Physics and Astronomy Publications
Physics and Astronomy
7-1970
Low-Energy Optical Absorption Peak in
Aluminum and Al-Mg Alloys
L. W. Bos
Iowa State University
David W. Lynch
Iowa State University, dlynch@iastate.edu
Follow this and additional works at:
http://lib.dr.iastate.edu/physastro_pubs
Part of the
Atomic, Molecular and Optical Physics Commons
The complete bibliographic information for this item can be found athttp://lib.dr.iastate.edu/physastro_pubs/24. For information on how to cite this item, please visithttp://lib.dr.iastate.edu/howtocite.html.
Low-Energy Optical Absorption Peak in Aluminum and Al-Mg Alloys
Abstract
Measurements of the absorptivity of Al and dilute Al-Mg alloys were made at 4.2 K in the 0.2- to 5-eV energy
region by a calorimetric technique. In addition to the well-known conductivity peak around 1.6 eV, another
peak has been found near 0.5 eV. These peaks are present in the alloys, broadened, but unshifted, as well as in
Al. The peaks are at about 2|V200|and 2|V111|, respectively, where the V's are Fourier coefficients of the
pseudopotential.
Keywords
pseudopotential, conductivity, Ames Laboratory
Disciplines
Atomic, Molecular and Optical Physics | Physics
Comments
VQLUME 2$,NUMBER $
PHYSICAL
REVIEW LETTERS
20Jvr.v 1970 LOW-ENERGY OPTICAL ABSORPTION PEAK IN ALUMINUM AND Al-Mg ALLOYS*L.
W. Bos1' and D. W. LynchDepartment of Physics and Institute for Atomic Research, Iowa State University, Ames, Iowa 50010 (Received 14May 1970)
Measurements ofthe absorptivity ofAl and dilute Al-Mg alloys were made at
4.
2 K in the 0.2-to 5-eV energy region by a calorimetric technique. In addition to the well-known conductivity peak around1.
6eV, another peak has been found near0.
5eV. These peaks are present inthe alloys, broadened, but unshifted, as well as in Al. The peaks are atabout 2''200( and 2k~~ql~ respectively, where the V's are Fourier coefficients ofthe
pseu-dopotential.
The optical absorption ofaluminum
is
rather featureless in the region of interband absorption. The only reported structure'is
aref
lectivity minimum at1.
5 eV, although suspicions of an absorption peak at lower energy have been re-ported in some experimental work.'
'
The strength of the1.
5-eV absorption peak was once a contro-versial subject,'
but recent calculations agree well with experiment.'
"
This peak was original-ly attributed to transitions near~
in the Bril-louin zone,"
but now it appears to be the result oftransitions occurring ina larger
volume of the zone, near, but not including W.Harrison" first
pointed out that for many polyvalent metals, a pseudopotential approach leads to parallel bands, partly empty, partly occupied, throughout ap-preciable regions ofk space. Such bands lead toa
sharp edge in the conductivity ata
photon en-ergy of 2~Vs~, where Vois a
Fourier coefficientof the pseudopotential. At higher energy, each edge is followed by
a
decayingtail.
Golovashkin, Kopeliovich, and Motulevich"carried
out similar calculations. Any pseudopotential calculation of the band structure for Al should give bands that will yield roughly the above results when theop-tical
propertiesare
calculated from them. Re-cent calculations,'
"
including calculated dipole matrix elements, predict two absorption peaks, corresponding to 2~V»,~(-1.
5 eV) and 2I V»&l(-0.
5eV). The lower energy peak has not been reported previously. The high-energy peak has recently been remeasured.'
%e have measured the absorptivity of Al and several alloys of Al with Mg at
4.
2 K usinga
calorimetric method.
'
'
' The aluminum wasa
99.
999% pure Cominco polycrystalline ba.r.
The alloys were produced by
arc
melting. The samples were spark cut, mechanically polished, electropolished, annealed in Ar at 550 Cfor
atleast
100h, then electropolished again justbe-fore use. The samples were exposed to
air
forless
than 2 h, and within another 3h were at a pressure of 10'
Torr
orless.
Data were takenby
a
method similar to that of Biondi,'
""
but with higher sensitivity and resolution. A prism double monochromator was used to ensure high spectral purity in the infrared. Since structure in the1.5-eV
absorptivity peak has been report-ed' for pure Al, points were taken every 0,02 eV (the spectral bandpass) in that region. Theer-rors
in the measurements of the absorptivity are estimated to be 2~/~ in the0.
5-
to3-eV
region, rising to 10% at0.
