Pentafluorophenylphosphonic Acid as a New Building Block for Molecular Crystal Fabrication

pubs.acs.org/crystal

Article

Pentaﬂuorophenylphosphonic Acid as a New Building Block for

Molecular Crystal Fabrication

Sylvain G. Dutremez,* Xavier Dumail, Sonia Mallet-Ladeira, Arie van der Lee, Dominique Granier,

Nathalie Masquelez, and Jean-S é bastien Filhol

Cite This: Cryst. Growth Des. 2021, 21, 2028 −2045

Read Online

sı

*

Metrics & More

Article Recommendations

Supporting Information

ACCESS

ABSTRACT: Molecular crystals have been prepared from pentaﬂuorophenylphosphonic acid (4). These include 2C₆F₅PO₃H₂·H₂O

+

(4·4·H₂O), [(C₆F₅PO₃H⁻)(H₃O⁺)]·C₆F₅PO₃H₂(4·4⁻·H₃O⁺), [(C₆F₅PO₃H⁻)(NH₄)]·C₆F₅PO₃H₂(4·4⁻·NH₄), [(C₆F₅PO₃H⁻)-

(Me₂NH₂)] (4⁻·DMA⁺), [(C₆F₅PO₃H⁻)(Me₂NH₂)]·H₂O (4⁻·DMA⁺·H₂O), [(C₆F₅PO₃H⁻)(Me₂NH₂)]·C₆F₅PO₃H₂(4·4⁻·

+

DMA⁺), [(C₆F₅PO₃H⁻)(Me₂NH₂)_0.5(NH₄)_0.5] (4⁻·DMA⁺·NH₄), and [(C₆F₅PO₃H⁻)(⁺H₃NCH₂CO₂H)] (4⁻·Gly⁺), where

DMA⁺= dimethylammonium and Gly⁺= glycinium. All of the assemblies incorporate an ammonium cation, a water molecule, or a

hydronium ion in their structure, and these included species act as adhesive agents. They interact via O/N−H···O hydrogen bonds

and C−H···O contacts with PO₃H₂or PO₃H⁻moieties located in polar sheets, forcing the building blocks to assemble with a 2D

layered arrangement. The robustness of this arrangement is reminiscent of that observed in metal arylphosphonates and guanidinium

sulfonates, yet the architectures described herein diﬀer signiﬁcantly from the second family of compounds in two aspects: the lack of

puckering of the layers and the absence of void spaces to accommodate solvent molecules. Interestingly, however, just as in

guanidinium sulfonates, two structural types have been recognized depending on the size of the included species: a single-layer

stacking motif and a bilayer stacking motif. Even though hydrogen bonding is the prevailing interaction in these systems, the use of

perﬂuoroaryl groups is also central, as these moieties bring about weak C−F···π, π···π, C−F···F−C, C−F···H−C, and O/N−H···F−

C interactions that help to increase the cohesion of the nonpolar regions. As a result, new architectures are created that signiﬁcantly

diﬀer in some cases from those prepared from phenylphosphonic acid. However, the weakness of perﬂuoroaryl-based interactions is

also their strength, as these interactions are suﬃciently ﬂexible to allow changes in the organization of the aromatic groups in

response to changes (ionization) occurring in the polar regions, as observed in the structures of 4·4·H₂O and 4·4⁻·H₃O⁺. This

situation is quite unusual and may be regarded as some kind of acidity-modulated polymorphism. The work presented here broadens

the knowledge on molecular crystal formation with arylphosphonic acids, a research area largely dominated by carboxylic and

sulfonic acids. Also, the computational evaluation of 4·4·H₂O and 4·4⁻·H₃O⁺described in this report suggests that this family of

ﬂuorinated compounds may show interesting prospects as molecular semiconductors, a research area that is currently receiving

increased attention.

+

INTRODUCTION

■

Metal phosphonates are an important class of materials with

Received: October 13, 2020

Revised: January 29, 2021

Published: February 24, 2021

applications in ﬁelds as diverse as oxidation catalysis,

magnetism, corrosion protection, surface modiﬁcation of

implants, nuclear waste treatment, and photoluminescence.¹

These hybrid materials exhibit a wide range of architectures,

from discrete species to clusters and extended networks. In this

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

2028

Crystal Growth & Design

pubs.acs.org/crystal

Article

Scheme 1. Synthesis of Pentaﬂuorophenylphosphonic Acid (4)

category of compounds, metal arylphosphonates occupy a

special place due to the fact that a large number of them have a

2D layered arrangement in the solid state. Owing to this

structural peculiarity, metal arylphosphonates are amenable to

intercalation type reactions, exhibit ion exchange properties,

can be used as heterogeneous acid catalysts, and show good

prospects as proton conductors.¹Yet, despite the large body of

literature available on these latter materials, only a limited

number of papers have been published concerning the use of

phenylphosphonic acid in crystal engineering studies.²⁻¹⁷Even

more surprisingly, just one report could be found concerning

the crystallographic characterization of ﬂuorinated arylphos-

phonic acids,¹⁸in sharp contrast with the fact that ﬂuorine-

containing molecules are now emerging as important building

blocks for molecular assembly fabrication.¹⁹We wish to

continue this theme here and report on the synthesis of

molecular crystals²⁰derived from pentaﬂuorophenylphos-

phonic acid (4) (Scheme 1). A related study was published

a few years ago by Desiraju and co-workers on ﬂuorobenzoic

acids.²¹

The ultimate goal of this work was to graft diacid 4 onto

metal oxide surfaces^22,23and then establish if these modiﬁed

surfaces could bind substrates via arene−perﬂuoroarene,

anion−π, and lone pair−π interactions.²⁴⁻³⁸Prior to doing

that, however, we set about studying the crystallization of 4

with guests susceptible to giving host−guest complexes

assembled by such interactions. Numerous attempts have

been made, but regrettably, such assemblies could not be

obtained. In particular, we have found that 4 did not form

multicomponent crystals with benzene, carbon tetrachloride,

halides, and oxo anions. Instead, assemblies organized by

hydrogen bonding were isolated that include 2C₆F₅PO₃H₂·

H₂O (4·4·H₂O), [(C₆F₅PO₃H⁻)(H₃O⁺)]·C₆F₅PO₃H₂(4·4⁻·

molecular orbital arguments that substantiate the validity of 4·

4⁻·H₃O⁺. This situation may be regarded as some kind of

acidity-modulated polymorphism. With regard to the ammo-

nium cations present in the other assemblies, they originate

from the salt or the guest molecule utilized in the

cocrystallization experiment or from decomposition of the

N,N-dimethylformamide (DMF) solvent. Importantly, these

included species act as adhesive agents, as they interact with

the PO₃groups and create polar sheets, forcing the building

blocks to assemble with a 2D layered arrangement. The

robustness of this arrangement is reminiscent of that observed

in metal arylphosphonates and guanidinium sulfonates, the

main diﬀerence from the second family of compounds being

the lack of puckering of the layers.^20,39−43Interestingly, just as

in guanidinium sulfonates, two structural types have been

recognized: the ﬁrst type consists of a single-layer stacking

motif made of an alternation of polar and nonpolar regions

along the stacking direction. In this case, aryl groups

protruding from a polar layer alternatively point up or down

that layer and interdigitate with aromatic moieties from

surrounding layers. The second structural type is bilayer

stacking. In this case, all of the pentaﬂuorophenyl groups

protruding from a polar layer point in the same direction and

do not interdigitate with neighboring aromatic moieties. As a

result, two consecutive polar regions are separated from one

another by two successive layers of C₆F₅groups, instead of just

one in the single-layer stacking motif.

Even though hydrogen bonding is the prevailing interaction

in these systems, the use of perﬂuoroaryl groups is clearly

central, as these moieties bring about weak C−F···π, π···π, C−

F···F−C, C−F···H−C, and O/N−H···F−C interactions that

help to increase the cohesion of the nonpolar regions. As a

result, new architectures are created that signiﬁcantly diﬀer in

some cases from those prepared from phenylphosphonic acid.

Also, compact structures are obtained that show no or

prohibitively small void spaces to accommodate solvent

molecules, in sharp contrast with guanidinium sulfonates.⁴⁴⁻⁴⁶

However, the weakness of perﬂuoroaryl-based interactions is

also their strength, as these interactions are suﬃciently ﬂexible

to allow changes in the organization of the aromatic groups in

response to changes (ionization) occurring in the polar

regions, as observed in the structures of 4·4·H₂O and 4·4⁻·

H₃O⁺. This attribute is linked to the ability of the perﬂuoroaryl

moieties to absorb part of the negative charge of the ionized

phosphonate groups and to interact more strongly with the

perﬂuoroaryl groups of the neutral phosphonic acid molecules.

+

H₃O⁺), [(C₆F₅PO₃H⁻)(NH₄)]·C₆F₅PO₃H₂(4·4⁻·NH₄),

[(C₆F₅PO₃H⁻)(Me₂NH₂)] (4⁻·DMA⁺), [(C₆F₅PO₃H⁻)-

+

(Me₂NH₂)]·H₂O (4⁻·DMA⁺·H₂O), [(C₆F₅PO₃H⁻)-

+

(Me₂NH₂₊)]·C₆F₅PO₃H₂(4·4⁻·DMA⁺), and [(C₆F₅PO₃H⁻)-

+

(Me₂NH₂)_0.5(NH₄)_0.5] (4⁻·DMA⁺·NH₄). Also, crystalliza-

tion of 4 in the presence of glycine gave the salt

[(C₆F₅PO₃H⁻)(⁺H₃NCH₂CO₂H)] (4⁻·Gly⁺), which shows

protonation of glycine instead of the expected COO⁻···C₆F₅

interaction. All of the molecular crystals incorporate an

ammonium cation, a water molecule, or a hydronium ion in

their structure. Isolation of the hydronium-containing structure

4·4⁻·H₃O⁺was totally unexpected due to the existence of the

neutral form 4·4·H₂O, yet computational evaluation of these

two polymorphs at the DFT-D3 level provides energy and

+

2029

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

syringe. A few drops of bromopentaﬂuorobenzene were added to the

ﬂask to initiate the formation of the magnesium derivative. The

bromopentaﬂuorobenzene left in the funnel was diluted with 200 mL

of anhydrous diethyl ether. This solution was added dropwise to the

ﬂask at such a pace that a gentle reﬂux was maintained. When the

addition was complete, the reaction mixture was reﬂuxed for 30 min,

and then it was allowed to reach room temperature with stirring.

In a second 500 mL ﬂask, a solution containing 27 mL of diethyl

chlorophosphite (188 mmol) in 150 mL of anhydrous diethyl ether

was prepared. The pentaﬂuorophenylmagnesium bromide solution

was added dropwise via a canula to the diethyl chlorophosphite

solution while the temperature was kept at −60 °C. The reaction

mixture was stirred at −60 °C for 2 h and then at room temperature

overnight. The solvent was removed in vacuo, resulting in the

formation of a gelatinous solid. n-Pentane was added to the solid to

precipitate the magnesium salts. The salts were removed by ﬁltration

through Celite, and then the solvent was removed in vacuo. This

procedure was repeated a second time so as to ensure removal of the

maximum amount of salts. The crude product was puriﬁed by vacuum

distillation (50−55 °C, 0.01 mbar). The target compound was

isolated as a colorless liquid in 63% yield. NMR (¹H, ³¹P{¹H}) and IR

data for 1 can be found elsewhere.⁵³¹⁹F NMR (376.5 MHz, CDCl₃,

ppm): δ −161.1 (m, 2F; meta), −151.0 (t, ³J_FF= 20.3 Hz, 1F; para),

−135.1 (m, 2F; ortho). ¹³C NMR (100.6 MHz, CDCl₃, ppm): δ 16.9

(d, ³J_PC= 6 Hz; CH₃), 64.4 (d, ²J_PC= 15.4 Hz; CH₂), 115.3 (dm, ¹J_PC

The work presented here broadens the knowledge of

molecular crystal formation with arylphosphonic acids, a

research area largely dominated by carboxylic and sulfonic

acids. In addition, the structural similarities that exist between

derivatives of 4 and metal arylphosphonates suggest that the

former materials may be worthy of further study with regard to

proton conduction and ion exchange applications. Proton

conductivity investigations of 4·4·H₂O and 4·4⁻·H₃O⁺are

currently underway and will be the focus of a separate report.⁴⁷

In connection with ion exchange applications, the exposed

results set the ground for further studies aimed at assessing to

what extent the use of other kinds of organic templates or

metal cations modiﬁes the architectures of the derivatives of 4

and whether the use of biologically relevant amines is suitable

for the preparation of multicomponent assemblies of

pharmaceutical signiﬁcance.^15,16,48,49Finally, the computa-

tional evaluation of 4·4·H₂O and 4·4⁻·H₃O⁺reported in this

work suggests that this family of ﬂuorinated compounds and

their derivatives may show interesting prospects as molecular

semiconductors, a research area that is currently receiving

increased attention.⁵⁰⁻⁵²

= 55.7 Hz; ipso), 137.6 (dm, ¹J_CF= 254 Hz; meta), 142.6 (dm, ¹J_CF

=

EXPERIMENTAL SECTION

■

256 Hz; para), 147.1 (dm, ¹J_CF= 249 Hz; ortho). EI+ MS (70 eV) m/

z (%): 288 (17) [M]⁺, 243 (38) [M − OCH₂CH₃]⁺, 168 (31)

[C₆F₅H]⁺, 121 (11) [P(OCH₂CH₃)₂]⁺.

