Quantitative single molecule measurements on the interaction forces of poly(L-glutamic acid) with calcite crystals

Article abstract of DOI:10.1021/ja074070i

The interaction between poly(L-glutamic acid) (PLE) and calcite crystals was studied with AFM-based single molecule force spectroscopy. Block copolymers of poly(ethylene oxide) (PEO) and PLE were synthesized and covalently attached to the tip of an AFM cantilever. In desorption measurements the molecules were allowed to adsorb on the calcite crystal faces and afterward successively desorbed. The corresponding desorption forces were detected with high precision, showing for example a force transition between the two blocks. Because of its importance in the crystallization process in biominerals, the PLE-calcite interaction was investigated as a function of the pH as well as the calcium concentration of the aqueous solution. The sensitivity of the technique was underlined by resolving different interaction forces for calcite (104) and calcite (100).

Full text of DOI:10.1021/ja074070i

Published on Web 11/16/2007
Quantitative Single Molecule Measurements on the Interaction
Forces of Poly(L-glutamic acid) with Calcite Crystals
Lars Sonnenberg,*,† Yufei Luo, Helmut Schlaad, Markus Seitz,
‡
‡
†
Helmut C o¨ lfen,*,‡ and Hermann E. Gaub
†
Contribution from the Lehrstuhl f u¨ r Angewandte Physik and Center for NanoScience,
Ludwig-Maximilians-UniVersit a¨ t M u¨ nchen, Amalienstrasse 54, 80799 M u¨ nchen, Germany, and
Max-Planck-Institute of Colloids and Interfaces, Colloid Chemistry, Research Campus Golm,
14424 Potsdam, Germany
Received June 5, 2007; E-mail: sonnenberg@physik.uni-muenchen.de; coelfen@mpikg.mpg.de
Abstract: The interaction between poly(L-glutamic acid) (PLE) and calcite crystals was studied with AFM-
based single molecule force spectroscopy. Block copolymers of poly(ethylene oxide) (PEO) and PLE were
synthesized and covalently attached to the tip of an AFM cantilever. In desorption measurements the
molecules were allowed to adsorb on the calcite crystal faces and afterward successively desorbed. The
corresponding desorption forces were detected with high precision, showing for example a force transition
between the two blocks. Because of its importance in the crystallization process in biominerals, the PLE-
calcite interaction was investigated as a function of the pH as well as the calcium concentration of the
aqueous solution. The sensitivity of the technique was underlined by resolving different interaction forces
for calcite (104) and calcite (100).
Introduction
called double hydrophilic block copolymers, which allow for a
precise control of crystallization because of the separation of
Biominerals are composite or hybrid materials consisting of
an inorganic crystal phase and an organic phase with outstanding
materials properties and hierarchical organization. These
advanced features are the result of a precise control of
crystallization and self-organization processes by soluble and
insoluble organic molecules. Biominerals are thus considered
as archetypes for future advanced materials.
10,11
binding and stabilizing moieties.
Nevertheless, detailed and
quantitative knowledge of the polymer-crystal interaction is
still lacking and with it a quantitative explanation for face-
specific polymer binding (e.g., certain acidic proteins interact
1
1
2,13
selectively in vitro with calcite crystal faces
), which could
verify computer modeling predictions on such interactions.14 It
is therefore not only important for the fundamental understand-
ing of biomineralization processes to quantify the polymer-
crystal interactions but also for the understanding of additive-
controlled crystallization in general concerning topics such as
Many of the soluble organic polymers, which can be extracted
2-5
from biominerals, are extraordinarily acidic. Matrix proteins
in biominerals typically contain 30-40% of aspartic acid (D)
or glutamic acid (E) residues.6 The acidic residues can interact
with calcium, which is the basis for many biominerals like
calcium carbonates and phosphates. Acidic macromolecules are
therefore considered to play a key role in biomineralization
,7
9,12,13
15
face-selective polymer adsorption,
oriented attachment,
16
or mesocrystal formation, where face-selective additive ad-
sorption is of importance for the subsequent morphogenesis
process by self-assembly to ordered nanoparticle superstructures
or by further crystal growth. The collection of such data is thus
a most important prerequisite to transfer the lessons from nature
into the synthetic materials world.
5
processes. On the basis of this knowledge, synthetic model
polymers have been designed and widely applied for the control
of crystallization processes. Especially successful are the so-
8,9
AFM-based single-molecule force spectroscopy allows for
the precise measurement of the interaction forces of single
†
Ludwig-Maximilians-Universit a¨ t M u¨ nchen.
Max-Planck-Institute of Colloids and Interfaces.
‡
17,18
(
(
(
(
(
1) Mann, S. Biomineralization: Principles and Concepts in Bioinorganic
Materials Chemistry; Oxford University Press: New York, 2001.
2) Lowenstam, H. A.; Weiner, S. On Biomineralization; Oxford University
Press: New York, 1989.
3) Simkiss, K.; Wilbur, K. M. Biomineralization: cell biology and mineral
deposition; Academic Press: San Diego, 1989.
4) Wheeler, A. P.; Sikes, C. S.; Williams, R. J. P., Ed.; VCH: Weinheim,
polymers with solid (mineral) surfaces.
In so-called desorp-
(10) C o¨ lfen, H. Macromol. Rapid Commun. 2001, 22, 219-252.
(11) Yu, S. H.; C o¨ lfen, H. J. Mater. Chem. 2004, 14, 2124-2147.
(12) Addadi, L.; Berkovitchyellin, Z.; Weissbuch, I.; Vanmil, J.; Shimon, L. J.
W.; Lahav, M.; Leiserowitz, L. Angew. Chem., Int. Ed. 1985, 24, 466-
485.
1
989; p 95.
