LJ14115E PRE July 17, 2015 14:14

Important Notice to Authors

Attached is a PDF proof of your forthcoming article in PRE. Your article has 7 pages and the Accession Code is LJ14115E.
Please note that as part of the production process, APS converts all articles, regardless of their original source, into standardized
XML that in turn is used to create the PDF and online versions of the article as well as to populate third-party systems such
as Portico, CrossRef, and Web of Science. We share our authors’ high expectations for the fidelity of the conversion into XML
and for the accuracy and appearance of the final, formatted PDF. This process works exceptionally well for the vast majority of
articles; however, please check carefully all key elements of your PDF proof, particularly any equations or tables.
Figures submitted electronically as separate PostScript files containing color appear in color in the online journal. However,
all figures will appear as grayscale images in the print journal unless the color figure charges have been paid in advance, in
accordance with our policy for color in print (http://journals.aps.org/authors/color-figures-print). For figures that will be color
online but grayscale in print, please ensure that the text and captions clearly describe the figures to readers who view the article
only in grayscale.
No further publication processing will occur until we receive your response to this proof.

Specific Questions and Comments to Address for This Paper

INSTRUCTIONS TO AUTHOR: Please check that color online figures are understandable also in the printed black-white
versions. Modify captions or figures, if appropriate.
1 Is ‘bath’ the correct word here and below?
2 Please check that the journal abbreviation is correct.
3 Please remove duplicate entry (i.e., Refs. [6] and [26]) from the reference list and renumber all subsequent reference citations
in both the text and in the reference list.
4 Please give city of publication; please check if page number is correct.
5 Please remove duplicate entry (i.e., Refs. [16] and [19]) from the reference list and renumber all subsequent reference
citations in both the text and in the reference list.
6 Please verify the page number in Ref. 17.
7 Please update Ref. [29] if possible.
8 Please provide page number.
9 Please remove duplicate entry (i.e., Refs. [34] and [36]) from the reference list and renumber all subsequent reference
citations in both the text and in the reference list.
10 Please give year and page number.
FQ: This funding provider could not be uniquely identified during our search of the FundRef registry. Please check information
and amend if incomplete or incorrect.
Q: This reference could not be uniquely identified due to incomplete information or improper format. Please check all
information and amend if applicable.
Titles in References: The editors now encourage insertion of article titles in references to journal articles and e-prints. This
format is optional, but if chosen, authors should provide titles for all eligible references. If article titles remain missing from
eligible references, the production team will remove the existing titles at final proof stage.
FundRef: Information about an article’s funding sources is now submitted to CrossRef to help you comply with current or future
funding agency mandates. Please ensure that your acknowledgments include all sources of funding for your article following any
requirements of your funding sources. CrossRef’s FundRef registry (http://www.crossref.org/fundref/) is the definitive registry
of funding agencies. Please carefully check the following funder information we have already extracted from your article and
ensure its accuracy and completeness:
CNPq (Brazil)
FAPESP (Brazil)

Other Items to Check

r Please note that the original manuscript has been converted to XML prior to the creation of the PDF proof, as described above.
Please carefully check all key elements of the paper, particularly the equations and tabular data.
r Please check PACS numbers. More information on PACS numbers is available online at http://journals.aps.org/PACS/.
r Title: Please check; be mindful that the title may have been changed during the peer review process.
r Author list: Please make sure all authors are presented, in the appropriate order, and that all names are spelled correctly.
r Please make sure you have inserted a byline footnote containing the email address for the corresponding author, if desired.
Please note that this is not inserted automatically by this journal.
r Affiliations: Please check to be sure the institution names are spelled correctly and attributed to the appropriate author(s).
LJ14115E PRE July 17, 2015 14:14

r Receipt date: Please confirm accuracy.

r Acknowledgments: Please be sure to appropriately acknowledge all funding sources.
r Hyphenation: Please note hyphens may have been inserted in word pairs that function as adjectives when they occur before a
noun, as in “x-ray diffraction,” “4-mm-long gas cell,” and “R-matrix theory.” However, hyphens are deleted from word pairs
when they are not used as adjectives before nouns, as in “emission by x rays,” “was 4 mm in length,” and “the R matrix is
tested.”
Note also that Physical Review follows U.S. English guidelines in that hyphens are not used after prefixes or before suffixes:
superresolution, quasiequilibrium, nanoprecipitates, resonancelike, clockwise.
r Please check that your figures are accurate and sized properly. Make sure all labeling is sufficiently legible. Figure quality in
this proof is representative of the quality to be used in the online journal. To achieve manageable file size for online delivery,
some compression and downsampling of figures may have occurred. Fine details may have become somewhat fuzzy, especially
in color figures. The print journal uses files of higher resolution and therefore details may be sharper in print. Figures to be
published in color online will appear in color on these proofs if viewed on a color monitor or printed on a color printer.
r Please check to ensure that titles are given as appropriate.
r Overall, please proofread the entire article very carefully.