1

Characterization Methods. Solution H, ¹³C, ¹⁹F, ²⁹Si, and ³¹P

NMR spectra were obtained on Bruker instruments of the following

types: Avance DPX 200, AC 250, and Avance DRX 400. Chemical

1

C₆F₅P(O)(OC₂H₅)₂(2). A 500 mL ﬂask equipped with a magnetic

stir bar was charged with a 24.906 g portion of phosphonite 1 (86.5

mmol). Compound 1 was diluted with 200 mL of acetone. An 8 mL

portion of hydrogen peroxide solution was added dropwise to the

acetone solution while the mixture was cooled with an ice bath. After

addition was complete, stirring was continued overnight. Acetone was

removed on a rotary evaporator. The remaining liquid was transferred

into a separatory funnel. Water was added so as to increase the

volume of the aqueous layer, and then the product was extracted with

dichloromethane (3 × 50 mL). The organic layers were combined

and dried over magnesium sulfate. After removal of the drying agent

and evaporation of the dichloromethane, the product was puriﬁed by

vacuum distillation (78−82 °C, 0.1 mbar). Yield: 50%. NMR (¹H, ¹⁹F,

³¹P{¹H}), IR, and MS data for 2 have been reported previously.^54,55

¹³C NMR (100.6 MHz, CDCl₃, ppm): δ 16.6 (d, ³J_PC= 7 Hz; CH₃),

shifts were referenced as follows: H (protio impurities of the NMR

solvents), ¹³C (NMR solvents), ¹⁹F (CFCl₃), ²⁹Si (Me₄Si), ³¹P (85%

H₃PO₄). Solution-state infrared spectra were recorded in the

transmission mode on a PerkinElmer 1600 Series FT-IR spectrometer

with a 4 cm⁻¹resolution. Solid-state infrared spectra were recorded

with a 4 cm⁻¹resolution on a PerkinElmer Spectrum Two FT-IR

spectrometer equipped with a diamond crystal attenuated total

reﬂectance (ATR) unit. Electron impact (EI) mass spectra were

obtained on a JEOL JMS-DX300 instrument and fast atom

bombardment (FAB) spectra on a JEOL JMS-SX102 A instrument.

Coupled TGA-DSC experiments were conducted under a nitrogen

ﬂow (100 mL min⁻¹) on a TA Instruments SDT-Q600 Simultaneous

TGA/DSC apparatus, with a heating rate of 5 K/min. X-ray powder

patterns were recorded on a PANalytical X’pert MPD-Pro

diﬀractometer in the Bragg−Brentano θ−θ reﬂection geometry with

Ni-ﬁltered Cu Kα radiation (λ = 1.5418 Å). Measurements were

performed at room temperature in the 3−60° 2θ range, using a step

size of 0.033° and a counting time per step of 240 s. Elemental

analysis of 3 was carried out at the Service Central de Microanalyse of

the Centre National de la Recherche Scientiﬁque (CNRS), Vernaison,

France, and that of 4 was performed in house using an Elementar

Vario Micro Cube apparatus.

2

1

64.1 (d, J_PC= 6 Hz; CH₂), 105.0 (dm, J_PC= 185 Hz; ipso), 138.1

1

(dm, J_CF≈ 253 Hz; meta), 144.4 (dm, J_CF≈ 260 Hz; para), 147.8

(dm, J_CF≈ 256 Hz; ortho).

1

C₆F₅P(O)(OSiMe₃)₂(3). A 250 mL ﬂask equipped with a magnetic

stir bar was charged with a 5.353 g portion of phosphonate 2 (17.6

mmol). Compound 2 was diluted with 100 mL of anhydrous

dichloromethane. A 6 mL portion (35 mmol) of iodotrimethylsilane

(purity 83%) was added dropwise to the dichloromethane solution

while the mixture was cooled with an ice bath. Stirring near 0 °C was

continued for 2 h, and then the mixture was warmed to room

temperature overnight. The volatiles were removed in vacuo with

Materials. The chemicals used in this study were obtained from

the following commercial sources: magnesium turnings (Alfa Aesar),

bromopentaﬂuorobenzene (Alfa Aesar), diethyl chlorophosphite (Alfa

Aesar), 30% w/w solution of hydrogen peroxide in water (Fluka),

iodotrimethylsilane (Alfa Aesar), magnesium sulfate (VWR), Celite

(Sigma-Aldrich), ammonium nitrate (Sigma-Aldrich), ammonium

chloride (Alfa Aesar), and glycine (Alfa Aesar). Solvents were

purchased from the following suppliers: diethyl ether (Sigma-

Aldrich), n-pentane (Carlo Erba), acetone (Sigma-Aldrich), dichloro-

methane (Carlo Erba), carbon tetrachloride (VWR), benzene (Alfa

Aesar), methanol (Carlo Erba), N,N-dimethylformamide (Carlo

Erba), and 37% w/w hydrochloric acid solution (VWR). Prior to

use, diethyl ether was distilled over sodium−benzophenone ketyl.

Synthesis. C₆F₅P(OC₂H₅)₂(1). A 500 mL ﬂask equipped with a

magnetic stir bar and a dropping funnel was charged with a 5.0 g

portion of magnesium (206 mmol). The setup was placed under an

inert atmosphere of argon. The magnesium turnings were covered

with 50 mL of dry diethyl ether. The dropping funnel was charged

with a 25 mL portion of bromopentaﬂuorobenzene (201 mmol) via a

1

heating to 70 °C, giving an oily brown solid. Yield: 86%. H NMR

(200.1 MHz, CDCl₃, ppm): δ 0.37 (s, 18H; CH₃). ¹⁹F NMR (235.4

MHz, CDCl₃, ppm): δ −160.7 (m, 2F; meta), −148.2 (broad, 1F;

para), −132.1 (m, 2F; ortho). ¹³C NMR (100.6 MHz, CDCl₃, ppm):

1

δ 1.1 (s; CH₃), 107.7 (dm, J_PC= 193 Hz; ipso), 137.9 (dm, J_CF

≈

1

254 Hz; meta), 143.7 (dm, J_CF≈ 259 Hz; para), 147.2 (dm, J_CF

254 Hz; ortho). ³¹P (81.0 MHz, CDCl₃, ppm): δ − 15.4 (m). ²⁹Si

(39.8 MHz, CDCl₃, ppm): δ +25.2 (s). IR (CCl₄, cm⁻¹): ν 1255 (s,

δ_s(CH₃)), 1228 (m, ν(PO)), 1057 (s, ν(P−O−Si)). EI+ MS (70

eV) m/z (%): 465 (8) [M + Si(CH₃)₃]⁺, 393 (4) [M + H]⁺, 377

(100) [M − CH₃]⁺, 305 (13) [M − CH₃− Si(CH₃)₃+ H]⁺, 73 (92)

[Si(CH₃)₃]⁺. Anal. Calcd for C₁₂H₁₈F₅O₃PSi₂(392.41): C, 36.73; H,

4.62; F, 24.21; P, 7.89. Found: C, 34.85; H, 4.32; F, 22.53; P, 8.05.

C₆F₅P(O)(OH)₂(4). A 250 mL ﬂask equipped with a magnetic stir

bar was charged with an 8.64 g portion of trimethylsilyl ester 3 (22.0

2030

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

2031

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

mmol). Compound 3 was solubilized in 125 mL of dichloromethane.

A 10 mL portion of water (556 mmol) was added to the

dichloromethane solution. The mixture was reﬂuxed for 1 h. Water

was added so as to increase the volume of the aqueous layer, and then

this layer was separated from the organic phase. The aqueous layer

was washed several times with dichloromethane. Phosphonic acid 4

was extracted by washing the aqueous phase with several portions of

diethyl ether. Subsequently, the aqueous layer was saturated with

sodium chloride and washed three more times with diethyl ether. The

ether fractions were combined and dried over magnesium sulfate. A

yellowish powder was recovered after removal of the drying agent and

evaporation of the solvent. A white solid was obtained upon washing

the yellowish powder with dichloromethane and drying in vacuo.

Yield: 70%. The X-ray powder pattern of the solid matches that of 4·

4·H₂O. DSC: 54.1−71.6 °C (endotherm, loss of water), 178 °C

(endotherm, triple point). Literature melting point: 141−142 °C.^56,57

were located in diﬀerence Fourier maps. O−H distances for H₃O⁺in

4·4⁻·H₃O⁺and water molecules in 4·4·H₂O were restrained to

0.85(2) Å with free U_iso(H) values. Likewise, O−H distances of

hydroxyl groups were restrained to 0.84(2) Å in 4⁻·DMA⁺and 4⁻·

Gly⁺, and N−H distances were restrained to 0.91(1) Å in 4⁻·Gly⁺.

For some hydrogen-bonded systems, the SADI instruction was used

to make O−H or N−H distances approximately equal within a certain

standard deviation. For 4⁻·DMA⁺, a model was found in which the

two independent pentaﬂuorophenyl groups are disordered over two

positions with occupancy factors of approximately 0.5. Similarity

restraints on bond lengths and angles as well as on displacement

parameters were used to model these disorders. For 4·4·H₂O, the

crystal investigated was twinned, and the twin law (100, 010, 1/201)

was determined by the TwinRotMat software implemented in

PLATON.⁶¹The two domains found were reﬁned to a volume

ratio of 30:70. Crystallographic parameters and basic information

pertaining to data collection and structure reﬁnement are summarized

in Table 1. Structure drawings were done with OLEX2.⁶²

1

The H, ¹³C{¹⁹F}, ¹⁹F, and ³¹P NMR data of 4 in DMSO-d₆can be

found in ref 18. IR (ATR, cm⁻¹): ν 3250 (v br, ν(O−H)_H2O), 2788 (v

br, ν(O−H)_POH), 2174 (v br, ν(O−H)_POH), 1234 (m, ν(P=O)), 1027

(s, ν(P−OH)), 963 (vs, ν(P−OH)). FAB+ MS (GT) m/z (%): 993

(1) [4M + H]⁺, 745 (5) [3M + H]⁺, 497 (33) [2M + H]⁺, 249 (100)

[M + H]⁺. Anal. Calcd for C₆H₂F₅O₃P·0.5H₂O (257.05): C, 28.04; H,

1.18; O, 21.78. Found: C, 27.82; H, 0.90; O, 21.52.

Calculation Method. Computation of very soft materials with a

cohesive energy mostly dominated by van der Waals forces, at ﬁnite

temperature, is still a great challenge. It requires the use of a

functional that includes dispersion contributions and the utilization of

zero-point energy and ﬁnite temperature corrections by means of a

statistical physics approach.

Crystallization Experiments. 4·4·H₂O. About 50 mg of acid 4

was dissolved in a 8 mL portion of diethyl ether, and a 4 mL portion

of CCl₄was added to the solution. The mixture was allowed to

evaporate slowly to dryness, giving colorless crystals. X-ray diﬀraction

(single crystal and powder) shows that 4·4·H₂O is the only phase

present. Alternatively, single crystals with the same lattice constants

can be obtained upon crystallization of 4 from a 2/1 v/v mixture of

diethyl ether and benzene.

A 148-atom unit cell, C₄₈H₂₄O₂₈F₄₀P₈, and a 222-atom unit cell,

C₇₂H₂₆O₄₂F₆₀P₁₂, were used to compute the structures of 4·4·H₂O

and 4·4⁻·H₃O⁺, respectively. Periodic electronic structure calculations

were performed using density functional theory (DFT) within the

PBE+D3 generalized gradient approximation.⁶³Projected augmented

wave (PAW) pseudopotentials^64,65were utilized as implemented in

the VASP code.^66,67Γ-centered (3 × 3 × 2) and (3 × 2 × 2) k-point

meshes with the high k-point density in the direction of the smallest

unit cell parameter were employed respectively for the 4·4·H₂O and

4·4⁻·H₃O⁺polymorphs. The structures were computed using a 550

eV cutoﬀ (to account for the more electronegative oxygen and

ﬂuorine atoms and to properly describe electronic polarization) with

optimization of the positions of all or just a part of the atoms present

in the unit cell, with or without optimization of the lattice constants.

For all of the calculations, the residual forces on optimized atoms

were lower than 0.01 eV Å⁻¹. Vibrations were computed using ﬁnite

displacements of 0.01 Å to derive the dynamic matrix that was

diagonalized to extract vibrational frequencies. Finite-temperature

corrections and zero-point energies were calculated from phonons

using a physical−statistical approach, and these data were used to

obtain enthalpic and entropic vibrational contributions.

4·4⁻·H₃O⁺. Compound 4 was dissolved in a few milliliters of a

concentrated hydrochloric acid solution. The clear solution was

allowed to evaporate slowly to dryness, giving colorless crystals. X-ray

diﬀraction (single crystal and powder) indicates that a mixture of 4·

4⁻·H₃O⁺and 4·4·H₂O is present.

+

4·4⁻·NH₄. Equimolar amounts of 4 and ammonium nitrate were

dissolved in methanol. The resulting solution was allowed to

evaporate slowly to dryness, giving colorless crystals.

4⁻·DMA⁺. Compound 4 was dissolved in DMF. The solution was

allowed to evaporate slowly to dryness, giving colorless crystals. X-ray

crystallography shows that the crystals contain dimethylammonium

cations due to decomposition of the solvent.

4⁻·DMA⁺·H₂O. Compound 4 was dissolved in DMF. The solution

was allowed to evaporate slowly to dryness, giving colorless crystals.