2 eV, not including anyerrors
due to imperfect samplesurfaces.
Figure 1shows the absorptivities of Al and three alloys of Al with Mg. The absolute values of the absorptivities in this figure are difficult to guarantee. Several samples with noticeably poorer electropolished surfaces yielded absorp-tivity spectra like those shown in Fig. 1, but shifted upward by about
0.01
in the infrared and0.
02 around 3eV. The structure near0.
5eV wasstill
discernible in these samples.Better
surfaces might yield still lower absorptivities in the infrared. However, the absorptivities for pure Al shown in Fig. 1
are
lower than any re-ported to date, 4indicating somewhat bettersur-faces on our samples.
A Kramers-Kronig inversion of the absorptivity was made to obtain the optical conductivity. Above 5 eV, published
data"
for
Al were used, while for the alloys above 3 eV, we used the Al data multiplied by afactor
between0.
8 and 1to make a smooth fitwith the alloy data at 3 eV. From 0 to0.05
eV the absorptivity of a freeelec-tron gas was used, and asmooth monotonic curve was drawn between
0.05
and0.
2 eV to connect to the data. Changes in the low-energy extrapola-tion had little effect on the computed conductivity, but changes in the high-energy extrapolation had a large effect on the magnitude, but not the shape, ofthe conductivity between0.
2 and 3 eV. The alloy data are particularly sensitive here because their poorer surfaces producederrors
in the ab-sorptivity. The conductivityspectra
have only qualitative significance for the alloys, but thatVo«ME 25,NUMB
"-
PHYS
+L
RE»E~
LETTERS
2QJUJ.Q197&I
~n~k
.03
0.20—.
pz-.PI -0 0. I-CL CL O
v&O.I0—
CQ
I
2S
I I8 2.0 2.2
Y (eV) l.6
PHOTON ENERG
I
0.8 I.O
Al+$.0% 04 06
Al. dashed curve&
o~est atlo~energy' ~
data on an expand Solid
c
t
hows the low4Kversus photo
5 5y Mg. The &n
Fgo
]
.
Absorp25gj Mg; solid cur Mgy shed curve
ed scale.
l2
IO—
Al
t4K
hm that for pure a It would seem
sis
should not
becorrect.
A Aas cut from one
'
it
for
optical mmeasuremenns.
s.
assume 000 indicating
a
'
i
"but
instead orle mu i ity,'
a
ocor
tivity using ebtain the (loca in sitions, to o
I
tivit
.
A Drude contribution to tlont}1 ' 1
for
pure Alis
us the conductivityfor
pd eaks shou e b
t
both interban pe—V
"k.
g---
ithroom-tem-b
ec
ause the1.6-e
pta
on Al,free
ofnon1occality problems,s
t
Al-I
llo,
'1
and both peaks
persist
in roblems. pcontribution o The interband co
zs
s
2for
Al and~ ~
is
shown in Fag.eak more vis&hie, ake the low-energy pe
order to m e
-IO I E O
8—
E O6—
b FO I OZ=0.00
b
O
0.0080
I 0
2.2 2.4
I 2.0 I 02 2 I I
08 IO . . 1.6 I.8
I
0
ON ENERGY (eV)
PHOT
nd
.
0 g ( sheds e curve obtained bya Kramers-mers-1t
e o vit was characterized byth
ttl
od
tit
Kronig inversion. T e suVoLUME 25,NUM@Em $
PHYSI
CAI
REVIEW
LETTERS
20 JU~v 1970a
Drude term has been subtracted from the totalconductivity. Because interband absorption con-tributes everywhere in our data range and
be-cause ofthe anomalous skin effect, there is noway to obtain the Drude parameters from our data.
For
both samples we used k~~ =12.7 eV.For
Al we used (~~T)'
=y=0.
0035and0.
0030,while
for
the alloy, y ranges from0.
0062 to0.008,
the former value being that expected from the dc conductivity."
It
is clear
that the main peak in Al seems structurefree
except for possible slope changes at1.
40 and1.
65 eV (better detected on plots with differentscales).
A recentset
ofmeasurements' by an ac technique" reports structure on the low-energy side of the1.
6-eV
peak, structure large enough to be discernible in our data, but whichis
not found. The reason for the disagreement is notclear.
Itis
either anartifact
of the ac methodor
apparatus,or
the result of inferior surfaces in our measurements.The new peak at
0.
50eV has been predicteda
number oftimes,'
"
but not observed before. According to Harrison's model, there should be sharprises
in the conductivity at Nv= 2iVoi.
