arXiv:1103.3345v1 [cond-mat.mes-hall] 17 Mar 2011 au faotafwhnrdBh antn rmun- from magnetons Bohr hundred few a expected about the than of larger value much are moments magnetic cle iae atcemgei oettrsott eabout be to out 2000 turns es- moment the magnetic instance, function For particle Langevin timated parameters. modified fit func- unphysical the in a to results field as magnetic nanoparticles of data NiO tion magnetization antiferromagnetic the bare func- fitting However, Langevin of modified 11–13]. a [6, by described tion be to expected is on conclusions 10]. [9, similar some independently reached work nanoparticles also our NiO After have workers expectation. other and contrary belief behavior superparamagnetic popular not temperature to low is the nanoparticles NiO that that, showed of behav- found also glass We spin we shows where ior. system this 8] temperatures, [7, some low reported reports nanoparticles at also We more NiO workers. many on other by work by followed mostly 6], been 5, has [2, different this of and nanoparticles sizes NiO magne- on reported 4–6]. measurements [4] [2, tization Milligan and system Richardson studied back, well compara- Long and a interesting is more NiO tively differ- Among nanoparticles, in antiferromagnetic 3]. [2, unobservable ent nanoparticles behavior ferrimagnetic show or unusual they ferro and because mainly surprising attention of some lot Re- a have quite particle. materials gained antiferromagnetic the of of nanoparticles surface cently the near directions in an different pointing of moments elementary because moment of cancellation magnetic imperfect net nonzero a then an have large can of sufficiently volume it size, made particle to is of surface particle reciprocal antiferromagnetic the the as If varies which [1]. ratio, superparamagnetism ferromagnetism as weak such N´eeland properties to ex- According magnetic should material hibit small. antiferromagnetic when an very shows of made them particles tiny is prop- of emergent collection size a the particle or to the particle due a physi- mainly that a For erties is late. attraction of interest the of cist lot a quite attracting been h antzto fatfroantcnanoparticles antiferromagnetic of magnetization The have size in nanometers few a of particles on Studies ffc fDsrbtdPril antcMmnso h Magne the on Moments Magnetic Particle Distributed of Effect µ B o . mprils[] uhvle o h parti- the for values Such [6]. particles nm 5.3 for ewrs aoatce ueprmgeim magnetizat superparamagnetism, 75.75.Fk nanoparticle, 75.75.-c, Keywords: 75.50.Tt, 75.50.Ee, this numbers: on PACS reported t observations disregarding anomalous that many find the We for moment. reason magnetic particle in tion antzto fnnprilso i r esrdadanal and measured are NiO of nanoparticles of Magnetization colo hsc n aeil cec,Taa University Thapar Science, Materials and Physics of School eateto hsc,Ida nttt fTcnlg Kanp Technology of Institute Indian Physics, of Department Nanoparticles .P Rajeev P. K. .D Tiwari D. S. ple antcfil [8]. field magnetic applied mbr i atce safnto fapidmag- be- applied ( well of N´eel temperature but function the temperatures, a high low sufficiently as at particles field NiO netic bare nm 5 ture inlmtosaeue oaayetemagnetization the analyze to above used tradi- data the are when methods nanoparticles find tional NiO to particle of attempt the moment for an numbers magnetic is unphysical getting work for This reason the again. nanoparti- once NiO moti- antiferromagnetic system observations the cles These revisit not magneti- to [6]. is us scaling cooled nanoparticles vated this NiO field show of to zero found magnetization and The cooled zation. field low tween yratn nauosslto iklntaeadsodium and nitrate nickel solution aqueous in reacting by the the for by account system can this workers. we earlier on that reported out observations turned anomalous by it surprised as pleasantly results were distribution We the a moment. account particle then into the in and taking moment analysis the magnetic repeated particle considering us- in without data distribution the function analyzed any Langevin We modified [14]. the K 523 ing about be to known ih( with eil sepce ob nyafnto fteapplied the of function a only be field to magnetic expected is terials [2]. no yet is model there this though for nanoparticles support mag- experimental NiO has particle for of model value moment larger large multisublattice netic this be A explain to to found