FEATURES since magnetic nanoparticles are less easily b) ~Fig. 1biTEMof a cell (h-tert human destroyed or inactivatedby cells than many fibrobtast)exposed to lactoferrin non-magentic ones there is the disadvan labellednanoparticles after 24 hours. tage that persistentparticles maycause later Theparticles are on the cellsurfaceor cell damage and death. The same consider closeoutside it) seearrows)and none ations apply to situations where magnetic havebeen endocytosed. Claimshave nanoparticles are beingusedfor generating been made inthe literature that hyperthermiaby the application ofexternal transferrin-labelled (a related protein) fields. particlesare endocytosed butthe Cartmell et al. (2003) have suggested evidenceforthis is often basedon light that mechanical stimulation could be microscopywhich does not have applied to cells by using an external mag sufficientresolution to locatethe netic field to applya twisting motion to the particlesprecisely.Arrowsmarkcell particles. Since mechanical stimulation is surfaceaccumulations. effective in activating cells this may be a Figsla and lb courtesyofOrA.K.Gupta, future area for therapeutic developments. UniversityofGlasgow. Oneinterestingif ratherworryingpossi bilityis that exposure ofcells to nanofeatures or nanoparticles can Cartmell, S.H. et aL (2003) Development ofmagnetic particle techniques cause large, if perhaps only temporary, changes in gene expres for long term culture ofbone cells with intermittent mechanical stimula sion. This would not necessarily be linked solely to the use of tion. IEEETransactions in Nanobioscience 1.92-97 magnetic nanoparticles because the work of Dalby et al. (2003) Cranfield,C et al. (in press) IEEETransactions in Nanobioscience shows that nanofeatures on polymer surfaces can cause change in gene expression. Dalby,M.J.,et al.(2003).Interaction ofhuman and tissue celltypeswith 95nm high nanotopography. IEEEThmsactions in Nanobioscience 1:18-23 Since excess particles may well be adsorbed onto the surfaces ofcells and intercellular material it is appropriate to consider how Kobayashi, H., et aL (2003). Micro-magnetic resonance lymphangiogra these adsorbed particles might affect cell behaviour.Thishashard phy in mice using a novel dendrimer-based magnetic resonance imaging lybeen examined but some recent work (Berry and Curtis 2003) contrast agent. Cancer Research 63(2): 271-276 suggests that theymay act much as nanofeatured surfaces do. Poduslo, J.P.et aL (2002). Molecular targeting ofAlzheimer's amyloid ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• n •••••••••••••• u ••• plaques for contrast-enhanced magnetic resonance imaging Neurobiolo References gy ofDisease 11,315-329 Berry, C and Curtis A. (2003) Applications ofMagnetic Nanoparticles in The J.Magnetism and Magnetic Materials Volume 194, Issue 1-3,April Biomedicine 1999published about 40 papers relevant to this area. A title list can be Journal ofPhysics D:Applied Physics found at http://www.infomag.ru:8082/dbase/J087E1010503-101.txt
solid is a metal or an insulator, what only matters is if the atoms either acquire an induced magnetic moment or they have a per Magnetostriction: manent one. By Hookes's law the deformation suffered is proportional to the material size, say 1,and therefore it is conve nient to express the deformation by the relative variation fundamental principles A= MIl, called the linearMS. In a crystal Ais doubly anisotropic: it depends on the crystallographic direction of measurement and on the direction along which the magnetization, Mis oriented by and novel magneto the applied magnetic field, if. Because MS depends on the crys talline directions of Aand M(~and a,respectively) we have to specifyMS as A(a,~)(Fig. I). But also in a poly-crystalline mate strictive materials rial, formed by small crystallites more or less randomly distributed,the MS measured along if, called the parallelMS, Allis A.delMoral quite different from thatmeasured alongthe perpendiculardirec Departamento de Magnetismo, Dpto.Ffsicade Materia tion, termed the perpendicular MS, Al.. It can happen that Condensada & lCMA, Universidadde Zaragoza & CSlC, 50009 A.l==- ~I/2, but usuallythis is not the case and then thevolume of Zaragoza,Spain the solid changes by the amountIN / V = ~I+ 2A.l, called the vol ume MS. On