1503/1503.01642.md

Photostrictive materials

B. Kundys

Institut de Physique et de Chimie des Matériaux de Strasbourg (IPCMS), UMR 7504 CNRS-ULP,
67034 Strasbourg, France

The full version of the article 'Photostrictive materials' is available at http://dx.doi.org/10.1063/1.4905505

Light-matter interactions that lead to nonthermal changes in size of the sample constitute a photostrictive effect in many compounds. The photostriction phenomenon was observed in four main groups of materials, ferroelectrics, polar, and non-polar semiconductors, as well as in organic-based materials that are reviewed here. The key mechanisms of photostriction and its dependence on several parameters and perturbations are assessed. The major literature of the photostriction is surveyed, and the review ends with a summary of the proposed technical applications.

TABLE OF CONTENTS

INTRODUCTION .....	1
FERROELECTRIC MATERIALS.....	1
SbSI single crystals.....	2
Modified by lanthanum $\backslash text\{Pb\}(\backslash text\{Zr\}\_x\backslash text\{Ti\}\_\{1-x\})\backslash text\{O\}\_3$     ferroelectric ceramics .....	3
The $\backslash text\{Sn\}\_2\backslash text\{P\}\_2\backslash text\{S\}\_6$ crystals and ceramics.....	3
The $\backslash text\{BiFeO\}\_3$ single crystals and films .....	3
$\backslash text\{PbTiO\}\_3$ thin films .....	4
A model description of photostriction in     ferroelectric materials .....	4
POLAR SEMICONDUCTORS .....	5
CdS single crystals .....	5
GaAs single crystals .....	5
NONPOLAR SEMICONDUCTORS.....	6
Germanium.....	6
Silicon.....	6
Carbon nanotubes .....	7
Chalcogenide glasses .....	7
ORGANIC POLYMERS.....	8
SPIN-CROSSOVER MOLECULAR CRYSTALS ...	8
APPLICATIONS .....	8
Microactuation.....	9
Microsensing .....	9
Energy harvesting structures .....	9
Photonics .....	9
Strain mediated magnetization control.....	9
Photostrictive-magnetostrictive magnetometers .	10
Sonic-ultrasonic emission .....	10
Light controlled gas storage .....	10
Concluding remarks and outlook .....	10

INTRODUCTION

Shape responsive solids that are sensitive to a variety of external excitations upon variation of environmental

conditions constitute an important class of functional smart materials.^1–3 Their subclass, involving light-matter interactions that result in non-thermal sample deformation, is termed photostrictive. The photostriction differs in origin depending on the type of investigated material. The photostriction in electrically polar solids can be defined as photo-induced deformation of the lattice associated with a change in the internal electric field leading to a converse piezoresponse in the photovoltaic compounds. In organic polymers, light can trigger a transformation in molecular structure within the same chemical formula inducing large volumetric changes. In polar semiconductors, light modulates an electric field via free charge generation at the surface and causes an elastic strain due to the converse piezoelectric effect. In non-polar semiconductors, such as Si or Ge light irradiation, usually larger than a band gap creates an excess of electron-hole pairs in the conduction band, leading to deformation of the sample directly or via a change in atomic bonds in covalently bonded amorphous semiconducting materials.

This paper reviews the work on photostriction since its discovery in these groups of materials, with a concern to unite, compare, and summarize the existing disparate findings and analyses that describe the photostrictive phenomenon.

FERROELECTRIC MATERIALS

The discovery of the photostriction phenomenon in electrically polar compounds dates back to the sixties, when the Japanese paper of Tatsuzaki et al. reported a photo-induced strain in single crystals of SbSI.⁴ This work was then followed by other authors investigating further the same material.^5–7 Afterward, investigations were focused on the light-induced ferroelectric bimorph bending by Brody.^8,9 Most of the studies including $\text{Pb}[\text{Zr}x\text{Ti}{1-x}]\text{O}_3$ (PZT-based) materials development and device designing for applications, belong, however to the group of Uchino. The renewed attention is currently devoted to $\text{BiFeO}_3$ (BFO) and $\text{PbTiO}_3$ ferroelectric compounds. The background mechanism of photostriction in ferroelectrics is acombination of a bulk photovoltaic effect,¹⁰ i.e., substantial voltage generation and converse piezoelectricity.

SbSI single crystals

The SbSI single crystals were pioneering objects for photostriction measurements among ferroelectric compounds.⁴ This compound is at the same time photosensitive¹¹ and ferroelectric¹² with a Curie point of 22°C. The wavelength of maximum photoconduction lies between 580 and 740 nm depending on the stoichiometric ratios between $\text{Sb}_2\text{S}_3$ and $\text{SbI}_3$ . It was reported that the length of an SbSI crystal along the polar c axis is changed when it is illuminated uniformly by visible light, in the presence of a dc electric field along the polar c axis. The reported photostriction change (order of $10^{-5}$ ) was found to be opposite to that of warming excluding ordinary thermal expansion due light absorption. Moreover, the photoinduced strain was found to be

positive below the $T_c$ and negative in some narrow temperature region just above $T_c$ . (Ref.4) With further heating, no photostriction was observed in the paraelectric phase confirming its relation to the electrically polar state only. The maximum photostriction was observed near $T_c$ in a polar phase.

Notably, the observed photostriction shows a discontinuity as a function of temperature near $T_c$ , passing through positive and negative maxima with a large electric field dependence

This effect was also found to be wavelength dependent by other authors¹³ with a monotonic dependence shown on Fig. 3(a).

The unexpected wavelength dependence of photostriction correlates with the photoconductivity for both ferroelectric (15°C) and near-paraelectric (25°C) states (Fig. 3(b)). The sign change for both states in photostriction is observed. At the near-paraelectric (25°C) state, the observed wavelength dependence was proposed to be predominantly a consequence of electric field screening by light generated charges as an electric polarization averages to zero in the near-paraelectric state. In the ferroelectric region, however, the wavelength dependence results from two competitive contributions: electric field redistribution leading to increment of electric field and electric field screening by the light generated charges leading to decrease of electric field in the sample. This suggestion seems to be in agreement with the Landau free energy expansion analysis¹⁴ with two competing contributions: the first contribution is proportional to square of polarization, while second one is proportional to the average light-induced charge

FIG. 3. (a) Photostriction due to illumination in ferro and near para-electric (just after the transition) states. Silver electrodes were used. (b) Photoconductivity as a function of wavelength after Ref. 13.concentration (see the model description). Obviously, the two contributions opposing each other as a function of wavelength give rise to the observed amonotonous dependence. In confirmation of this assumption, the photostriction in the ferroelectric state changes its sign at $\sim 650$ nm where maximum of photocurrent is observed (Fig. 3).

Despite its extraordinary photovoltaic and photostrictive properties, the SbSI single crystals are not attractive for room temperature applications due to low ferroelectric critical temperature ( $T_c \approx 295$ K). In that regard, the possibility of $T_c$ increase towards 330 K in the modified SbSI ceramics^15–18 can stimulate a farther research in this direction.

Modified by lanthanum $\text{Pb}(\text{Zr}x\text{Ti}{1-x})\text{O}_3$ ferroelectric ceramics

The photostriction studies were further focused on the room temperature ferroelectric lanthanum modified PZT (PLZT) by Brody⁸ and Uchino.^19–22,27 As expected for ferroelectric ceramics, the preparation methods used to synthesize samples affect largely the photostrictive properties due to changes in homogeneity, stoichiometry, and grain sizes.²³ Remarkably, with grain size optimization, the photostriction as high as 0.01% was reported to be achievable.²⁴ Such a large value is very encouraging and already comparable to electrical field-induced strain of common piezoelectric materials. Due to the light penetration depth dependence, the photostriction logically increases as sample thickness of the $\text{WO}_3$ doped PLZT ceramics decreases²⁵ with an optimal sample thickness determined to be 33 $\mu\text{m}$ . Moreover, the increase of photoelastic effect can be expected for coherent light using the model of validation and analysis.²⁶ Unfortunately, the response time being tens of seconds²³ is still slow due to high dielectric permittivity values in PZT based compounds. This issue can be understood via simple empirical formula for linear photostriction as a function of time ( $t$ ), piezoelectric coefficient $d_j$ , and light-induced electric field $E_j$ ²⁸

$${\lambda }_{hv}=\frac{\mathrm{\Delta }L}{L}=d_j{E}_j(1-e^{-t\mathrm{/}RC}),(1)\backslash lambda\_\{hv\}\; =\; \backslash frac\{\backslash Delta\; L\}\{L\}\; =\; d\_j\; E\_j\; (1\; -\; e^\{-t/RC\}),\; \backslash quad\; (1)$$

where $R$ and $C$ are resistance and capacitance of the sample. For $t \ll 1$ , one can write $\Delta L/L \cong d_j E_j (t/RC)$ . Where low dielectric permittivity (capacitance) is preferable to have faster response speed. By contrast, large piezoelectric coefficient and large internal electric field (polarization) are needed for large photostriction. The PLZT exhibits a large photostriction under near-ultraviolet (UV) illumination due to its wide band-gap. Although no direct wavelength dependent photostriction measurement is present in the literature, the maximum photostriction can be expected at the 366 nm wavelength where the maximum photovoltaic properties in the PLZT samples²⁹ are observed.