Ashcroft'svalues"
of iVoi, which yield a Fermisurface in substantial agreement with
experi-ment, give 21 V2oo1=1 52 eVand 21ViiiI=0 485
eV. The conductivity should then fall offas
'(kv-2i
Voi)"'
beyond the sharprise.
"
Dresselhaus' andBrust'"
have recently calcu-lated the conductivity more realistically, using the above pseudopotential coefficients. Their peak values of the conductivity are in substantial agreement with those ofFig.
2. They point out that the low-energy peak does notrise
at 2iV»,
ibut begins at zero photon energy,
a
result ofthe crossing of two bands at the Fermi level on cer-tain planes in the Brillouin zone. This low-en-ergy absorption must be considered when obtain-ing the electron effective mass from infrared data on Al, as recognized by Golovashkin, Motu-levich, and Shubin' and by Gurzhi and Motulevich.'
Adding Mg, which simply. lowers the Fermi en-ergy according to
a
rigid-band model, has little effect on the positions of the peaks. The peaks simply broaden and become lower, an effect of the increased scattering in the alloys. Thus, the two values of iVoi change very little. A shiftin the Fermi level
is
not expected to alter the position ofthe conductivity peak, becausea
low-er Flow-ermi
level only makes the regions in &space where transitions occur at a particular
photon energy move slightly. Since the bands are nearly parallel in both regions, no appreci-able change in the conductivity results.
Measurements further into the infrared
are
in progress, and a more detailed separation of the absorptivity intofree-carrier
and interband terms will be reportedlater.
We wish to acknow-ledge profitable discussions and correspondence with Dr. D. Brust, Dr. G. Dresselhaus, andPro-fessor
K.I.
Kliewer, and the sample preparation by H. H. Baker andF.
A. Schmidt,*Work performed atthe Ames Laboratory ofthe U.
S.
Atomic Energy Commission.
)Present address: Western Electric Company, Engineering Research Center, Box 900, Princeton,
N.
J.
08540.H.
E.
Bennett, M.Silver, andE.
J.
Ashley,J.
Opt.Soc.Amer. 53, 1089 (1963).
A.
I.
Golovashkin, G.P.
Motulevich, and A.A.Shubin, Zh. Eksp. Teor. Fiz.38, 51 (1960)ISov. Phys. JZTP 11,38 (1960)].
R.N. Gurzhi and G.
P.
Motulevich, Zh. Eksp. Teor. Fiz. 51, 1220 (1966) [Sov. Phys. JZTP 24, 818(1967)].M. A. Biondi and A.
I.
Guobadia, Phys. Bev. 166, 667 (1968).C.
J.
Powell,J.
Opt. Soc.Amer. 60, 78 (1970).J.
C.Phillips, Solid State Phys. 18, 55 (1966). 'A.J.
Hughes, D. Jones, and A. H. Lettington,J.
Phys. C: Proc. Phys. Soc., London 2, 102 (1969). G.Dresselhaus, M.
S.
Dresselhaus, and D. Beagle-hole, in Proceedings ofthe Symposium onthe Elec-tronic Density ofStates, Washington, D.C., 3-6November 1969(to be published).
D. Brust, Phys. Lett. 31A, 289 (1970). D. Brust, Phys. Rev. B(to be published).
E.
Ehrenreich, H.B.
Phillipp, andB.
Sega11, Phys. Rev. 132, 1918 (1963).W.A. Harrison, Phys. Rev. 147, 467 (1966). A.
I.
Golovashkin, A.I.
Kopeliovich, and G.P.
Motulevich, Zh. Eksp. Teor. Fiz. 53, 2053 (1967) [Sov. Phys. JZTP 26, 1161 (1968)].'
K.G.Bamanthan, Proc. Phys. Soc.London, Sect.A 65, 532 (1952).' M. A. Biondi, Phys. Rev. 102, 964 (1956).
M. A. Biondi and
J.
A. Rayne, Phys. Rev. 115,1522 (1959).B.
P.
Madden,L.
B.
Canfield, and G.Hass,J.
Opt.Soc.Amer. 53, 620 (1963).
G.
E.
H. Reuter andE.
H.Sondheimer, Proc.Boy. Soc., Ser.A 195, 336 (1948);R.B.
Dingle, Physica 19, 311(1953);R.Fuchs and K.L.
Kliewer, tobe published.~A.
T.
Robinson andJ.
E.
Dorn,J.
Metals 3, 457 (1951).D.Beaglehole, Appl. Opt. 7, 2218 (1968).
¹ W.Ashcroft, Phil. Mag. 8, 2055 (1963).