Generation and search of axion-like light particle using intense crystalline

field

Wei Liao
Center for High Energy Physics, Peking University,
Beijing 100871, P. R. China

Institute of Modern Physics, East China University of Science and Technology,

130 Meilong Road, Shanghai 200237, P.R. China

Intense electric field ∼ 1010 − 1011 V/cm in crystal has been known for a long time and has
arXiv:1011.6460v2 [hep-ph] 19 Jul 2011

wide applications. We study the conversion of axion-like light particle and photon in the intense
electric field in crystal. We find that the conversion of axion-like particle and photon happens
for energy larger than keV range. We propose search of axion-like light particle using the intense
crystalline field. We discuss the solar axion search experiment and a variety of shining-through-wall
experiment using crystalline field. Due to the intense crystalline field which corresponds to magnetic
field ∼ 104 − 105 Tesla these experiments are very interesting. In particular these experiments can
probe the mass range of axion-like particle from eV to keV.

PACS numbers: 14.80.Va, 29.90.+r

The search of axion-like particle(ALP) has at- fect of photon-ALP conversion, to happen in the
tracted a lot of interests after the axion was in- crystalline field.
vented in a mechanism solving the strong CP prob- In this Letter we study the ALP-photon con-
lem in the Standard Model [1–4]. The search of version in intense crystalline field and propose to
ALP is based on a mechanism that ALP and pho- do laboratory search of ALP using the crystalline
ton can transform to each other in external electro- field. We first briefly review the intense electric
magnetic field, e.g. in external magnetic field [5]. field in crystal. We study the ALP-photon conver-
Many varieties of ALP search experiments based sion in crystalline field. We propose to do ALP
on this mechanism have been proposed and been search experiment using intense crystalline field.
done [6, 7]. So far, no evidence of ALP has been The structure of a single crystal is described by
found. One of the crucial difficulties in laboratory the Bravais lattice
search of ALP is that the external magnetic field
is maximally around a few Tesla in laboratory. ~ = l1~a1 + l2~a2 + l3~a3 ,
R (1)
It is well known for a long time that extremely where ~ai is a primitive vector and li an integer.
strong electric field exists in crystal. For a charged The crystal is periodic, e.g. the atomic density
particle incident on a crystal with a direction al- ~ = ρ(~x). The same is true for the electric
ρ(~x + R)
most parallel to a crystallographic plane the strong
field and the potential in crystal. Using the recip-
electric field of nuclei add constructively so that
rocal lattice vector ~q the potential and the electric
a macroscopic and continuous electric field, the
field can be Fourier transformed and be expressed
plane continuum electric field, is obtained. As a
as
consequence, positrons can be trapped in a poten-
tial well of depth ∼ 102 V between two crystal- U (~x) =
X
Uq~ e−i~q·~x , (2)
lographic planes with lattice distance ∼ 0.1 nm
q
~
and are channeled between two parallel crystallo-
~ x) = ~ q~ e−i~q·~x ,
X
graphic planes. The plane continuum electric field E(~ E (3)
is estimated ∼ 1010 −1011 V/cm which corresponds q
~
to magnetic field ∼ 104 − 105 T.
where E ~ q~ = −i~q Uq~. ~q = 2π P3 ni~bi where ni
Charged particles when channeling in crystal os- i=1
cillate in the transverse direction and can pro- is an integer, ~bi = 12 ǫijk (~aj × ~ak )/(~a1 · (~a2 × ~a3 ))
duce coherent radiation which has been intensively and ~ai · ~bj = δij . In Fig. 1 we give a plot for
explored theoretically and experimentally [8, 9]. the cubic lattice. A crystallographic plane labeled
A high energy photon ( > ∼ GeV) incident on a (010) is shown in the plot. The plane is parallel
crystal is also expected to be affected by the to the h100i and h001i axis and perpendicular to
strong crystalline field and to produce electron- h010i axis.
positron pairs [9] since the radiation and the pair- In a good approximation the plane continuum
production processes are related by the crossing electric field can be considered constant in the lon-
symmetry. It’s natural to expect other quantum gitudinal direction and periodic in the transverse
process of photon, such as the Primakoff type ef- direction. For example, the plane continuum po-
 2

FIG. 2: ALP-photon conversion in intense crystalline

field. Plane continuum electric field in y direction is
plotted.