The $\text{Sn}_2\text{P}_2\text{S}_6$ crystals and ceramics

The sulfide ferroelectric $\text{Sn}_2\text{P}_2\text{S}_6$ with a band gap near 2.3 eV undergoes a second order³⁰ ferroelectric phase transition at 339 K with the crystal symmetry change from polar

monoclinic $P_c$ to nonpolar $P_{21/c}$ .^31,32 A prominent photoconductivity,³³ large piezoelectric coefficients, and electro-optic effects³⁴ made this compound a candidate for photostriction studies.³⁵ The photovoltaic currents were found to be 5 times smaller for ceramics than for single crystals, directly correlating with the value of the electric polarization. Using the general formula (1) where photostriction is a product of light generated electric field and piezoelectric coefficient, the photostriction was evaluated from the current voltage characteristic under infrared light illumination. The photostriction estimate for single crystals gives a quite small value of 0.0047 ppm under the infrared light intensity of 2.1 W/cm².

The $\text{BiFeO}_3$ single crystals and films

Multiferroic $\text{BiFeO}_3$ compound is an extraordinary system containing several built-in order parameters that can lead to exotic cross-functionality. In addition to electromagnetic effect,³⁶ it shows also optical functionalities³⁷ that, thanks to ferroelectricity, can interact with strain.³⁸ As mentioned above, the photostriction in ferroelectric materials is known to originate indirectly from the superposition of photovoltaic and stricative effects resulting in a very slow response time (typically tens of seconds) in classical ferroelectrics.²³ In this respect, the report on the fast (below 100 $\mu\text{s}$ ) photostriction in a single crystal of $\text{BiFeO}_3$ has recently attracted an attention from the optical scientific community.^39–43 Because photostriction depends on the light penetration depth, its response time and a magnitude can still be improved in thin films. Indeed, the photostriction response in BFO films was reported to be in the picoseconds range.³⁹ As the electric polarization and piezoelectric coefficients play an important role to determine the magnitude of the effect (Eq. (1)), the fast photostriction in BFO films can yet be improved via doping induced polarization increase³⁹ and enhancement of the piezoelectric coefficient. For example, BFO doping with Tb,⁴⁴ Sm,⁴⁵ or La,⁴⁶ was reported to increase piezoelectric coefficients significantly. Taking into account recently reported quasi-instantaneous photostriction response,⁴⁷ the exciting opportunity to design fast responsive light-controlled devices can be realized. As light energy decreases, the photostriction effect in a single crystal of BFO also shows amonotonous dependence⁴⁸ and follows the photocurrent reported in Ref. 49 with a some delay (Fig. 4). This feature appears to be common for polar semiconductors also (see the Polar Semiconductors section) and is connected with amonotonous light-induced charge excitation with an explanation similar to that of SbSI.

Multiferroic BFO also possesses a magnetic degree of freedom as it is antiferromagnetic. This can potentially couple photostriction to magnetic order via elastic interactions. Interestingly, the photostriction in $\text{BiFeO}_3$ single crystal was found to decrease in magnetic field.³⁸ Although this effect was checked not to be connected with magnetostriction or visible magnetic field induced changes in polarization, its exact origin remains to be investigated. As discussed above the photostriction may consist of two competing mechanisms. Indeed, the evidence for both optical rectification³⁹ and light-induced change generation⁴⁰ mechanisms ofFIG. 4. Wavelength dependence of photostriction response in the single crystal of $\text{BiFeO}_3$ (Ref. 48) and photocurrent of $\text{BiFeO}_3$ .

photostriction in $\text{BiFeO}_3$ were reported. However, the free charge contribution is under debate.⁴⁷ As far as thin films are concerned, it is also important to sort out any possible stray free charges as leakage currents in thin films of BFO, screening the electric polarization, are known to be a serious challenge.^50,51

$\text{PbTiO}_3$ thin films

The fast (tens on ps time scale) photostrictive effect has also been recently seen in thin films of the prototypical ferroelectric $\text{PbTiO}_3$ and explained also to result from the coupling to its intrinsic photovoltaic response.⁵² The light-induced changes in the unit-cell tetragonality were observed using the time-resolved X-ray scattering reaching 0.25% of maximum of photoinduced strain. The photostriction depends nonlinearly on the light fluence (Ref.52) showing a saturation that is inconsistent with dominant thermal effects.

After ultrafast pulse excitation of 50 fs at 400 nm, the changes in tetragonality lasting several nanoseconds are observable. The concluding mechanism was proposed to originate from the coupling of the photogenerated changes in the internal electric fields to the converse piezostain within the film. This is consisted with the qualitative theory described above where the wavelength dependence of photostriction and its correlation with photocurrent is a key in determining dominant contribution to the photoelastic effect.

A model description of photostriction in ferroelectric materials

While formula (1) is successful in describing qualitatively the photostriction and its response time dynamics at saturated ferroelectric state, it also assumes that the piezoelectric coefficient is light independent that strictly speaking not always the case. An analytical description containing the strain-induced effects involves additional terms in describing the light-induced charge generation $\sum_j n_j \epsilon_j \Omega_j$ to the free energy of the system¹⁴

$$F=\frac{1}{2}\alpha P^2+\frac{1}{2}\beta P^4+\frac{1}{2}\gamma P^6-\frac{1}{2}\sum _i\sum _j{k}_{ij}{\mathrm{\Omega }}_i{\mathrm{\Omega }}_j-P^2\sum _j{\xi }_j{\mathrm{\Omega }}_j+\sum _j{n}_j{ϵ}_j{\mathrm{\Omega }}_j,(2)F\; =\; \backslash frac\{1\}\{2\}\; \backslash alpha\; P^2\; +\; \backslash frac\{1\}\{2\}\; \backslash beta\; P^4\; +\; \backslash frac\{1\}\{2\}\; \backslash gamma\; P^6\; -\; \backslash frac\{1\}\{2\}\; \backslash sum\_i\; \backslash sum\_j\; k\_\{ij\}\; \backslash Omega\_i\; \backslash Omega\_j\; -\; P^2\; \backslash sum\_j\; \backslash xi\_j\; \backslash Omega\_j\; +\; \backslash sum\_j\; n\_j\; \backslash epsilon\_j\; \backslash Omega\_j,\; \backslash quad\; (2)$$

where $k_{ij} = \partial^2 F / \partial \Omega_i \partial \Omega_j$ components of the elastic stiffness tensor. $\Omega_{ij}$ is the strain tensor. $\xi_j = \partial^3 F / \partial P^2 \partial \Omega_j$ is the electrostriction tensor. $\alpha$ , $\beta$ , and $\gamma$ are the known coefficients of ferroelectric free energy expansion. $n_j$ is an average concentration of electron subsystem charges at the energies levels $\epsilon_j$ (conduction band levels of traps and recombination) close to the bandgap energy. Taking the derivative with respect to polarization, which is the internal electric field, one can get

$$\frac{\mathrm{\partial }F}{\mathrm{\partial }P}=\alpha P+\beta P^3+\gamma P^5-2P\sum _j{\xi }_j{\mathrm{\Omega }}_j=E_j\mathrm{.}(3)\backslash frac\{\backslash partial\; F\}\{\backslash partial\; P\}\; =\; \backslash alpha\; P\; +\; \backslash beta\; P^3\; +\; \backslash gamma\; P^5\; -\; 2P\; \backslash sum\_j\; \backslash xi\_j\; \backslash Omega\_j\; =\; E\_j.\; \backslash quad\; (3)$$

The components of the deformation tensor $\tau_j$ can be defined as $\partial F / \partial \Omega_j$

$$\frac{\mathrm{\partial }F}{\mathrm{\partial }{\mathrm{\Omega }}_j}=-\frac{1}{2}\sum _i{k}_{ij}{\mathrm{\Omega }}_i-P^2{\xi }_j+\sum _j{n}_j{ϵ}_j={\tau }_j\mathrm{.}(4)\backslash frac\{\backslash partial\; F\}\{\backslash partial\; \backslash Omega\_j\}\; =\; -\backslash frac\{1\}\{2\}\; \backslash sum\_i\; k\_\{ij\}\; \backslash Omega\_i\; -\; P^2\; \backslash xi\_j\; +\; \backslash sum\_j\; n\_j\; \backslash epsilon\_j\; =\; \backslash tau\_j.\; \backslash quad\; (4)$$

Putting $\Omega_i = 0$ in Eq. (1), one can simplify the equation to

$${\tau }_j=\sum _j{n}_j{ϵ}_j-P^2{\xi }_j\mathrm{.}(5)\backslash tau\_j\; =\; \backslash sum\_j\; n\_j\; \backslash epsilon\_j\; -\; P^2\; \backslash xi\_j.\; \backslash quad\; (5)$$

