Characterization of proton conducting polyphosphate composites

1
Characterization of proton conducting
polyphosphate composites
D. Freude2, S. Haufe3, D. Prochnow2, H.Y. Tu1, U.
Stimming11Technische Universität München,
2Universität Leipzig ,3Proton Motor Fuel Cell
GmbH, Germany
1894 Wilhelm Ostwald demonstrates that fuel
cells are not limited by the Carnot efficiency.
2001 Composite electrolytes preparation,
characterization and investigation of the
conductivity PhD thesis by Stefan Haufe
2002 Solid-state MAS NMR studies of composite
material were performed in the high field up to
17 T (750 MHz) and at temperatures of about 530 K
(maximum 850 K by laser heating), PhD thesis by
Daniel Prochnow.
2
Synthesis of polyphosphate composite
silicon
nitrogen phosphorus oxygen
XRD-structure of NH4PO3
XRD-structure of (NH4)2SiP4O13
Preparation of NH4PO3 NH4H2PO4 (NH2)2
NH4PO3 (modification I)
NH4PO3 (modification II)
Preparation of composite 10 NH4PO3 SiO2
6 NH4PO3 / (NH4)2SiP4O13
3
Characterization by XRD, CA, REM
Chemical analysis Composition of the material is
3.7 wt H,11.5 wt N, 29.6 wt P and 2.9 wt
Si.It yields NH4PO36(NH4)2SiP4O131.
XRD
REM
X-ray diffraction indicates the presence of
NH4PO3 in modifications I and IIand
(NH4)2SiP4O13 as well.
Particle size 5 15 mm
C.Y. Shen, N.E. Stahlheber and D.R. Dyroff, J.
Am. Chem. Soc. 91 (1969) 62-67
4
Characterization by TG
Termogravimetry was performed with a heating rate
of 10 K/min and a helium flow of 100 mL/min.
After an initial mass loss (mostly NH3) of 7
the material is thermally stable upon cycling
between 50 C and 300 C.
5
Conductivity measurements

• Increase in conductivity after heating from room
  temperature up to 300 C parallelto the mass loss
  of NH3 observed by thermal gravimetric analysis.
• The conductivity does not exhibit any significant
  changes with further heating-cooling cycles. The
  values reach from 110-7 S/cm at 50 C to 210-2
  S/cm at 300 C.
• The temperature dependent dc conductivity
  measurements in a two chamber hydrogen cell
  reveal that the ionic conductivity is a proton
  conductivity. The conductivities measured by ac
  and dc techniques coincide.

Arrhenius plot of conductivity measured by ac
impedance spectroscopy in dry hydrogen
6
Gas variation

• Varying the gas environment from dry to humid
  hydrogen has a dramatic effect. Due to water
  uptake of the sample, the conductivity increases
  reversibly by almost an order of magnitude.
• Activation energies vary from 0.5 eV to 1.0 eV in
  dry atmosphere and 0.1 eV to 0.2 eV in humid
  atmosphere at 300 C and 50 C, respectively.

Arrhenius plot of conductivity after activation
of composite material measured in dry hydrogen,
dry oxygen, dry argon and humid hydrogen
7
NMR measurements
NomenclatureQ0 isolated PO4-tetrahedrons, Q1
chain end groups, Q2 middle groups in chain
anions
31P MAS NMRT 297 K
31P MAS NMR spectrum of APP-II at ?rot  10 kHz.
Asterisks denote spinning side bands.
31P MAS NMR spectrum of ASiPP at ?rot  10 kHz.
Asterisks denote spinning side bands.

• Four Q2-signals due to four non-crystallographic
  sites in ASiPP (cf. XRD)
• Chain length about 500 Q-units in ASiPP

• One Q2-signal according to one non-crystallographi
  c site in APP-II (cf. XRD)
• Chain length about 150 Q-units

• Q0-signal due to impurities

8
NMR spectra of the composite at T 297 K
31P MAS NMR
1H MAS NMR
Sum of the spectra of APP-II and ASiPP
Composite (non-activated)
ASiPP
Composite activated
Composite (non-activated)
APP
1H MAS NMR spectrum of non-activated composite
and its single components
31P MAS NMR spectrum of non-activated composite
compared to the spectral addition of single
components

• Proton resonance in spectra of APP is assigned to
  NH4 species (d 7.0 ppm)
• Additional resonance at d 9.0 ppm in spectra of
  ASiPP is due to protons in hydrogen bridges
• Only one signal at d 7.3 ppm in the spectrum of
  the non-activated composite

• Spectrum of (non-activated) composite shows the
  same 31P resonance positions with the same
  chemical shift anisotropies as observed in the
  single components.
• Chain length dramatically decreased upon
  composition (5 Q-units) and increases again after
  activation up to 50 Q-units.