5) Gotliv, B. A.; Kessler, N.; Sumerel, J. L.; Morse, D. E.; Tuross, N.; Addadi,
L.; Weiner, S. ChemBioChem 2005, 6, 304-314.
(13) Addadi, L.; Weiner, S. Angew. Chem., Int. Ed. 1992, 31, 153-169.
(14) Shi, W. Y.; Wang, F. Y.; Xia, M. Z.; Lei, W.; Zhang, S. G. Acta Chim.
Sinica 2006, 64, 1817-1823.
(15) Wang, T. X.; Reinecke, A.; C o¨ lfen, H. Langmuir 2006, 22, 8986-8994.
(16) C o¨ lfen, H.; Antonietti, M. Angew. Chem., Int. Ed. 2005, 44, 5576-5591.
(17) Seitz, M.; Friedsam, C.; J o¨ stl, W.; Hugel, T.; Gaub, H. E. ChemPhysChem
2003, 4, 986-990.
(
6) Weiner, S. Calcif. Tissue Int. 1979, 29, 163-167.
(
7) Wheeler, A. P.; Low, K. C.; Sikes, C. S. ACS Symposium Series 444;
American Chemical Society: Washington, DC, 1991; pp 72-84.
8) Xu, A. W.; Ma, Y. R.; C o¨ lfen, H. J. Mater. Chem. 2007, 17, 415-449.
9) C o¨ lfen, H. Top. Curr. Chem. 2007, 271, 1-77.
(
(
15364
9
J. AM. CHEM. SOC. 2007, 129, 15364-15371
10.1021/ja074070i CCC: $37.00 © 2007 American Chemical Society

Interaction Forces of Poly(L-glutamic acid) with Calcite Crystals
A R T I C L E S
tion measurements single molecules are covalently attached to
an AFM tip and are allowed to adsorb on the mineral substrate.
Then, they are successively desorbed in an equilibrium process
by separating the AFM tip from the substrate surface (Figure
3). The measured force-distance traces exhibit plateaus of
constant force, from which the specific interaction of the
polymeric molecules and the mineral surface can be obtained
in a quantitative manner.
In order to elucidate the yet unknown polymer interaction
with the mineral phase, we have applied AFM-desorption
measurements to quantitatively investigate the interaction of
poly(L-glutamic acid) (PLE) with calcite (CaCO3) crystal faces
on the single molecule level. For this, a triblock copolymer
consisting of a polyamine block for chemical coupling to the
AFM tip, a PEO block as hydrophilic spacer, and the PLE block
as interacting block was synthesized and coupled to the AFM
tip. In the context of this study, two key aspects were
addressed: (a) The impact of experimental conditions (pH, Ca-
concentration) on the molecular interaction of PLE with the
calcite surface, and (b) the impact of crystal face on the
molecular interaction of PLE with the calcite surface.
Materials and Methods
Polymer Synthesis (Polyamine-b-PEO-b-PLE). A double hydro-
philic block copolymer consitsting of poly(ethylene oxide) (PEO) and
poly(L-glutamic acid) (PLE) and which also carries primary amine
functions was designed for the AFM-based single molecule force
spectroscopy measurements. The polyamine block allows for the
covalent attachment of the block copolymer to the AFM tip with high
efficiency and long-term stability because of the multiple binding sites
(see also further below). For this study the PLE was of main importance
as its interaction with calcite crystal surfaces was investigated. The
PEO has been included for several reasons: (i) PEO-b-PLE form
together a double hydrophilic block copolymer, which are of special
importance for drug delivery, as crystallization additive or particle
stabilizer in water (ref 10), (ii) in single molecule force spectroscopy
experiments by means of an AFM, PEO is often used as a spacer to
avoid interactions of the molecules of interest with the (tip) surface.
All reagents and solvents were purchased from Sigma-Aldrich. The
synthetic pathway is visualized in Figure 1.
Figure 1. Schematics of the synthesis of PEO-b-PLE carrying primary
amine functions.
1
9
bromide (0.02 g) were suspended in 10 mL of N,N-dimethylformamide
(DMF) under vigorous stirring. The mixture was gradually warmed to
1
00 °C in an oil bath and stirred at this temperature overnight. After
cooling down to room temperature, the mixture was condensed to 2
mL and precipitated into diethyl ether. The obtained yellow solid was
dissolved in CHCl
PB-b-PEO-phthalimide) was then precipitated again in diethyl ether
and dried in vacuum.
c) PB-b-PEO-phthalimide (0.2 g) was dissolved in a mixture of
3
and the solution passed through a filter. The product
ω-Hydroxy-polybutadiene-block-poly(ethylene oxide) (PB-b-PEO-
OH). PB-b-PEO-OH was synthesized by anionic polymerization of
ethylene oxide (EO) using an ω-hydroxy-1,2-polybutadiene (average
number of repeating units: 40), as reported by Justynska et al.20 The
characterization with nuclear magnetic resonance spectroscopy (NMR)
and size exclusion chromatography (SEC) yielded the following
(
(
dioxane (20 mL) and methanol (10 mL), and then hydrazine mono-
hydrate (1 mL) was added. The solution was heated to reflux with
stirring overnight. After the solvent was evaporated, 0.5 N aqueous
HCl was added and the mixture stirred for 15 min. After dilution with
1
results: H NMR: mole fraction EO ) 0.87, SEC:PDI ) 1.05 (apparent
polydispersity index).
ω-Amino-polybutadiene-block-poly(ethylene oxide) (PB-b-PEO-
NH ). (a) PB-b-PEO-OH (0.5 g) was dissolved in 20 mL of CH Cl ,
2 2 2
1
5 mL of water and heating for 1 h, the solution was filtered and
dialyzed against deionized water (MWCO: 3500 Da). The PB-b-PEO-
NH was then isolated by freeze-drying.