the other hand, the shape (or form of the unit cell) of the material is modified if ~I*-A.l' and a measure of this effect Magnetostriction and magnetoelastic c-oupling is the difference At= All- A.L,called the shape MS. Both MS's hen any magnetic solid acquires a m~netization,Munder appear in crystalline, polycrystalline and amorphous solids. Wthe application of a magnetic field, H at the same time its The key questions are: why is this phenomenon so general, crystalline (or amorphous) lattice is deformed, a phenomenon and what is the physical mechanism of MS? We know that atoms called magnetostriction (MS from hereinafter). Therefore this is a possess two main sources of magnetic moment the orbital elec general effect suffered by all solid matter, once it becomes mag tronic motion around the nucleus, which gives rise to the orbital netized. It therefore appears for any kind of magnetic materials: angular momentum,Land the spin,S.Classically the orbital cur diamagnetic, paramagnetic, ferromagnetic, anti-ferromagnetic, rent produces a magnetic field, 13L where the spin magnetic ~Thi· ferrimagnetic, superconductors, etc. It does not matter if the moment , ms...... = - SflB has an energy - ms.....• tlL. S IS a re1'"atiVlStic
europhysics news NOVEMBER/DECEMBER 2003 211 Article available at http://www.europhysicsnews.org or http://dx.doi.org/10.1051/epn:2003603 FEATURES
effect, and because BL-L, the energy becomes Eso =~L'S,which links the spins to the orbits, e.g. in a ferromagnet. This interac tion is called the spin-orbit coupling,and is the first ingredient of MS. By it when S (or 11)is rotated by the torque r =msxB,L is dragged on. In a ferromagnet this is very important because the spins become ordered below the Curie temperature. But there is another equally important ingredient, the interaction which cou ples the orbit to the lattice, ifwe want the lattice to deform when their atoms are magnetized. Let us consider a ferromagnet (FM) or antiferromagnet (AF) in order to fix ideas. For transition metals (TM), notably Fe, Co, Ni and Mn, or their ions (in FM insulators) and for rare earth (RE) metals and insulators, their respective atomic 3d and 4fshells are incompletely filled, and therefore they possess spin and orbital momenta, giving rise to a magnetic moment, m = mL+ ms. In magnetic solids there exists a strongly inhomogeneous electric field, called the crystal electricfield (CEF), whose gradients inter act with the ion magnetic electrons, giving rise to a splitting ofthe A B!h.:L A([111],(111))MSisotherms vs.appliedmagneticfield ion energy levels while still keeping some degeneracy (CEF ener for______the cubicintermetallicferromagnetTbAh (Te=115K)[1]. ---1 gy levels). The splitting is purely dictated by symmetry: e.g. in cubic symmetrythe levels are eg(x2) and t2g(x3). This remaining A=BI c, is attained when the full energy is minimized. The Bij degeneracy is quite important for MS, because otherwise (L) is are also called MEL constants and the magnetic moment, m, cou zero. In the solid, 3d and 4f electrons have orbital wave-func pling to the strainedlattice, is called MEL coupling.The number of tions, eporb,rather different from those ofthe free atoms (or ions), strains, and so ofMEL constants, is dictated by the crystal sym as a consequence ofthe CEF potential which admixes the free metry alone. The irreduciblestrains are the minimum number of atom states JML>(MLis the quantized projection ofL along them, spanning all unit cell deformations compatible with sym some crystal symmetry axis, 02). So the electronic charge distri metry. They are four for cubic symmetry (see fig. 2): volume bution (i.e.leporbI2) is determined by the symmetry entourage of dilation E (l (a), tetragonal distortion along a cubic axis EY (b), the ionic site and is very anisotropic. Most importantly is that orthorhombic deformation of{lOO}planes E1(not shown in Fig. eporbchanges when L rotates, with the result that the CEF energy 2), and shear ofthe <100> axes EE(c). changes, giving rise to the magnetocrystalline anisotropy (MCA), bywhich the crystal energy changes with the Mdirection. 