X-ray crystallography shows that the structure contains dimethylam-

monium cations and water molecules.

RESULTS

4·4⁻·DMA⁺. Compound 4 was dissolved in DMF, and then a few

drops of a concentrated hydrochloric acid solution were added. The

mixture was allowed to evaporate slowly to dryness, giving colorless

crystals.

■

Synthesis. Pentaﬂuorophenylphosphonic acid (4) was

synthesized as shown in Scheme 1. In the ﬁrst step,

pentaﬂuorophenylmagnesium bromide was allowed to react

with diethyl chlorophosphite in anhydrous diethyl ether to

produce diethyl pentaﬂuorophenylphosphonite 1. In the

second step, phosphonite 1 was transformed into phosphonate

2 by oxidation with H₂O₂. Dealkylation of phosphonate 2 was

next achieved by reaction with 2 equiv of iodotrimethylsilane

to generate bis(trimethylsilyl) ester 3. In the last step,

hydrolysis of 3 smoothly gave phosphonic acid 4. The

synthesis of phosphonite 1 was inspired by previous work by

Namestnikov and colleagues.⁵³Phosphonate 2 is also a known

compound that has been prepared in the past via a

photochemical modiﬁcation of the Arbuzov reaction using

(EtO)₃P and C₆F₅I.⁵⁴Alternatively, a two-step process can be

used to make 2 that involves a photochemical reaction between

(EtO)₂POP(OEt)₂and C₆F₅I to produce 1, followed by

oxidation of 1 with Me₃COOH.⁵⁵With regard to diacid 4, this

compound can be obtained by hydrolysis of tetraﬂuoro-

(pentaﬂuorophenyl)phosphorane.⁵⁶In addition, mechanistic

+

4⁻·DMA⁺·NH₄. Equimolar amounts of 4 and ammonium chloride

were dissolved in DMF. The resulting solution was allowed to

evaporate slowly to dryness, giving colorless crystals.

4⁻·Gly⁺. Equimolar amounts of 4 and glycine were dissolved in

water. The resulting solution was allowed to evaporate to dryness,

giving colorless crystals.

Single-Crystal X-ray Diﬀraction. X-ray diﬀraction data were

collected at low temperature (100(2) or 180(2) K) on a Bruker-AXS

APEX II QUAZAR diﬀractometer equipped with a 30 W air-cooled

microfocus source (for 4·4⁻·DMA⁺, 4⁻·Gly⁺, 4⁻·DMA⁺·H₂O, 4·4⁻·

+

H₃O⁺, and 4·4⁻·NH₄), or an Oxford Diﬀraction GEMINI EOS

+

instrument (for 4⁻·DMA⁺·NH₄, 4⁻·DMA⁺, and 4·4·H₂O). Mo Kα

radiation (λ = 0.71073 Å) was used in each experiment. The

structures were solved by direct methods using SHELXS-97⁵⁸or

SIR92,⁵⁹and reﬁnement was carried out by the least-squares method

on F²with use of the SHELXL program.⁶⁰All non-hydrogen atoms

were reﬁned with anisotropic displacement parameters. Hydrogens

bonded to carbon atoms were reﬁned isotropically at calculated

positions using the riding model. All N-bound and O-bound H atoms

2032

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

studies show that 4 forms when solutions of C₆F₅PF₂or

C₆F₅P(OCH₃)₂are treated with HSO₃F, followed by

hydrolysis.⁵⁷Last, quite recently, an alternative strategy to

prepare 4 was reported by a German group.¹⁸The procedure

involves the synthesis of bis(diethylamino)-

pentaﬂuorophenylphosphane by a reaction between

C₆F₅MgBr and ClP(NEt₂)₂, followed by hydrolysis of

C₆F₅P(NEt₂)₂with aqueous HCl to produce the correspond-

ing phosphinic acid. In the last step, mild oxidation of the

phosphinic acid with DMSO and catalytic quantities of iodine

in methyl tert-butyl ether (MTBE) aﬀords 4. Interestingly, in

our hands, we have found that 4 crystallizes with 1/2 mol of

water per mole of acid, as proven by coupled TGA-DSC,

infrared, and elemental analyses. Furthermore, the X-ray

powder pattern of the as-synthesized material matches that

of 4·4·H₂O (vide infra).

assembly. As can be seen from Figure 1, facing sheets of

PO₃H₂groups in the polar region are separated by one sheet of

water molecules. Hydrogen-bonding interactions are observed

between PO₃H₂groups and water molecules (see Table S1 in

the Supporting Information), but no interaction is detected

between PO₃H₂groups from facing layers. In the a direction,

C₆F₅PO₃H₂units arrange in columns due to P−OH···OP

hydrogen-bonding interactions (Table S1), and these columns

are connected with one another by rows of water molecules

along b. Two types of hydrogen-bond networks involving water

molecules are found: P−OH···H₂O···HO−P and P=O···

HOH···OP networks. The separation distance between

water molecules along a alternates between 4.794(4) and

4.801(4) Å. The spacing between these molecules along b is

either 5.247(7) or 5.270(7) Å. Thus, water molecules are

isolated and show no interaction with their neighbors (no H-

bonded water chains are noticed). The separation distance

between mean PO₃H₂planes from facing layers in the polar

region is 4.056 Å, and the separation distances between the

two mean PO₃H₂planes and the mean water plane are about

half of that distance: i.e., 2.004 and 2.052 Å. Furthermore, the

aforementioned O−H···O hydrogen-bond networks are

supplemented by C−F···H−O interactions and C₆F₅···C₆F₅

interactions. The geometrical parameters of the C−F···H−O

interactions (Table S1) are in line with values previously

reported for organics.²⁶With regard to C₆F₅···C₆F₅inter-

actions, these interactions are of two types: C−F···F−C

contacts and C−F···π_Fcontacts.^26,69−72The former inter-

actions exist not only between C₆F₅fragments located in the

same layer but also between C₆F₅fragments situated in facing

layers. Type I and type II C−F···F−C interactions are both

observed (Table S1).^26,69−71Interestingly, it must be pointed

out that intermolecular interactions involving ﬂuorine atoms,

especially those of the F···F type, are more repulsive than

attractive in nature, and this issue has been discussed in detail

in two recent publications.^73,74With regard to C−F···π_F

contacts, these interactions seem to be weak (Table S1).

Certainly, they are much weaker than those observed in the

structure of the anhydrous form of C₆F₅PO₃H₂, for which a

C−F···Cg distance of 3.279 Å is measured with a C−F···Cg

angle of 135.23°.¹⁸

Single-Crystal X-ray Diﬀraction Studies. Structure of 4·

4·H₂O. Crystals of 4·4·H₂O resulted from attempts to prepare

inclusion compounds showing intermolecular C−Cl···π_Fand

π_F···π_Hcontacts,^{24−27,35−38}by crystallization of 4 in the

presence of CCl₄and benzene. X-ray diﬀraction reveals that

the desired assemblies did not form; instead, crystals with

2C₆F₅PO₃H₂·H₂O as a formula moiety were obtained. This

compound crystallizes in the triclinic space group P1 with Z =

4 (Table 1). There are four C₆F₅PO₃H₂molecules in the

asymmetric unit and two water molecules. All of the

components occupy general positions. Variable-temperature

XRD measurements indicate that no phase transition occurs in

the range 100−298 K.

When the structure is viewed down the a axis, a bilayer

stacking motif is evident (Figure 1). In this motif, each polar

Serendipitous Discovery of a Polymorphic Structure of 4·

4·H₂O. The ability of 4 to self-assemble with chloride ions was

tested by crystallization of the latter from a concentrated

hydrochloric acid solution. Interestingly, single-crystal X-ray

diﬀraction showed that the expected assembly did not form

and, instead, crystals with C₆F₅PO₃H₂·C₆F₅PO₃H⁻·H₃O⁺as a

formula moiety were obtained. The unexpected compound, 4·

4⁻·H₃O⁺, crystallizes in the triclinic space group P1 with Z = 6

(Table 1). There are three C₆F₅PO₃H₂molecules in the

asymmetric unit, three C₆F₅PO₃H⁻anions, and three hydro-

nium cations. All of the components occupy general positions.

Variable-temperature XRD measurements indicate that no

phase transition occurs in the range 100−298 K.

Figure 1. Unit cell of 4·4·H₂O viewed along a showing a bilayer

4·4⁻·H₃O⁺is a polymorph of 4·4·H₂O due to protonation of

the included water molecule by one phosphonic acid group.

When the structure is viewed down the a axis (Figure 2), a

bilayer stacking motif is visible which is similar to that

encountered in the structure of 4·4·H₂O. In this motif, each

polar region made up of PO₃H₂groups, PO₃H⁻anions, and

hydronium cations is separated from the next polar region by

two noninterpenetrated layers of C₆F₅fragments. In the a

direction, C₆F₅PO₃H₂and C₆F₅PO₃H⁻units arrange in

stacking motif.

region made up of PO₃H₃groups and water molecules is

separated from the next polar region by two noninterpene-

trated layers of C₆F₅fragments. This stacking motif is

reminiscent of that found in lipid bilayers. Also, a similar 2D

arrangement has been reported recently for cadmium

benzylphosphonate and several ﬂuorinated derivatives.⁶⁸

Water molecules are essential to ensure cohesion of the

2033

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

moieties. A layered A−B−A−B arrangement is evident in a

view of the structure of anhydrous C₆F₅PO₃H₂along a, the

layers piling up along b (Figure 3). This time, however, each

Figure 3. View of anhydrous C₆F₅PO₃H₂along a showing a layered

A−B−A−B arrangement.¹⁸

Figure 2. View along a of 4·4⁻·H₃O⁺showing a bilayer stacking motif.

columns due to P−OH···OP/⁻O−P hydrogen-bonding

interactions (see Table S2 in the Supporting Information).

Interestingly, all of the columns are not identical: one column

out of three is made up solely of C₆F₅PO₃H₂molecules, the

next column is made up of C₆F₅PO₃H⁻anions, and the third

column is a mix of C₆F₅PO₃H₂and C₆F₅PO₃H⁻units. These

three types of columns alternate along b. The columns are

connected with one another by two types of hydrogen-bond

networks: one involving exclusively phosphonate groups and

one involving hydronium cations (Figure 2). Hydronium

cations arrange in columns along a, just as water molecules do

in the structure of 4·4·H₂O. Their spacing in this direction is

either short (3.396(3) or 3.604(3) Å), or long (6.002(3) or

6.247(3) Å). In the b direction, two types of hydronium rows

are visible, one that alternates short (5.398(2) Å) and long

(10.370(2) Å) spacings, and one that has only very long

spacings (15.734(1) Å). However, hydronium cations also

function as a glue between opposing sheets of PO₃H₂/PO₃H⁻

groups in the polar region. As can be seen from Figure 2, there

is no direct hydrogen-bonding interaction between facing

phosphonate groups in the polar region, but the layers are

separated by and hydrogen-bonded to one sheet of hydronium

cations. The separation distance between mean phosphonate

planes from facing layers in the polar region is 4.385 Å, and the

separation distances between the two mean phosphonate

planes and the mean hydronium plane are about half of that

distance: i.e., 2.176 and 2.209 Å. In addition, the structure

exhibits numerous C−F···H−O interactions, C−F···F−C

contacts, and C−F···π_Fcontacts that are comparable in

strength to those encountered in the structure of 4·4·H₂O

(Table S2).

layer consists of columns of hydrogen-bonded C₆F₅PO₃H₂

molecules. In each layer, the columns are all parallel and tilted

with respect to the b direction of the cell. In one layer out of

two, the columns are tilted clockwise by 90°, and in the other

layers the columns are tilted anticlockwise by the same

amount. The main interactions that ensure cohesion between

layers are C−F···π_Fcontacts (Figure 3). Clearly, the structural

diﬀerences that exist between 4·4·H₂O and anhydrous

C₆F₅PO₃H₂are a consequence of the presence of water

molecules in the former compound that keep the PO₃H₂

groups in the same plane.

The structure of 4·4·H₂O is also quite diﬀerent from that of

C₆H₅PO₃H₂. First, unlike 4·4·H₂O, C₆H₅PO₃H₂crystallizes

with a single-layer stacking motif: i.e., the structure consists of

a succession of polar regions made of PO₃H₂moieties and

nonpolar regions made of phenyl groups piling up along c

(Figure 4). Second, there is no water molecule in the crystal

lattice of C₆H₅PO₃H₂; thus, the polar region is quite compact.

Speciﬁcally, in that region, the mean separation distance

between facing phosphonate planes is 2.151 Å at 183 K and

2.204 Å at 295 K.^2,3However, in a fashion analogous to that

for 4·4·H₂O, the structure of C₆H₅PO₃H₂exhibits numerous

C−H···π interactions that ensure cohesion of the nonpolar

region.

Structure of 4·4⁻·NH₄. The ability of 4 to interact with

+

nitrate ions was also tested by crystallization of the latter in the

presence of ammonium nitrate.^28,29,33Crystals of 4·4⁻·NH₄

+

were obtained that have C₆F₅PO₃H₂·C₆F₅PO₃H⁻·NH₄as an

+

empirical formula. 4·4⁻·NH₄crystallizes in the triclinic space

+

group P1 with Z = 2 (Table 1). There is one C₆F₅PO₃H₂

molecule in the asymmetric unit, one C₆F₅PO₃H⁻anion, and

one ammonium cation. All of the components occupy general

positions.