2
and then p-tosyl chloride (0.19 g) and triethylamine (0.095 g) were
added under an argon atmosphere. The solution was heated to reflux
overnight. The solution was afterward concentrated to about 5 mL and
precipitated into diethyl ether. The product (PB-b-PEO-OTos) was
collected through centrifugation, reprecipitated three times, and dried
in vacuum.
γ-Benzyl L-Glutamate-N-carboxyanhydride (BLE-NCA). γ-Ben-
zyl L-glutamate (5 g) was suspended in a mixture of R-pinene (6.57
mL, 5.74 g) and ethyl acetate (20 mL) (freshly distilled from CaH ).
2
The suspension was heated to reflux (76 °C), and a solution of
triphosgene (3.13 g) in ethyl acetate (9.5 mL) was slowly added under
stirring. After a clear solution was formed, 75 mL of heptane was slowly
added under good stirring. The solution was cooled to -10 °C, kept
overnight at this temperature, and then rapidly filtered. The solid NCA
was recrystallized twice from ethyl acetate/heptane 1:4 (v/v), washed
(b) In a 25 mL three-necked round-bottom flask, PB-b-PEO-OTos
(0.2 g), potassium phthalimide (0.093 g) and tetra-n-butylammonium
(
18) Smith, B. L.; Schaffer, T. E.; Viani, M.; Thompson, J. B.; Frederick, N.
A.; Kindt, J.; Belcher, A.; Stucky, G. D.; Morse, D. E.; Hansma, P. K.
Nature 1999, 399, 761-763.
2
1
twice with heptane, and dried in vacuum at room temperature.
Polybutadiene-block-poly(ethylene oxide)-block-poly(γ-benzyl L-
glutamate) (PB-b-PEO-b-PBLE). PB-b-PEO-NH and BLE-NCA
2
(
19) Kataoka, K.; Harada, A.; Nagasaki, Y. AdV. Drug DeliVery ReV. 2001, 47,
1
13-131.
(
20) Justynska, J.; Hordyjewicz, Z.; Schlaad, H. Polymer 2005, 46, 12057-
2064.
1
(21) ISOCHEM. French Patent EP 1 201 659 A1, 2002.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 49, 2007 15365

A R T I C L E S
Sonnenberg et al.
on a solid surface and afterward successively desorbed by separating
tip and surface. When the tip is approaching the bare surface, polymers
will adsorb onto it. When retracting the AFM tip, the molecules
successively desorb from the surface. Therefore, desorption takes place
under equilibrium conditions because of the much faster dynamics of
the polymer-surface contacts in comparison to the retraction velocity
of the AFM tip (ca. 1 µm/s). In other words, the polymers are highly
mobile on the surface and are able to rearrange on the surface rather
than being pinned to the surface.26 As a consequence, the polymer chains
are simply “pulled-off” the surface and the desorption process is not
loading-rate-dependent, i.e., the detected force is not a function of the
retraction velocity of the cantilever. Under these conditions the
desorption process manifests itself as a plateau of constant force in the
force-distance curves. The height of the force plateau represents the
magnitude of the desorption force between the polymer chain and the
surface, and the mean desorption force is obtained by a statistical
analysis. Note that the desorption force can only be determined with
high accuracy from the last plateau (Fdes in Figure 3), as no
intermolecular interactions are contributing here.
Figure 2. Attachment of PEO-b-PLE carrying primary amine functions at
the R-chain end to an epoxy-functionalized AFM tip.
were weighted in separate two-neck flasks and dried overnight at room
temperature in vacuum. The flasks were filled with argon and DMF
2 2 5
(distilled from CaH and from P O ) was added (3 mL and 5 mL,
respectively). The mixtures were stirred at room temperature until
complete dissolution, and then the solution of the NCA was added to
the polymer via a transfer needle. The combined solution was stirred
for 3 days at 60 °C, and then the solution was cooled down to room
temperature and concentrated to 3 mL. The product was afterward
precipitated into diethyl ether and dried under reduced pressure.
In the case of low lateral mobility of the polymers, the observed
force-extension traces are significantly altered; the flat desorption
plateaus pass into stretching curves (from which the elasticity of single
26,27
polymer chains can be deduced) or a combination of both.
However,
if the lateral mobility of the polymers is high as throughout in the
experiments described here, the molecules rearrange on the surface to
maximize the adsorption enthalpy. Flat force plateaus in the force-
extension spectra are then obtained even on rough surfaces (e.g., etched
silicon nitride).2 That is, the surface roughness or surface structures
1
H NMR: 819 BLE repeating units, SEC: PDI ) 1.4 (apparent
polydispersity index) (see Supporting Information, Figures SI-1 and
SI-2).
Poly(4-(2-(amino hydrochloride)-ethylthio)-butylene)-block-poly-
8,29
(e.g., steps on the calcite surface) are not reflected in the measurements
(ethylene oxide)-block-poly(γ-benzyl L-glutamate) (PA-b-PEO-b-
curves under these conditions.
PBLE). In a 25 mL three-necked round-bottom flask, PB-b-PEO-b-
PBLE (0.150 g) and 2-aminoethanethiol hydrochloride (0.023 g) were
dissolved in 5 mL of DMF. Under stirring, the solution was saturated
with nitrogen for 30 min. Azoisobutyronitrile (AIBN, 0.004 g) was
then added, and the solution was stirred for 24 h at 60 °C. The product
was precipitated into diethyl ether, filtered, redissolved in DMF/water
Moreover, the distance at which the plateau force drops to zero
(plateau length) denotes the complete detachment of the polymer from
the surface, and thus the number of observed plateaus equals the number
of desorbed polymer chains.