3d Invareffect, magnetostrictive superlattices and "giant" shells are little screened from the CEF ("medium" to "strong" magnetostriction materials CEF) whereas 4fones are much more screened ("weak" CEF). For There is a second kind of MS due to the spatial dependence of the former, the ground state eporbis formed by almost pure 50% the exchange (EX) interaction, i.e. aJ {r)ldr, called exchange (EX) admixtures ofI±ML>states and therefore the quantum mechan MS. Because JSi'Sj is isotropic, EX MS produces just a volume ics expectation value (Lz) is rather weak (an effect called deformation (VMS), given by IDex - {aJ {r)/ar)/B, where Bis the quenching). Contrarilyfor the RE metals quenching is veryweak, bulk modulus. Since this VMS shows up spontaneouslyinthe ther i.e., (Lz)is weakly reduced. FM orAF order in solids is the result of mal expansion (THE) when the temperature approaches from the exchange interaction between electrons, with energy above the Curie temperature, Te it gives an additional contribu E""= - 2JSi'Sj,which is clearly isotropic. Very important for MS is tion to the lattice THE, that manifests itself as an anomaly ofthe the fact thatJ (exchange integral) depends on the atom distance. THE coefficient, a V = 3 (lIL)(aL/aT). It also manifests itselfin an This interaction gives rise to the spontaneous magnetization, M,. increase of Abeyond its technical saturation, As,a phenomenon Therefore when Ms (or S) is rotated within the crystal by the r called forced MS,where fundamental many-body electron repul-
torque, the ionic charge cloud is dragged on and also distorted, ,;,;------~1 with the consequent distortion of the ion entourage: so the crys ..0/~.,. 7.:7 .:I tal is deformed. This deformation is the so-called single-ion CEF I I ' I I magnetostriction and it requires the rotation ofMsin some way: I I either by rigid vector rotation or by domain wall displacements I that entrain Ms rotation. This happens for 70S, 90°, and 109.50 :/,// domain walls (DW) in cubic crystals (1800 DW displacement can '-- ..J
212 europhysicsnews NOVEMBER/DECEMBER 2003 FEATURES
12.-- ...... ---,---.,.---,----r--, .-«l p... ~ .eo
A flg,~.;Invar effect in the thermal expansion (THE)ofthe from the bulk Ho one, 8~Ulk(x points) is due to interface and non tetragonal hard ferromagnet Nd2Fe14B.Alsoshown is the lattice linear MELstresses. Alsoshown is the quotient BY/BYbulk (€ THEextrapolation (line) {1]. . points) {1]. sion is involved and which varies with H linearly. It can happen form BNLEmf (the MEL free energyis now quadratic in the strain), that the spontaneous VMS is positive and of such a magnitude which can be as substantial as Bvol. that it fully or partially cancels the lattice THE, a phenomenon The NLMEL coupling is also well manifested in the elasticcon known as the invar effect,and bywhich the material does not con stants, where it gives rise to the appearance of a non-symmetricor tract belowTc• Obviously this effect has enormous consequences rotational strains, ffi;j, which differ from the usual symmetric ones for applications, where materials with "null" effective a are (E ij = E ji) in that ffi;j *'COji. These strains give rise to the rotational required (watches, optical interferometers, condensers, standards, invariance of the MEL energy, in the sense that the MEL etc.). Examples of invar materials are the Fe-Ni, Fe-Ni-Cr, Fe-Pt anisotropy that originates from the Mrotation is the same as the alloys, certain compositions of Fe-Ni-Mn alloys, NhAl, MnSi, one produced by an imposed external rotational deformation of amorphous Fe-Band intermetallics RECo2,RE2Fe14Band DY2Fe17 opposite sense. The existence of the three MEL stresses has been 2 (Fig.3). Theoretically it can be shown that COmag- T ,precisely the experimentallyobserved in hexagonal symmetry,SL,RE/SP (SP= same dependence as that for the lattice thermal contraction, Y,Lu, Sc) (see Figo4) and also in SL and ML made of cubic transi ffiIatt(T). Therefore compensation ofthe lattice THE takes place, or tion metal, such as TMlM, where TM= Fe,Co,Ni and M is a noble in other words av=o. metal or Cu. The RE/SP SL are important from a basic point of Currentlyvigorous research is been undertaken on MS in arti view, because RE magnetic moments are well localized and CEF ficially periodic structures known as multilayers (ML) and MS is veryamenable to study. However,because REmetals are FM superlattices (SL), formed by alternative deposition of thin layers or helically ordered