Ways to Respond
r Web: If you accessed this proof online, follow the instructions on the web page to submit corrections.
r Email: Send corrections to preproofs@aptaracorp.com
Subject: LJ14115E proof corrections
r Fax: Return this proof with corrections to +1.703.791.1217. Write Attention: PRE Project Manager and the
Article ID, LJ14115E, on the proof copy unless it is already printed on your proof printout.
r Mail: Return this proof with corrections to Attention: PRE Project Manager, Physical Review E,
c/o Aptara, 3110 Fairview Park Drive, Suite #900, Falls Church, VA 22042-4534, USA.
LJ14115E PRE July 17, 2015 14:14

1 PHYSICAL REVIEW E 00, 002700 (2015)

2 Environment-induced anisotropy and sensitivity of the radical pair mechanism in the avian compass

3 Alejandro Carrillo,1 Marcio F. Cornelio,2 and Marcos C. de Oliveira1

1
4 Instituto de Fı́sica Gleb Wataghin, Universidade Estadual de Campinas, CEP 13083-859, Campinas, São Paulo, Brazil
2
5 Instituto de Fı́sica, Universidade Federal de Mato Grosso, 78068-900, Cuiabá MT, Brazil
6 (Received 25 November 2014; published xxxxxx)

7 Several experiments over the years have shown that the earth’s magnetic field is essential for orientation
8 in birds’ migration. The most promising explanation for this orientation is the photo-stimulated radical pair
9 (RP) mechanism. In order to define a reference frame for the orientation task radicals must have an intrinsic
10 anisotropy. We show that this kind of anisotropy and consequently the entanglement in the model are not necessary
11 for the proper functioning of the compass. Classically correlated initial conditions for the RP, subjected to a fast
12 decoherence process, are able to provide the anisotropy required. Even a dephasing environment can provide the
13 necessary frame for the compass to work and also implies fast decay of any quantum correlation in the system
14 without damaging the orientation ability. This fact significantly expands the range of applicability of the RP
15 mechanism providing more elements for experimental search.

16 DOI: 10.1103/PhysRevE.00.002700 PACS number(s): 87.50.−a, 03.67.−a, 03.65.Yz, 82.30.−b

17 I. INTRODUCTION weak for each individual pair, can effectively account for the 57

required anisotropy. In addition, although some discussion has 58

18 The ability of birds to use the earth’s magnetic field to
been given recently about the importance of entanglement 59
19 orientate themselves in the correct direction for migration [1,2]
in the magneto-perception process [23,24], this still remains 60
20 has originated several experimental works devoted to under-
obscure. In this sense, we also find that entanglement is neither 61
21 standing of the main features of the underlying mechanism
necessary in our model of isotropic molecules nor in the 62
22 [3–6]. One of the first proposals for modeling the appearance
anisotropic ones. Furthermore, we also verify the functioning 63
23 of the magnetic compass was that magneto-perception operates
of the compass in the presence of artificial radio frequency 64
24 by means of anisotropic magnetic field effects on the rate of
fields. We find that the isotropic model cannot work in the 65
25 production of yields of a photo-stimulated radical pair reaction
presence of such a field in agreement with experimental 66
26 [7–9]. Although other models have been proposed, such as a
findings [13,25,26]. However, the anisotropic model can work 67
27 magnetite-based magneto-perception [10,11], the proposal of
under some environmental conditions not unlikely in an open 68
28 a radical pair mechanism (RPM) has been recently reinforced
system, which notwithstanding disagree with experimental 69
29 [12], and strong experimental and theoretical evidences have
observation. This fact gives one more piece of evidence in 70
30 been presented in its favor [13–18]; suitable molecular candi-
favor of the present model of an isotropic molecule together 71
31 dates for mechanism are cryptochrome photoreceptors [19,20].
with the environment induced anisotropy. 72
32 The solid state RPM model can be summarized as follows [12]:
We begin by reviewing in Sec. II the basic model for 73
33 a molecular precursor reacts to form a pair of radicals due to
the avian compass based on a photo-stimulated radical pair 74
34 photochemically driven electron transfer. Taking into account
reaction [24]. In Sec. III we describe the models employed 75
35 that both radicals are created in a single event, it is natural to
for environmental noise. In Sec. IV we discuss our main 76
36 assume that the electron spins are initially entangled [in the
results regarding the environmental induction of anisotropy, 77
37 following we assume that the radical pair (RP) is created in a
and finally in Sec. V we conclude the paper. 78
38 singlet s state, although working with a triplet t state is also
39 possible]. This singlet state evolves under the influence of a II. MODEL 79
40 Hamiltonian containing an hyperfine interaction term between
41 the nuclei and their electrons and a Zeeman interaction term Let us consider the Hamiltonian of the RP, neglecting for 80

42 between the unpaired electrons. Due to the anisotropy of the the moment all other possible interactions, such as exchange 81

43 hyperfine tensor [21], the interconversion between entangled and dipolar, to be 82