The AFM desorption measurements were conducted with a home-
built instrument. Nominal spring constants of the cantilevers were
calibrated using the thermal oscillation method. Because of the
calculated spring constants may have deviations of up to 10%, a single
tip was used for an entire set of measurement series to monitor minute
force differences. Thereby, every measurement series consists of 750-
1000 force-distance curves. As a consequence, the binding of the
polymers to the tip had to be very robust because for every tip about
20000 force-distance curves had to be recorded. The usage of a
polyamine-block for attaching the molecules to the tip proved to be
highly efficient and have long-term stability because of the multiple
binding sites.
1
:1 (v/v), and dialyzed against deionized water. After freeze-drying,
the triblock copolymer with pendant amine groups was obtained as a
white powder.
About 30 amino groups should be attached to the R-chain end of
PEO-b-PBLE, considering an efficiency of thiol addition of ∼75% (see
ref 20).
Poly(4-(2-amino-ethylthio)-butylene)-block-poly(ethylene oxide)-
block-poly(L-glutamate) (Polyamine-b-PEO-b-PLE). The polyamine-
b-PEO-b-PBLE (0.100 g) was dissolved in a mixture of DMF (5 mL)
and water (5 mL), and the solution was saturated with nitrogen for 10
min. A 0.1 N NaOH (0.5 mL) solution was added under stirring, and
the solution was stirred overnight at room temperature. The solution
was then dialyzed and freeze-dried to obtain the triblock copolymer
polyamine-b-PEO-b-PLE.
The covalent attachment of the block copolymer PEO273-b-PLE819
to the AFM tip allows for studying the interaction of both the PEO
and the PLE with the calcite surface in a single experiment. Figure 4
shows a force plateau as obtained by continuously desorbing the block
copolymer from the calcite (104) face in an aqueous solution of 10
Modification of AFM Tips. Epoxy-functionalized AFM tips for the
covalent attachment of polyamines were obtained by treating the Si N
3 4
tip (Microlever, Veeco Instruments) with (3-glycidyloxypropyl)tri-
methoxysilane for 30 s and rinsing with toluene and water three times.
Afterward the polymers were immediately attached to the AFM tip by
soaking it in an aqueous solution of polyamine-b-PEO-b-PLE (con-
centration 0.25 mg/mL) for 90 min (Figure 2). After rinsing with water,
the modified tip was dried with nitrogen.
mM CaCl . It may be important to point out that because of the
2
attachment of the polymer to the AFM tip via the amine functions, the
PEO block is first desorbed and then the PLE block (see Figures 2 and
3). This means that at distances larger than the contour length of the
PEO273 (∼100 nm), the measured desorption force exclusively originates
from the desorption of PLE monomers.
AFM Desorption Experiments. In desorption experiments1
7,22-25
The interaction force of the PLE is slightly higher than that of PEO,
as qualitatively expected, because PEO is not supposed to adsorb
single polymer chains covalently attached to the AFM tip are adsorbed
(
(
(
(
22) Hugel, T.; Grosholz, M.; Clausen-Schaumann, H.; Pfau, A.; Gaub, H.; Seitz,
M. Macromolecules 2001, 34, 1039-1047.
(26) K u¨ hner, F.; Erdmann, M.; Sonnenberg, L.; Serr, A.; Morfill, J.; Gaub, H.
E. Langmuir 2006, 22, 11180-11186.
23) Friedsam, C.; Becares, A. D.; Jonas, U.; Seitz, M.; Gaub, H. E. New J.
Phys. 2004, 6, 9.
(27) Hugel, T.; Seitz, M. Macromol. Rapid Commun. 2001, 22, 989-1016.
(28) Sonnenberg, L.; Parvole, J.; K u¨ hner, F.; Billon, L.; Gaub, H. E. Langmuir
2007, 23, 6660-6666.
24) Cui, S. X.; Liu, C. J.; Wang, Z. Q.; Zhang, X. Macromolecules 2004, 37,
9
46-953.
25) Chatellier, X.; Senden, T. J.; Joanny, J. F.; di Meglio, J. M. Europhys.
Lett. 1998, 41, 303-308.
(29) Sonnenberg, L.; Parvole, J.; Borisov, O.; Billon, L.; Gaub, H. E.; Seitz,
M. Macromolecules 2006, 39, 281-288.
15366 J. AM. CHEM. SOC.
9
VOL. 129, NO. 49, 2007

Interaction Forces of Poly(L-glutamic acid) with Calcite Crystals
A R T I C L E S
Figure 3. Illustration of a typical single molecule desorption experiment with polymer chains covalently attached to the tip of an AFM cantilever.
Figure 4. Two force versus distance curves of PEO-b-PLE obtained upon
successive desorption from the calcite (104) surface. At a distance of about
1
00 nm the transition from desorbing ethylene oxide monomers to glutamate
monomers can be seen. If the transition occurs at distances smaller than
00 nm, this is most probably because of the attachment point of the polymer
1
at the curved AFM tip. The force regime is about 60 pN, the difference in
force about 5 pN.
Figure 5. Measurement series consisting of 1000 force-distance curves
10 mM CaCl2, pH 6). Note that only plateaus originating from desorption
(
strongly onto a calcite crystal.30 This is reflected by a sharp transition
in force at a distance that coincides with the contour length of the
PEO273 block (∼100 nm). Note that in Figure 4 the desorption of a
single PEO-b-PLE polymer can be seen, whereas the staircase-like
pattern in Figure 3 arises from the desorption of several polymer chains.