proposed value. been been expected also esti- the has moment than method the the this However get by to mated particles moment. the of magnetic Here number is particle estimated temperature [2]. the low by used at divided magnetization been saturation also of moment has value magnetic nanoparticles particle Another NiO of of particles. estimation of the for surface method the on spins compensated o,susceptibility ion, eew rsn h antzto esrmnson measurements magnetization the present we Here i aoatce eepeae yaslglmethod sol-gel a by prepared were nanoparticles NiO h antzto fnnprilso antcma- magnetic of nanoparticles of magnetization The T bf B T bv h iucto eprtr ( temperature bifurcation the above ) o h rsn ytmi bu 9 n10G 100 in K 295 about is system present the for ytmi h literature. the in system i itiuini h nlssi the is analysis the in distribution his zdtkn noacutadistribu- a account into taking yzed T bf h au ftebfraintempera- bifurcation the of value The . ail 404 India 147004, Patiala , B n temperature and r281,India 208016, ur T N iaino NiO of tization ftesse hc is which system the of ) T n hudscale should and T bf be- ) 2 hydroxide at pH = 12 at room temperature as described superparamagnetic cannot be described by Equation (1). elsewhere [4–6]. We used nickel (II) nitrate hexahydrate Rather, they are well described by an altered form known (99.999%), sodium hydroxide pellets (99.99%), both from as the modified Langevin function [11–13], which con- Aldrich, and triple distilled water to make nickel hydrox- tains an extra linear term in B i.e. ide. The sample of nickel oxide nanoparticles was pre- pared by heating the nickel hydroxide at 523 K for 3 M = M0L(x)+ χaB, (2) hours in flowing helium gas (99.995%). The sample was characterized by x-ray diffraction and transmission elec- where χa is the susceptibility of randomly oriented anti- tron microscopy. The average crystallite size as well as ferromagnetic particle cores. We fitted the data shown in the particle size were found to be about 5 nm. The de- Figure 1 to the above equation and the resulting fits are tails of sample synthesis and structural characterization shown as solid lines in the figure. Values of fit parameters have been reported in one of our earlier works [8]. All M0, µp and χa obtained are presented in Table I. the magnetic measurements were done with a commercial SQUID magnetometer (Quantum Design, MPMS XL5). TABLE I: Values of fit parameters M0, µp and χa to Equa- tion (2) and the corresponding values of the goodness of fit 2 parameter R for the 5 nm NiO nanoparticles at different 3.0 temperatures.
310 K
320 K 2.5 −6 2 T M0 µp χa (10 R
330 K (K) (emu/g) (µB) emu/g Oe) 2.0 340 K 310 1.30 1967 34 0.9993 350 K 320 1.18 1841 32 0.9994 1.5
0.15 330 1.05 1825 31 0.9995
1.0 340 0.94 1734 29 0.9996
M (emu / g) (emu M 350 0.82 1621 27 0.9997 M (emu / g) / (emu M
0.5
0.00
0 500
B (G) From Figure 1 we see that the fits to Equation (2)
0.0 more or less pass through the measured data points which
0 10 20 30 40 50 means the fit quality is quite good. A statistical measure of the goodness of a fit is the coefficient of determination B (kG) R2. The closer R2 is to unity the better the fit. Values of the coefficient of determination R2 for the fits are shown FIG. 1: (Color online) Magnetization as a function of applied 2 magnetic field for 5 nm NiO particles at different tempera- in Table I along with the fit parameters. The R values tures. Solid lines show fits to Equation(2). The inset shows a are greater than 0.999 in all cases and this once again magnified view of the data along with the fits at lower mag- confirms the good quality of the fits. netic fields which makes it clear that the fit quality is poor From Table I we see that the particle magnetic mo- there. ment µp is about two thousand Bohr magnetons and it decreases with increasing temperature. Somewhat simi- lar results have been reported by others on NiO nanopar- We measured the magnetization M of the 5 nm NiO ticles [6] as well as on NiO nanorods [16]. However there particles as a function of external applied magnetic field is a glaring inconsistency between the numbers we get B at different temperatures T (Tbf