44 singlet and triplet states depends on the direction of the applied 

45 magnetic field through the Zeeman term in the Hamiltonian. Ĥ k = Îik · Aik · Ŝk + ωe B · Ŝk , (1)
46 It is necessary also to assume that the radicals are almost i

47 immobile, without significant diffusive motion, in order to where the first term is the hyperfine contribution and the 83

48 avoid the anisotropy present in the system to be averaged second one is the Zeeman contribution, with i labeling the ith 84

49 away. The RP yields depend on the relative alignment of the nucleus in the kth radical. Earth’s magnetic field is given by 85

50 magnetic field in relation to the sample [21,22], so that it can B = B0 (sin θ cos φ, sin θ sin φ, cos θ ), and A is the hyperfine 86

51 work as a compass. tensor. Î and Ŝ are the spin operators for the nucleus and 87

52 In this work, we show that the anisotropy in the molecule the electron, respectively, ωe = ge μB /, ge is the electron g 88

53 can be replaced by an anisotropic environment. This fact factor, B0 = 47μT , and μB is the Bohr magneton. We assume 89

54 allows for a free isotropic molecule in a diffusive environment that the electron ge tensor is isotropic and close to that of a 90

55 to work as a compass. The model is derived through the free electron, and that the hyperfine tensor A is, for simplicity, 91

56 inclusion of dipole-dipole interaction, which, although being diagonal. Additionally, it can be either isotropic or anisotropic 92

1539-3755/2015/00(0)/002700(7) 002700-1 ©2015 American Physical Society

ACC. CODE LJ14115E AUTHOR Carrillo
LJ14115E PRE July 17, 2015 14:14

CARRILLO, CORNELIO, AND DE OLIVEIRA PHYSICAL REVIEW E 00, 002700 (2015)

93 according to the model. The direction of the applied magnetic

94 field in relation to the RP fixed axis system is defined in
95 terms of the polar angles (θ,φ); without losing generality we
96 are going to assume φ = 0 in order to simplify procedures.
97 The most appropriate units to work with are those of nuclear
98 magnetic resonance (angular frequency units).
99 In order to fully determine the dynamics of our system, we
100 are going to use a master equation approach [24]. To account
101 for the singlet or the triplet yields, we add their formation rate
102 to the master equation as a dissipative process modulated by
103 a factor k in angular frequency units. The simplest model we
104 can consider is one nucleus coupled with its unpaired electron
105 plus a free electron, with Hamiltonian Ĥ = Î1 · A · Ŝ1 + ωe B ·
106 (Ŝ1 + Ŝ2 ). The evolution of the system can be written as
d k
ρ(t) = i[ρ(t),Ĥ ]/ + [2Pn ρ(t)Pn −Pn ρ(t)−ρ(t)Pn ] FIG. 1. (Color online) Angle sensitivity of the radical pair with
dt n
2 an isotropic hyperfine tensor A = aI in the presence of the amplitude
γ damping noise described by the Lindblad operator (4) for different
+ [2B̂ρ(t)B̂ † − B̂ † B̂ρ(t) − ρ(t)B̂ † B̂]. (2) increasing values of the coupling constant γ from top to bottom. The
2 first curve at the top is a reference curve for γ = 0, k = 0.01 MHz,
107 The operators Pn are projectors over the singlet or triplet shells for the anysotropic hyperfine tensor and initial singlet state.
†
108 as described in Ref. [24]; as with any projector, Pn = Pn
†
109 and Pn Pn = Pn = Pn . The sum is taken over all possible
2
As a result of its action, the populations of the density operator 141
110 electronic and nuclear spin states, i.e., |l,m ⊗ |α, where l will have an additional term of the form ρ̇↑↑ = −γρ↑↑ and 142
111 is the electronic singlet or triplet state and m its projection, ρ̇↓↓ = γρ↑↑ . In Fig. 1 we show the behavior of the singlet 143
112 and α represents the nuclear spin state, either ↑ or ↓. The yield when the system is in the presence of the noise process 144
113 process mediated by k can be thought of as a measurement described above. It is evident that, as in the case of the 145
114 on the RP state, giving information about the amount of dephasing noise, the damping process described here suffices 146
115 the singlet or triplet yield. However, the density operator to generate an angular sensitivity in the radical pair, although 147
116 formalism enables us to take into account the environment in this should not be a surprise: the action of the operator on the 148
117 which it lives. A wide range of noise processes tested to model states (3) of the system makes clear that the interconversion 149
118 environmental interactions resulted in decoherence times of †
process is anisotropic, given that σ2 |T ,1 = 0. 150
119 hundreds of microseconds [24,27], although there has been
120 some controversy [28] over the numerical results found in Ref.
121 [27]. The noise affecting the system can be either amplitude B. Pure dephasing noise 151