This equation shows that the deformation does not depend on the sign of polarization as expected and should change in the presence of the generated charges. On the other hand, the piezoelectric coefficient determined from Eq. (1) using Eqs. (3) and (4) becomes

$$d_j=\frac{-P^2{\xi }_j+\sum _j{n}_j{ϵ}_j}{\alpha P+\beta P^3+\gamma P^5-2P\sum _j{\xi }_j{\mathrm{\Omega }}_j}\mathrm{.}(6)d\_j\; =\; \backslash frac\{-P^2\; \backslash xi\_j\; +\; \backslash sum\_j\; n\_j\; \backslash epsilon\_j\}\{\backslash alpha\; P\; +\; \backslash beta\; P^3\; +\; \backslash gamma\; P^5\; -\; 2P\; \backslash sum\_j\; \backslash xi\_j\; \backslash Omega\_j\}.\; \backslash quad\; (6)$$

Importantly, this free energy expansion analysis shows that the photostriction (Eq. (5)) itself has two competingcontribution: one can arise from direct change in polarization (for example, via optical rectification effect) and another from the light generated charges. Such a competition can indeed be confirmed by experimental results on the wavelength dependence obtained for SbSI single crystals (Fig. 3(a)) and therefore manifests a qualitative agreement. However, a quantitative verification of the theory performed on SbSI single crystals has not been found to agree with experimental results.¹⁴ It was suggested that changes in ferroelectric domain state under illumination (electric field changes and leads to domain movement) must be taken into account. Indeed, in ferroelectric samples, the both piezo (odd with electric field) and electrostrictive (even with electric field) effects are present. Such an interplay together with charged domains movement at low subcoercive electric fields can lead to the exiting multilevel hysteresis behavior in both bulk⁵³ and surface⁵⁴ effects and the theory must be extended to include them.

POLAR SEMICONDUCTORS

CdS single crystals

First photostriction study in polar semiconductors started with the photomechanical effect observation in non-centrosymmetric CdS single crystals by Lagowski and Gatos.⁵⁵ It was observed that a visible illumination induces an elastic deformation resulting in the mechanical vibrations of CdS cantilevers with an orientation perpendicular to the crystallographic axis (001) with the two types of possible deformations (Fig. 6). Since the cantilevers are normally longer than wider, one can assume that the deflection of the type shown on Fig. 6 (left side) is dominant. Under such assumption, one can perform an estimation of the magnitude of the photostriction effect. For small deflection ( $\Delta$ ), large radius ( $R$ ) of deformation and assuming that $\alpha \approx L/R$ , where $L$ is the initial sample length one can write

$$\mathrm{\Delta }\cong L^2\mathrm{/}2R\mathrm{.}(7)\backslash Delta\; \backslash cong\; L^2/2R.\; \backslash quad\; (7)$$

Taking the reported⁵⁵ relation of deflection of the free end of the cantilever to its length (10 mm) of $5 \times 10^{-2}$ , the radius of the curvature estimate gives approximately 100 mm. The photostriction can then be calculated using the cantilever thickness ( $t$ ) and the bending radius via the relation

$$t\mathrm{/}2R(8)t/2R\; \backslash quad\; (8)$$

and gives approximately 75 ppm.

The effect was explained to originate from the combination of light-induced changes to the surface electric field and

FIG. 7. Wavelength dependence of photostriction for CdS cantilever and its correlation with the photoconductivity spectrum.

elastic strain due to the converse piezoelectric effect. Although the wavelength dependence of the amplitude of vibrations exhibited by the CdS wafers was observed very similar to their surface-photovoltage spectrum, it was found different considerably from the photoconductivity wavelength dependence (Fig. 7). This issue is in agreement with experimental observations in SbSI single crystals (Fig. 3) originating from the possible interplay of the two competing mechanisms in the photoelastic effect (Eq. (5)).

GaAs single crystals

Due to its noncentrally symmetric crystallographic structure, the similar photomechanical vibrations were also reported for (111) GaAs single crystals.⁵⁶ With the chopped light, the cantilever was excited to their resonant vibration. As in the case of CdS photostriction effect, the wavelength dependence of the amplitude of this vibration was found to be different to that of photoconductivity (Fig. 8). The results were interpreted within a model based on light-induced modulation of the piezoelectric surface stresses. The amplitude of the vibration was found to increase with decreasing pressure, apparently due to a decrease of the damping factor.

For incident photon energies larger than the bandgap, the amplitude of the photomechanical vibration versus illumination intensity exhibited a saturation that can be considered typical for surface photovoltage behaviour (Fig. 8 (inset)). The similar slope showing photostriction saturation versus light intensity was recently observed for PbTiO₃

FIG. 6. Crystallographic orientation of the CdS cantilever with two possible types of light-induced deformations.FIG. 8. Wavelength dependence of photostriction for (111)GaAs cantilever and its correlation with the photoconductivity spectrum. Inset shows photostriction amplitude as a function of the light intensity for the light 1.45 eV.

ferroelectric film (see above) and is inconsistent with dominant thermal heating effect. Importantly, no photostriction was observed in the both CdS and GaAs wafers with prismatic surfaces. For both (001) CdS and (111) GaAs wafers, it was concluded that contributions from thermoelastic or pyroelectric effects can rather be excluded due to low intensity of sub-bandgap monochromatic radiation. In addition, the amplitude of the photomechanical vibration was found to decrease significantly under superimposed steady-state illumination. Interference by light pressure effects was also excluded since the photostriction found only when surfaces perpendicular to the polar axis are illuminated. It was therefore concluded that photostriction should occur in other non-centrosymmetric semiconductors with a depletion surface layer. Furthermore, it can also exist in the depleted regions in the bulk materials and, for example, in rectifying junctions perpendicular to the polar axis. The rough estimation of the photostriction in the GaAs wafers for $10\ \mu\text{m}$ thick sample employing analysis shown on Fig. 6 and data from Ref. 56 gives 0.4 ppm, which is more than 2 orders of magnitude less than for CdS wafers. The possible explanation of this difference is the large difference in the absolute value of the piezoelectric coefficients of these two compounds^57,58 in agreement with empirical formula (1).

NONPOLAR SEMICONDUCTORS

Germanium

Generated by monochromatic light excess electron hole pairs were reported to create the lattice dilatation in nonpolar semiconducting germanium.⁵⁹ Enlightened semiconductor at the wavelength corresponding to photon energies larger than the band gap develops electron-hole pairs throughout the surface at the thickness of the light penetration depth. The light generated electrons in the conduction band can contribute to the atomic bonding energy, whereas holes in the valence band can decrease the energy of covalent bonds. Because a volume of semiconductor can depend on the occupation of the electronic energy levels,⁶⁰ the material can deform in the illuminated region. The sign of the photostriction is

determined by the pressure susceptibility of the energy gap which is positive in germanium leading to lattice dilation.^61–63 The spectral response of the photostriction and photoconductivity was measured in Ref. 59 As energy decreases, the photostriction follows the photoconductivity with a small delay. Notably, this behavior is inverse to the one observed for polar semiconductors and ferroelectrics (see Figs. 4, 7, and 8) indicating a different photostricative origin. To explain this feature, the existence of the $\langle 100 \rangle$ exciton states with relatively long apparent lifetimes has been postulated.⁵⁹

The photostriction was found to be a linear function of the change carriers concentration with the maximal reported value of $1.12 \times 10^{14}\ \text{cm}^{-3}$ . Taking this value, the photostriction can be estimated to be $7.84 \times 10^{-4}$ ppm for the reported sample thickness of 0.5 mm that is three orders of magnitude smaller than in the low photostricative polar GaAs semiconductor.

Silicon

Contrary to the light-induced lattice dilation in germanium, a light-induced contraction effect was observed in phosphorus-doped n-type silicon ((111) face).⁶⁴ The response of the strain quartz sensor glued on the Si crystal due to laser light of 1.17 eV with a pulse duration of 25 ns has shown the behavior opposite to that expected for thermal expansion implying the initial lattice contraction. A similar explanation of this phenomenon in terms of a light-induced charge generation was proposed. The contraction agrees well with the negative sign of pressure susceptibility of the energy gap, which is indeed negative for silicon. Photoexcitation of electron-hole pairs therefore contracts the lattice under the illumination in silicon. Similar as for germanium, the photostriction was found to be linear function of light intensity. The magnitude of photostriction at room temperature can be evaluated to be $-0.414$ ppm for the light fluence of $0.03\ \text{J/cm}^2$ using data from Ref. 64. The photostriction studies in the Sicantilevers have also revealed a linear dependence upon light power (not shown) and wavelength (Ref. 65). The photostrictive deflection increases linearly with increasing wavelength up to the drop at $\sim 1.1 \mu\text{m}$ .

Using data from Ref. 65 and Eqs. (7) and (8), one can estimate the photostriction of $\sim 1.25 \text{ ppm}$ at the 780 nm excitation for the reported sample thickness of $0.5 \mu\text{m}$ . Yet, six times larger effect was found by time resolved X-ray diffraction measurements at 123 K using much larger light intensity.⁶⁶ Observing the shift of a Bragg peak during laser illumination at 123 K, the photostriction of 6.4 ppm has been found for the sample thickness of 0.5 mm with the fluence of $127 \text{ mJ cm}^{-2}$ at the wavelength of $2480 \text{ \AA}$ .

