradual change of angle of the plane of the po-larized wave the crystal gets dark and bright repeatedly. No dis-persion effect is observed.

R O TAT I O N [ ] C O L O U R

0 red10 dark aqua20 pale bluegreen30 yellow green40 yellow50 yellow orange60 orange70 deep orange80 orange red90 red100 dark aqua

Table 12: Polarized light microscopy study of a single untwinned crystal ofpurified AMPPO3. Having been rotated from to 100 with thestep of 10, the crystal changes its colour.


Figure 31: When rotating the crystal of purified AMPPO3 with respect tothe plane of polarized light, the strong light dispersion (typicalfor highly biaxial monoclinic crystals) is observed. The single un-twinned crystal was rotated from 0 to 100 with the step of 10.The colour at the beginning was dark red, at 90 the crystal is redagain.

9S U M M A RY O F T H E E X P E R I M E N TA L PA RT

One of the key points of focus of this thesis was preparation of newmaterials with potential for nonlinear optics. In the past years, severalsuch materials had been prepared and characterised within our group.[38, 65, 45] The mentioned materials with nonlinear response to the aminopyrimidineexternal electric field are all based on 2aminopyrimidine and thecomplexicity of the structure arises from the H-bonds mediated viaaminogroup of the pyrimidine ring. The next step in complexicityshould have been induced by the use of 2,4diaminopyrimidine, e. g.by introducing second aminopgroup. However, it turned out that thetendencies of the organic base to form base-pairs overtakes other pos-sible H-bond interaction. Most of the products successfully prepared diaminopyrimi-

dinehaving used DAP as a starting material include protonized DAP(1+)with inorganic acid anion as a counterpart. Apparently, only very few or-ganic acids are suitable to form crystalline salt or co-crystal with DAPdue to their large size and lower dipole moment, hence also lower abil-ity to stabilize the structure via H-bonds.To sum up, seven new compounds were prepared and characterized.Unfortunately, all of them are centrosymmetric hence not usable fornonlinear optics. The single crystal X-ray analysis was performed inall cases except from one with worse crystallinity, whereas for all theseven compounds the powder X-ray analysis was run and IR and Ra-man spectra were measured and assigned.One of the formerly prepared materials in our research group is crystal growth

AMPPO3 that is proven to show good SHG efficiency and phasematching when measured on a powder sample. [45] However, the ef-ficiency of SHG generation is orientation-dependant and therefore islowered when a powder sample was used. Having known the advan-tageous physical properties of AMPPO3, the single crystal growth hasbeen performed to be eventually able to evaluate the true, orientationdependent SHG efficiency and also material properties of AMPPO3 ina single crystal. After initial purification the crystal growth attemptswere successful and the first generation bulk crystals have been re-ported in this thesis. Unfortunately the material has strong tendenciesfor twin crystals formation and none of the attempts led to an individ-ual single crystal, it was always a cluster which can, however, be cutinto pieces. Those were used as seeds for a new growth and by thetime of finishing this thesis, the second generation of the large crystalsis still being grown as the planned growth duration is six months.

61

Part III

C O M P U TAT I O N A L C H E M I S T RY

10M O D E L L I N G O F E X P E R I M E N TA L P R O P E RT I E S

Molecular modelling is concerned with mimicking the behaviour of moleculesor molecular systems, often with the help of a simplified description of the sys-tem, employing calculations to predictions. [41]

The ultimate goal of the computational part of this thesis was to ob-tain a model vibrational spectra, interpretation and assignment of thevibrational modes and better understanding the structure and bond-ing in the solids we have been focused on. The main effort was to cre-ate such a well-fitted model that none or minimal scaling of the spectrawould be needed. However, dipole and polarizability changes calcu-lation might be difficult, the intramolecular weak interactions (repre-sented mainly by hydrogen bonds) have to be taken into account aswell, while all of this has to be put into balance with confined CPUtime resources.