122 damping or dephasing noise [29]. Our purpose is in the study of pure dephasing channels as 152

a cause of angular sensitivity. This kind of channel would not 153

123 III. ENVIRONMENTAL NOISE affect the interconversion rate between singlet and triplet yields 154

124 A. Amplitude damping noise and in consequence is not imposing explicitly nonuniform 155

decay rates. The operators originating pure dephasing noise 156

125 The main characteristic of amplitude damping noise is commute with the Hamiltonian, giving rise to effects only on 157
126 that the operators that originate it do not commute with the the coherences of the density matrix. This channel could be 158
127 Hamiltonian of the system; this has several implications on the originated due to dipole interactions, magnetic fluctuations 159
128 way the channel affects its dynamics, the most obvious being in the biological environment [24], or even energy transfer 160
129 that the energy is not preserved during its evolution. Moreover, through spin hopping [30] among others. Dipolar and exchange 161
130 amplitude damping introduces a somewhat obvious anisotropy interactions between the electron pair in the radical have little 162
131 in the system by promoting the interconversion between singlet influence on the angular sensitivity [31]. However, a net effect 163
132 and triplet states and imposing nonuniform decay rates on the of the influence of the molecules surrounding the radical pair 164
133 state coherences (nondiagonal terms). can be a source of noise. From the dipolar interaction alone, 165
†
134 As an example we use the operator σ2 , which describes three dephasing noise terms can emerge, although here we are 166
135 absorption of energy for the unpaired electron, raising its spin. going to focus our analysis on only one of them. The dipolar 167
136 Acting on the singlet state S| it generates a triplet state with interaction can be written as 168
137 angular projection m = 1, i.e., μB
†
√ Hdip = [3(m1 · r)(m2 · r) − r 2 m1 · m2 ]
σ2 |S = 2|T ,1. (3) 4π r 5
μB g1 g2 
138 This operator does not conserve the energy of the system. = [3(σ1 · r̂)(σ2 · r̂ − σ1 · σ2 ], (5)
139 When used as a perturbative noise, it produces a Lindblad-like 8π r 3
140 term in the density matrix of the form where gi are the electronic factors of each electron, σ = 169

γ † † † êx σ̂x + êy σ̂y + êz σ̂z , and r̂ is the unitary normal vector 170
Lρ = (2σ2 ρσ2 − σ2 σ2 ρ − ρσ2 σ2 ). (4) in spherical coordinates. If we consider N environmental 171
2

002700-2
LJ14115E PRE July 17, 2015 14:14

ENVIRONMENT-INDUCED ANISOTROPY AND . . . PHYSICAL REVIEW E 00, 002700 (2015)

172 electronic spins interacting through Eq. (5) with the unpaired an important experimental observation: It was observed, in a 213

173 electron in the radical pair, we will have a net dipolar set of experiments with European robins, that an oscillating rf 214

174 contribution: magnetic field B rf , perpendicular to the earth’s, disrupts avian 215
 μB g1 g2 compass functioning [13,25], leaving the birds without a sense 216
Hdip = [3(σ1 · r̂)(σi · r̂) − σ1 · σi ]. (6) of direction. So when the magnitude of k exceeds a threshold,
i
8π ri3 217

the influence of the measurement process should overwhelm 218

175 Before proceeding any further, we must note that due to the the action of the rf field [24], and as a result, the sensitivity 219

176 intermolecular distances, the magnitude of the rate governing to the variations of earth’s magnetic field would be present 220

177 the interaction is going to be small compared with the despite the disrupting effect. This fact can be contrasted to 221

178 production rate of the radical products k described and the experimental results to pick a suitable upper bound to k 222

179 proportional to 1/r 3 . We are interested in showing how from as k = 0.01 MHz, which will be used unless stated otherwise. 223

180 this interaction a dephasing noise process may be obtained. The processes mediated by γ show the same behavior; i.e., 224

181 To do that let us consider the Lindblad-like part of the master their influence will allow the compass to work in spite of the 225

182 equation in the interaction picture: presence of the rf field if their magnitudes are high enough. 226

 ∞ As the experimental observation tells us that there will not 227

d
ρ(t) = − trB {[Hdip (t),[Hdip (t − s),ρ(t)]]} ds, (7) be a compass if the birds are subject to this rf field, we use 228
dt 0 this to set upper bounds for the noise amplitudes in the same 229

1 183 where trB indicates the tracing of all the bath degrees of order of magnitude of k. This fact also implies a lower limit 230

184 freedom. If the correlations of the system go to zero much for the decoherence time of the system, because an upper 231

185 faster than the natural time of the system τS , we can write bound to the noise amplitudes implies that this time cannot 232

186 ρ(t) = ρS (t) ⊗ ρB , where ρB and ρS (t) are the density matrix be arbitrarily small, and we are going to have at least tens of 233

187 of the bath and the system, respectively. For simplicity in the microseconds until the loss of all coherences in the system. 234

188 notation let us write Âi = σi · r̂ and  = σ1 · r̂. One of the We use a magnitude Brf = 150 nT as proposed in Ref. [24]. 235