Carbon nanotubes

Semiconducting carbon nanotubes possess a wide range of direct bandgaps in the solar spectrum range, and high carrier mobility at low scattering ratio, which make themselves ultimate photovoltaic materials.⁶⁷ Especially, single wall carbon nanotubes (SWNT) manifest a record power conversion efficiencies in the form of p-n junctions with Si.⁶⁸ The first observation of photostrictive effect in low dimensional carbon structures was reported by Zhang and Iijima⁶⁹ and was found to be too large to be explained by thermal expansion alone. It was concluded that electrostatic interactions are important in the observed phenomenon. For low dimensional samples, the photon pressure can be important but was sorted out by the fact that the photostrictive-like movement direction was different from the light propagation direction in many cases. The white light with an intensity of $20 \text{ mW/cm}^2$ induced a displacement of $170 \mu\text{m}$ of the 1 mm long filament with $10 \mu\text{m}$ diameter in a time scale of 100 ms. However, the exact evaluation of the photostrictive contribution requires comparison with thermal expansion measurements. To address this issue, the SWNT suspended in air were studied by other authors⁷⁰ who determined that electron-phonon coupling explains the experimental photostrictive observations, neglecting thermal expansion. In contrast, a dominating thermal expansion was concluded for nanotubes in surfactant or

in bundles. The photostriction of the both SWNT and multi wall carbon nanotubes (MWNT) was studied in the form of free standing thin films in Ref. 71. It was concluded that the photostriction is orders of magnitude larger than the expected thermal expansion effect. For MWNT films, the 675 nm light at the irradiance of $1.5 \text{ mW/cm}^2$ provokes an initial subordinate contractive deformation with the stress change of $-5.5 \text{ kPa}$ in 1 s and then expansion is observed with the maximum of photoinduced stress of $\sim 0.3 \text{ MPa}$ for longer time scales. Assuming the smallest estimated Young's modulus value of 270 GPa (Ref. 72) and zero thermal expansion contribution, the photostrictive deformation is $\sim 1.1 \text{ ppm}$ . The SWNT films, however, have shown a dominant light-induced contraction without deformation sign change with a smaller effect ( $-57.3 \text{ kPa}$ ). More generally, the carbon nanotubes in different composites structures constitute a promising class of photoactuation systems that are in detail reviewed in Ref. 73.

Chalcogenide glasses

The covalently bonded amorphous semiconducting materials containing considerable amount of chalcogen atoms were reported to show photostriction in 1974.⁷⁴ Contrary to abovementioned elastic photostriction in classical semiconductors here light-induced deformation has a plastic character and persist after light is off. Only a thermal heating above certain critical temperature can recover the initial sample shape. The phenomenon was first shown in thin films of $\text{As}2\text{Se}3$ , $\text{As}2\text{S}3$ , $\text{As}{40}\text{Se}{50}\text{Ge}{10}$ , $\text{As}{40}\text{Se}{25}\text{S}{25}\text{Ge}{10}$ , and $\text{As}{40}\text{Se}{10}\text{S}{40}\text{Ge}_{10}$ deposited in the mica substrates that were found to bend under heating and become flat under light. The deformation level was evaluated from the change in the radius of bending to be of order of $10^{-3}$ . Unfortunately, the process of the light-induced deformation is slow (2h timescale). While As-based samples show positive photostriction, the Ge based glasses ( $\text{GeSe}_2$ and $\text{GeS}_2$ ) show light-induced contraction.⁷⁵ The photostriction reported $\text{As}_2\text{S}3$ films range from 0.42% (Ref. 76) to $\sim 6%$ (Refs. 77 and 75) in response to light illumination of a bandgap energy. Yet, a possibility of an order of magnitude improvement in photostriction has been reported for $\text{As}x\text{S}{100-x}$ ( $5 \leq x{\text{As}} \leq 40$ ) bulk glasses⁷⁸ and $\text{As}_2\text{S}_3$ thin films.⁷⁹ The photostriction in these materials was reported to be nonthermal in origin^79,80 as well as the optomechanical response of the thin film cantilever (configuration similar to the one shown in Fig. 6) was reported to be accurately controlled by varying the angle between light polarization and the cantilever axis.⁸¹

Several mechanisms were proposed^82,83,88 to describe photostriction in the chalcogenide glasses and it seems to be reasonable to assume that the phenomenon is generally determined by light efficiency in modifying the bond between atoms in such a way that it finds another equilibrium state leading to a different atomic configuration. This scenario is possible thanks to amorphous nature of the chalcogenide glasses and is usually accompanied with a optical absorption edge shift (photodarkening effect).^84-87 Although the exact mechanism of the light-induced change in the atomic bond configuration is still unclear to a large extent, itseems that light generated charges exited from the valence to conductance band are responsible for atomic bond changes.⁸⁸ Indeed, some experimental results indicate that the number of light generated charges is related to the photoinduced volume expansion.⁸⁰

ORGANIC POLYMERS

Unlike other photostrictive compounds, the organic based materials showing photostrictive effect are multiple. A photosensitive organic molecules resulting in the light-induced reorientation or ionization reaction are generally responsible for the photostriction mechanism in these materials. Because many organic based materials were already reviewed elsewhere,^89–91 we will only briefly list the most important compounds in which photostriction was reported to differ from direct light heating effect as many reports risk to be obscured by dominant thermal expansion.^92,93

The first report on the photostriction effect in polymer dates back to 1971 when Van der Veen and Prins have reported photomechanical energy conversion in a polymer membrane.⁹⁴ These poly-(4,4′-diaminoazobenzene pyromellitimide) films were known to undergo isomerization transition under UV light and a volume photostrictive effect of 3.6% was observed (1.2% of linear contraction). The response and recovery times were in the range of tens of minutes that is correlated with the time needed for chemical structure transition. Although a heating contribution may be present, it was not found to be dominant as the independent measurements as a function of temperature had shown that contraction induced by light is much larger than a thermal deformation.

The related compound of poly(ethylacrylate) networks with azo-aromatic crosslinks shows an UV induced a geometrical change in the chemical structure of the containing azobenzene unit. These compounds in a form of film have shown photoelongation of 0.25% in 1980 (Ref. 95) due to so called light-induced photoisomerization transition.

The presence of azobenzene units can also trigger photostriction in the azobenzene-containing liquid-crystalline

polymers (LCPs) via similar photochemical structure changes of $x^2$ containing molecules. While the molecular formula remains the same, the exact arrangements of the exited molecules change. The recent research progress in the area of photosensitive and photomobile applications is in detail reviewed in Ref. 90.

The polymer materials manifesting both viscosity and elasticity (elastomers) were reported to develop the largest photostriction—between 10% and 400%.⁹⁶ Molecular shape in these materials is tied to macroscopic sample shape and depends significantly on the state of nematic order. The background mechanism is a light-induced movement of a photosensitive cluster of atoms that changes a chemical bond and produces a twist in a polymer chain. Although the change in chemical bond is not directly connected with volume photostriction, it induces nematic order-disorder transition leading to the enormous photostrictive effect. An elegant demonstration of photostriction was also reported for crystalline elastomer films based on azobenzene derivative.⁹⁷ The fascinating bending of film repeatedly and precisely along any chosen direction was reported. Using the linearly polarized light, the bending can be controlled by light polarization rotation. The above mentioned photostriction effects are unfortunately in the seconds or minutes response time scales. In that respect, the recent reports on the ms⁹⁸ and even $\mu$ s⁹⁹ photostrictive speed time scales should trigger a new research effort.

SPIN-CROSSOVER MOLECULAR CRYSTALS

These molecular magnets¹⁰⁰ exhibit a high spin (HS, $S=2$ ) to low spin (LS, $S=0$ ) state transition that can be switched by pressure, temperature, or light irradiation.¹⁰¹ Owing to their organic content and magnetic metal atoms, these materials can change their structural arrangements under light leading to the change in the magnetic state. The generally accepted mechanism of photostriction here is a light sensitive electronic subsystem that intrinsically lead to the change in the bond configuration to form another metastable structure with different elastic,¹⁰² optical, and magnetic properties.¹⁰³

In the representative iron (II) molecular complexes $[\text{Fe}(\text{phen})_2(\text{NCS})_2]$ , for example, the photostriction happens together with a change in magnetic properties and can reach volumetric change of 1.1% at 30 K (Ref. 104) in connection with the light-induced excited spin state trapping effect.^105,106 A spin state photoswitching was also reported to be followed by the volume expansion in the molecular Fe(III) spin-crossover monoclinic polymorph $[(\text{TPA})\text{Fe}(\text{III})(\text{TTC})]\text{PF}_6$ with a different time scales for thermally and light-induced regimes.¹⁰⁷ Other photostrictive candidates of this rich family of compounds can be found in Refs. 101 and 103. Unfortunately, a temperature range where photostriction can be expected is well below the room temperature in most of these materials.