FIG. 1: Crystal lattice

If the path length in crystal is a constant the
cross-section can be expressed as σ = SP where P
tential can be expressed as [8] is the probability of ALP-photon conversion and S
p p is the geometric cross section of the target crystal.
U (y) = V [cosh(δ( 1 + η 2 − s + η 2 )) − 1], (4) A complete expression of P should sum contribu-
~ q~ in crystal. For simplicity we concen-
tions of all E
where s = 2|y ′ |/d and |s| ≤ 1. d is the interplanar
distance and y ′ = y − ymid the transverse coordi- trate on the case that ALP interact efficiently with
nate relative to the middle between two neighbor- a plane electric field which is periodic in y direc-
ing planes ymid . η, V , δ are parameters. For exam- tion and constant in x and z direction, as shown in
ple, δ = 3.85, V = 6.4 V, η 2 = 0.0007, d = 0.119 Fig.2. The relevant reciprocal vector is perpendic-
nm for the (110) plane of the W crystal at room ular to the crystallographic plane: ~qT = qT ŷ with
temperature [8]. The potential height U0 is around qT = 2πnT /d where nT 6= 0 is an integer.
130 V. For (110) plane of the Ge crystal, δ = 3.2, Implementing (3) into (6) and integrating over
V = 4.5 V, η 2 = 0.0052, d = 0.2 nm, U0 = 40 V. x and y coordinate we find that coherent ALP-
The plane continuum electric field can be derived photon conversion happens for
from (4). kfx = kix = 0, kiy − kfy − qT = 0. (7)
The Lagrangian of the photon and the ALP is
1 1 Then after integration over z coordinate we find
L = ∂µ φ∂ µ φ − m2φ φ2
2 2 1 sin2 (∆L/(4|~ki |))
1 1 P = ( gφ |ET |L cos θγ cos θi )2 G ,(8)
− F µν Fµν + gφ φFµν F µν , (5) 2 (∆L/(4|~ki |))2
4 4
where φ is the field of ALP, mφ the mass of where L is the path length in crystal as shown in
ALP, F µν the field strength of photon and F µν = Fig. 3. θγ is the angle between photon direction
1 µνρσ ρσ and z axis as shown in Fig. 2. θi is the angle of
2ε F . gφ is the coupling constant.
For a uniform incident flux the cross section for the initial particle. For γ → φ process, θi = θγ . ∆
the ALP-photon conversion in crystal is is

1
Z 3
d kf 1 1
σ = 2πδ(Ei − Ef ) ∆ = m2f − m2i − 2qT (kiy − qT ), (9)
2Ei vi (2π)3 2Ef 2
Z 2 where mf and mi are masses of the final and initial
3 ~ ~ i(~
ki −~
kf )·~
x
 
× d x gφ (kγ × E) · ǫ̂ e

 , (6) particles. For γ → φ process, mf = mφ and mi =
Ω
ωp where ωp is the plasma frequency of photon
where Ei and Ef are the energies of the initial and in crystal and is around tens eV. G = 1 − ωp2 /E 2
final particles respectively, vi the velocity of the where E is energy. G can be taken as one for ωp ≪
initial particle relative to the crystal detector, ~ki E. ET = −iqT UqT is the Fourier transform of
and ~kf the initial and final momenta, E ~ = E(~
~ x) the electric field in y direction. Using (4) ET is
~ expressed as
the electric field in the crystal, kγ the momentum
of the photon, ~ǫ the polarization vector of photon: 1 d
Z
ǫµ = (0,~ǫ). Ω is the volume of the crystal. One ET = −iqT dy U (y)eiqT y . (10)
d 0
can clearly see in (6) that photon polarized normal
to the plane of ~k and E ~ involve into the transfor- For example, for (110) plane of W crystal
mation with ALP. Photon polarized in the plane and qT = 2π/d we find |ET | = 1.7 × 1011
of ~k and E
~ does not interact with ALP. V/cm. For (110) plane of the Ge crystal
 3

γ We see that due to the periodic electric field

in crystal the final particle is reflected by
the crystal. The resonant conversion requires
|∆L/(4E)| < ∼ 1 and we find accordingly that
the width of resonance in energy and inci-
dent angle is: |δE| < ~ −1
φ
0
∼ 2|ki |L /(qT sin θi ) and
|δθi | < −1 ~ 0
∼ 2EL /(qT |ki | cos θi ).
Note that in the limit (ωp , mφ ) ≪ (qT , E) or if
L mφ ≈ ωp , (14) and (15) are reduced to the Bragg
scattering angle:
qT qT
sin θi0 = , sin θf0 = − . (16)
~ 2|~kf |
γ 2|ki |