189 terms emerging from the double commutator is

[Hdip ,[Hdip ,ρS (t) ⊗ ρB ]] A. Hyperfine tensor 236

As stated by Schulten [7], for the compass to work some 237

= Â2 ρS (t)Âi Âj ρB + ρS (t)Â2 ρB Âi Âj
kind of anisotropy must be present in the system. If, as usual, 238

−2ÂρS (t)ÂÂi ρB Âj . (8) we choose the source of anisotropy in the hyperfine tensor, 239

there is a sensitivity in the RPM to all initial states. To show 240

190 Due to the cyclic permutation of the trace, trB {Âi Âj ρB } = that, we start by observing that for some values of the rates, for 241

191 trB {ρB Âi Âj } = trB {Âi ρB Âj } = trB {ρB } = 1. The remain- example, γ = 2k, the influence of the rf field over the compass 242

is strengthened; and if γ is of the order of 20k the compass- 243

192 ing term of the double commutator after taking the trace is
disrupting effect by the rf field can still be observed. It is 244
193 then
interesting to note that even if in this situation the environment 245

Â2 ρS (t) + ρS (t)Â2 − 2ÂρS (t)Â contributes to the insensibility of the compass by means of 246

the rf field, in other circumstances it can increase it [33].

= −2[2(1 − 2Â)ρS (t)(1 − 2Â)
247

One of the consequences of using these processes is that the 248

− (1 − 2Â)2 ρS (t) − ρS (t)(1 − 2Â)2 ]. (9) decoherence times are short. To give a measure of those times 249

for testing the model with anisotropy in the hyperfine tensor 250
194 This means that the dipolar interaction between the subsystems
we use quantum discord (QD) [34], which is able to signal the 251
195 of the radical pair and the environmental degrees of freedom
presence of any kind of quantum correlations, and concurrence 252
196 is a generator of dephasing channels. The procedures are
[35], which measures quantum correlation as entanglement 253
197 similar for the remaining five operators that commute with the
only. The former is defined as 254
198 radical pair Hamiltonian, giving raise to five more dephasing
199 channels. Finally then we introduce the dephasing channel in δ←− = IAB − J←− ,
AB AB
200 the Lindblad form in the last term of Eq. (2), where the operator
B̂ commutes with the where IAB is the mutual information between A and B and
√ Hamiltonian (1); from (9) we have that
201 255

202 B̂ = (1̂ − 2σ1 · n̂)/ 2, where σi = êx σ̂x + êy σ̂y + êz σ̂z and n̂ is defined as IAB = SA + SB − SAB , Sx is the von Neumann 256

entropy of system x, J← − = max{B } [S(ρA ) − x

203 is the unitary normal vector in spherical coordinates. The rate AB x x px S(ρA )] 257

204 of the process is given by γ , which represents a measure of is the classical correlation between the subsystems, and 258

205 the strength of the interaction between the system and the px = T rA {Bx ρAB Bx } and ρAx = T rB {Bx ρAB Bx }/px ; the 259

206 environment. All results presented in this paper were obtained maximum is taken over the positive measurements {Bx } made 260

207 by direct numerical integration of Eq. (2). over the system B. The evolution of both QD and concurrence 261

can be seen in Fig. 2. The reason for employing the QD 262

is that it signals the presence of quantum correlations even 263

208 IV. ANISOTROPIES AND ANGULAR SENSITIVITY
when there is no entanglement [36]. Therefore if there is 264

209 The yields’ measurement process, regardless of its magni- any quantum correlation preserving over long times, which 265

210 tude, is not going to affect the performance of the compass. could then be relevant, it would be signaled by QD. However, 266

211 Therefore k could be arbitrarily high and the compass would what both measures show is quite the opposite: i.e., they show 267

212 still work. However, an upper bound to k can be determined by that in general there is a rapid loss of any form of coherence 268

002700-3
LJ14115E PRE July 17, 2015 14:14

CARRILLO, CORNELIO, AND DE OLIVEIRA PHYSICAL REVIEW E 00, 002700 (2015)

FIG. 3. (Color online) Entanglement density for different values

of the rates γ and k, varying the angles θ of inclination of the earth’s
magnetic field (47μT ). The entanglement was measured by the use
of concurrence. In the left panel, the rate values are k = 0.05 MHz,
γ = 6k, and in the right panel the values are k = 0.1 MHz, γ =
4k. It is interesting to note that with smaller angles there are more
entanglement sudden deaths and revivals, and that for angles around
π/4 the decoherence time is shorter. The nuclear √ initial condition
used in both cases was |ψNuclear  = (|0 + |1)/ 2.