In the curve presented in Figure 3 the force transition between the two
blocks as shown in Figure 4 is not visible, as it is generally only rarely
observed. This is first of all because PEO often does not adsorb on the
calcite surface as was shown in desorption experiments with polyamine-
b-PEO273 on calcite (data not shown). However, if a desorption plateau
was observed, the desorption force equals the values as found for the
PEO in the block copolymer. The finding of rare desorption events
but still with “high” desorption force has already been reported in the
literature previously for the adsorption of poly(acrylic acid) on self-
of PLE can be seen here because the unspecific adsorption peak at small
distances hides desorption of PEO. In all curves the same molecule is probed.
In addition from time to time a second molecule participates. The mean
interaction force is obtained via statistical analysis of the plateau forces
(see inset). A typical value for fwhm of the force histograms is 6 pN.
transition distance is shifted to shorter lengths because of the attachment
point of the polymer on the curved AFM tip. Moreover, as soon as
several polymers are desorbed in parallel, intramolecular interactions
also contribute, smearing out the force transition such that it is not
detectable anymore.
The strength of single molecule desorption measurements, namely
the measurement of interaction forces with high precision and the
possibility to resolve minute force differences, is illustrated in Figure
5. Therein a series of force-distance curves is shown, in which the
same PEO-b-PLE molecule was probed over a thousand times. Calcite
2
3
assembled monolayers. In addition to the reduced PEO adsorption,
the “unspecific” peak at small distances in the force-extension curves
often extends to the distance regime where the force transition is
expected and therefore covers the force transition. This is in particular
true when several polymers are adsorbed to the surface and/or the
(104) was used as the solid surface. In the case of occasional
contribution of a second chain, two plateaus can be seen in the spectra.
The force transition between the two blocks is not observed here, as
the unspecific adsorption peak at small distances exceeds the transition
regime.
(30) C o¨ lfen, H.; Qi, L. M. Chem.-Eur. J. 2001, 7, 106-116.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 49, 2007 15367

A R T I C L E S
Sonnenberg et al.
The mean desorption force of the glutamic acid with the calcite
crystal is obtained by a statistical analysis of the plateau force of all
curves that exhibit plateaus longer than 200 nm. In this case the PEO
block with its contour length of 100 nm is not able to alter the
interaction of the PLE with the calcite surface. The resulting force
histogram is shown in the inset of Figure 5. A typical value for the
full width at half-maximum (fwhm) is 6 pN; thus, the quantification
of the mean desorption forces is possible with a precision of 1-2 pN.
Calcite Crystals. Calcite (104) was provided by R. U. Barz (Institut
f u¨ r Mineralogie, LMU M u¨ nchen) as well as crystallized by Y. Luo
(MPI-KG). In the desorption measurements no differences in the
desorption forces were found for the different samples. Calcite (100)
was purchased from MTI Corporation (Richmond, VA).
The Calcite (104) surface and (100) surface both consist of alternating
Figure 6. Mean desorption forces in dependence of the pH of the aqueous
CaCl2 solution. Each circle represents an independent measurement series
as shown in Figure 3. Error bars represent the fwhm of the force histograms.
An increased force is observed only between pH 8 and 9. In that regime,
the calcite point of zero charge (pzc) is expected.
Ca atoms and CO
exist in the distance of neighboring Ca-CO
CO groups (see Supporting Information, Figures SI-5 and SI-6).
3
groups. Differences between the two crystal faces
3
and the orientation of
3
However, calcite surfaces are not atomically flat in solution and can
reform in solution by etch pits or dissolution-recrystallization. Even
stable (104) cleavage planes reconstruct in moist air through pit
electrostatic contribution to the desorption force. As the pKa of
36
the PLE is <6, an electrostatic attraction between the positively
37-40
charged calcite surface
and the negatively charged polymer
3
1
formation and film growth.
should be expected at pH 6-8, resulting in the highest
Control Measurements with Poly(L-glutamic acid). Poly(L-
glutamic acid) was obtained from Sigma-Aldrich as sodium salt with
a molar mass of 750-5000 g/mol (CAS Number 26247-79-0). Because
of the broad molar mass distribution, mass/volume was chosen as the
polymer concentration unit instead of molar concentration to have a
constant number of monomer units independent of the polymer molar
2+
desorption force in absence of Ca . However, this is clearly
not observed in Figure 6, as our experiments were conducted
2+
in an excess of 10 mM Ca , and more importantly, the
desorption force stays constant below pH 7 and above pH 10.
On the basis of the interaction forces between PLE and calcite
(104) in dependence of the salt concentration of the aqueous
mass. A 1 mg/mL polymer solution was titrated with 10 mM CaCl
solution (see Supporting Information, Figure SI-3). ú-potentials were
determined with a Zetasizer (Malvern Instruments). Added CaCl
concentrations were chosen for a Ca2+ saturated polyanion solution on
2
solution (see further below, Figure 7), some additional comments
can be given on the pH-dependent interaction forces. First, when
screening all electrostatic interactions between polymer and
calcite surface at high salt concentrations, the desorption force
equals the desorption force found at pH 8-9. One can therefore
conclude that no electrostatic interactions are contributing to
the overall desorption force in this regime. Second, when
decreasing the salt concentration (increasing electrostatic inter-
actions) at pH 6, the desorption force decreases. That is, the
electrostatic interaction between polymer and surface is repul-
sive. From the shape of the curve in Figure 6, one would then
conclude that the electrostatic interaction at pH > 9 is repulsive,
too. However, note that because of the small force differences
obtained as a function of the pH, electrostatic interactions are
supposed to be weak. This corresponds either to a weakly
charged polymer, a weakly charged surface, or both.
2
the basis of the CaCl
Figure SI-4).
2
titration results (see Supporting Information,
Results and Discussion
pH-Dependent Interaction Forces between PLE and
Calcite (104). The crystallization process in the presence of
polymeric additives is strongly dependent on the experimental
conditions such as pH or ionic strength of the solution. This
was empirically demonstrated for BaSO4 which has the advan-
tage of a single polymorph with constant morphology over a
wide range of pH and experimental conditions. However, the
crystal morphologies could be varied to a large extent by
32
different polymer additives and different pH conditions,
especially for polycarboxylates.33 A general correlation between
pH-variation and crystal morphology has not been established
so far.