When the structure is viewed down the a axis (Figure 5), a

bilayer stacking motif is visible that is similar to those

encountered in the structures of 4·4·H₂O and 4·4⁻·H₃O⁺. In

this motif, each polar region made up of PO₃H₂groups,

Comparison of the Structure of 4·4·H₂O with Those of

Anhydrous 4 and C₆H₅PO₃H₂. Evidently, the structure of 4·4·

H₂O is quite diﬀerent from that of anhydrous C₆F₅PO₃H₂. The

structure of 4·4·H₂O consists of a succession of layers that are

all parallel and perpendicular to the c axis (Figure 1). The

layers are made of PO₃H₂groups, water molecules, or C₆F₅

2034

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

cations arrange in columns along a, just as hydronium cations

do in the structure of 4·4⁻·H₃O⁺. Their spacing in this

direction is 6.226(1) Å. In the b direction, the separation

distance between ammonium cations is 7.917(2) Å. As can be

seen from Figure 5, each ammonium cation in the polar region

is kept in place tightly by four hydrogen bonds with nearby

oxygen atoms: three strong hydrogen bonds are detected and a

weaker one is found (Table S3). Thus, similarly to 4·4⁻·H₃O⁺,

ammonium cations serve as bridges between C₆F₅PO₃H₂and

C₆F₅PO₃H⁻columns and keep opposing sheets of PO₃H₂/

PO₃H⁻groups in close vicinity in the polar region. No

interaction is detected between phosphonate groups from

facing layers. The separation distance between mean

phosphonate planes from facing layers in the polar region is

4.474 Å, and the separation distance between each mean

phosphonate plane and the mean ammonium plane is half of

that distance: i.e., 2.237 Å. In addition, the structure exhibits

several C−F···H−O interactions, C−F···F−C contacts, and

C−F···π_Fcontacts that are comparable in strength to those

encountered in the structures of 4·4·H₂O and 4·4⁻·H₃O⁺

(Table S3).

Two structures are known in which the ammonium

benzenephosphonate entity is present. The ﬁrst structure was

Figure 4. View along a showing the single-layer stacking motif present

reported in 2008 by Ng et al. and has C₆H₅PO₃H⁻·NH₄as a

+

in C₆H₅PO₃H₂.³

formula moiety.⁵This compound crystallizes in the ortho-

rhombic space group Pbcn with Z = 8, and its 3D arrangement

is a bilayer stacking motif similar to that found in 4·4⁻·NH₄.

+

The second structure was reported in 2005 by an Indian team,

4

and that structure is almost identical with that of 4·4⁻·NH₄.

+

Speciﬁcally, its formula moiety is C₆H₅PO₃H₂·C₆H₅PO₃H⁻·

+

NH₄, it crystallizes in the triclinic space group P1 with Z = 2,

and its cell dimensions are close to those of 4·4⁻·NH₄.

+

Furthermore, its 3D arrangement is a bilayer stacking motif

similar to that found in 4·4⁻·NH₄, with layers consisting of an

+

alternation of C₆H₅PO₃H₂columns and C₆H₅PO₃H⁻columns.

Thus, in this particular case, replacement of the C₆H₅groups

by C₆F₅groups does not generate any signiﬁcant change in the

organization of the building units; this is presumably because

their organization is governed by the phosphonate−phospho-

nate and ammonium−phosphonate hydrogen-bond networks,

not by interactions between aromatic groups.

Structure of 4⁻·DMA⁺. Crystals of 4⁻·DMA⁺were obtained

by crystallization of 4 from N,N-dimethylformamide (DMF).

The formula moiety of 4⁻·DMA⁺is C₆F₅PO₃H⁻·C₂H₈N⁺. This

compound crystallizes in the triclinic space group P1 with Z =

4 (Table 1). There are two pentaﬂuorophenylphosphonate

anions in the asymmetric unit and two dimethylammonium

cations. All of the components occupy general positions.

The dimethylammonium cation present in the structure

does not originate from the salt used during crystallization. It

originates from decomposition of the solvent by the strongly

acidic phosphonic acid. A similar situation has been observed

previously in the synthesis of several hybrid iodoplumbates, in

which hydrolysis of an amide with hydriodic acid was used to

generate the organic cation incorporated in the structure.⁷⁵

Interestingly, unlike the previous compounds, 4⁻·DMA⁺

crystallizes with a single-layer stacking motif: i.e., the structure

consists of a succession of polar regions made of PO₃H⁻

moieties and dimethylammonium cations, separated along b by

interpenetrated layers of pentaﬂuorophenyl groups (Figure 6).

The latter groups are disordered over two positions in a ratio

close to 50:50.

Figure 5. View along a of 4·4⁻·NH₄⁺showing a bilayer stacking motif.

PO₃H⁻anions, and ammonium cations is separated from the

next polar region by two noninterpenetrated layers of C₆F₅

fragments. Organization of the building blocks in the lattice of

4·4⁻·NH₄⁺is reminiscent of that found in 4·4⁻·H₃O⁺. Similarly

to 4·4⁻·H₃O⁺, C₆F₅PO₃H₂and C₆F₅PO₃H⁻units arrange in

columns in the a direction, with one column out of two made

up solely of C₆F₅PO₃H₂molecules, and the next column

containing only C₆F₅PO₃H⁻anions. These two types of

columns alternate along b. In these columns, C₆F₅PO₃H₂

molecules and C₆F₅PO₃H⁻anions are all parallel but do not

interact with another. The separation distance between them is

6.226(1) Å: i.e., the length of the a axis. The columns are

connected with one another by two types of hydrogen-bond

networks, one involving exclusively phosphonate groups and

one involving ammonium cations (Figure 5). Ammonium

2035

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

near the C₆F₅moieties. Thus, for layered materials such as

those studied here, the single-layer stacking motif is the best-

suited arrangement for accommodation of cations with large

sizes.

One structure is known of a phenylphosphonic acid salt

incorporating dimethylammonium cations.⁷The latter struc-

ture is diﬀerent from that of 4⁻·DMA⁺. First, it is a

multicomponent system of empirical formula C₂H₈N⁺·

C₆H₅PO₃H⁻·2C₆H₅PO₃H₂. Second, inspection of its packing

indicates that it has a partially interpenetrated bilayer structure,

with nonpolar regions in which phenyl groups are not all

parallel but some groups are oriented perpendicular to other

groups.

Structure of 4⁻·DMA⁺·H₂O. Crystals of 4⁻·DMA⁺·H₂O

were obtained by crystallization of 4 from DMF. The formula

moiety of 4⁻·DMA⁺·H₂O is C₆F₅PO₃H⁻·C₂H₈N⁺·H₂O. This

compound crystallizes in the orthorhombic space group

P2₁2₁2₁with Z = 4 (Table 1). There is one pentaﬂuor-

ophenylphosphonate anion in the asymmetric unit, one

dimethylammonium cation, and one water molecule. All of

the components occupy general positions.

Figure 6. View along c of 4⁻·DMA⁺showing a single-layer stacking

motif.

As already pointed out in the case of 4⁻·DMA⁺, the

dimethylammonium cation present in the structure originates

from decomposition of the DMF solvent by the strongly acidic

phosphonic acid. Beyond that, three major structural diﬀer-

ences exist between 4⁻·DMA⁺·H₂O and 4⁻·DMA⁺: ﬁrst, one

water molecule is found in the lattice. Second, the compound

crystallizes in the chiral space group P2₁2₁2₁with a Flack

parameter of 0.02(3), meaning that the structure is

homochiral. Third, the pentaﬂuorophenyl groups are not

disordered.

The structure of 4⁻·DMA⁺is organized by numerous O/N−

H···O, C−F···F−C, and C−F···π interactions, as noticed in the

previous structures. Only one C−F···H−O interaction is

detected, but a large number of C−H···O and C−F···H−C

interactions involving the methyl groups of the dimethylam-

monium cations are observed (Table S4). Dimethylammonium

cations are sandwiched between two layers of PO₃H⁻groups

in the polar region. Given the bulkiness of these cations, it was

anticipated that the separation distance between facing PO₃H⁻

planes would be much greater than the distances observed in

the previous structures. Interestingly, however, the opposite

situation is observed: i.e., the separation distance between

mean PO₃H⁻planes is only 3.737 Å. Furthermore, the

separation distances between the two mean PO₃H⁻planes and

the mean dimethylammonium plane are about half of that

distance: i.e., 1.867 and 1.870 Å. Close examination of the

structure of 4⁻·DMA⁺reveals that this is because the methyl

groups protrude in the PO₃H⁻layers and occupy void spaces

The structure of 4⁻·DMA⁺·H₂O looks very similar to that of

4⁻·DMA⁺: it crystallizes with a single-layer stacking motif: i.e.,

the structure consists of a succession of polar regions made of

PO₃H⁻moieties and dimethylammonium cations separated in

the c direction by interpenetrated layers of pentaﬂuorophenyl

groups (Figure 7). Interestingly, unlike Naﬁon, water

molecules are not located in the polar regions but occupy

void spaces between pentaﬂuorophenyl groups in the nonpolar

regions. These molecules show no disorder, as they are at the

Figure 7. View along a of 4⁻·DMA⁺·H₂O showing a single-layer stacking motif.

2036

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

heart of an intricate hydrogen-bond network. First, they display

short O−H···F−C interactions with nearby pentaﬂuorophenyl

groups (Table S5). Second, they interact with phosphonate

groups via O−H···O hydrogen bonds and with dimethylam-

monium cations via C−H···O contacts. Third, they are bonded

to one another and form chains that run along a. The presence

of water molecules in the nonpolar regions has also an eﬀect on

C₆F₅···C₆F₅interactions: because of the increased distance

between aromatic groups, fewer C−F···F−C and C−F···π

interactions are observed in comparison with the previous

structures (Table S5). The separation distance between mean

PO₃H⁻planes from facing layers is quite short, 2.698 Å,

because of the protrusion of the methyl groups of the

dimethylammonium cations in the PO₃H⁻layers. No

interaction is detected between phosphonate groups from

facing layers.

contains only C₆F₅PO₃H⁻anions, resulting in an A⁻A⁻−C⁺−

A⁻A⁻−C⁺stacking pattern (A⁻A⁻, anionic layer made of

phosphonate groups arranged in a head-to-tail fashion; C⁺,

cationic layer). In 4·4⁻·DMA⁺, one phosphonate layer out of

two contains only C₆F₅PO₃H₂molecules, and the next one is

made solely of C₆F₅PO₃H⁻anions, resulting in a NN−C⁺−

A⁻A⁻−C⁺−NN−C⁺arrangement (NN, neutral layer made of

phosphonic acid molecules arranged in a head-to-tail fashion).

Such a remarkable arrangement is not observed in the structure

of the hydro counterpart C₂H₈N⁺·C₆H₅PO₃H⁻·2C₆H₅PO₃H₂.⁷

The architecture of 4·4⁻·DMA⁺is assembled by numerous O/

N−H···O, C−H···O, C−F···H−O/N, and C−F···H−C hydro-

gen-bonding interactions (Table S6), among which quite a few

involve dimethylammonium cations. Interestingly, only a

limited number of C−F···F−C and C−F···π contacts are

detected between aromatic rings. Certainly, the reason for this

situation is not to be sought in the presence of water molecules

between aromatic rings in the nonpolar regions, as no water is

present in 4·4⁻·DMA⁺. Close examination of the structure

reveals the existence of two π-stacking interactions between

ﬂuorinated rings with geometries similar to those found in

regular arenes.^25,27There is one face-to-face interaction and

one parallel-displaced interaction (Table S6). This arrange-

ment is unusual for a perﬂuoroarene and could just be the

result of enforced proximity created by the packing of the

building blocks. Nevertheless, by comparison with the π···π

interactions found in 2,3,4,5,6-pentaﬂuorophenol (Cg···Cg =

4.313 Å)⁷⁶and those found in the bis(2,3,4,5,6-pentaﬂuor-

ophenol)·2,3,4,5,6-pentaﬂuoroaniline molecular crystal (Cg···

Cg ≥ 4.246 Å),^77,78these interactions look reasonably strong.

The separation distance between mean PO₃H₂/PO₃H⁻planes

from facing layers is very short, 1.774 Å, because of the

protrusion of the methyl groups of the dimethylammonium

cations in the layers. Furthermore, unlike the previous

structures, hydrogen-bonding interactions are detected be-

tween PO₃H₂/PO₃H⁻groups from facing layers.

Structure of 4·4⁻·DMA⁺. Crystals of 4·4⁻·DMA⁺resulted

from attempts to crystallize 4 from a DMF solution containing

a few drops of a concentrated hydrochloric acid solution. The

formula moiety of 4·4⁻·DMA⁺is C₆F₅PO₃H₂·C₆F₅PO₃H⁻·

C₂H₈N⁺. This compound crystallizes in the triclinic space

group P1 with Z = 2 (Table 1). There is one pentaﬂuor-

ophenylphosphonic acid molecule in the asymmetric unit, one

pentaﬂuorophenylphosphonate anion, and one dimethylam-

monium cation. All of the components occupy general

positions. Again, the dimethylammonium cation present in

the structure originates from decomposition of the DMF

solvent by strong acid.

As previously noticed in the structures of 4⁻·DMA⁺and 4⁻·

DMA⁺·H₂O, a single-layer stacking motif is evident in the

structure of 4·4⁻·DMA⁺. The layers pile up along c (Figure 8).