π/4, the decoherence time is shorter than for angles near 0 281

and π/2. 282

B. The relevance of entanglement for angular sensitivity 283

Following the discussion in Ref. [23], the initial state in 284

the RPM is not a perfect singlet (or triplet) state, so a natural 285

way to test if the quantum correlations play a fundamental 286

role in the working of the compass is to choose nonentangled 287

initial conditions. In Ref. [23] several random initial conditions 288

FIG. 2. (Color online) Concurrence (top oscillating plot) and were used, and the conclusion was that there is not a crucial 289

quantum discord (lower oscillating plot) evolution for different values dependence on the entanglement of the initial state and that 290

of the rates and magnetic field inclination angles. (a) Measurement even RPs with initial separable states with only classical 291

rate k = 0.1 MHz, and dissipation rates γ = 2k, with an angle θ = 0; correlations can produce an inclination sensitivity in the 292

in the inset an angle of θ = π/2 was used. (b) Measurement rate singlet yield. We perturbed 1000 singlet initial conditions, 293

k = 0.01 MHz, and dissipation rates γ = 10k with and angle θ = 0; some of them entangled, evolving under the density operator 294

in the inset an angle of θ = π/2 was used. Is worth noting that the described in Eq. (2) with the hyperfine tensor anisotropic, 295

angle affects the amount of quantum correlations measured by the and with the decoherence processes turned off, i.e., γ = 0. To 296

QD and the concurrence; for smaller angles the concurrence values ensure real quantum states we distributed half of the sample 297

are not only higher, but predict the same time of decoherence as the using Bures’s measure [37] defined as 298
QD; however, for higher angles, the entanglement is less important
than the classical correlations, making the QD to be higher and to ρS1 = (I + U † )ρ˜S ρ˜S † (I + U † ), (10)
predict longer correlation times. The decoherence was computed by
means of the fidelity [32] between the evolving state and the initial ρS1
ρS = . (11)
singlet state. tr(ρS1 )
Here U is a random unitary matrix distributed according to 299

the Haar measure [38], I is the identity matrix, and ρ˜S is the 300

269 compared to the results in Ref. [24]. The fast loss of coherence perturbed singlet initial matrix. We proceeded in a similar way 301

270 is not a surprise: having an open environment like the one with the other half of the samples, distributing the perturbed 302

271 we can expect in the eye of the bird should lead naturally state using the Hilbert-Schimdt measure: 303

272 to a fast loss of quantum correlations. This result should not ρ˜S ρ˜S †
273 compromise the correct behavior of the compass; one of the ρS = . (12)
274 reasons, concerning the role of entanglement in the system, tr(ρ˜S ρ˜S † )
275 will be presented below. It is also interesting to note that the With both renormalizations we ensure that our perturbed 304

276 change in the inclination of the magnetic field also affects density matrix describes a true quantum state. The results of 305

277 the loss of coherence. This can be seen in Fig. 3, where the the simulations can be seen in Fig. 4. As a first observation, 306

278 change of concurrence with time and inclination angle, for γ we can see that the mechanism is robust. Even if some of the 307

279 fixed, is shown. For small angles θ there are going to be more yields had no resemblance with the unperturbed system (black 308

280 collapses and revivals of entanglement, and for angles near line in the figure), the mean gave the same amount of yield 309

002700-4
LJ14115E PRE July 17, 2015 14:14

ENVIRONMENT-INDUCED ANISOTROPY AND . . . PHYSICAL REVIEW E 00, 002700 (2015)

FIG. 4. (Color online) Perturbations in a singlet initial condition
with an anisotropic hyperfine tensor without environmental noise. The
red (lower continuous) curve represents the mean of 1000 perturbed
singlet initial conditions. The mechanism appears quite robust, as the
differences in the yield production remain unchanged when compared
with the unperturbed singlet initial condition, given by the black (top
continuous) curve in the figure). The dashed lines are four random
yields taken from the sample.
FIG. 5. (Color online) Angle sensitivity for non singlet or triplet
initial conditions, specifically ρ0 = |αααα| with isotropic hyperfine
tensors. Black (top curve): nuclear spin initially set to | ↑. Blue
√(lower
curve): nuclear spin initially set to a mixed state (| ↑ + | ↓)/ 2. Red
(middle curve): nuclear spin initially set to | ↓.
C. Environmental induced anisotropy 342
If the Hamiltonian is isotropic, with singlet or triplet initial 343

conditions, the expected behavior is an absence of change 344

310 for each inclination of the field. This implies that the amount in the production rates of the yields, and there is going to 345

311 of entanglement is not a necessary condition to obtain angular be sensitivity only if there is a decoherence process present. 346

312 sensitivity, given that none of the random initial conditions This can be understood as another class of anisotropy induced 347

313 were in its maximum of entanglement and some of them by the environment, which chooses a preferred direction 348

314 were not entangled at all. The results show that the amount for the system through the dissipation. This can open the 349

315 of entanglement is not a fundamental factor for the sensitivity search for a suitable chemical species responsible for the 350

316 in the change of the angle of the applied magnetic field. A RP creation, because the molecule does not need to present 351

317 more specific example is presented in Fig. 6 (red curve), with anisotropic hyperfine or Zeeman interactions, and the degree of 352