only at rather low temperatures (Gd, Dy, Tb (=5-103 A)oftwo magnetic materials. Because ofthe modification and Ho, although more complex modulated magnetic structures of the electronic structure, an interface MS is manifested at the do appear), they are of not much use for applications. Instead the interface (IF) of the two layers (in the SL the layers grow coher ferromagnetics TM/M ML, with high Tc and weak magnetic ently, forming an artficial macroscopic lattice). Usually only one anisotropy, may have wide application as MEL transducers layer is magnetic (m),the m-m exchange beingtransmittedbythe (actuators and sensors) within nanostructured devices. non-magnetic (nm) one. Because ML and SL are grown upon a MS in soft transition metals and their alloys is relatively small substrate, this restrains the MS distortion, and therefore experi (expressed in 10-6 = IflSt units), with As=At= - 8, -34 for Fe and mentally one measures the MEL stress. Neel (1954) showed that Ni, - 22 for (FeCo)soB20, and --30 for spinel ferrites (TMxFel-x04, the IF MEL stress has the form BIF/tm,where tmis the layer thick TM = Fe,Ni,Mn). An exception is the Co ferrite that has the very ness. Therefore this stress becomes large value 1..([100], [100])=-700,where all values are at RT.How as important as the bulk contribu ever, for the RE metals and their intermetallic compounds, MS is tion for very thin m-layers, and can generally very large,up to = 1% (thousands of 1 fJSt) at low tem be of opposite sign. When the peratures, because of the RE3+ion large quadrupolar electric growth is epitaxial (Le. the lattice moment [Q - J(J-1/2)],which interacts with the inhomogeneous constants am= anmin the SL),which CEF, albeit this is more more shielded than in TM. The inter can be achieved usingsophisticated metallics, RETM2 (TM=Fe,Co,Ni and Mn) with Laves-phase fabrication techniques (molecular structure (the RE3+ form a diamond lattice), showthe largest At beam epitaxy, laser ablafion and ever found (the so-called"giant"MS).Among these are, for exam sputtering), the m-layer is under ple: TbNh (0.23% at 4.2K), NdCo2 (-0.17% at 77K) and TbFe2' mechanical stress, which induces with the largest room-temperature Atknown (0.2%), reaching 004 enormous misfit strains (up to % at 4.2K The problem for use in applications of these inter- =10%), when the bulk materials have different lattice constants. In ~fllk~;ATerfenol magnetostrictive device for generation of such a situation the magnetoelastic seismic waves (from Handbook of Giant Magnetostrictive coupling becomes non-linear, and Materials, G.Engdahl, Academic,NewYork,2000).. Bvol is modified by a stress of the
europhysics news NOVEMBER/DECEMBER 2003 213 fEATURES metallics is the large saturation magnetic field required, of '" IT. taken together with electronic structure, are involved in magne However, combining the opposite sign anisotropies of Tb3+ and tostriction, and from which fundamental information can be Df+ in the compound TbO.27DYo.73F~(a material commercially extracted. Magnetostriction is a ubiquitous phenomenon in solid known as Terfenol@),the saturation field is reduced down to matter, spanning a wide range of values, between "'10-8 (strongly '" 2kOe , still with a room temperature MS of 0.12%. This mater correlated systems) and "'10-2 (rare earth and other inter ial is currently that most used for MEL transducers, and can be metallics). Applications in sensors and actuators, among many prepared in engineering amounts (Fig.S). But the largest MS ever others, is an active reality and their use in nanostructured measured is in TbMn2, where At(40K, H=lST)=0.6% and, which devices is also promising. COt(20K,H=lST)=-1.6%are certainly hugevalues.But otheralloys not containingRE also show giantvolume MS values of0.5-1%, at near room temperature and at rather modest fields ("'1-5 kOe), notably FeRh, FeRh1-xPtxand Hf1-xTaxFe2_y,when the applied magnetic field induces an AF to FM transition. This makes them References also very promising for applications. [1] A.del Moral, Magnetostriction;BasicPrinciplesand Applications, As concluding remarks, we have seen that the most funda Inst.ofPhysics Pub.,Bristol,to appear (2004) (references to original mental interactions in solids (exchange, CEF and spin-lattice), papers are given here).