2+
It is known that complexation of Ca and polymers bearing
carboxylic groups can occur as a function of the accessible
4
1,42
carboxylate functions
introducing positive charges into the
2
+
Desorption measurements were conducted in aqueous 10 mM
CaCl2 solution at varying pH between 6 and 12, and the mean
desorption forces between PLE and calcite (104) were analyzed
side chains. We have therefore looked at the Ca binding of
PLE at the relevant pH values by titration experiments and found
the expected higher Ca binding at higher pH as compared to
the lower pH (see Supporting Information). Electrokinetic
measurements further showed that the ú-potential of poly(L-
glutamic acid) saturated with Ca in solution is still negative
(-10 mV at pH 6 and -35 mV at pH 11) but considerably less
negative than that of the polymer itself (-65 mV) (see
2
+
(see Figure 6). Thereby, the same mean desorption force of ca.
64 pN was obtained for pH 6 and pH 11 and a slight increase
2
+
of ca. 4 pN was visible around pH 8.6.
This maximum position of the desorption force in Figure 6
corresponds to the literature-documented point of zero charge
3
4,35
(pzc), which was reported to be between pH 8.2 and 9.5.
(
36) Schlaad, H. AdV. Polym. Sci. 2006, 202, 53-73.
Therefore, at pH > 10 the calcite surface and the PLE are
expected to be negatively charged, resulting in a repulsive
(
37) Vancappellen, P.; Charlet, L.; Stumm, W.; Wersin, P. Geochim. Cosmochim.
Acta 1993, 57, 3505-3518.
(
38) Fenter, P.; Geissbuhler, P.; DiMasi, E.; Srajer, G.; Sorensen, L. B.; Sturchio,
N. C. Geochim. Cosmochim. Acta 2000, 64, 1221-1228.
(
31) Kendall, T. A.; Martin, S. T. J. Phys. Chem. A 2007, 111, 505-514.
32) Qi, L. M.; C o¨ lfen, H.; Antonietti, M. Angew. Chem., Int. Ed. 2000, 39,
(39) Pokrovsky, O. S.; Mielczarski, J. A.; Barres, O.; Schott, J. Langmuir 2000,
16, 2677-2688.
(40) Lin, Y. P.; Singer, P. C. Geochim. Cosmochim. Acta 2005, 69, 4495-
(
6
04-607.
(33) Qi, L. M.; C o¨ lfen, H.; Antonietti, M. Chem. Mater. 2000, 12, 2392-2403.
(34) Moulin, P.; Roques, H. J. Colloid Interface Sci. 2003, 261, 115-126.
(35) Churchill, H.; Teng, H.; Hazen, R. M. Am. Mineral. 2004, 89, 1048-1055.
4504.
(41) Madsen, F.; Peppas, N. A. Biomaterials 1999, 20, 1701-1708.
(42) Kriwet, B.; Kissel, T. Int. J. Pharm. 1996, 127, 135-145.
15368 J. AM. CHEM. SOC.
9
VOL. 129, NO. 49, 2007

Interaction Forces of Poly(L-glutamic acid) with Calcite Crystals
A R T I C L E S
expected. However, this conclusion is drawn from a separate
consideration of the properties of the polymer and calcite surface
but not as a whole. In agreement to our findings for pH < 8,
namely that the polymer interacts with a seemingly negatively
charged calcite surface at pH < pzc, Seitz et al. found an
attractive/repulsive electrostatic interaction between poly(vinyl
amine)/poly(acrylic acid) and calcite in NaCl solution at pH 6.
This also corresponded to a seemingly negatively charged calcite
1
7
surface.
When comparing desorption measurement series in depen-
dence of the pH, qualitative differences in adsorption behavior
were observed. At pH > 10 the probability of observing
desorption events in the force-extension spectra is decreased
(e.g., the consecutive observation of 1000 desorption plateaus
as shown in Figure 3 is extremely unlikely), which is also
reflected in the smaller number of data points in this pH regime
in Figure 6. That is, for PLE molecules adsorption on the calcite
surface is less favorable at pH > 10 than at pH < 8. These
different adsorption characteristics may also be regarded as a
hint that the mechanisms that result in the repulsion between
PLE and calcite are different for the two pH regimes. For pH
> 10, PLE and calcite are similarly charged and therefore
experience “real” electrostatic repulsion, which may act as an
adsorption barrier. For pH < 8, PLE and calcite are expected
to be oppositely charged so that adsorption may be advanta-
geous. In this case one would have to conclude that the measured
electrostatic repulsive force is induced by the adsorption of the
Figure 7. Mean desorption forces in dependence of Ca-concentration in
solution (pCa ) -log[Ca]) for A) calcite (104) and B) calcite (100) at pH
6
. Two independent measurement series as shown in Figure 3 were
performed at every pCa value (small symbols). In the regime pCa < 2.5
the measured desorption forces are highly reproducible; large symbols
represent weighted averages of the two data points. Data points were fit by
-
0.5
F ) F0 + Rστ[Ca]
with R ) const., the surface potential σ ) const.
and the polymer charge density τ. For τ ) 1 (dotted curve) the fit function
shows the dependence of the desorption force on electrostatic screening
Debye screening length κ ∝ [Ca] ); a variable charge density due to,
2+
PLE on the calcite surface. ú-Potentials indicate further Ca
2
+
-
1
-0.5
coordination by the polymer at Ca excess, which would be
expected near the calcite surface under the conditions at pH <
pzc.