Nonetheless, the stacking motif of 4·4⁻·DMA⁺is slightly

diﬀerent from that found in 4⁻·DMA⁺and 4⁻·DMA⁺·H₂O: in

4⁻·DMA⁺and 4⁻·DMA⁺·H₂O, each phosphonate layer

+

Structure of 4⁻·DMA⁺·NH₄. Crystals of 4⁻·DMA⁺·NH₄

were obtained by crystallization of 4 from a DMF solution

containing equimolar amounts of ammonium chloride. The

formula moiety of 4⁻·DMA⁺·NH₄⁺is C₆F₅PO₃H⁻·0.5C₂H₈N⁺·

+

0.5NH₄. This compound crystallizes in the monoclinic space

group C2/c with Z = 8 (Table 1). There is one

pentaﬂuorophenylphosphonate anion in the asymmetric unit,

half of a dimethylammonium cation, and half of an ammonium

cation. The pentaﬂuorophenylphosphonate anion occupies a

general position, whereas the ammonium and dimethylammo-

nium cations sit on 2-fold rotation axes. Dimethylammonium

cations originate from decomposition of the DMF solvent by

the phosphonic acid.

As can be seen in Figure 9, 4⁻·DMA⁺·NH₄⁺crystallizes with

a single-layer stacking motif. The layers pile up along a. This

situation is remarkable, as it shows that when small cations and

larger ones are present at the same time in a layered system like

the ones studied here, the resulting stacking motif is that

imposed by the larger cations. However, this assertion should

be tempered, as the single-layer stacking motif observed in 4⁻·

DMA⁺·NH₄is slightly diﬀerent from those observed up to

+

now. Speciﬁcally, in the present case, each nonpolar region

does not consist of pentaﬂuorophenylphosphonate anions

pointing alternatively upward or downward. Instead, penta-

ﬂuorophenylphosphonate ions assemble in pairs, and each pair

points alternately upward or downward. Consequently, large

hydrophilic channels delineated by NH₂⁺and PO₃H⁻moieties

Figure 8. View along b of 4·4⁻·DMA⁺showing a single-layer stacking

motif.

2037

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

metric unit and two glycinium cations. All of the components

occupy general positions.

A single-layer stacking motif was anticipated for 4⁻·Gly⁺on

the basis of the size of glycine. X-ray diﬀraction shows that this

is true; the direction of the layers coincide with the diagonal of

the ac face of the cell (Figure 10). The arrangement of the

+

Figure 9. View along c of 4⁻·DMA⁺·NH₄showing a single-layer

stacking motif.

are created in the c direction that contain rows of NH₄⁺cations

(Figure 9). This arrangement allows for maximization of

hydrogen-bonding interactions between ammonium and

phosphonate groups (Table S7). Interestingly, it is also

Figure 10. View along b of 4⁻·Gly⁺showing a single-layer stacking

+

reminiscent of the arrangement found in 4·4⁻·NH₄, for

motif.

which a bilayer stacking motif was observed (see above). Thus,

even though the overall structure of 4⁻·DMA⁺·NH₄shows a

+

single-layer stacking motif enforced by the large dimethylam-

monium cations, some bits of it recall the bilayer stacking motif

layers is the same as those found in 4⁻·DMA⁺and 4⁻·DMA⁺·

H₂O: i.e., each nonpolar region consists of pentaﬂuorophe-

nylphosphonate anions pointing alternatively upward or

downward and also all of the layers are identical in the

stacking direction. The nonpolar region shows no abnormal

feature: aromatic groups interact with one another through C−

F···F−C and C−F···π interactions, and the strengths of these

interactions are in line with those detected in the previous

structures. Furthermore, the distance between aromatic groups

is 4.880 Å, suggesting the absence of any signiﬁcant π···π

interaction between them.

The polar region is made up of phosphonate groups and

glycinium ions. The assembly is held together by an intricate

network of hydrogen bonds (Table S8). These interactions are

supplemented by C−F···H−O and C−F···H−C interactions

with nearby C₆F₅groups. As can be seen from Figure 10,

phosphonate groups and glycinium ions assemble in groups of

four to give R²₄(14) and R⁴₄(12) hydrogen-bonded rings. These

rings are connected to one another and form a strip. The width

of the strip is large, 4.119 Å, thereby preventing interaction

between phosphonate groups from facing layers.

A large body of literature is available on the preparation of

self-assembled structures between glycine and a mono-, di-, or

tricarboxylic acid. On the other hand, reports about molecular

crystal formation between glycine and a phosphorus-based acid

are less numerous, and only a handful of crystal structures have

been published on such compounds. For instance, assemblies

prepared from glycine and mono- and diphenyl phosphates

have been characterized crystallographically,^79,80as well as

those obtained from cocrystallization of glycine with

+

observed in the structure of 4·4⁻·NH₄made solely of small

+

NH₄cations. Unsurprisingly, no interaction is detected

between phosphonate groups from facing layers because

these groups are kept apart by ammonium ions. The separation

distance between mean phosphonate planes from facing layers

in the polar region is 4.382 Å, and the separation distance

between each mean phosphonate plane and the mean nitrogen

plane is half of that distance: i.e., 2.191 Å. Hydrophilic

channels are surrounded by hydrophobic channels that are

delineated by (CH₃)₂N and C₆F₅moieties (Figure 9). In this

way, a maximum number of C−F···H−C contacts are

generated (Table S7). Moreover, the channels are empty.

C₆F₅···C₆F₅interactions are also present, but their contribution

to the organization of the structure seems to be much less

signiﬁcant than hydrogen-bonding interactions: no Cg···Cg

contact smaller than 4.0 Å is noticed, and only one C−F···F−C

and three C−F···π interactions are detected (Table S7).

Structure of 4⁻·Gly⁺. The goal of this crystallization

experiment was to force the pentaﬂuorophenyl moiety of 4

to interact with the carboxylate group of an amino acid. The

amino acid that was chosen is glycine. X-ray diﬀraction shows

that a two-component assembly does form, but disappoint-

ingly, the sought-after interaction is not observed. Instead,

protonation of the carboxylate group of the amino acid has

taken place. This result is another demonstration of the

strongly acidic character of 4. The formula moiety of 4⁻.Gly⁺is

C₆F₅PO₃H⁻·HO₂CCH₂NH₃. This compound crystallizes in

+

the monoclinic space group P2₁with Z = 4 (Table 1). There

are two pentaﬂuorophenylphosphonate anions in the asym-

2038

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

Table 2. Computed Energy Values for the H₃O⁺(4·4⁻·H₃O⁺) and H₂O (4·4·H₂O) Forms of Pentaﬂuorophenylphosphonic

Acid

exptl

structure

H-optimized exptl

structure

optimized in exptl

unit cell

fully

optimized

zero-point energy

(harmonic)

ﬁnite-temperature

free energy at

300 K

corrections

+

E_{H O}(eV/

−229.455

−226.490

−2.966

−233.959

−234.046

0.087

−234.293

−234.281

−0.011

−234.293

−234.283

−0.010

5.218

−0.733

−0.724

−0.009

−229.807

−229.772

−0.034

3

fu)

E_{H O}(eV/

fu)

5.233

2

ΔE (eV/fu)

−0.014

phosphoric and phosphorus acids.⁸¹⁻⁸⁶Interestingly, similarly

to 4⁻·Gly⁺, all of these structures show a layered arrangement.

Another single-crystal X-ray diﬀraction analysis of 4⁻·Gly⁺

was carried out at 293 K. At this temperature, the space group

is the same as that found at 180 K, i.e. P2₁, and the structure is

indistinguishable from the low-temperature structure. These

results suggest the absence of a ferroelectric phase transition in

the 180−293 K range, in contrast to the situation observed for

other systems containing glycine.⁸³⁻⁸⁶

that the energy diﬀerence between the fully optimized H₃O⁺

and H₂O phases is again quite small, around 10 meV/fu, but

this time the H₃O⁺phase is slightly more stable. With such a

small energy diﬀerence, zero-point energies and ﬁnite-temper-

ature contributions cannot be neglected and must therefore be

computed.

The H₃O⁺phase is further stabilized by 14 meV/fu relative

to the H₂O phase by zero-point energy (Table 2). The main

reason for this stabilization is the presence of H₃O⁺ions in the

H₃O⁺phase that have, on average, weaker O−H bonds in

comparison to H₂O molecules in the H₂O phase, thereby

resulting in a lower zero-point energy contribution. Finite-

temperature eﬀects associated with enthalpic and entropic

changes also stabilize the H₃O⁺phase by 9 meV/fu; this

stabilization stems from the fact that the H₃O⁺phase has a

slightly higher density of low-energy phonons in comparison to

the H₂O phase. A combination of all of these eﬀects leads to a

34 meV/fu (<1 meV/atom) stabilization of the H₃O⁺phase

relative to the H₂O phase. Clearly, the energy diﬀerence

between the two phases is tiny; thus, it is diﬃcult to conclude

deﬁnitely which phase is more stable because of the inherent

uncertainties of the DFT method. However, it is evident that

the two polymorphs are very close energetically, consistent

with the fact that they both can be isolated experimentally and

suggesting that they may compete with each other depending

on crystallization conditions.

Computational Evaluation of the Structures of

2C₆F₅PO₃H₂·H₂O (4·4·H₂O) and [(C₆F₅PO₃H⁻)(H₃O⁺)]·

C₆F₅PO₃H₂(4·4⁻·H₃O⁺). Energy Calculations. Unlike crystal

structures obtained from neutron diﬀraction experiments, X-

ray crystal structures suﬀer from the large uncertainty on H

atom positions, which is linked to the low diﬀracting power of

the unique electron of hydrogen. Thus, evaluation of the

quality of DFT calculations by direct comparison with

experimental X-ray structures is not possible, as the poor

positioning of H atoms dominates the relaxation energy. As a

result, in a ﬁrst step, we have optimized the positions of the H

atoms in the experimental unit cells, while keeping the

coordinates of the heavier atoms unchanged. By doing so,

stabilization of the systems by up to 7.5 eV/formula unit (fu),

nearly 1 eV/H atom, is observed (Table 2) for 4·4⁻·H₃O⁺

(hereafter called the H₃O⁺form) and 4·4·H₂O (hereafter

called the H₂O form). These calculations conﬁrm the poor

positioning of the hydrogen atoms in the experimental X-ray

structures and, at the same time, allow for a more accurate

picture of the hydrogen-bonding schemes to be obtained. In

addition, for this partial relaxation, it is found that the energy

diﬀerence between the H₃O⁺and H₂O polymorphs is small, 87

meV/fu, with the H₂O phase being more stable. The CIF ﬁles

of the computed structures with optimized H atom positions

are provided in the Supporting Information paragraph.

Then, the overall quality of the calculations relative to the X-

ray structure can be evaluated by optimization of all atom

positions in the experimental unit cell, starting from the H-

optimized structure. In this approach, the energy variation

comes solely from the heavier atoms that are correctly

positioned by X-ray diﬀraction, a large relaxation implying a

poor description of the chemical bonds by DFT (provided that

complicating factors such as phase transitions⁸⁷are absent).

This optimization leads to structures that are only 300 meV/fu

(9 meV/atom) more stable than those derived from experi-

ments. Furthermore, optimization of the unit cells shifts the

experimental lattice parameters by less than 0.6% and leads to

negligible energy changes. These ﬁndings strongly suggest that

the DFT-D3 method used in this work provides a correct

description of the systems under consideration and makes

possible a direct comparison with the experimental results. The

CIF ﬁles of the fully optimized structures are available as in the

Supporting Information paragraph. Furthermore, it is found

Eﬀect of Ionization on Phase Stability. The H₂O and

H₃O⁺phases are polymorphs with two salient diﬀerences: (i)

ionization of one C₆F₅PO₃H₂molecule according to the

equation C₆F₅PO₃H₂+ H₂O ⇄ C₆F₅PO₃H⁻+ H₃O⁺; (ii) the

dissimilar organizations of the C₆F₅groups in the nonpolar

regions (Figure 11).

In the case of the H₃O⁺phase, half of the C₆F₅PO₃H₂

molecules present in the unit cell are ionized. This creates a

mix of neutral C₆F₅PO₃H₂entities and negatively charged

C₆F₅PO₃H⁻species. Ionized species are usually disfavored

relative to neutral species, as charge separation has an

important energy cost. This cost can be estimated by

Figure 11. Organization of the C₆F₅moieties in the computed H₂O

(A) and H₃O⁺(B) polymorphs. Rows of C₆F₅moieties with the same

orientation are colored in blue or yellow.

2039

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

computing the structure of a hypothetical ionized phase

created by hydrogen transfer from half of the C₆F₅PO₃H₂

entities to H₂O molecules in the H₂O phase and calculating

the energy of the local minimum. In this case, it is found that

the ionized structure is 400 meV/fu less stable than the neutral

structure. Conversely, it is possible to compute the structure of

a hypothetical enforced neutral phase created by H atom

transfer from H₃O⁺cations to C₆F₅PO₃H⁻anions in the H₃O⁺

phase. Under such circumstances, the neutral structure is

found to be unstable and converges back to the ionized H₃O⁺

phase, no local minimum being reached. Thus, in the H₃O⁺

phase, there must be some eﬀect that substantially stabilizes

the ionized form relative to the neutral form and that

compensates for the 400 meV/fu associated with ionization.

As the two polymorphs show similar H-bond networks with

about four hydrogen bonds per PO₃group, it is clear that the

origin of this stabilization has to be found elsewhere.