In general case (14) and (15) are required to be

away from the Bragg scattering angle.
FIG. 3: Reflecting through the wall. Photons are sent We stress that according to (14) and (15) a care-
to crystal. Axion-like particles produced in crystal are ful arrangement of the incident angle with energy is
reflected by the crystal and after passing through the
required for resonant ALP-photon transformation
wall generate photons in another crystal target.
to happen. Another condition is that the energy
E should be larger than qT /2:
and qT = 2π/d we find |ET | = 4.0 × 1010 π 0.2 nm
V/cm. We see that when |∆L/(4|~ki |)| > 1 E> = 3.09 keV × . (17)
d d
the conversion probability is oscillatory with
an amplitude 4gφ2 |ET |2 |~ki |2 cos2 θγ cos2 θi /∆2 . Solar axions have average energy hEi ≈ 4.2 keV
[10] and have a significant fraction of flux with
When |∆L/(4|~ki |)| < ∼ 1 the probability is
resonantly enhanced with a value about E> ∼ 3 keV . So crystal detector, e.g. making use of
1 2 2 2 2 2 (110) plane in Ge crystal described above, can be
4 gφ |ET | L cos θγ cos θi . used to detect solar axion. The experiment can be
It’s easy to find that ∆ = 0 corresponds to kfz =
z done by carefully arranging the crystal target such
ki and together with (7) we see that the resonance
that the solar axion flux incident on the crystal
limit corresponds to the case with
with a specific angle to a chosen crystallographic
~ki − ~kf − ~ plane. Due to the condition (14), for a chosen an-
qT = 0. (11)
gle only axions of selected energy contribute to the
~ q~ associated with resonant ALP-photon conversion. The spectrum
In this limit the component E T
of solar axion can be scanned by adjusting the an-
q~T in (3) add coherently and give an resonant en-
gle of crystal target to solar axion flux. Note that
hancement to the cross-section. When this res-
for a different crystallographic plane the resonance
onant enhancement happens the cross-section is
~ q~ |2 and only photon polarized condition may also be satisfied for a different en-
proportional to |E T ergy. Since the spectrum of solar axion is contin-
in ~kγ × ~qT direction involves into the coherent con- uous many crystallographic planes can contribute
version with ALP. In the resonance region E ~ q~ asso- to the resonant ALP-photon conversion. Detailed
ciated with all other reciprocal vectors effectively proposal of solar axion search experiments using
contribute zero after integration over space. crystalline field should take all possible planes into
In the resonant limit we find that account. This is beyond the scope of this Letter
Another interesting experiment is a reflecting-
1 m2f − m2i through-wall experiment, a variety of the shining-
kiy = qT + , (12)
2 2qT through-wall experiment [11]. The experimental
1 m2f − m2i setup can be arranged as shown in Fig. 3. Photons
kfy = − qT + . (13) such as hard X-rays which have large enough pen-
2 2qT
etration depth in crystal are sent to a crystal. Due
In this limit the incident angle and outgoing angle to the condition of the resonant conversion ALPs
are produced are reflected by the crystal. After pass-
ing through a wall ALPs convert back to photons
qT m2f − m2i in crystalline field of another crystal target. This
sin θi0 = + , (14) experiment can be done using mono-energetic X-
2|~ki | 2|~ki |qT
rays with careful angular arrangement of crystal
qT m2f − m2i targets and incident angle of X-rays. In a sym-
sin θf0 = − + . (15)
2|~kf | 2|~kf |qT metric experimental setup the probability finding
 4

X-rays through the wall is In conclusion we find that coherent ALP-photon

conversion can happen in crystalline field if the en-
1 sin4 (∆L/(4E)) ergy is larger than about keV and a condition of
P = ( gφ |ET |L cos2 θγ )4 . (18)
2 (∆L/(4E))4 the energy and the incident angle is satisfied. We
propose to do ALP search experiments using in-
(18) is estimated as tense electric field in crystal. These experiments
have the following virtues: 1) Since the crystalline
−20 4 sin4 (∆L/(4E)) field is several orders of magnitude stronger than
P = 7.5 × 10 × cos θγ
(∆L/(4E))4 the magnetic field available in laboratory, ALP