M = M0L(x), (1) data to Equation (2). In the next paragraph we shall give another argument against the numbers presented in 1 where L(x) = [coth (x) ( x )] is the Langevin function Table I. µpB − and x = . Here M0 is the saturation magnetization, N´eel had discussed various ways by which a magnetic kBT µp is the particle magnetic moment and kB is the Boltz- moment can appear on an NiO particles due to incom- mann constant. However, magnetization of small parti- plete compensation of atomic moments in different sub- cles of antiferromagnetic materials, even though they are lattices [1, 5]. According to N´eel the particle magnetic 3 moment of an NiO nanoparticle can be written as TABLE II: Values of fit parameters N, s, n and χa to Equa- 2 µp = p µA µB. (3) tion (5) and the values of R for the 5 nm NiO particles at different temperatures. The mean magnetic moment of a par- Here p is a number which depends on the size, crystal ticle estimated from the fit parameters are also shown. structure and form or shape of the particle and µA is 2+ the magnetic moment of Ni ion which is known to be 17 −6 2 T N (10 s n χa(10 µmean R 3.2 µB [17]. N´eel proposed a total of five different pos- (K) /g) (µB) emu/g Oe) (µB) sibilities for the origin of magnetic moment on an NiO particle of approximately cubic shape. Out of the five, 310 8.7 1.34 116.5 21.4 287 0.999997 two possibilities give the values of particle magnetic mo- 320 7.3 1.30 130.3 21.4 305 0.999998 ment to be zero corresponding to p = 0. The other three 330 6.5 1.27 136.8 20.9 308 0.999997 1 1 2 possible values of p are about n 3 , n 2 and n 3 where n is 340 6.9 1.28 118.9 19.8 271 0.999998 the number of Ni2+ ions in the particle. Now, there can 350 6.8 1.26 114.4 18.9 254 0.999998 be many other possible set of values of p corresponding to shapes such as spheres, tetrahedrons or, more real- main panel one can see that the situation is much better istically, irregular shapes. But it seems safe to assume 2 at higher fields. This low field misfit is an indication of that the upper limit of p will be about n 3 and the av- the role a distribution in the particle magnetic moment erage will be considerably less. Making use of the fact plays on the magnetization of a system. In this case the that the crystal structure of NiO is face centered cubic modified Langevin function has clearly underestimated with lattice constant 4.176 A˚, the value of n for a 5 nm the contribution of the larger moments. diameter spherical particle turns out to be about 3592. Following Silva et. al. we assume that in a sample of Using the information presented above along with Equa- nanoparticles the distribution in particle magnetic mo- tion (3) we get the particle magnetic moment to be about ment, µ, can be described by a log normal distribution 49 µB, 191 µB and 750 µB corresponding to p values of 1 1 2 function of the form [18] n 3 , n 2 and n 3 respectively. Thus according to N´eel’s 2 picture the maximum possible value of particle magnetic 1 [ln( µ )] moment for a 5 nm NiO particle would be about 750 µ f(µ)= exp n , (4) B µs√2π − 2s2 which is much smaller than the values shown in Table I. While using Equation (2) to fit the magnetization data where n and s are parameters that characterize the dis- we made a tacit assumption that all the particles have tribution. The mean particle magnetic moment µmean is the same magnetic moment. But, this is not true. A equal to n√es2 . Now, Equation (2) takes on the new sample of NiO nanoparticles has a distribution of particle form magnetic moments not only due to distribution in size ∞ and shape but also due to the way in which the imbalance M(B,T )= N µL(x)f(µ)dµ + χaB, (5) Z of spins arises in the particle as pointed out by N´eel. We 0 hazard the guess that disregarding the distribution in where N is the total number of particles contributing to particle moments is perhaps the reason for the unphysical magnetization and L(x) is the Langevin function with fit parameters obtained in Table I. Now we would like x = µB . We used this equation to fit the magnetization kBT to carry out this fitting taking into account a moment data of Figure 1 and the fit parameters thus obtained distribution. Making our path easier is a precedent in along with the values of coefficient of determination R2 this kind of analysis set by Silva et. al. in analyzing the are shown in Table II. magnetization of ferritin, a biological antiferromagnetic We see in Table II that the values of the coefficient of nanoparticle system [18]. determination R2 are surprisingly greater than 0.99999 In a sample that has a distribution in particle mag- in all the cases. These values of R2 are much better than netic moments the low field magnetization is governed the