Interestingly, the two polymorphs show very diﬀerent

arrangements of the C₆F₅rings that might account for this

stabilization (see Figure 11).

Eﬀect of Deprotonation on Electronic Structure. We next

investigated the eﬀect of the diﬀering arrangements of the C₆F₅

groups on the electronic structure in the two polymorphs. To

this end, we have computed the electron density variation

associated with the transformation of a neutral H₂O phase into

an ionized H₃O⁺phase in two ways, one using the

experimental H₂O phase (4·4·H₂O) as a starting point and a

hypothetical ionized H₃O⁺phase as the end product (Figure

12, top line), and one using a hypothetical neutral H₂O phase

as a starting point and the experimental ionized H₃O⁺

polymorph (4·4⁻·H₃O⁺) as the end product (Figure 12,

bottom line). The hypothetical ionized H₃O⁺phase used in the

top transformation of Figure 12, abbreviated H₃O⁺(hypo.),

was created by forcing ionization of half of the C₆F₅PO₃H₂

molecules present in the experimental H₂O phase. The neutral

H₂O phase used in the bottom transformation of Figure 12,

abbreviated H₂O(hypo.), was obtained by forced proton

transfer from H₃O⁺cations to C₆F₅PO₃H⁻anions in the

experimental H₃O⁺phase. From the perspective of the C₆F₅

rings, the formation of C₆F₅PO₃H⁻and H₃O⁺species by

protonation of H₂O molecules by PO₃H₂moieties induces an

overall gain in electron density for the cycles connected to

PO₃H⁻groups (see blue isosurfaces in Figure 12).

Concomitantly, a slight electron density loss is observed for

the C₆F₅rings connected to neutral PO₃H₂groups upon

formation of H₃O⁺species: a fraction of the charge density of

the neutral molecule is transferred to H₃O⁺ions, probably

through hydrogen bond strengthening (see yellow isosurfaces

in Figure 12). With ionization, there is also an increase in

electron density between aromatic rings, suggesting the

formation of weak intermolecular bonds (see blue isosurfaces

in Figure 12). Nevertheless, these new bonds are quite

diﬀerent in the two structures. For H₃O⁺(hypo.), the response

to ionization is the formation of pairs of C₆F₅rings (red

ellipse), while for the experimental H₃O⁺structure, there is a

strong delocalization of the electron density over several

aromatic rings that generates some kind of π-conjugated

polymer (red wavy line). This suggests that, in the

experimental H₃O⁺form, stronger intermolecular bonding

interactions are created in comparison with H₃O⁺(hypo.).

Therefore, in 4·4⁻·H₃O⁺, ionization is compensated by (i)

charge delocalization over the pentaﬂuorophenyl moieties, (ii)

a stabilizing electrostatic interaction between positively

charged C₆F₅moieties of C₆F₅PO₃H₂molecules and negatively

charged C₆F₅moieties of C₆F₅PO₃H⁻ions, and (iii) stronger

interactions between the aromatic rings of C₆F₅PO₃H⁻and

C₆F₅PO₃H₂species by intermolecular weak bond formation.

As seen in Figure 12, charge transfer and electrostatic

interactions are also observed in H₃O⁺(hypo.) upon forced

ionization of the experimental H₂O phase, but electron

delocalization and intermolecular bond formation are far

more limited. In this case, formation of a C₆F₅−C₆F₅dimer is

observed, instead of a network.

The increased interaction between C₆F₅PO₃H⁻and

C₆F₅PO₃H₂fragments can be rationalized using the computed

molecular orbital diagram for the free species shown in Scheme

2. Aromatic ring stacking is controlled mainly by the

interaction between associated frontier orbitals with the

correct local symmetry. The relevant orbitals for C₆F₅PO₃H₂

are the LUMO and HOMO-1 (as the HOMO is associated

with the oxygen lone pair), and those for C₆F₅PO₃H⁻are the

LUMO and HOMO-4 (as the HOMO to HOMO-3 orbitals

are associated with the oxygen lone pairs and the P−O bonds).

Upon formation of C₆F₅PO₃H⁻by ionization, the electron-

enriched species sees a shift of its molecular orbitals to higher

energies because of stronger electron−electron repulsion.

Therefore, the HOMO-4 of C₆F₅PO₃H⁻and the LUMO of

C₆F₅PO₃H₂become energetically closer and can interact

strongly, the latter orbital having the correct local symmetry to

interact with the HOMO-4 of C₆F₅PO₃H⁻. In H₃O⁺(hypo.),

only dimers are created, which is not suﬃcient to compensate

for the energy penalty associated with ionization. In the

experimental H₃O⁺polymorph, the arrangement of the C₆F₅

groups is conducive to the formation of a pseudo-1D π-

connected system that takes advantage of charge delocalization

over the neutral molecules and beneﬁts from intermolecular

bond strengthening between aromatic moieties, thereby

Figure 12. Charge density variation upon ionization of half of the

C₆F₅PO₃H₂entities into C₆F₅PO₃H⁻and H₃O⁺. (top row)

Transformation of the experimental H₂O phase into H₃O⁺(hypo.).

(bottom row) Ionization of H₂O(hypo.) into the experimental H₃O⁺

phase. (left column) Yellow isosurfaces showing the localization of

electron-deﬁcient areas. (right column) Blue isosurfaces reproducing

an increase in electron density. Ionization induces the formation of

intermolecular bonds that generate either a dimer (red ellipse) or a

polymer (red wavy line).

2040

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

Scheme 2. Molecular Orbital Diagram Showing the Interactions between C₆F₅PO₃H⁻and C₆F₅PO₃H₂Fragments

compensating for the energy cost caused by ionization. As the

widths of bands pertaining to molecular orbitals are directly

linked to orbital overlaps in crystal structures, delocalization in

the two polymorphs was investigated by looking at their

density of states (DOS). The band at the top of the

conduction band and that at the bottom of the valence band

for both polymorphs are associated mainly with the π and π*

carbon-dominated molecular orbitals of the aromatic moiety

(see Figure S9). The two bands for the H₃O⁺form show larger

widths (by about 100 meV) in comparison to those of the

H₂O form. This suggests a stronger overlap between the

orbitals of the molecules in the H₃O⁺polymorph and a more

eﬃcient delocalization. This greater overlap can also be

checked by computing the Fukui functions (f⁺and f⁻).^88,89

Consistently, the computed Fukui functions (Figure S10)

show that the H₃O⁺phase has more delocalized electrons and

holes in comparison to the H₂O form. This is indicative of the

existence of quite diﬀerent electronic mobilities: the H₂O

polymorph tends to form more localized electrons or holes,

while the H₃O⁺polymorph exhibits more delocalized electrons

or holes. As the last occupied and ﬁrst empty bands of the

H₃O⁺phase have larger widths, this induces a reduction in the

band gap: i.e., 3.10 eV for the H₃O⁺form and 3.21 eV for the

H₂O form.

In summary, subtle structural changes have a strong

inﬂuence on the ionization of pentaﬂuorophenylphosphonic

acid and modify its pK_avalue. Indeed, polymorphism turns

C₆F₅PO₃H₂into a stronger acid in comparison to water in the

H₃O⁺form. This situation is a consequence of eﬃcient charge

reorganization with proton transfer and is linked with a

strengthening of intermolecular interactions. The ionization

cost induced by protonation of water by the acid is

compensated by the more favorable stacking of the C₆F₅

moieties made possible by a better electronic delocalization.

Thus, this family of ﬂuorinated compounds may show

interesting prospects as molecular organic semiconductors,

with their delocalization properties controlled by poly-

morphism or by exchange of the protons with other cations.

Interestingly, this dichotomy in the arrangement of the C₆F₅

groups is observed exclusively in the structures of 4·4·H₂O and

4·4⁻·H₃O⁺.

DISCUSSION

■

Unlike guanidinium sulfonate systems in which guanidinium

cations are essential to promote the crystallization of the

sulfonate anions with a layered arrangement, the construction

from pentaﬂuorophenylphosphonic acid 4 of layered archi-

tectures with alternating polar and nonpolar domains is not

contingent on the presence of a structuring cation. In this

respect, this acid is clearly a unique building block.

Furthermore, it has the amazing ability to cocrystallize with

neutral or cationic species, yielding polar regions where only

neutral 4, only 4⁻, or mixtures of 4 and 4⁻species coexist. This

leads to many diﬀerent structures that are organized by ionic

and H-bonding interactions depending on the protonation

state of the PO₃moieties. Furthermore, as shown computa-

tionally, the polar and nonpolar regions are interdependent, as

ionized PO₃H⁻moieties induce electron donation to the

aromatic rings they are connected to, while PO₃H₂groups are

associated with electron-accepting phenyl rings. As a result, a

change in the protonation state of the phosphonate groups in

the polar region modiﬁes the interactions between aromatic

rings in the nonpolar region, thereby inducing structural

changes. Such changes are evident when the structure of 4·4·

H₂O is compared with that of its 4·4⁻·H₃O⁺polymorph. This

structural versatility suggests that a large variety of assemblies

based on 4 should be accessible, the present work reporting on

just a few of them. In addition, the electronic structure of 4-

based crystals can be tuned to incorporate purely donor, purely

acceptor, or mixed donor/acceptor aromatic rings, giving rise

to electron localization/delocalization phenomena. This makes

these materials potentially useful for certain types of

technological applications. Especially, derivatives of 4 could

be of interest in the ﬁeld of molecular electronics, as shown in

the Computational Evaluation section. Some of the crystals

exhibit electron delocalization and probably electron con-

duction throughout the nonpolar domains; thus, doping could

be achieved by modifying the ionization of 4. Finally, 4-based

materials could also be suitable for molecular recognition

through a modiﬁcation of the ionic and hydrogen bond

networks in the polar regions, causing measurable changes in

the electronic properties of the nonpolar domains. Compound

4 is a new brick for a new class of hydrogen-bonded organic

frameworks (HOFs) featuring the coupling between an ionic

2041

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

state in the polar region and an electronic state in the nonpolar

region.

A great deal remains to be done with 4 and its derivatives in

terms of crystal engineering studies. Especially, we have tried

to cocrystallize phosphonate 2 with halide anions and benzene

molecules, but there again multicomponent crystals could not

be obtained. Perhaps the replacement of the ethyl groups of 2

with substituents known to promote crystallization such as

phenyl and tert-butyl groups will give better results.

Furthermore, work is planned to prepare metal complexes

with 4 to see how the structures of these complexes compare

with those of the phenylphosphonate analogues. Various tests

will then be conducted on these complexes to see if the

absence of groups capable of hydrogen bonding allows for the

existence of arene−perﬂuoroarene, anion−π, and lone pair−π

interactions and if these architectures are capable of including

solvent molecules. Finally, of special note is the fact that the

para position of the perﬂuorophenyl ring is activated;¹⁸thus, it

is possible to replace the ﬂuorine atom borne by that carbon by

other groups, which makes easy derivatization of 4 or

complexed 4 possible.

CONCLUSION

■

In conclusion, the solid-state structures of pentaﬂuorophenyl-

phosphonic acid and its salts are organized primarily by

hydrogen bonding. The chief hydrogen-bonding interactions

are O/N−H···O interactions, but C−H···O contacts are also

important in stabilizing structures that contain organic cations.

Interactions involving perﬂuorophenyl groups are central as

well, and these interactions can have various forms: namely,

C−F···H−O/N, C−F···H−C, C−F···F−C, C−F···π_F, and

occasionally π_F···π_Fcontacts. Although the energies of these

interactions are low in comparison with O/N−H···O contacts,

their large number undoubtedly helps to increase the stability

of the architectures by enhancing cohesion in the nonpolar

regions. Thus, novel architectures diﬀering signiﬁcantly from

those prepared from phenylphosphonic acid are sometimes

obtained. Furthermore, perﬂuoroaryl-based interactions are

suﬃciently ﬂexible to allow changes in the organization of the

aromatic groups in response to changes (ionization) occurring

in the polar regions, as observed in the structures of 4·4·H₂O

and 4·4⁻·H₃O⁺. This situation is quite unusual and may be

regarded as some kind of acidity-modulated polymorphism.

Clearly, the use of perﬂuorophenyl moieties is an asset in

crystal engineering and supramolecular chemistry studies due

to the plasticity of this grouping.

ASSOCIATED CONTENT

* Supporting Information

The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acs.cgd.0c01402.