APPLICATIONS

For the reason that photostrictive materials involve a coupling between optical and mechanical functionalities,they can serve as converters between light and mechanical energy. Hypothetically, photostrictive materials can be used ubiquitously where piezoelectric or magnetostrictive materials are used with an important wireless advantage and without a need of magnetic field. Although the relative photostrictive displacements at fast speed are small, their repetitive operation can, in principle, be used to develop acoustic or even electrical power. Up to this date, the most of the technical developments have been proposed for PZT based ceramics²⁶ and polymers based materials.⁹⁰ Here, we point out some motivating applications combining our suggestions with already reported ones, where the aim is to underline the potential advantages that photostrictive materials possess. The utilization of the photostrictive effect in already demonstrated and prospective technical applications includes as follows.

Microactuation

Using a photostrictive bimorph configuration, a photo-driven relay,²¹ light-driven micro walking mechanism,¹⁰⁸ and photoacoustic devices¹⁰⁹ have been demonstrated on PZT based ceramics.¹¹⁰ Other applications may be optically controlled electric delay lines, robots,¹¹¹ optical micro positioning, wireless control of flexible structures,¹¹² vibrations,¹¹³ and photophones. In Ref. 20, it was also proposed to construct a sun chasing mechanisms for photovoltaic panels using photostrictive films on the flexible substrates. As far as polymers are concerned, light controlled biomimetic actuators become an interesting possibility as well as light assisted crack self healing effects in these materials.¹¹⁴

Microsensing

Microsensing can include tunable sensors for incident radiation using strong photostriction dependence on the wavelength and power. So called indirect microsensors utilizing photodetection when the device is loaded with condensing moisture (microhygrometer) or incident interstellar dust (microbalance).²⁰ Since photostriction can depend on the electric field applied to the sample (Fig. 2) and polarization it can be also used to determine the polarization state without measuring charge or capacitance directly. The chemical gas sensing was proposed using a photostrictive cantilever device¹¹⁵ utilizing the general cantilever approach for chemical and biological sensors.¹¹⁶

Energy harvesting structures

The combination of photostriction materials with piezoelectric materials in multilayers or other hybrid structures may lead to light energy conversion into electricity similar to that reported for piezoelectric¹¹⁷ and piezoelectric-magnetostrictive structures.¹¹⁸ Adding magnetostrictive materials to such a structures can even offer an additional degree of freedom to control their performance by magnetic fields.

Photonics

In optics, a fast and large photostriction can be a missing property to achieve all-optical control in optical logic

elements.¹¹⁹ For example, one can imagine photostrictive transistors that can control a light flow in a similar fashion as a transistor controls an electric current flow. There can be also a great deal of interest in substituting electro-optical components with photostrictive devices. For example, electrically controlled piezoelectric fibers¹²⁰ can be substituted by the polymer fibers¹²¹ with a possible photostrictive effect for all optical operation. The performance is based on the possible optical control of the refractive index and birefringence because photostriction is closely linked to photoinduced refractive index change. For example, azobenzene containing polymers⁹⁰ and chalcogenide glasses Ge-As-Se,¹²² As₆₀Se₄₀, and Sb₄₅Se₅₅¹²³ do show this property. As well as in As₂S₃ films, the photostriction is actually accompanied with an increase of refractive index.¹²⁴ Yet, many electrically polar and photovoltaic materials¹²⁵ and spin-crossover molecular crystals¹²⁶ do also show photorefractive effect. Consequently, holographic recording is another possible application.

Strain mediated magnetization control

Controlling the magnetization direction in low dimensional magnetic materials is a key issue for magnetic recording. This is conventionally performed by applying external magnetic fields. In spintronics, for example, the magnetic field can be generated by passing large currents to micro strips generally made of gold. However, these currents generate heating effect that becomes increasingly important as dimensions are reduced below a micrometer, making this approach challenging for high-density magnetic memories. Use of photostriction here in a combination with magnetic overlayer having inverse magnetostriction effect can constitute an alternative and wireless approach. This idea is analogous to the piezoelectric magnetic anisotropy control via inverse magnetoelastic effect. The magnetization control was proven to be feasible using piezoelectric materials and recently it was also reported for photostrictive matrix with magnetoelastic elements.¹²⁷ The preliminary experiments with CoFe/BiFeO structures have also demonstrated a possibility of photoelastic-photoresistive-magnetoelastic coupling.¹²⁸ If a large and fast coupling is achieved it makes possible the integration of the light controlled magnetic properties with spintronics. The possible prototype memory tunneling magnetoresistive (TMR) cell is shown in Fig. 12. A highly magnetoelastic layer deposited onto photostrictive substrate changes its magnetic anisotropy in response to photostrictive effect of the substrate thanks to inverse magnetostrictive effect.¹²⁹ On the top of the magnetoelastic film, the tunneling layer is deposited to form TMR junction. Contrary to the lower ferromagnetic layer, the upper one must be weakly magnetostrictive to keep its magnetization insensitive to the light generated strain through the substrate. By the virtue of the TMR effect, this change can reflect itself in the vertical resistance change (Fig. 12(b)).

Taking into account that the bulk strain can show a well defined hysteresis loop behavior at subcoercive electric field region,⁵³ the multistate nonvolatile memory operation becomes possible. This should lead to an additional degreeFIG. 12. TMR junction (a) and its expected light controlled operation (b). Red arrows show possible magnetization rotation as a function of elastic distortion induced by light.

of freedom in hybrid straintronic-spintronic devices where stable changes in magnetization can be created by light and then maintained at zero power.

Photostrictive-magnetostrictive magnetometers

The construction of photoelastic-piezomagnetic magnetometers can be another possibility. Its design involves a deposition of the magnetoelastic material on a top of the photoelastic plate (Fig. 13). A light beam with an alternating power or/and wavelength illuminated on the photoelastic plate can generate a photostrictive strain. This dynamic strain is transferred to the magnetoelastic layer. By the virtue of the inverse magnetostrictive effect, this dynamic strain generates a small magnetization rotation in the magnetic layer which can induce electromotive force in a pick-up coil. The low frequency signal field can then be extracted from a magnitude of the modulated electromotive force using conventional phase sensitive detection techniques.

Similar configurations involving a piezoelectric element have been used for low frequency magnetic field detection in Ref. 130 with a reported sensitivity of $6.9 \times 10^{-6} \text{ A}\cdot\text{m}^{-1}/\sqrt{\text{Hz}}$ . This value may now be improved owing to the technical progress in numerical processing in the phase sensitive detection. Moreover, it might be advantageous to modulate light power and frequency compared to an amplitude and frequency of the electric field in piezoelectric samples, often limited by electric breakdown and fatigue.

FIG. 13. Schematic diagram of the magnetostrictive-photostrictive sensor for a DC magnetic field detection. White arrows show possible magnetization modulation as a function of elastic distortion induced by light.

Sonic-ultrasonic emission

Sonic-ultrasonic emission induced by light is another possible application. These devices take advantage of their linear response under oscillation of light frequency or/and power. In this case, magnetostrictive/photoelastic multilayers may respond to both magnetic field and light, possibly leading to enhanced low frequency sonic emission. The currently used piezoelectric materials are limited by breakdown voltage limit, providing a room for magnetostrictive/photoelastic components improvements. Recently reported ultrafast GHz coherent shear phonons photogeneration and photodetection in the $\text{BiFeO}_3$ with an ultrahigh efficiency may contribute to an essential breakthrough in this direction.¹³¹

Light controlled gas storage

The use of photostriction effect in porous organometallic materials for gas storage is particularly interesting, owing to the high technological importance of this topic and the multi-fold energy (hydrogen) storage impact. Due to the peculiar structure of this class of materials, the volume appears to be a key parameter for the storage properties, controlling the pore size. The hydrogen binding energy is extremely sensitive to the inter-atomic distances, due to the steep dependence of the van der Waals interaction on atomic separation, thus providing a possible way for deformation dependent control. For example, strain effects have been found to affect dramatically the gas storage properties in Mg.¹³² The key challenge here is to elaborate the porous materials with large photostriction at room temperature opening a way to fast optical pore size control.

Concluding remarks and outlook

We reviewed the current research progress in the field of photostriction and photostrictive materials touching some basics of the photoelastic phenomenon, related compounds, and mechanisms for light triggered strain control, with particular emphasis on finding the general rules in each exacting group of materials and comparing the magnitude of the effect. Table I summarizes the reported maximal values of photostriction for different compounds. For the reason that photostriction differs intrinsically from compound to compound and depends on thickness and light intensity, one can introduce a photostrictive efficiency at the given wavelength as

$${\eta }_{\text{eff}}=t\ast \frac{{\lambda }_{h\nu }}{I},(9)\backslash eta\_\{\backslash text\{eff\}\}\; =\; t\; *\; \backslash frac\{\backslash lambda\_\{h\backslash nu\}\}\{I\},\; \backslash quad\; (9)$$

where $t$ is the thickness along the illumination direction, $\lambda_{h\nu} = \Delta L/L$ is a linear photostriction, and $I$ is a light irradiance. Due to the linear approximation, the values of photostrictive efficiency have a rather indicative character, and the possible surface roughness dependence¹³³ is not taken into account by Eq. (9). The available data nevertheless show that the optimized PLZT ceramics tend to have the best photostrictive efficiency among the reported photostrictive ferroelectrics. The second place, however, belongs to the $\text{BiFeO}_3$ single crystal which might be more suitable for applications due to the much faster response time. The $\text{BiFeO}_3$ thin filmsTABLE I. Comparison of photostrictive properties for different compounds.