(
for example, charge regulation mechanisms was modeled with a titration
behavior, τ ) 1/(1 + exp[(pCa - pD)/t]) (solid line).
In general, the properties of surface-adsorbed polyelectrolytes
can be significantly changed with respect to the same molecules
in solution, e.g., the dissociation equilibrium can vary upon
Supporting Information). This is because the partial neutraliza-
tion of the polymer is due to the Ca2+ coordination. Neverthe-
less, a further shift of the ú-potential occurs upon addition of a
4
3-45
adsorption.
Such effects may be even more pronounced at
2+
large excess of Ca , from -35 mV to -20 mV at pH 11 (see
dielectric boundaries because of image charge interactions,
which are for example opposing at low dielectric surfaces.4
In particular, charge reversal has been predicted for semiflexible
Figure SI-4).
6,47
2
+
The Ca coordination to the carbonate groups can be
expected in a manner similar to that for the polymer carboxyl
groups, which leads to the repulsion of a polymer with pH-
dependent negative charge from an almost neutral crystal
surface. This could be an explanation for the repulsive com-
ponents yielding the reduced desorption force observed at pH
4
8,49
polyelectrolytes on charged planar surfaces.
Our findings reveal that the simple epitaxial or even
electrostatic view on polymer-crystal interactions is too simpli-
fied to explain the experimental findings and that quantitative
AFM force spectroscopy can yield atomic details of the additive
interaction with a crystal surface in a quantitative manner.
Moreover, they point out the ability of this technique to monitor
even small differences in molecular properties arising from the
adsorption of polyelectrolyte molecules on charged surfaces.
However, further experiments will be needed to elucidate the
detailed individual contributions of the polymer upon interaction
with a crystal surface.
>
10.
At the pzc, one expects an equal distribution of calcium and
carbonate ions at the surface. When decreasing the pH from
pH > 10 to pzc, the calcite surface becomes more and more
uncharged and with it the electrostatic repulsion between PLE
and calcite less and less pronounced. As a consequence, the
measured desorption forces rise in that pH regime. For pH <
pzc, the electrostatic interaction between PLE and calcite is
repulsive again, such that the maximum of the desorption force
in Figure 6 reflects the pzc of the calcite surface.
pCa-Dependent Interaction Forces between PLE and
Calcite (104)/Calcite (100). Concluded from the above results,
the synthesis of biomimetic materials on the basis of calcium
carbonates may further be crucially dependent on the calcium
The observed results for the desorption force at pH < 8,
namely a weak electrostatic repulsion between PLE and calcite,
were the most unexpected findings of the pH-dependent
(
43) Burke, S. E.; Barrett, C. J. Langmuir 2003, 19, 3297-3303.
measurements. At this pH (pH < pzc), the calcite surface has
(44) Sukhishvili, S. A.; Granick, S. Langmuir 2003, 19, 1980-1983.
to be expected as Ca2+ terminated, but the polymer will still be
(45) Xie, A. F.; Granick, S. Nat. Mater. 2002, 1, 129-133.
(
46) Netz, R. R. J. Phys.: Condens. Matter 2003, 15, S239-S244.
neutral to slightly negatively charged (see above). In other
words, even in the presence of Ca2+ in solution a marginally
higher desorption force than that at pH ) pzc would be
(47) Netz, R. R. Phys. ReV. E 1999, 60, 3174-3182.
(48) Netz, R. R.; Joanny, J. F. Macromolecules 1999, 32, 9013-9025.
(49) Boroudjerdi, H.; Kim, Y. W.; Naji, A.; Netz, R. R.; Schlagberger, X.; Serr,
A. Phys. Rep. 2005, 416, 129-199.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 49, 2007 15369

A R T I C L E S
Sonnenberg et al.
concentration. Among the polymer-crystal interactions, elec-
trostatic ones should be most affected by a variation of the ionic
strength of the solution. The importance of electrostatic interac-
tions was recently pointed out by Volkmer et al., who found
that the charge density of an organic template layer is of
importance for tuning the CaCO3 crystallization process with
respect to polymorph selection and crystal orientation under
monolayers.50
of the carboxylic groups of the poly(L-glutamic acid). As
electrostatic interactions are stronger at low salt concentrations
(less screening), charge regulation is more pronounced, too.
Thus, the polymer charge density decreases with decreasing salt
concentration. Using the same argument, the charge density of
a surface adsorbed polymer is furthermore altered by the surface
charges and vice versa.
In order to fit the experimental data the polymer charge
density was modeled with a titration behavior in dependence
of pCa (τ ) 1/(1 + exp[(pCa - pD)/t]), solid lines in Figure
7). Deflection points of the titration functions were found to be
at pD104 ) 1.6 for the calcite (104) and at pD100 ) 2.0 for the
calcite (100) surface. This finding is in qualitative agreement
with the surface charge densities of the calcite surfaces. The
larger charge density of the calcite (104) surface in comparison
to the calcite (100) surface leads to a reduction of the polymer
charge already at higher salt concentrations, i.e., lower pCa
values.
Desorption experiments were performed at pH 6 of an
aqueous solution containing between 0.1 mM and 1 M CaCl2.
The corresponding mean desorption forces plotted versus pCa
2
+
)
-log[Ca], with [Ca] being the concentration of Ca ions in
solution, are shown in Figure 7 for two very similar calcite
faces: (104) and (100). For both surfaces a similar behavior
and force differences smaller than 10 pN were observed. It was
found that the desorption forces are highly reproducible at salt
concentrations 0 < pCa < 2.5 and nonreproducible at pCa >
2
.5 (see Figure 7).