■

sı

OLEX2 drawings showing the contents of the

asymmetric units with atom-labeling schemes, tables of

distances and angles for intermolecular hydrogen-

bonding interactions and short contacts, projected

densities of states on the carbon atoms of the aromatic

ring for the H₂O and H₃O⁺phases, and Fukui functions

(f⁺and f⁻) associated respectively with electrons (in the

conduction band) or holes (in the valence band) for the

H₃O⁺and H₂O phases (PDF)

DFT-optimized structure H2O.Hopt (CIF)

DFT-optimized structure H2O.allopt (CIF)

DFT-optimized structure H3O.optH (CIF)

DFT-optimized structure H3O.allopt (CIF)

The size of included species (cation or neutral molecule) is

important as well, and this situation is reminiscent of that

observed in guanidinium sulfonates. With small cations or

water molecules, bilayer stacking is possible because of the

compact size of the cementing agent and the ability to form

multiple hydrogen bonds with nearby phosphonate groups in

the polar region. With bulky cations, bilayer stacking is

unfavorable because the hydrogen-bond network in the polar

region would be loose, and the separation distance between

phosphonate layers would be quite large. Instead, single-layer

stacking is observed. Such an arrangement creates void spaces

near the pentaﬂuorophenyl groups where the bulky cation can

nest. In addition, this arrangement takes advantage of the

numerous hydrogen bonds between the methyl groups of the

cations and the ﬂuorine atoms of the perﬂuorophenyl moieties

located nearby. Ultimately, one can surmise that, with very

bulky cations, the layered arrangement will be disrupted and

reduced dimensionality will result (columns or discrete

assemblies will be produced instead of sheets). Such a situation

has been observed previously for the hydrogenated counter-

parts: namely, 1,4-diazabicyclo(2.2.2)octane bis-

(phenylphosphonic acid),¹⁰dicyclohexylammonium hydrogen

phenylphosphonate,⁹diisopropylammonium hydrogen phenyl-

phosphonate,⁸and (triphenylmethyl)ammonium hydrogen

phenylphosphonate.⁶

In addition, assemblies with compact nonpolar regions are

obtained that show no or prohibitively small void spaces to

accommodate solvent molecules, in sharp contrast with

guanidinium sulfonates. According to PLATON analyses,⁶¹

none of the structures described herein contain residual

solvent-accessible voids. Thus, it is no surprise to ﬁnd that it is

impossible to incorporate halides, oxo anions, carbon

tetrachloride, or benzene molecules in them. If such an

endeavor could be achieved, it would certainly lead to

complete structural rearrangement to compensate for the

large energy cost it would entail.

Accession Codes

CCDC 1882394−1882401 contain the supplementary crys-

tallographic data for this paper. These data can be obtained

free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by

emailing data_request@ccdc.cam.ac.uk, or by contacting The

Cambridge Crystallographic Data Centre, 12 Union Road,

Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

AUTHOR INFORMATION

Corresponding Author

■

Sylvain G. Dutremez − Institut Charles Gerhardt, UMR 5253

CNRS-ENSCM-UM, Equipe CMOS, Université de

Montpellier, 34095 Montpellier Cedex 5, France;

orcid.org/0000-0003-0765-3779; Phone: +33-4-67-14-

42-23; Email: dutremez@univ-montp2.fr; Fax: +33-4-67-

14-38-52

Authors

Xavier Dumail − Institut Charles Gerhardt, UMR 5253

CNRS-ENSCM-UM, Equipe CMOS, Université de

Montpellier, 34095 Montpellier Cedex 5, France

Sonia Mallet-Ladeira − Institut de Chimie de Toulouse (FR

2599), Université Paul Sabatier, 31062 Toulouse Cedex 9,

France

2042

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

Hydrogenphosphonate and Phenylphosphonate. Jordan J. Chem.

2011, 6, 113−121.

(13) Singh, U. P.; Kashyap, S.; Butcher, R. J.; Singh, H. J.; Mishra, B.

K. Effect of electron donating substituents on supramolecular

structure of salts having phenylphosphonic acid and pyrazoles. Struct.

Chem. 2011, 22, 931−941.

(14) Singh, U. P.; Narang, S. A supramolecular approach towards the

construction of molecular salts using phosphonic acid and pyrazole.

CrystEngComm 2014, 16, 7777−7789.

(15) Weakley, T. J. R. The Crystal Structures of Guanidinium

Phosphonate - Acetone (1/1) and Guanidinium Phenylphosphonate -

Water (1/2). J. Chem. Soc. Pak. 1980, 2, 145−150.

Arie van der Lee − Institut Européen des Membranes, UMR

5635 CNRS-ENSCM-UM, Université de Montpellier, 34095

Montpellier Cedex 5, France; orcid.org/0000-0002-4567-

1831

Dominique Granier − Institut Charles Gerhardt, UMR 5253

CNRS-ENSCM-UM, Réseau Rayons X et Gamma,

Université de Montpellier, 34095 Montpellier Cedex 5,

France

Nathalie Masquelez − Institut Européen des Membranes,

UMR 5635 CNRS-ENSCM-UM, Université de Montpellier,

34095 Montpellier Cedex 5, France

(16) Latham, K.; Downs, J. E.; Rix, C. J.; White, J. M. Interactions of

guanidinium with benzene-sulphonic, -phosphonic and -arsonic acids

and several of their nitro-derivatives. J. Mol. Struct. 2011, 987, 74−85.

(17) Gregson, R. M.; Glidewell, C.; Ferguson, G.; Lough, A. J. meso-

5,5,7,12,12,14-Hexamethyl-1,4,8,11-tetraazacyclotetradecane as a

building block in supramolecular chemistry; salts formed with 2,2 ′-

biphenol, 4,4 ′-biphenol, 4,4 ′-thiodiphenol, 4,4 ′-sulfonyldiphenol, 3-

and 4-hydroxybenzoic acids, 3,5-dihydroxybenzoic acid and phenyl-

phosphonic acid; supramolecular structures in zero, one, two and

three dimensions. Acta Crystallogr., Sect. B: Struct. Sci. 2000, 56, 39−

57.

Jean-Sébastien Filhol − Institut Charles Gerhardt, UMR

5253 CNRS-ENSCM-UM, Equipe CTMM, Université de

Montpellier, 34095 Montpellier Cedex 5, France;

orcid.org/0000-0002-3681-9267

Complete contact information is available at:

https://pubs.acs.org/10.1021/acs.cgd.0c01402

Notes

The authors declare no competing ﬁnancial interest.

(18) Alter, C.; Neumann, B.; Stammler, H.-G.; Hoge, B. Bis-

(diethylamino)pentafluorophenylphosphane as Valuable Precursor for

the Design of Tetrafluorophenylphosphanes, Tetrafluorophenylphos-

phinic and -phosphonic Acids. Eur. J. Inorg. Chem. 2018, 867−875.

(19) Panini, P.; Chopra, D. Understanding of Noncovalent

Interactions Involving Organic Fluorine. Lect. Notes Chem. 2015,

87, 37−67.

ACKNOWLEDGMENTS

■

The computational work described in this study was performed

using HPC resources from GENCI-CINES (Grant No. 2020-

A0080910369).

REFERENCES

(20) Russell, V. A.; Ward, M. D. Molecular Crystals with

Dimensionally Controlled Hydrogen-Bonded Nanostructures. Chem.

Mater. 1996, 8, 1654−1666.

■

(1) Clearﬁeld, A.; Demadis, K. Metal Phosphonate Chemistry: From

Synthesis To Applications; RSC Publishing: Cambridge, U.K., 2012.

(2) Weakley, T. J. R. Benzenephosphonic Acid. Acta Crystallogr.,

Sect. B: Struct. Crystallogr. Cryst. Chem. 1976, 32, 2889−2890.

(3) Mahmoudkhani, A. H.; Langer, V. Phenylphosphonic acid as a

building block for two-dimensional hydrogen-bonded supramolecular

arrays. J. Mol. Struct. 2002, 609, 97−108.

(4) Rao, K. P.; Vidyasagar, K. Syntheses, Structure and Intercalation

Properties of Low-Dimensional Phenylphosphonates, A(HO₃PC₆H₅)-

(H₂O₃PC₆H₅) (A = Alkali Metal, NH₄and Tl). Eur. J. Inorg. Chem.

2005, 4936−4943.

(5) Lin, Z.; Lei, X.-Q.; Bai, S.-D.; Ng, S. W. Ammonium

benzenephosphonate. Acta Crystallogr., Sect. E: Struct. Rep. Online

2008, 64, No. o1607.

(6) Yuge, T.; Kai, N.; Hisaki, I.; Miyata, M.; Norimitsu, T.

Topologically Different Two Pseudo-decahedral Hydrogen-bond

Networks Composed of Triphenylmethylammonium Phosphonates.

Chem. Lett. 2007, 36, 1390−1391.

(7) Beckmann, J.; Dakternieks, D.; Duthie, A. Dimethylammonium

phenylphosphonate ·2(phenylphosphonic acid). Appl. Organomet.

Chem. 2003, 17, 817−818.

(8) Sarr, M.; Boye, M. S.; Diasse-Sarr, A.; Grosjean, A.; Guionneau,

P. Diisopropylammonium hydrogen phenylphosphonate. Acta Crys-

tallogr., Sect. E: Struct. Rep. Online 2012, 68, No. o3078.

(9) Diop, T.; Diop, L.; Maris, T.; Stoeckli-Evans, H. Dicyclohex-

ylammonium hydrogen phenylphosphonate. Acta Crystallogr., Sect. E:

Struct. Rep. Online 2012, 68, No. o1432.

(10) Ferguson, G.; Glidewell, C.; Gregson, R. M.; Meehan, P. R.

Chain, Sheet and Framework Structures in the Adducts of

Phenylphosphonic Acid with 4,4 ′-Bipyridyl (1/1), Piperazine (2/1)

and 1,4-Diazabicyclo[2.2.2]octane (2/1). Acta Crystallogr., Sect. B:

Struct. Sci. 1998, 54, 129−138.

(11) Fuller, J.; Heimer, N. E. Substituted imidazolium phenyl-

phosphonate salts for nonlinear optical applications. J. Chem.

Crystallogr. 1995, 25, 129−136.

(12) Abu-Salem, Q.; Sweidan, K.; Mallah, E.; Joshi, R.; Mubarak, M.

S.; Steinmann, M.; Voelter, W. Hydrogen-Bonded Phosphorous

Acids. Synthesis and Structure of Imidazolium-Containing Salts of

(21) Dubey, R.; Pavan, M. S.; Desiraju, G. R. Structural landscape of

benzoic acid: using experimental crystal structures of fluorobenzoic

acids as a probe. Chem. Commun. 2012, 48, 9020−9022.

(22) Li, H.; Ratcliff, E. L.; Sigdel, A. K.; Giordano, A. J.; Marder, S.

R.; Berry, J. J.; Brédas, J.-L. Modification of the Gallium-Doped Zinc

Oxide Surface with Self-Assembled Monolayers of Phosphonic Acids:

A Joint Theoretical and Experimental Study. Adv. Funct. Mater. 2014,

24, 3593−3603.

(23) Sang, L.; Mudalige, A.; Sigdel, A. K.; Giordano, A. J.; Marder, S.

R.; Berry, J. J.; Pemberton, J. E. PM-IRRAS Determination of

Molecular Orientation of Phosphonic Acid Self-Assembled Mono-

layers on Indium Zinc Oxide. Langmuir 2015, 31, 5603−5613.

(24) Patrick, C. R.; Prosser, G. S. A Molecular Complex of Benzene

and Hexafluorobenzene. Nature 1960, 187, 1021.

(25) Meyer, E. A.; Castellano, R. K.; Diederich, F. Interactions with

Aromatic Rings in Chemical and Biological Recognition. Angew.

Chem., Int. Ed. 2003, 42, 1210−1250.

(26) Reichenbächer, K.; Suss, H. I.; Hulliger, J. Fluorine in crystal

̈

engineering-“the little atom that could ”. Chem. Soc. Rev. 2005, 34, 22−

30.

(27) Salonen, L. M.; Ellermann, M.; Diederich, F. Aromatic Rings in

Chemical and Biological Recognition: Energetics and Structures.

Angew. Chem., Int. Ed. 2011, 50, 4808−4842.

(28) Quinonero, D.; Garau, C.; Rotger, C.; Frontera, A.; Ballester,

̃

̀

P.; Costa, A.; Deya, P. M. Anion-π interactions: do they exist? Angew.

Chem., Int. Ed. 2002, 41, 3389−3392.

(29) Schottel, B. L.; Chifotides, H. T.; Dunbar, K. R. Anion-π

interactions. Chem. Soc. Rev. 2008, 37, 68−83.

(30) Robertazzi, A.; Krull, F.; Knapp, E.-W.; Gamez, P. Recent

advances in anion-π interactions. CrystEngComm 2011, 13, 3293−

3300.

(31) Gale, P. A.; Busschaert, N.; Haynes, C. J. E.; Karagiannidis, L.

E.; Kirby, I. L. Anion receptor chemistry: highlights from 2011 and

2012. Chem. Soc. Rev. 2014, 43, 205−241.

(32) Giese, M.; Albrecht, M.; Rissanen, K. Anion-π Interactions with

Fluoroarenes. Chem. Rev. 2015, 115, 8867−8895.

2043

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

(33) Giese, M.; Albrecht, M.; Rissanen, K. Experimental

investigation of anion-π interactions - applications and biochemical

relevance. Chem. Commun. 2016, 52, 1778−1795.

oligophenyl organic semiconductors: a design approach based on

experimental and computational perspectives. J. Mater. Chem. C 2019,

7, 3881−3888.

(52) Cagardová, D.; Matuska, J.; Michalík, M.; Poliak, P.; Lukes, V.

Theoretical comparative study of promising semiconducting aromatic

molecules and their fluorinated counterparts. Synth. Met. 2020, 260,

No. 116263.

(53) Namestnikov, V. I.; Trishin, Yu. G.; Bel’skii, V. K. Synthesis of

substituted 5,6,7,8-tetrafluoro-1,4-dihydrobenzo[e]-1,2,4-diazaphos-

phorin by the reaction of diethyl pentafluorophenylphosphonite

with C-(p-nitrophenyl)-N-phenylnitrilimine. Russ. J. Gen. Chem. 1998,

68, 1334−1335.

(54) Burton, D. J.; Flynn, R. M. A Facile Synthesis of Trifluoro-

methane- and Pentafluorobenzenephosphonates. Synthesis 1979, 615.