Compounds		Excitation (nm)	Thickness along illuminated direction	Light irradiance/fluence	Maximum $\backslash lambda\_\{h\backslash nu\}$ reported (%)	$\backslash eta\_\{eff\}$ (m3/W)	References
Ferroelectrics	SbSI single crystals	450	...	...	-0.007	...	4
	PLZT ceramics	365	0.5 mm	150 W/m2	0.01	$3.3\; \backslash times\; 10^\{-10\}$	24
	BiFeO3 crystal	365	90 $\backslash mu$m	326 W/m2	0.003	$8.2\; \backslash times\; 10^\{-12\}$	48
	BiFeO3 film	400	35 nm	2 mJ/cm2	0.46	$4\; \backslash times\; 10^\{-25\}$	47 and 134
	PbTiO3 film	400	20 nm	4.8 mJ/cm2	0.25	$5.2\; \backslash times\; 10^\{-26\}$	52 and 134
	Sn2P2S6 crystal	Infrared	...	210 kW/m2	$4.7\; \backslash times\; 10^\{-7\}$	...	35
Polar semiconductors	CdS crystal	496	15 $\backslash mu$m	...	0.0075	...	55 and 135
	GaAs crystal	855	10 $\backslash mu$m	...	$4\; \backslash times\; 10^\{-5\}$	...	56
Semiconductors	Germanium crystal	1069	0.5 mm	...	$7.84\; \backslash times\; 10^\{-8\}$	...	59
	Silicon crystal	248	0.5 mm	127 mJ/cm2	$-6.4\; \backslash times\; 10^\{-4\}$	$-3.7\; \backslash times\; 10^\{-20\}$	66 and 136
	Carbon nanotubes (MWNT)	675	$\backslash sim\; 10$ nm	15 kW/m2	$1.1\; \backslash times\; 10^\{-4\}$	$7.3\; \backslash times\; 10^\{-19\}$	71
Chalcogenide glasses	As40Se25S25Ge10 film	...	2.6 $\backslash mu$m	...	0.045	...	74
	As2Se3 film	...	...	400 W/m2	6.4	...	75
	As2S3 film	...	...	400 W/m2	5.4	...	75
	As40Se50Ge10 film	...	1.6 $\backslash mu$m	...	0.19	...	74
	GeSe2 film	...	...	400 W/m2	-5.6	...	75
	GeS2 film	...	...	400 W/m2	-11	...	75
Organic polymers	Poly-(4,4'-diamino-azobenzene-epyrromellitimide) films	400	0.1 mm	...	-1.2	...	94
	Nematic elastomers	365	...	...	20	...	96
	Poly(ethylacrylate) networks with azo-aromatic crosslinks	365, 436	...	...	0.25	...	95
	[Fe(phen)2(NCS)2] crystal	...	...	...	1.1 (Ref. 137)	...	104

are more efficient than the PbTiO₃ ones probably due to the unsaturated photostriction slope versus light fluence in the case of the BFO film.⁴⁷ In both groups of materials, ferroelectrics and polar semiconductors, the photostrictive mechanism is basically the same, involving direct light coupling to polarization and a possibility of light-induced free charge generation. Consequently, the exact contribution of each mechanism can depend on the excitation wavelength and can qualitatively be described by the Landau formalism with light generated charges term. In these materials, a general feature is the change in the photostriction trend versus wavelength that is correlated with a maximum in the photoconductivity.

Due to the better insulating nature and domain structure, the ferroelectric materials, however, can also be ferroelastic, meaning that a remnant strain levels induced by light can be created, involving even irreversible domain dynamics, at subcoercive electric fields. Therefore, a photoferroelastic effect can be considered as a third possible contribution to photostriction in photovoltaic ferroelectrics, involving an even larger deformation effect than expected by theory.

The photostriction in nonpolar semiconductors is generally determined by the amount of light generated charges.¹³⁸ The effect is relatively smaller and depends on the sign of the energy gap pressure susceptibility. The positive or negative susceptibilities mean positive or negative photostriction, respectively. In amorphous semiconductors, however, the excess of light-generated charges can induce plastic

deformation on the atomic bonds leading to much larger photostriction. Although chalcogenide glasses show large photostriction, its plastic nature and long response time are quite limiting factors for most technical implementations besides hologram recording. The reported electron-phonon-mediated photostriction in the multiwall carbon nanotubes fits well into the class of nonpolar photostrictive semiconductors, with an efficiency better than the one of silicon. In combination with thermal effects in composite structures, the carbon nanotubes manifest an important photoactuation application potential.¹³⁹

The photostriction in organic materials accounts for many prominent examples with large effects, where a core mechanism is either light-induced chemical structure transitions or different kinds of molecular reorientation (cooperative motion and alignment) under light irradiation.

While the largest photostriction is observed in organic based materials, their response time is generally slow limiting technical applications. However, photovoltaic ferroelectrics have an opposite problem: the response time is fast but the magnitude of the effect is small. In that respect, uniting the both advantages in a single material seems to be a promising research undertaking. Hypothetically, one can look for materials that are ferroelectric-photovoltaic and organic-based at the same time. To address this issue, one can focus on metal-organic systems, some of them were indeed reported to be ferroelectric.^140,141 Another promising candidate for photostriction can be molecular crystals; some of them were reported to be highly light-sensitive and electrically polar at the sametime.¹⁴² The spin crossover molecular systems are particularly interesting, because photostriction is accompanied with a change in magnetic properties. Thus, it might be interesting to study magnetostrictive-photostrictive coupling in some recently reported compounds.¹⁴³

As far as ferroelectric materials are concerned, it appears to be extremely important to distinguish between the photostriction in saturated and spontaneous ferroelectric states. Based on recent experiments, the analogy to remnant strain effects^53,54 can also be seen in photostriction.¹²⁸ From a practical and short-term perspective, it appears to be the most promising to test creating ferroelastic remnant states by light and erasing them by electric fields. This can open an avenue for the light-controlled logic and memory applications in both optics and magnetic straintronics with photostrictive elements.

¹Z. L. Wang and Z. C. Kang, Functional and Smart Materials (Plenum Publishing Corp., Spring Street, New York, 1998), p. 233.

²R. Bogue, "Smart materials: A review of recent developments," Assem. Autom. 32(1), 3–7 (2012).

³M. Shahinpoor and H.-J. Schneider, Intelligent Materials (The Royal Society of Chemistry, 2008).

⁴I. Tatsuzaki, K. Itoh, S. Ueda, and I. Ishindo, Phys. Rev. Lett. 17, 198 (1966).

⁵V. M. Fridkin, I. I. Groshik, V. A. Lakhovizkaya, M. P. Mikhailov, and V. N. Nosov, Appl. Phys. Lett. 10, 354 (1967).

⁶A. A. Grekov, P. E. Pasynkov, and E. D. Rogach, Fiz. Tverd. Tela 12, 2216 (1972) (in Russian).

⁷A. A. Grekov, E. D. Rogach, and L. N. Syrkin, Fiz. Tverd. Tela 14, 2768 (1972) (in Russian).

⁸P. S. Brody, Ferroelectrics 50(1), 27 (1983).

⁹P. S. Brody, U.S. patent 4,524,294 (18 June 1985).

¹⁰V. M. Fridkin, Photoferroelectrics (Springer, New York, 1979).

¹¹R. Nitsche and W. J. Merz, J. Phys. Chem. Solids 13, 154 (1960).

¹²E. Fatuzzo, G. Harbeke, W. J. Merz, R. Nitsche, H. Roetschi, and W. Ruppel, Phys. Rev. 127, 2036 (1962).

¹³A. A. Grekov, E. D. Rogach, and A. G. Sukniazov, Fiz. Tverd. Tela 14, 3559 (1970) (in Russian).

¹⁴R. E. Pasynkov, Ferroelectrics 6, 19 (1973).

¹⁵J. Grigas, A. Kajokas, A. Audzijonis, and L. Zigas, J. Eur. Ceram. Soc. 21, 1337 (2001).

¹⁶A. Audzijonis, L. Zigas, R. Sereika, and A. Kvedaravicius, Ferroelectrics 425, 45 (2011).

¹⁷B. Garbarz-Glos, Ferroelectrics 292, 137 (2003).

¹⁸B. Garbarz-Glos and J. Grigas, Ferroelectrics 393, 38 (2009).

¹⁹K. Uchino and M. Aizawa, Jpn. J. Appl. Phys., Part 1 24, 139 (1985).

²⁰K. Uchino, M. Aizawa, and S. Nomura, Ferroelectrics 64, 199 (1985).