. derivatives and organic radicals like the famous BEDT-TTF, well known to form conducting and superconductingmaterials. They Molecularmagnetism: obtainedby electrocrystallisation a compound of general formu la [BEDT-TTFh[MnCr(oxalatoh]. The structure comprises honeycomb inorganic layers of [MnCr(oxalatoh]- separated by mesoscopic and stacks of the organic radicals, as shown in Figure 1. The average charge on the BEDT-TIF molecules is +0.34. The inorganic layer is insulating while the organic moiety is a conductor. The mag nanoscopic structures netic coupling is fairly strong within the inorganic layers and this, coupled to the fact that weak interactions are operative D. Gatteschi between the layers determined a transition to bulk ferromagnetic Department ofChemistry, University ofFlorence,INSTM behaviour below 5.5 K. Therefore below the critical temperature ...... the materialbehaves as a ferromagnetic conductor.Although sys tems like this are well known starting from iron itself, in the ll the magnets currently used are based on metallic or ionic molecular derivative the magneticelectrons are different from the Alattices. Since about two decades chemists have started a conducting electrons, thus offering the possibility of observing thorough project of using molecular chemistry techniques to new phenomena. develop new classes of magnets based on molecules rather than Another important feature of molecular magnets is that they on metals or oxides. The idea behind this is the challenge of cre are in general ins~ators,therefore they are much more transpar ating new classes ofmaterials from which new exciting properties ent to UV-visible light than classic magnets. Therefore it is may be expected. In a sense this research is the continuation of possible to use light to induce magnetic transitions. This thatwhich was successfully developed when itwas discovered that approach has been used by the groups of Verdaguer and organic compounds can behave as conductors and superconduc Hashimoto for instance.[5,6] Prussian blue derivatives are com tors like the classic inorganic materials. In a few years it has been plex cyanides of general formula ABC(CN). When B= Fe2+ and discovered that purely organic magnets are indeed possible, C= C03+the compound is diamagnetic because both ions are in although the critical temperatures are still very low. The most their low spin, non-magneticstate. By illuminating with red light promisingresults have been obtained using sulfur-nitrogenbased organic radicals which behave as weak ferromagnets below 35 K ... [1]. On the other hand, using a mixture of transition metal ions and organic radicals it has been possible to obtain a room tem perature ferrimagnet [2], and similar results have been obtained ... '. using derivatives ofthe oldPrussian Blue compounds [3). Beyond the chemical challenge of assembling new structures .... using moderately stable building blocks like organic radicals, the ...... factors suggesting that molecular magnetism can indeed provide ...... new interesting classes ofmaterials are: ...... • The possibility of fine tuning the properties of the materials ...... ,...... by using flexible molecular techniques .... • The possibility of building a la carte magnetic molecules of .... increasing size which behave as molecular nanomagnets (b) (ol • The possibility of obtaining multifunctional materials ... fl!L..1.;Thestructure of [BEDT-TIFMMnCr{oxalatoh]. An example of the last possibility has been recently reported by a) overviewofthe lattice;b)the organic layers;c)the inorganic Coronado et al. ['4) They used a hybrid approach, assembling layers. together inorganic building blocks like transition metal oxalato
214 europhysics news NOVEMBER/DECEMBER 2003