9
1H MAS NMR between 297 K and 580 K
Second cycle
Activation in the MAS rotor
T 580K
T 297K
T 297K
First heating and subsequent cooling observed by
1H MAS NMR. During the activation process a
second signal arises due to the ammoniac loss.
This new signal, which is assigned to protons in
bridging positions, seems to be responsible for
the high protonic conductivity.
No further signals arise or vanish during cycling
after activation. The 1H MAS NMR spectrum is
reversible.
10
Chemical exchange and line merging
Theoretical dependence of the line shape on the
exchange rate k for a two-spin-system Three
cases for k Dn two lines are observed (slow
exchange), for k ? Dn one very broad signal that
often cannot be observed for k Dn one narrow
signal at the averaged line position is observed
(fast exchange).

• 1H MAS NMR spectrum of activated composite shows
  two signals at 297 K.
• At higher temperatures the signals are broadened
  and merge to one line.
• It can be concluded that a chemical exchange
  takes place between the two species.

11
Determination of exchange rates
Peak intensities in deopendence on themixing time
(T320 K)
2D-EXSY spectrum of an activated composite. T
297K, ?mix  10 ms.

• Exchange rates k were measured between 297 K and
  440 K using 1D NOESY NMR.
• The analysis of the peak intensities in
  dependence on the mixing time gives the exchange
  rates.

• An Arrhenius-plot of k for temperatures above 370
  K yields an activation energy of 0.8 eV

• The presence of cross peaks indicates the
  chemical exchange.

12
Diffusion measurements with PFG and SFG NMR
sequence te Adiff attenuation due to diffusion ArAr1Ar2attenuation due to relaxation
PFG 2t1t2 exp-g2G2Dd2(D-d/3) exp-2t1/T2-t2/T1
SFG 2t1t2 exp-g2G2D t12(t2 2t1/3) exp-2t1/T2-t2/T1
T / K

• Proton diffusion measurements were performed by
  means of PFG (Pulsed Field Gradient) NMR at nL
  400 MHz up to 450 K and SFG (Stray Field
  Gradient) NMR at nL 118 MHz up to 600 K.
• The activation energy of the diffusion
  coefficient (about 0.3 eV) is to compare with the
  ac conductivity activation energies varying from
  0.5 eV to 1.0 eV in dry atmosphere.

13
Conclusions

• It is well-known that ammonium polyphosphate
  composites combine the high protonic conductivity
  and mechanical stability and exhibit interesting
  properties as an electrolyte in the
  intermediate-temperature fuel cells.
• The prepared ammonium polyphosphate composites
  contain the phases of (NH4)2SiP4O13 as well as of
  NH4PO3, modification I and II. The composite
  shows thermo-chemical stability after the first
  heating cycle.
• The composite also exhibits high conductivity in
  humid atmosphere. The change from humid to dry
  atmosphere causes a reversible decrease in the
  electrical conductivity by some orders of
  magnitude.
• A comparison of ac and dc experiments reveals
  that the electrical conductivity relates to
  proton conductivity.
• 1H MAS NMR measurements demonstrate that
  (non-ammonium) bridging protons are created by
  the activation procedure of the composite.
• 31P MAS NMR measurements show that the
  phosphorous chain length of about 500 Q-units in
  APP decreases upon composition to a value of 5
  for ASiPP and increases again after activation up
  to 50.
• A chemical exchange between ammonium and bridging
  protons can be observed. Above 380 K the
  activation energy of the exchange rate amounts to
  0.8 eV.
• NMR diffusion coefficients yield an activation
  energy of about 0.3 eV. This is to compare with
  the ac conductivity activation energies varying
  from 0.5 eV to 1.0 eV in dry atmosphere.

T. Kenjo and Y. Ogawa, Solid State Ionics 76
(1995) 29-34
S. Haufe, Thesis, Technical University of Munich,
2002
D. Prochnow, Thesis in preparation, University of
Leipzig