²¹T. Sada, M. Inoue, and K. Uchino, J. Ceram. Soc. Jpn., Int. Ed. 95, 499 (1987) (in Japanese).

²²S.-Y. Chu and K. Uchino, Ferroelectrics 174, 185 (1995).

²³A. Dogan, P. Poosanaas, I. R. Abothu, S. Komarneni, and K. Uchino, J. Ceram. Soc. Jpn. 109, 493 (2001).

²⁴K. Takagi et al., J. Am. Ceram. Soc. 87(8), 1477 (2004).

²⁵P. Poosanaas, A. Dogan, S. Thakoor, and K. Uchino, J. Appl. Phys. 84, 1508 (1998).

²⁶P. Poosanaas, K. Tonooka, and K. Uchino, Mechatronics 10, 467–487 (2000).

²⁷A. Dogan, A. V. Prasadarao, K. Uchino, P. Poosanaas, and S. Komarneni, J. Electroceram. 1, 105 (1997).

²⁸P. Poosanaas and K. Uchino, Mater. Chem. Phys. 61, 36 (1999).

²⁹K. Nonaka, M. Akiyama, A. Takase, T. Baba, K. Yamamoto, and H. Ito, Jpn. J. Appl. Phys., Part 1 34, 5380 (1995).

³⁰A. A. Molnar, Yu. M. Vysochanskii, A. A. Horvat, and Yu. S. Nakonechnii, Ferroelectrics 192, 137 (1997).

³¹R. Nitsche and P. Wild, Mater. Res. Bull. 5, 419 (1970).

³²C. D. Carpentier and R. Nitsche, Mater. Res. Bull. 9, 1097 (1974).

³³S. Odoulov, A. Shumelyuk, U. Hellwig, R. Rupp, and G. Brost, Jpn. J. Appl. Phys., Part 1 35, 5154 (1996).

³⁴J. Kroupa, Y. Tyagur, A. A. Grabar, and Y. M. Vysochanskii, Ferroelectrics 223, 421 (1999).

³⁵Y. W. Cho, S. K. Choi, and Y. M. Vysochanskii, J. Mater. Res. 16(11), 3317 (2001).

³⁶D. Lebeugle, D. Colson, A. Forget, M. Viret, A. M. Bataille, and A. Gukasov, Phys. Rev. Lett. 100, 227602 (2008).

³⁷J. Kreisel, M. Alexe, and P. A. Thomas, Nat. Mater. 11, 260 (2012).

³⁸B. Kundys, M. Viret, D. Colson, and D. O. Kundys, Nat. Mater. 9, 803 (2010).

³⁹Z. Jin et al., Appl. Phys. Lett. 101, 242902 (2012).

⁴⁰H. Wen et al., Phys. Rev. Lett. 110, 037601 (2013).

⁴¹Editorial, "A historical bond," Nat. Photonics 4, 663 (2010).

⁴²L. Y. Chen et al., Appl. Phys. Lett. 101, 041902 (2012).

⁴³P. Ruello, T. Pezeril, S. Avanesyan, G. Vaudel, V. Gusev, I. C. Infante, and B. Dkhil, Appl. Phys. Lett. 100, 212906 (2012).

⁴⁴X. Chen, G. Hu, W. Wu, C. Yang, and X. Wang, J. Am. Ceram. Soc. 93, 948 (2010).

⁴⁵S. Fujino, M. Murakami, V. Anbusathaiah, S.-H. Lim, V. Nagarajan, C. J. Fennie, M. Wuttig, L. Salamanca-Riba, and I. Takeuchi, Appl. Phys. Lett. 92, 202904 (2008).

⁴⁶I. O. Troyanchuk, D. V. Karpinsky, M. V. Bushinsky, V. A. Khomchenko, G. N. Kakazei, J. P. Araujo, M. Tovar, V. Sikolenko, V. Efimov, and A. L. Kholkin, Phys. Rev. B 83, 054109 (2011).

⁴⁷D. Schick et al., Phys. Rev. Lett. 112, 097602 (2014).

⁴⁸B. Kundys, M. Viret, C. Meny, V. Da Costa, D. Colson, and B. Doudin, Phys. Rev. B. 85, 092301 (2012).

⁴⁹H. T. Yi, T. Choi, S. G. Choi, Y. S. Oh, and S.-W. Cheong, Adv. Mater. 23, 3403 (2011).

⁵⁰G. W. Pabst, L. W. Martin, Y.-H. Chu, and R. Ramesh, Appl. Phys. Lett. 90, 072902 (2007).

⁵¹C.-F. Chung and J.-M. Wuz, Electrochem. Solid-State Lett. 8, F63–F66 (2005).

⁵²D. Daranciang et al., Phys. Rev. Lett. 108, 087601 (2012).

⁵³B. Kundys, V. Iurchuk, C. Meny, H. Majjad, and B. Doudin, Appl. Phys. Lett. 104, 232905 (2014).

⁵⁴V. Iurchuk, B. Doudin, and B. Kundys, J. Phys.: Condens. Matter 26, 292202 (2014).

⁵⁵J. Lagowski and H. C. Gatos, Appl. Phys. Lett. 20(1), 15 (1972).

⁵⁶J. Lagowski and H. C. Gatos, Surf. Sci. 45, 353 (1974).

⁵⁷H. Jaffe, D. Berlincourt, H. Krueger, and L. Shiozawa, in Proceedings of the 14th Annual Symposium On Frequency Control, U.S. Army Research and Development Lab., Fort Monmouth, New Jersey, 31 May 1960; M. Aven and J. S. Prener, Physics and Chemistry of II–VI Compounds (North-Holland, Amsterdam, 1967), p. 603.

⁵⁸Handbook of Electronic Materials, III–V Semiconducting Compounds, edited by M. Neuberger (IFI/Plenum Press, New York, 1972), Vol. 2.

⁵⁹T. Figielski, Phys. Status Solidi 1, 306 (1961).

⁶⁰R. W. Keyes, IBM J. Res. Dev. 5, 266 (1961).

⁶¹J. C. North and R. C. Buschert, Phys. Rev. Lett. 13, 609 (1964).

⁶²J. C. North and R. C. Buschert, Phys. Rev. 143, 609 (1966).

⁶³J. C. North and R. C. Buschert, J. Phys. Chem. Solids 27, 489 (1966).

⁶⁴W. B. Gauster and D. H. Habing, Phys. Rev. Lett. 18, 1058 (1967).

⁶⁵P. G. Datskos, S. Rajic, and I. Datskou, Appl. Phys. Lett. 73, 2319 (1998).

⁶⁶J. R. Buschert and R. Colella, Solid State Commun. 80(6), 419 (1991).

⁶⁷I. A. Levitsky, B. J. Landi, R. P. Raffaele, S. L. Castro, and S. G. Bailey, Prog. Photovoltaics 13(2), 165 (2005).

⁶⁸J. Yeonwoong, X. Li, N. K. Rajan, A. D. Taylor, and M. A. Reed, Nano Lett. 13(1), 95 (2013).

⁶⁹Y. Zhang and S. Iijima, Phys. Rev. Lett. 82, 3472 (1999).

⁷⁰S. B. Cronin et al., Phys. Rev. Lett. 96, 127403 (2006).

⁷¹S. V. Ahir, E. M. Terentjev, S. X. Lu, and B. Panchapakesan, Phys. Rev. B 76, 165437 (2007).

⁷²M.-F. Yu, O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly, and R. S. Ruoff, Science 287(5453), 637 (2000).⁷³I. A. Levitsky, W. B. Euler, and V. A. Karachevtsev, Photophysics of Carbon Nanotubes Interfaced with Organic and Inorganic Materials (Springer-Verlag, London, 2012).

⁷⁴T. Igo, Y. Noguchi, and H. Naga, Appl. Phys. Lett. 25, 193 (1974).

⁷⁵Y. Kuzukawa, A. Ganjo, and K. Shimakawa, J. Non-Cryst. Solids 227–230, 715 (1998).

⁷⁶H. Hamanaka, K. Tanaka, A. Matsuda, and S. Iijima, Solid State Commun. 19, 499 (1976).

⁷⁷M. Frumar, A. P. Firth, and A. E. Owen, Philos. Mag. B 50, 463 (1984).

⁷⁸F. Kyriazis and S. N. Yannopoulos, Appl. Phys. Lett. 94, 101901 (2009).

⁷⁹H. Hisakuni and K. Tanaka, Appl. Phys. Lett. 65, 2925 (1994).

⁸⁰A. Ganjo, Y. Ikeda, and K. Shimakawa, Appl. Phys. Lett. 74, 2119 (1999).

⁸¹M. Stuchlik, P. Kremer, and S. R. Elliott, J. Optoelectron. Adv. Mater. 3(2), 361 (2001).

⁸²S. R. Elliott, J. Non-Cryst. Solids 81, 71 (1986).

⁸³J. D. Musgraves, K. Richardson, and H. Jain, Opt. Mater. Express 1(5), 921 (2011).