M O D E L S T R U C T U R E S : Several methods of calculations were ex-plored via test jobs run on model structures. The model compoundswere chosen so that they represent the class of compounds of focus,that is organic or semiorganic materials with complex network of hydro-gen bonds. At first DFT calculations using Gaussian package [30], thenplane wave ones were performed. (Chapter 12, Chapter 13, respec-tively) Before the plane wave simulation, the models were exploitedto estimate the speedup of the parallelized software. (Section 13.1 )

The computational study itself is mainly focused on AMPPO3 aminopyrimidinium hydrogenphosphite, Figure 32c and AB aminopyrim-idine boric acid 3:2, Figure 32d, as these are the materials intensivelystudied by our group. [38] Their structure is unfortunately a bit toomuch complicated so simpler model structures are introduced, too.Urea (Figure 32a represents simple organic molecular crystal with ex-tensive network of N-H..O hydrogen bonds, which is a typical patternfor AMPPO3 and AB as well. Therefore Urea has a structure simpleenough to be used for testing but still represents the type of bondingin the studied complex compounds. On the top of that GUHP (guany-lurea hydrogenphosphite, Figure 32b) was selected as a model semior-ganic material of smaller size than AP and AB (hence supposedly notso computationally expensive) and represents another compound pre-pared and studied in our research group. [8]

65

66 M O D E L L I N G O F E X P E R I M E N TA L P R O P E R T I E S

(a) Urea

(b) GUHP

(c) AMPPO3

(d) AB

Figure 32: Selected model structures - Only illustrative figures - (a), (b):Model structures for organic and semi-organic materials stabilizedby hydrogen bonds. (c), (d): Experimental structure of compoundsstudied via computational chemistry.

11P L A N E WAV E S V S . AT O M I C B A S I S S E T

There are two approaches to simulate molecules or condensed matter,either via atomic-like basis functions or plane waves.

AT O M I C B A S I S S E T : As an approximation it may be claimed thatmolecules are only a set of slightly distorted free atoms. Within thismodel a one-particle basis function can be used, in molecular case itis centered on the nucleus and consists of a radial and a spherical part(respectively)

(r, ,) = R(r) Ylm(,) (8)

where the radial part R depends on the distance from the nucleus (r)whereas the angular part on the direction represented by the angles.

The most accurate model for an atomic orbital is a Slater type func-tion in the form er. However, this model is not efficient for calcula-tion unlike a Gaussian type fuction er

2. On the contrary, Gaussian

function cannot describe properly the region around r = 0 as it hasno cusp. Therefore GTO (Gaussian Type Orbitals) refer to contractedGaussians - a linear combination of Gaussians as close as possible tothe shape of Slater type orbital, STO. The model is never perfectly fit-ted to STO unless the set of Gaussians is infinite, which in reality neverhappens, but in spite of that it is still feasible. [15]

The advantage of the so-called atomic basis sets is they match thelinear combination of atomic orbitals ansatz and fit well with thechemical picture, but it is obvious that they are localised sets with noimplicit periodicity and that they depend strongly on the position ofthe atom. The diffuse part (shallow Gaussians) and polarization func-tions (that is, functions of higher angular quantum number than thehighest occupied orbital in the system) must be expressed via addi-tional terms. Also, there is no direct means of adjusting the particularbasis set, for example if one needs to increase the number of basis func-tions then one needs to find another basis set. Moreover, the core elec-trons are treated explicitly, which might make the calculation overlytedious.

Regarding the forces calculation, which is among other things thebase for theoretical vibrational spectra, the combination of incompletebasis set and dependance of atomic position creates additional force

67

68 P L A N E WAV E S V S . AT O M I C B A S I S S E T

terms (Pulay forces) [42] when calculating the energy derivatives. Itmakes the calculation more demanding. These terms arise from thenonorthonormality of GTO as the first two terms do not vanish in righthand side of the Hellman-Feynman Theorem Equation 9 which statesthe change of the energy (expectation value) with respect to any pa-rameter . The terms would vanish if we used a complete basis set oran atomic-position-independent one.

E

= E

[

+ ]

+

H (9)

P L A N E WAV E S : On the other hand, one can simulate molecules asslightly distorted free electrons and describe the basis set as a linear com-bination of plane waves. The waves are depe

Date post:	10-Oct-2015
Category:	Documents
Upload:	janamathauserova
View:	46 times
Download:	1 times

New molecular materials for nonlinear optics: Preparation and detailed characterisation

Documents