(55) Nair, H. K.; Burton, D. J. Facile Synthesis of Fluorinated

Phosphonates via Photochemical and Thermal Reactions. J. Am.

Chem. Soc. 1997, 119, 9137−9143.

́

̌

(34) Bauzá, A.; Mooibroek, T. J.; Frontera, A. Towards design

strategies for anion-π interactions in crystal engineering. CrystEng-

Comm 2016, 18, 10−23.

(35) Wang, H.; Wang, W.; Jin, W. J. σ -Hole Bond vs π -Hole Bond: A

Comparison Based on Halogen Bond. Chem. Rev. 2016, 116, 5072−

5104.

(36) Egli, M.; Sarkhel, S. Lone Pair-Aromatic Interactions: To

Stabilize or Not to Stabilize. Acc. Chem. Res. 2007, 40, 197−205.

(37) Mooibroek, T. J.; Gamez, P.; Reedijk, J. Lone pair-π

interactions: a new supramolecular bond? CrystEngComm 2008, 10,

1501−1515.

(38) Neel, A. J.; Hilton, M. J.; Sigman, M. S.; Toste, F. D. Exploiting

non-covalent π interactions for catalyst design. Nature 2017, 543,

637−646.

(39) Russell, V. A.; Etter, M. C.; Ward, M. D. Layered Materials by

Molecular Design: Structural Enforcement by Hydrogen Bonding in

Guanidinium Alkane- and Arenesulfonates. J. Am. Chem. Soc. 1994,

116, 1941−1952.

(40) Russell, V. A.; Etter, M. C.; Ward, M. D. Guanidinium Para-

Substituted Benzenesulfonates: Competitive Hydrogen Bonding in

Layered Structures and the Design of Nonlinear Optical Materials.

Chem. Mater. 1994, 6, 1206−1217.

(41) Russell, V. A.; Ward, M. D. Solid-state structure of a layered

hydrogen-bonded salt: guanidinium 5-benzoyl-4-hydroxy-2-methox-

ybenzenesulfonate methanol solvate. Acta Crystallogr., Sect. B: Struct.

Sci. 1996, 52, 209−214.

(42) Russell, V. A.; Ward, M. D. Two-dimensional hydrogen-bonded

assemblies: the influence of sterics and competitive hydrogen bonding

on the structures of guanidinium arenesulfonate networks. J. Mater.

Chem. 1997, 7, 1123−1133.

(43) Burke, N. J.; Burrows, A. D.; Mahon, M. F.; Warren, J. E.

Incorporation of dyes into hydrogen-bond networks: The structures

and properties of guanidinium sulfonate derivatives containing ethyl

orange and 4-aminoazobenzene-4 ′-sulfonate. Cryst. Growth Des. 2006,

6, 546−554.

(44) Horner, M. J.; Holman, K. T.; Ward, M. D. Lamellae-nanotube

isomerism in hydrogen-bonded host frameworks. Angew. Chem., Int.

Ed. 2001, 40, 4045−4048.

(45) Ward, M. D. Design of crystalline molecular networks with

charge-assisted hydrogen bonds. Chem. Commun. 2005, 5838−5842.

(46) Horner, M. J.; Holman, K. T.; Ward, M. D. Architectural

Diversity and Elastic Networks in Hydrogen-Bonded Host Frame-

works: From Molecular Jaws to Cylinders. J. Am. Chem. Soc. 2007,

129, 14640−14660.

(47) It is worth noting that the proton conductivity of anhydrous 4

has just been reported. See: Alter, C.; Wiesemann, M.; Neumann, B.;

Stammler, H.-G.; Hoge, B. Tetrafluoroaryl Phosphonic Acid

Functionalized Polyphosphazenes - Synthesis, Characterization, and

Evaluation of Proton Conductivity. Eur. J. Inorg. Chem. 2020, 4401−

4408.

(56) Furin, G. G.; Terent’eva, T. V.; Rezvukhin, A. I.; Yakobson, G.

G. Aromatic fluoro derivatives. XI. Polyfluorinated phenylphospho-

nium ions. Zh. Obshch. Khim. 1975, 45, 1473−1479.

(57) Furin, G. G.; Krupoder, S. A.; Rezvukhin, A. I.; Kilina, T. M.;

Yakobson, G. G. Aromatic fluoroderivatives. XCV. The investigation

of the behaviour of the polyfluoroaromatic compounds containing

group VA elements in acid media. J. Fluorine Chem. 1983, 22, 345−

375.

(58) Sheldrick, G. M. A short history of SHELX. Acta Crystallogr.,

Sect. A: Found. Crystallogr. 2008, 64, 112−122.

(59) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.

Completion and Refinement of Crystal Structures with SIR92. J. Appl.

Crystallogr. 1993, 26, 343−350.

(60) Sheldrick, G. M. Crystal structure refinement with SHELXL.

Acta Crystallogr., Sect. C: Struct. Chem. 2015, 71, 3−8.

(61) Spek, A. L. Single-crystal structure validation with the program

PLATON. J. Appl. Crystallogr. 2003, 36, 7−13.

(62) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A.

K.; Puschmann, H. OLEX2: a complete structure solution, refinement

and analysis program. J. Appl. Crystallogr. 2009, 42, 339−341.

(63) Grimme, S.; Ehrlich, S.; Goerigk, L. Effect of the damping

function in dispersion corrected density functional theory. J. Comput.

Chem. 2011, 32, 1456−1465.

(64) Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B:

Condens. Matter Mater. Phys. 1994, 50, 17953−17979.

(65) Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the

projector augmented-wave method. Phys. Rev. B: Condens. Matter

Mater. Phys. 1999, 59, 1758−1775.

(66) Kresse, G.; Hafner, J. Ab initio molecular-dynamics simulation

of the liquid-metal-amorphous-semiconductor transition in germa-

nium. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 49, 14251−

14269.

(67) Kresse, G.; Furthmuller, J. Efficiency of ab-initio total energy

̈

calculations for metals and semiconductors using a plane-wave basis

set. Comput. Mater. Sci. 1996, 6, 15−50.

(68) Akhmetova, I.; Beyer, S.; Schutjajew, K.; Tichter, T.; Wilke, M.;

Prinz, C.; Martins, I. C. B.; Al-Sabbagh, D.; Roth, C.; Emmerling, F.

Cadmium benzylphosphonates - the close relationship between

structure and properties. CrystEngComm 2019, 21, 5958−5964.

(69) Pavan, M. S.; Durga Prasad, K.; Guru Row, T. N. Halogen

bonding in fluorine: experimental charge density study on

intermolecular F ···F and F ···S donor-acceptor contacts. Chem.

Commun. 2013, 49, 7558−7560.

(48) Sopkova-de Oliveira Santos, J.; Montouillout, V.; Fayon, F.;

̀

Fernandez, C.; Delain-Bioton, L.; Villemin, D.; Jaffres, P.-A. Assembly

of benzene-1,3,5-tris(methylenephosphonic acid) and guanidinium

salt: Single crystal-X-ray characterization and ³¹P solid state NMR

investigations. New J. Chem. 2004, 28, 1244−1249.

(49) Plabst, M.; Stock, N.; Bein, T. Hydrogen-Bonded Networks of

Ammonium Bis- and Tetrakisphosphonates-Synthesis, High-Through-

put Screening and Structural Trends. Cryst. Growth Des. 2009, 9,

5049−5060.

(70) Mukherjee, A.; Tothadi, S.; Desiraju, G. R. Halogen Bonds in

Crystal Engineering: Like Hydrogen Bonds yet Different. Acc. Chem.

Res. 2014, 47, 2514−2524.

(50) Maiti, B.; Schubert, A.; Sarkar, S.; Bhandari, S.; Wang, K.; Li,

Z.; Geva, E.; Twieg, R. J.; Dunietz, B. D. Enhancing charge mobilities

in organic semiconductors by selective fluorination: a design approach

based on a quantum mechanical perspective. Chem. Sci. 2017, 8,

6947−6953.

(71) Yamamura, M.; Ikuma, S.; Nabeshima, T. Herringbone

structures of 2,7-dihalogenated acridine tailored by halogen-halogen

interactions. J. Mol. Struct. 2015, 1093, 59−64.

(72) Prasanna, M. D.; Guru Row, T. N. C-halogen ···π interactions

and their influence on molecular conformation and crystal packing: a

database study. Cryst. Eng. 2000, 3, 135−154.

(51) Maiti, B.; Wang, K.; Bhandari, S.; Bunge, S. D.; Twieg, R. J.;

Dunietz, B. D. Enhancing charge mobilities in selectively fluorinated

2044

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Crystal Growth & Design

pubs.acs.org/crystal

Article

(73) Vasylyeva, V.; Merz, K. Crystal architecture of the low melting

nitrogen heterocycles tetrafluoropyrimidine and trifluorotriazine. Z.

Kristallogr. - Cryst. Mater. 2014, 229, 603−608.

(74) Shishkin, O. V.; Merz, K.; Vasylyeva, V.; Zubatyuk, R. I.

Isotypic Transformation Principle in Molecular Crystals. Analysis of

Supramolecular Architecture of Fluorinated Benzenes and Pyridines.

Cryst. Growth Des. 2018, 18, 4445−4448.

(75) Eppel, S.; Fridman, N.; Frey, G. Amide-Templated

Iodoplumbates: Extending Lead-Iodide Based Hybrid Semiconduc-

tors. Cryst. Growth Des. 2015, 15, 4363−4371.

(76) Das, D.; Banerjee, R.; Mondal, R.; Howard, J. A. K.; Boese, R.;

Desiraju, G. R. Synthon evolution and unit cell evolution during

crystallization. A study of symmetry-independent molecules (Z ’ > 1)

in crystals of some hydroxy compounds. Chem. Commun. 2006, 555−

557.

(77) Gdaniec, M. On the polymorphs of pentafluorophenol and its

2:1 co-crystal with pentafluoroaniline. CrystEngComm 2007, 9, 286−

288.

(78) Kirchner, M. T.; Bläser, D.; Boese, R.; Desiraju, G. R. Additive

induced polymorphism. The pentafluorophenol-pentafluoroaniline

system. CrystEngComm 2009, 11, 229−231.

(79) Głowiak, T.; Szemik, A. W. Crystal structure and spectroscopic

properties of glycinium monophenylphosphate. J. Crystallogr.

Spectrosc. Res. 1986, 16, 79−89.

(80) Głowiak, T.; Wneķ , I. Crystal and molecular structure of

glycinium diphenylphosphate. J. Crystallogr. Spectrosc. Res. 1985, 15,

157−171.

(81) Thulasidhass, R.; Mohana Rao, J. K. The crystal structure of

Glycine orthophosphate. Curr. Sci. 1979, 48, 626−627.

(82) Averbuch-Pouchot, M. T.; Durif, A.; Guitel, J. C. Structures of

Glycine Monophosphate and Glycine cyclo-Triphosphate. Acta

Crystallogr., Sect. C: Cryst. Struct. Commun. 1988, 44, 99−102.

(83) Averbuch-Pouchot, M.-T. Structures of Glycinium Phosphite

and Glycylglycinium Phosphite. Acta Crystallogr., Sect. C: Cryst. Struct.

Commun. 1993, 49, 815−818.

́

(84) Baran, J.; Łukaszewicz, K.; Pietraszko, A.; Sledz, M. Structural

investigations of the ferroelectric glycinium hydrogenphosphite (GPI)

and its deuterated analogue (DGPI) I. X-ray diffraction studies of the

crystal structure of paraelectric and ferroelectric phases. J. Mol. Struct.

2002, 611, 155−168.

(85) Shikanai, F.; Komukae, M.; Czapla, Z.; Osaka, T. Crystal

Structure of NH₃CH₂COOHH₂PO₃in the Ferroelectric Phase. J.

Phys. Soc. Jpn. 2002, 71, 498−503.

(86) Lapshin, A. E.; Smolin, Yu. I.; Pankova, G. A. Crystal structure

of glycine phosphite C₂H₅NO₂·H₃PO₃. J. Struct. Chem. 2005, 46,

315−319.

(87) Duarte Rodrigues, A.; Fahsi, K.; Dumail, X.; Masquelez, N.; van

der Lee, A.; Mallet-Ladeira, S.; Sibille, R.; Filhol, J.-S.; Dutremez, S. G.

Joint Experimental and Computational Investigation of the Flexibility

of a Diacetylene-Based Mixed-Linker MOF: Revealing the Existence

of Two Low-Temperature Phase Transitions and the Presence of

Colossal Positive and Giant Negative Thermal Expansions. Chem. -

Eur. J. 2018, 24, 1586−1605.

(88) Parr, R. G.; Yang, W. Density Functional Approach to the

Frontier-Electron Theory of Chemical Reactivity. J. Am. Chem. Soc.

1984, 106, 4049−4050.

(89) Filhol, J.-S.; Doublet, M.-L. Conceptual Surface Electro-

chemistry and New Redox Descriptors. J. Phys. Chem. C 2014, 118,

19023−19031.

2045

https://dx.doi.org/10.1021/acs.cgd.0c01402

Cryst. Growth Des. 2021, 21, 2028−2045

Article Doi

Pentafluorophenylphosphonic Acid as a New Building Block for Molecular Crystal Fabrication

DOI: 10.1021/acs.cgd.0c01402

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.cgd.0c01402

Full text of DOI:10.1021/acs.cgd.0c01402

Products guided by the article

R&D Labs maybe for 14655-90-4

Relevant to this article

Hot Product