For interpretation of the findings, two salt-dependent contri-
The desorption forces in the nonreproducible regime at low
salt concentrations (pCa ≈ 3) are dependent on the AFM tip,
i.e., the set of molecules. This may again be correlated with
butions to the desorption force have to be considered: electro-
17
static screening and polymer charge density. In Debye-H u¨ ckel
approximation the electrostatic contribution to the desorption
force is proportional to the surface potential σ, the Debye
2+
the complexation of Ca ions by the PLE polymer. Because
of the attachment process of the triblock polymer to the AFM
tip, the number of bound polymer chains can only be roughly
controlled and is not known, as generally much more polymers
are bound to the tip than seen in the experiments (e.g., because
-
1
-0.5
screening length κ (∝ [Ca] ) and the linear (polymer) charge
-
1
density τ: Fel ∝ σκ τ. The desorption force then consists of a
non-Coulomb contribution F0 and the electrostatic desorption
force Fel. The experimental data were therefore fit by the
following formula: Fdes ) F0 + Rστ[Ca]-0.5 with R being a
constant and [Ca] the fit variable. The surface potential σ was
also regarded as constant.
2+
of the tip’s curvature). Thus, the degree of Ca complexation
at low salt concentrations, when the number of carboxylic groups
in the PLE polymers is comparable to the number of available
2+
Ca ions in solution, may depend on the AFM tip used.
At high salt concentrations the polymer is maximally charged
τ ) 1) and the desorption force is a function of the ionic
However, the effect may also be caused by the calcite surface
itself, as in electrophoretic mobility studies a variation of the
ú-potential of calcite (104) was observed at a similar salt
(
strength of the solution. However, τ ) 1 does not correspond
to a fully charged PLE polymer under the given experimental
conditions because of Ca2 complexation (see above). At pCa
34
concentration.
+
The results on the two calcite faces (104) and (100) shown
in Figure 7 indicate that the desorption forces are delicately
dependent on electrostatic interactions. It is a remarkable
observation that significant differences of the desorption forces
could be detected for the structurally similar calcite faces (see,
e.g., pCa ) 2 in Figure 7).
)
0 electrostatic interactions between polymer and surface are
almost completely screened, i.e., Fdes ≈ F0. When decreasing
the calcium concentration to pCa ) 1, the PLE is still charged
(
τ ) 1) but electrostatic interactions with the calcite are less
screened and not negligible any more, i.e., Fel * 0. In this regime
of “high” ionic strengths, the desorption force decreases
proportional to the Debye screening length (dotted lines in
Figure 7). This means that a repulsive interaction, Fel < 0,
between PLE and calcite surface has to be assumed at pH 6! It
may be interesting to note that the ratio of the fitted surface
potentials σ104/σ100 ) 1.7 is of the same order of magnitude as
the ratio of the density of surface ions (number of ions divided
by the surface area; see also Supporting Information), F104/F100
2.1. This may be interpreted as a hint that the differences in
the measured desorption force between the two crystal surfaces
are caused by different electrostatic interactions.
With decreasing salt concentrations, a phenomenon known
as charge regulation has to be taken into account.49,51 The charge
Conclusions
In polymer-controlled crystallization processes, small varia-
tions on the molecular level can result in large morphological
effects on the macroscopic level. We quantitatively investigated
the molecular forces between poly(L-glutamic acid) and calcite
on the single molecule level and showed that a variation of pH
and calcium concentration of the solution as well as a variation
of the crystal face led to variations of the interaction forces of
comparable strength. The observations underscore that a com-
plex interplay of molecular interactions has to be assumed for
polypeptide-mineral systems. We demonstrated that quantita-
tive AFM-based desorption measurements are able to enlighten
this interplay on the basis of the molecular interaction forces
with very high precision although the nature of the detected
interactions is not yet clear in all cases. In particular, this
approach should also be able to give insight into the dynamics
of crystal growth in the presence of polymers, as it involves
not only the adsorption but also the (transient) desorption and
conformational rearrangements of polymer chains at the mineral
)
density of all objects with dissociable groups, as for example
weak polyelectrolytes, is a function of the salt concentration
because of repulsive charge interactions within the polymer.
These intramolecular repulsions counter the dissociation reaction
(
50) Volkmer, D.; Fricke, M.; Agena, C.; Mattay, J. J. Mater. Chem. 2004, 14,
249-2259.
51) Friedsam, C.; Gaub, H. E.; Netz, R. R. Europhys. Lett. 2005, 72, 844-
50.
2
(
8
15370 J. AM. CHEM. SOC.
9
VOL. 129, NO. 49, 2007

Interaction Forces of Poly(L-glutamic acid) with Calcite Crystals
A R T I C L E S
surface. This could help to correlate molecular interactions and
conformations with macroscopic crystal properties and thus to
control the crystallization and materials properties.
the Deutsche Forschungsgemeinschaft (grants CO 194/1-1, SE
888/3-1, and SFB 486) and the Fonds der chemischen
Industrie.
Acknowledgment. Helpful discussions with Dr. Paul Bowen
EPFL, Lausanne), Prof. Lennart Bergstr o¨ m (University of
1
Supporting Information Available: Polymer synthesis ( H
(
NMR, SEC), titration of poly(L-glutamic acid) with CaCl2 at
different pH, ú-potential measurements of poly(L-glutamic acid),
and illustrations of calcite (104) and (100) surfaces. This material
is available free of charge via the Internet at http://pubs.acs.org.
Stockholm), Prof. Thomas Vilgis (MPI Mainz), and Prof. Roland
Netz (TU M u¨ nchen) are gratefully acknowledged. We thank
Margit Barth (MPI-KG) for the assistance with the CaCl2
titrations of poly(L-glutamic acid) and Heidi Zastrow (MPI-KG)
for ú-potential measurements. This work was supported by
JA074070I
J. AM. CHEM. SOC.
9
VOL. 129, NO. 49, 2007 15371