318 the initial state entanglement is not going to be crucial. The only requirement 353

for the correct functionality of the compass is then the 354

decoherence, which is necessary in such an open system. Both 355

ρ0 = 1
2
(|αβαβ| + |βαβα|); (13) cases, anisotropy induced by the environment and by the initial 356

conditions, can be seen in Fig. 6. 357

319 this state gives an appreciable change in the yields (and

320 therefore allow sensitivity) for different angles θ . This result
V. DISCUSSION 358
321 implies that with an unentangled initial condition there is still a
322 fully functional compass; in other words, the initial condition Our findings go beyond establishing the role of entangle- 359

323 can be a source of anisotropy: if it is not an entangled one, ment in the compass. We found that any kind of correlation, 360

324 the system is still sensitive to changes in the inclination of the quantum or classical, is sufficient for the RPM if the molecule 361

325 field. This observation about nonentangled initial conditions is anisotropic, and that an isotropic molecule can have 362

326 was first made by Hogben et al. [39]. In other words, in sensitivity for variations in the field even if it has classically 363

327 the absence of an explicit anisotropy in the hyperfine tensor correlated, separable initial conditions. One example of such 364

328 or in the g electronic factor, the sensitivity depends on the a state is given by Eq. (13). Furthermore, even if they lack any 365

329 inhomogeneity of the populations in the density matrix. As kind of correlation but are anisotropic, giving an unbalanced 366

330 an example consider the state described in the blue curve in weight to some populations over others, like the states used 367

331 Fig. 5, i.e., a state with only diagonal terms in its density in Fig. 5, we can also expect a working compass. A careful 368

332 matrix, evolving under an isotropic hyperfine coupling: Its analysis has to be made in order to transparently identify 369

333 yield distribution shows a high variation with the angle. the characteristics of the initial conditions that lead to a 370

334 Moreover, an initial state like ρ0 = |αααα|, which produces dependence of the singlet yield with the inclination angle of 371

335 a lower but still appreciable angle sensitivity, generates a the field. 372

336 different distribution for the singlet yield depending on the To avoid explicit environmental anisotropies, we used in 373

337 nuclear spin state, exemplifying the reach and relevance of the calculations only pure dephasing noises, avoiding the 374

338 this source of anisotropy. Surprisingly, after several numerical amplitude damping and thus changes in the energy levels of 375

339 tests, only maximally entangled states (Bell states) will not the system forced externally. 376

340 generate any kind of sensitivity when the hyperfine tensor is As discussed in Refs. [21,22], the anisotropy of a molecule 377

341 isotropic. can be averaged away if it has significant diffusive motion, 378

002700-5
LJ14115E PRE July 17, 2015 14:14

CARRILLO, CORNELIO, AND DE OLIVEIRA PHYSICAL REVIEW E 00, 002700 (2015)

or even rotations. Our findings of unexpected sources of 379

anisotropy relax this immobility requirement. This is highly 380

positive for the model given the wild conditions involving 381

actual biophotochemical processes. Along with this, isotropic 382

molecules seem to be more robust to environmental effects; 383

in the presence of an rf field the control experimental data 384

[13,25] show that the compass will no longer work. However, 385

if there is a strong enough dissipation the compass will 386

work normally [24]. An isotropic molecule can handle higher 387

noise magnitudes than an anisotropic one without jeopardizing 388

control experimental data. 389

ACKNOWLEDGMENTS 390

FIG. 6. (Color online) Angle sensitivity of isotropic Hamiltoni-

ans with k = 0.01 MHz. Black (middle curve): Evolution of the We thank M. Plenio and P. Hore for important comments 391 FQ
singlet yield for different angles with an entangled initial condition. and helpful discussion on our model. This work is supported 392

Blue (lower curve): Evolution of the singlet yield for different by the Brazilian funding agencies CNPq and FAPESP through 393

angles with an entangled initial condition and γ = 2k. Red (top the Instituto Nacional de Ciência e Tecnologia–Informação 394

curve): Evolution of the singlet yield for different angles with Quântica (INCT-IQ). M.C.O. wishes to thank the hospi- 395

the unentangled initial condition for both electronic and nuclear tality of the Institute for Quantum Information Science at 396

spins ρ0 = 14 (|αβαβ| + |βαβα|) ⊗ (| ↑↑ | + | ↓↓ |) and with the University of Calgary, where part of this work was 397