⁸⁴K. Shimakawa, A. Kolobov, and S. R. Elliott, Adv. Phys. 44(6), 475 (1995).

⁸⁵S. A. Keneman, Appl. Phys. Lett. 19, 205 (1971).

⁸⁶H. Hamanaka, K. Tanaka, and S. Iizima, Solid State Commun. 23, 63 (1977).

⁸⁷T. Mori, N. Hosokawa, and K. Shimakawa, J. Non-Cryst. Solids 354, 2683 (2008).

⁸⁸S. Kugler, J. Hegedus, and K. Kohary, in “Light-induced volume changes in chalcogenide glasses,” Optical Properties of Condensed Matter, edited by J. Singh (Wiley, 2006), pp. 143–158.

⁸⁹M. Irie, Bull. Chem. Soc. Jpn. 81(8), 917 (2008).

⁹⁰H. Yu, J. Mater. Chem. C 2, 3047 (2014).

⁹¹D. Iqbal and M. H. Samiullah, Materials 6, 116 (2013).

⁹²L. Matejka, M. Ilavsky, K. Dusek, and O. Wichterle, Polymer 22, 1511 (1981).

⁹³L. Matejka, K. Dusek, and M. Ilavsky, Polym. Bull. 1, 659 (1979).

⁹⁴G. Van der Veen and W. Prins, Nature 230, 70 (1971).

⁹⁵C. D. Eisenbach, Polymer 21(10), 1175 (1980).

⁹⁶H. Finkelmann, E. Nishikawa, G. G. Pereira, and M. Warner, Phys. Rev. Lett. 87, 015501 (2001).

⁹⁷Y. Yu, M. Nakano, and T. Ikeda, Nature 425(6954), 145 (2003).

⁹⁸M. Camacho-Lopez, H. Finkelmann, P. Palfy-Muhoray, and M. Shelley, Nat. Mater. 3, 307 (2004).

⁹⁹S. Kobatake, S. Takami, H. Muto, T. Ishikawa, and M. Irie, Nature 446, 778 (2007).

¹⁰⁰Molecular Magnets Recent Highlights: Recent Highlights, edited by W. Linert and M. Verdaguer (Springer, 2003).

¹⁰¹P. Gütllich, A. Hauser, and H. Spiering, Angew. Chem., Int. Ed. Engl. 33, 2024 (1994).

¹⁰²K. Lüghausen, H. Siegert, Th. Woike, and S. Haussühl, J. Phys. Chem. Solids 56, 1291 (1995).

¹⁰³P. Gütllich, Y. Garcia, and T. Woike, Coord. Chem. Rev. 219, 839 (2001).

¹⁰⁴M. Marchivie, Ph. Guionneau, J. A. K. Howard, G. Chastanet, J.-F. Letard, A. E. Goeta, and D. Chasseau, J. Am. Chem. Soc. 124, 194 (2002).

¹⁰⁵S. Decurtins, P. Gutlich, C. Kolher, and H. Spiering, Chem. Phys. Lett. 105, 1 (1984).

¹⁰⁶S. Decurtins, P. Gutlich, K. M. Hasselbach, A. Hauser, and H. Spiering, Inorg. Chem. 24(14), 2174 (1985).

¹⁰⁷M. Lorenc, J. Hébert, N. Moisan, E. Trzop, M. Servol, M. Buron-Le Cointe, H. Cailleau, M. L. Boillot, E. Pontecorvo, M. Wulff, S. Koshihara, and E. Collet, Phys. Rev. Lett. 103, 028301 (2009).

¹⁰⁸K. Uchino, J. Rob. Mechatronics 1(12), 24 (1989); http://www.fujipress.jp/finder/xslt.php?mode=present&inputfile=ROBOT000100020007.xml.

¹⁰⁹S.-Y. Chu and K. Uchino, in Proceedings of the 9th IEEE International Symposium on Applications of Ferroelectrics, ISAF '94 (1994).

¹¹⁰K. Uchino, Mater. Res. Innovations 1, 163 (1997).

¹¹¹M. Yamada, M. Kondo, R. Miyasato, Y. Nake, J. Mamiya, M. Kinoshita, A. Shishido, Y. Yu, J. Barrett, and T. Ikeda, J. Mater. Chem. 19, 60–62 (2009).

¹¹²D. Sun and L. Tong, Int. J. Solids Struct. 44, 672 (2007).

¹¹³D. Sun and L. Tong, J. Sound Vib. 312, 182 (2008).

¹¹⁴D. Habault, H. Zhang, and Y. Zhao, Chem. Soc. Rev. 42, 7244 (2013).

¹¹⁵P. G. Datskos, S. Rajic, M. J. Sepaniak, N. Lavrik, C. A. Tipple, L. R. Senesac, and I. Datskou, J. Vac. Sci. Technol., B 19, 1173 (2001).

¹¹⁶N. V. Lavrik, M. J. Sepaniak, and P. G. Datskos, Rev. Sci. Instrum. 75, 2229 (2004).

¹¹⁷A. Toprak and O. Tigli, Appl. Phys. Rev. 1, 031104 (2014).

¹¹⁸T. Lafont, L. Gimeno, J. Delamare, G. A. Lebedev, D. I. Zakharov, B. Viala, O. Cugat, N. Galopin, L. Garbuio, and O. Geoffroy, J. Micromech. Microeng. 22, 094009 (2012).

¹¹⁹See http://www.nlosource.com/PhotoMechHistory.html for more on photostriction and all-optical technologies.

¹²⁰S. Egusa et al., Nat. Mater. 9, 643 (2010).

¹²¹M. G. Kuzyk, Polymer Fiber Optics: Materials, Physics, and Applications (CRC Press, Taylor and Francis, Boca Raton, 2007).

¹²²L. Calvez, Z. Yang, and P. Lucas, J. Phys. D: Appl. Phys. 43, 445401 (2010).

¹²³K. K. Shvarts, J. A. Teteris, I. P. Manika, M. J. Reinfelde, and V. I. Gerbreder, J. Non-Cryst. Solids 90, 509 (1987).

¹²⁴A. Zoubir, M. Richardson, C. Rivero, A. Schulte, C. Lopez, K. Richardson, N. Hô, and R. Vallée, Opt. Lett. 29, 748 (2004).

¹²⁵P. J. Sturman, Photovoltaic and Photo-refractive Effects in Noncentrosymmetric Materials (Gordon and Breach Science Publishers, OPA, Amsterdam, 1992).

¹²⁶M. Imlau, Th. Woike, R. Schieder, and R. A. Rupp, Phys. Rev. Lett. 82, 2860 (1999).

¹²⁷H.-J. Liu, L.-Y. Chen, Q. He, C.-W. Liang, Y.-Z. Chen, Y.-S. Chien, Y.-H. Hsieh, S.-J. Lin, E. Arenholz, C.-W. Luo, Y.-L. Chueh, Y.-C. Chen, and Y.-H. Chu, ACS Nano 6(8), 6952 (2012).

¹²⁸B. Kundys, C. Meny, M. R. J. Gibbs, V. Da Costa, M. Viret, M. Acosta, D. Colson, and B. Doudin, Appl. Phys. Lett. 100, 262411 (2012).

¹²⁹E. du Tremolet de Lacheisserie, Magnetostriction: Theory and Application of Magnetoelasticity (CRC Press, BocaRaton, FL, 1993).

¹³⁰M. D. Mermelstein and A. Dandridge, Appl. Phys. Lett. 51, 545 (1987).

¹³¹M. Lejman, G. Vaudel, I. C. Infante, P. Gemeiner, V. E. Gusev, B. Dkhil, and P. Ruello, Nat. Commun. 5, 4301 (2014).

¹³²A. Baldi, M. Gonzalez-Silveira, V. Palmisano, B. Dam, and R. Griessen, Phys. Rev. Lett. 102, 226102 (2009).

¹³³P. Poosanaas and K. Uchino, Nonlinear Optics: Materials, Fundamentals and Applications (Optical Society of America, Conference paper, Kauai-Lihue, Hawaii, 2000).

¹³⁴The 50 fs illumination time was assumed to evaluate the irradiance.

¹³⁵Using data from Ref. 55, an approximate estimate was made (see Fig. 6 and text).

¹³⁶The 15 ns illumination time was assumed to evaluate the irradiance.

¹³⁷This value indicates the volumetric photostrictive change.

¹³⁸N. Inui, Jpn. J. Appl. Phys., Part 1 45, 1675 (2006).

¹³⁹X. Zhang et al., Nat. Commun. 5, 2983 (2014).

¹⁴⁰B. Kundys et al., Phys. Rev. B 81, 224434 (2010).

¹⁴¹N. Leblanc, N. Mercier, L. Zorina, S. Simonov, P. Auban-Senzier, and C. Pasquier, J. Am. Chem. Soc. 133, 14924 (2011).

¹⁴²Ch. Bosshard, R. Spreiter, and P. Günter, Ferroelectrics 258(1), 89 (2001).

¹⁴³S. Ohkoshi, K. Imoto, Y. Tsunobuchi, S. Takano, and H. Tokoro, Nat. Chem. 3, 564 (2011).