values of the same shown in Table I. This means that by the particles with larger magnetic moments. The con- the quality of the fits using Equation (5) is much better tribution of particles of lower magnetic moments to the than that using Equation (2). In Figure 2 we show the magnetization becomes important only at higher applied new fits which clearly look good. The inset clearly shows fields where the high field forces the moments to align that the fits are very good for lower values of applied with the field. The effect of a distribution in the parti- magnetic fields also where the effect of distribution in cle magnetic moment on the magnetization of the system the particle magnetic moment on the magnetization of will show up in the Langevin or modified Langevin fit. the system is important as already discussed. We can To see this dependence let us take a look at the com- now conclude that the fits of the magnetization data to parison of the measured magnetization to the modified Equation (5) are much better than that to Equation (2). Langevin fit at low fields in the inset of Figure 1. It is From Table II we find that the mean particle mag- clear that the fits are no good at low fields while from the netic moments, which are equal to n√es2 , turns out to 4
3.0 sideration of a distribution in particle magnetic moment
310 K could be the reason for the anomalously large values of
320 K 2.5 magnetic moment of NiO nanoparticles reported in the literature. 330 K
2.0
340 K
350 K
1.5
0.15 [1] L. N´eel, in Low Temperature Physics, edited by C. De-
1.0 witt, B. Dreyfus and P. D. de Gennes (Gordan and Beach,
M (emu / g) (emu M New York, 1962), p. 413. M (emu / g) / (emu M
0.5 [2] R. H. Kodama, S. A. Makhlouf, and A. E. Berkowitz, 0.00 0 500 Phys. Rev. Lett. 79, 1393 (1997). B (G) [3] S. Mørup and C. Frandsen, Phys. Rev. Lett. 92 217201 0.0 (2004) . 0 10 20 30 40 50 [4] J. T. Richardson and W. O. Milligan, Phys. Rev. 102,
B (kG) 1289 (1956). [5] J. T. Richardson, D. I. Yiagas, B. Turk, K. Foster, and 70 FIG. 2: (Color online) Magnetization as a function of applied M. V. Twigg, J. Appl. Phys. , 6977 (1991). magnetic field for 5 nm NiO particles at different tempera- [6] S. A. Makhlouf, F. T. Parker, F. E. Spada, and A. E. 81 tures. Solid lines show the fits of data to Equation (5). The Berkowitz, J. Appl. Phys. , 5561 (1997). 72 inset shows the fits at low magnetic fields. [7] S. D. Tiwari and K. P. Rajeev, Phys. Rev. B , 104433 (2005). [8] S. D. Tiwari and K. P. Rajeev, Thin Solid Films 505, 113 (2006). be about a few hundred Bohr magnetons, rather than a [9] E. Winkler, R. D. Zysler, M. V. Mansilla and D. Fioraniv, Phys. Rev. B 72, 132409 (2005). few thousand Bohr magnetons as found in the earlier fit, [10] Vijay Bisht and K P Rajeev, Journal of Physics Con- and are also well below the maximum possible value of densed Matter 22, 16003 (2010). about 750 µB according to the models proposed by N´eel [11] S. H. Kilcoyne and R. Cywinski, J. Magn. Magn. Mater. for the origin of magnetic moment in NiO particles. 140-144, 1466 (1995). In conclusion, we reported magnetization as a function [12] S. A. Makhlouf, F. T. Parker, and A. E. Berkowitz, Phys. 55 of applied magnetic field for 5 nm NiO particles at differ- Rev. B , R14717 (1997). ent temperatures above the bifurcation temperature T . [13] M. S. Seehra, V. S. Babu, A. Manivannan, and J. W. bf Lynn, Phys. Rev. B 61, 3513, (2000). Fitting the magnetization data to the modified Langevin [14] J. S. Smart and S. Greenwald, Phys. Rev. 82, 113 (1951). function without considering any distribution in the par- [15] I. S. Jacobs and C. P. Bean, in Magnetism,Vol.III edited ticle magnetic moments yields very large and unphysical by G. T. Rado and H. Suhl (Academic Press Inc., New values for the particle magnetic moment. However we York, 1963), p. 271. got reasonable values for the particle magnetic moment [16] M. S. Seehra, P. Dutta, H. Shim, and A. Manivannan, 129 if the modified Langevin function is used to fit the mag- Solid State Communications , 721 (2004). Introduction to Solid State Physics netization data taking into account a distribution in the [17] C. Kittel, (John Wi- ley and Sons, Inc., Seventh Edition, 1996), p. 426. particle magnetic moment. The distribution in the parti- [18] N. J. O. Silva, V. S. Amaral, and L. D. Carlos, Phys. cle magnetic moment arises from a distribution in particle Rev. B 71, 184408 (2005). size and form. This work clearly shows that the non con-