|ET |
4 
gφ
4 
l
4 search experiment using crystalline field has the
× V
, (19) potential to reach sensitivity beyond the present
2. × 1011 cm 10−8 GeV−1 5cm experiments based on ALP-photon conversion in
magnetic field. 2) ALP search experiments using
where l = L cos θγ is the length of crystal. The
crystalline field can probe wide range of mφ , from
event rate is
eV to keV, by adjusting the incident angle of ini-
W 10 keV F P tial flux to crystal when using hard X-ray or using
N = 19.7 year−1 , (20) solar axion source in experiments. This range of
10 W E 1% 10 −20
mφ cannot be probed with good sensitivity in ALP
where W is the power of the X-ray source, F the search experiments using external B field.
efficiency of the coherent conversion of ALP and
X-rays. F arises due to the fact that X-ray source
has an angular spread and energy spread and only Acknowledgment: I would like to thank X. Y.
parts of X-rays can satisfy the condition of reso- Li, Y. Liao and Y. Q. Chen for fruitful discussions
<
nant conversion: |∆L/(4E)| ∼ 1. We can see that on physics in intense external fields. This work
this experiment is very interesting. If 1% efficiency is supported by Science and Technology Commis-
can be achieved with a X-ray source of 10 W power, sion of Shanghai Municipality under contract No.
a detector using the (110) plane of W crystal with 09PJ1403800 and National Science Foundation of
−8 −1
5cm length can probe g to 10 GeV . It would China(NSFC), grant No. 10975052.
φ
be much better than the shining-through-wall ex-
periment using B field [7].
Note added After the submission of this Letter
A very interesting aspect of ALP search experi-
we received Refs. [12–18]. The idea of photon-
ment using crystal is that the range of mφ can be
axion conversion in crystal has been studied in
scanned by varying the incident angle. For exam-
[12, 13] and solar axion search experiments based
ple, for a fixed incident angle θi in φ → γ process
on [13] have been done [14–18]. We note that
the resonance(∆ = 0) happens for mφ :
the result in this Letter is different from the result
1 in [12, 13]: 1) The probability (8) has a different
m2φ = ωp2 − 2qT (|~ki | sin θi − qT ). (21) dependence on the angle θγ than in [12, 13]. 2)
2
The result in this Letter is valid for all possible
mφ range can be scanned in experiment by care- angles but the result in [12, 13] works only for the
fully adjusting the incident angle θi . This is an Bragg scattering angle which is a special case of our
important virtue of the ALP search experiment us- result, as shown in (16). 3) The resonance condi-
ing crystal. Using hard X-rays or the solar axion tion for non-zero mass case, (14), was not found in
source, ALP search experiments using crystalline [12, 13]. According to result in this Letter previ-
field can probe the range of mφ up to keV scale ous experiments [14–18] based on [13] should be
without loss of sensitivity. re-examined.

[1] R. D. Peccei and H. Quinn, Phys. Rev. Lett. 38, Rev. D32, 2988(1985).
1440(1977); Phys. Rev. D16, 1791(1977). [6] R. Bradley et al., Rev. Mod. Phys. 75, 777(2003).
[2] S. Weinberg, Phys. Rev. Lett. 40, 223(1978); F. [7] for a recent review see J. Jaeckel and A. Ringwald,
Wilczek, Phys. Rev. Lett. 40, 279(1978). Ann. Rev. Nucl. Part. Sci. 60, 405(2010).
[3] J. Kim, Phys. Rev. Lett. 43, 103(1979); M. A. [8] Electromagnetic processes at high energies in ori-
Shifman, A. I. Vainshtein and V. I. Zakharov, ented single crystals, V. N. Baier, et al., World
Nucl. Phys. B166, 493(1980). Scientific, Singapore, 1998.
[4] M. Dine, W. Fischler and M. Srednicki, Phys. [9] U. I. Uggerhoj, Rev. Mod. Phys. 77, 1131(2005).
Lett. B104, 199(1981); A. P. Zhitnitskii, Sov. J. [10] K. van Bibber, et al., Phys. Rev. D39, 2089(1989).
Nucl. Phys. 31, 206(1980). [11] L. B. Okun, Sov. Phys. JETP56, 502 (1982); A.
[5] P. Sikivie, Phys. Rev. Lett. 51, 1415(1983); Phys. A. Anselm, Yad. Fiz. 42, 1480(1985); Phys. Rev.
 5

D37, 2001(1988); K. Van Bibber, et al., Phys. (1998) .

Rev. Lett. 59, 759(1987). [16] H. M. Chang et al., Phys. Rev. D75, 052004
[12] W. Buchmuller and F. Hoogeveen, Phys. Lett. (2007).
B237, 278(1990). [17] R. Bernabei et al., Phys. Lett. B515, 6 (2001); R.
[13] E. A. Paschos and K. Zioutas, Phys. Lett. B323, Bernabei et al., Riv. Nuovo Cim. 26N1, 1 (2003).
367 (1994). [18] Z. Ahmed et al.,Phys. Rev. Lett. 103, 141802
[14] A. Morales et al., Astropart. Phys. 16, 325 (2002). (2009).
[15] F. T. Avignone et al., Phys. Rev. Lett. 81, 5068