γ = 0. developed. 398

[1] W. Wiltschko and F. W. Merkel, Verh. dt. zool. Ges. 59, 362
2 (1966).
[2] W. Wiltschko and R. Wiltschko, Science 176, 62 (1972).
[3] W. Wiltschko, J. Traudt, O. Gunturkun, H. Prior, and R.
Wiltschko, Nature (London) 419, 467 (2002).
[4] W. Wiltschko and R. Wiltschko, J. Comp. Physiol. A 191, 675
(2005).
[5] W. Wiltschko and R. Wiltschko, Naturwissenschaften 89, 445
(2002).
[6] R. Wiltschko, K. Stapput, P. Thalau, and W. Wiltschko, J. R.
3 Soc. Interf. 7, S163 (2010).
[7] K. Schulten, C. E. Swenberg, and A. Weller, Z. Phys. Chem.
Neue Fol. 111, 1 (1978).
[8] H. J. Werner, Z. Schulten, and K. Schulten, J. Chem. Phys. 67,
646 (1977).
[9] K. Schulten and A. Windemuth, in Biophysical Effects of Steady
Magnetic Fields, Vol. 25, edited by J. K. G. Maret and N. Boccara
4 (Springer-Verlag, 1986), p. 99–106.
[10] J. L. Kirschvink, M. M. Walker, and C. E. Diebel, Curr. Opin.
Neurobiol. 11, 462 (2001).
[11] S. Johnsen and K. J. Lohmann, Nat. Rev. Neurosci. 6, 703
(2005).
[12] T. Ritz, S. Adem, and K. Schulten, Biophys. J. 78, 707 (2000).
[13] P. Thalau, T. Ritz, K. Stapput, R. Wiltschko, and W. Wiltschko,
Naturwissenschaften 92, 86 (2005).
[14] F. Cintolesi, T. Ritz, C. W. M. Kay, C. R. Timmel, and P. J. Hore,
Chem. Phys. 294, 385 (2003).
[15] C. R. Timmel, F. Cintolesi, B. Brocklehurst, and P. J. Hore,
Chem. Phys. Lett. 334, 387 (2001).
[16] T. Ritz, R. Wiltschko, P. Hore, C. T. Rodgers, K. Stapput, T.
Thalau, C. R. Timmel, and W. Wiltschko, Biophys. J. 96, 3451
5 (2009).
[17] L. E. Foley, R. J. Gegear, and S. M. Reppert, Nat. Commun. 2,
356 (2011). 6
[18] K. Maeda, K. B. Henbest, F. Cintolesi, I. Kuprov, C. T. Rodgers,
P. A. Liddell, D. Gust, C. R. Timmel, and P. J. Hore, Nature
(London) 453, 387 (2008).
[19] T. Ritz, R. Wiltschko, P. J. Hore, C. T. Rodgers, K. Stapput, P.
Thalau, C. R. Timmel, and W. Wiltschko, Biophys. J. 96, 3451
(2009).
[20] T. Ritz, Procedia Chem. 3, 262 (2011).
[21] S. G. Boxer, C. E. D. Chidsey, and M. G. Roelofs, J. Am. Chem.
Soc. 104, 2674 (1982).
[22] S. G. Boxer, C. E. D. Chidsey, and M. G. Roelofs, Pub. Natl.
Acad. Sci. USA 79, 4632 (1982).
[23] J. Cai, G. G. Guerreschi, and H. J. Briegel, Phys. Rev. Lett. 104,
220502 (2010).
[24] E. M. Gauger, E. Rieper, J. J. L. Morton, S. C. Benjamin, and
V. Vedral, Phys. Rev. Lett. 106, 040503 (2011).
[25] T. Ritz, P. Thalau, J. B. Phillips, R. Wiltschko, and W. Wiltschko,
Nature (London) 429, 177 (2004).
[26] R. Wiltschko, K. Stapput, P. Thalau, and W. Wiltschko, J. R.
Soc. Interface 7(Suppl. 2), S163 (2010).
[27] J. N. Bandyopadhyay, T. Paterek, and D. Kaszlikowski, Phys.
Rev. Lett. 109, 110502 (2012).
[28] E. M. Gauger and S. C. Benjamin, Phys. Rev. Lett. 110, 178901
(2013).
[29] A. P. F. Marquardt, arXiv:0809.4403. 7
[30] M. B. Plenio and S. F. Huelga, New J. Phys. 10 (2008). 8
[31] O. Efimova and P. J. Hore, Biophys. J. 94, 1565
(2008).
[32] M. A. Nielsen and I. L. Chuang, Quantum Computation and
Quantum Information, 10th ed. (Cambridge University Press,
New York, 2011).

002700-6
LJ14115E PRE July 17, 2015 14:14

ENVIRONMENT-INDUCED ANISOTROPY AND . . . PHYSICAL REVIEW E 00, 002700 (2015)

[33] J. Cai, F. Caruso, and M. B. Plenio, Phys. Rev. A 85, 040304 [37] D. J. C. Bures, Trans. Am. Math. Soc. 135. 10
(2012). [38] M. Pozniak, K. Zyczkowski, and M. Kus, J. Phys. A 31, 1059
9 [34] H. Ollivier and W. H. Zurek, Phys. Rev. Lett. 88, 017901 (2001). (1998).
[35] W. K. Wootters, Phys. Rev. Lett. 80, 2245 (1998). [39] H. J. Hogben, T. Biskup, and P. J. Hore, Phys. Rev. Lett. 109,
[36] H. Ollivier and W. H. Zurek, Phys. Rev. Lett. 88, 017901 (2001). 220501 (2012).

002700-7