DTIC ADA195483: Pseudo-Stark Effect and FM/Stark Double Modulation Spectroscopy for the Detection of Statistical Fine Structure in Alexandrite.

Survival, Water, Medical Field Manuals

Military Manuals

Defense Technical Information Center

Document text

EPROOUCED  AT  GOVERNMENT  EXPENSE 


w 


AD-A1S5  483 

UNCLASSIFIED  H 


2a.  SECURITY  CLASSIFICATION  AUTHOR 


2b.  0ECLASSIFICAT10N  /  OOWNGRADI 


4.  PERFORMING  ORGANIZATION  REPORT  NUMBER(S) 
Technical  Report  #16,  IBM  RJ 


6a.  NAME  OF  PERFORMING  ORGANIZATION 
IBM  Research  Division 
Almaden  Research  Center 


6c  ADDRESS  (G<y,  State,  and  ZIP  Code) 

650  Harry  Road 

San  Jose,  CA  95120-6099 


CUMENTATION  PAGE 


lb.  RESTRICTIVE  MARKINGS 


3.  DISTRIBUTION/ AVAILABILITY  OF  REPORT 
This  document  has  been  approved  for 
public  release  and  sale;  its 
distribution  is  unlimited. 


S.  MONITORING  ORGANIZATION  REPORT  NUMBER(S) 


OFFICE  SYMBOL  I  7a.  NAME  OF  MONITORING  ORGANIZATION 
(If  applicable)  I  Office  of  Naval  Research 


7b.  ADDRESS  (C/ty,  State,  and  ZIP  Code)  ,  „  „  „  _ 

Chemistry  Division  Code  1113 

Arlington,  VA  22217 


8a.  NAME  OF  FUNDING/SPONSORING 
ORGANIZATION 

Office  of  Naval  Research 


8b.  OFFICE  SYMBOL  9.  PROCUREMENT  INSTRUMENT  IDENTIFICATION  NUMBER 
* '"’**“"  H00014— 84-C-0708 


10.  SOURCE  OF  FUNDING  NUMBERS 


PROGRAM 
ELEMENT  NO. 


PROJECT 

TASK 

NO. 

NO. 

8c.  ADDRESS  (City,  State,  and  ZIP  Code) 

Chemistry  Division,  Code  1113 
Arlington,  VA  22217 


11.  TITLE  (Include  Security  Classification) 

Pseudo-§tark  Effect  and  FM/§tark  Double  Modulation  Spectroscopy  for  the  Detection  of 
Statistical  Fine  Structure  in  Alexandrite 


12.  PERSONAL  AUTHORS)  -  — ... 

T.P.  Carter,  D.E.  Horne,  and  W.E.  Moerner  " 


13a.  TYPE  OF  REPORT 
interim  tecnhical 


WORK  UNIT 
ACCESSION  NO 


14.  DATE  OF  REPORT  (Year,  Month.  Day)  15.  PAGE  COUNT 
1988  June  20  /  21 


16.  SUPPLEMENTARY  NOTATION  / 

Submitted  for  publication  in  Chemical  Physics  Letters 


17. 

COSATI  COOES  | 

\  FIELD 

GROUP 

SUB-GROUP  1 

18.  SUBJECv|/ERMS  ( Continue  on  reverse  :f  necessary  and  identify  by  block  number) 
Statistical  Fine  Structure  >  Modulation  Spectroscopy^ 

Alexandrite  >  FM  Spectroscopy— 

Stark  Effect ;  \ 


19.  ABSTRACT  (Continue  on  reverse  it  necessary  and  identify  by  block  number)  ^ 

Using  laser-FM  spectroscopy  and  transient  spectral  bole,-burning,  the  R  transition  of  Cr  r 

_ _  lm 


ions  in  alexandrite  is  shown  to  exhibit  a  linear  pseudo-Stark  effect  with  coefficients  of 
0.141  MHz-cm^V  *  and  0.0538  MHz-cm/v"”^  at  1.6  K  for  'electric  fields  along  the  crystal 
a  and  c  axes,  respectively.  .This  result  was  used  to  develop  a  sensitive  FM/Stark  double 


modulation  method  for  the  detection  of  statistical  fine  structure  (SFS)  on  the  inhomo- 
geneously  broadened  R  line.  While  the  final  signal-to-noise  ratio  is  not  much  greater 
than  unity,  autocorrelation  of  the  measured  spectra  provides  the  first  strong  evidence 
for  the  presence  of  SFS  in  an  inorganic  material.  a  r 


20  DISTRIBUTION/ AVAILABILITY  OF  ABSTRACT  21  ABSTRACT  SECURITY  CLASSIFICATION 

C3  unclassified, 'unlimited  Q  same  as  rpt  □  one  USERS  unclassified 


Ira  N AMe  OF  ne  jeONyBLE  NOlVtOUAL 

Dr .  W.  E .  Moerner 


00  FOf!M  1473,  84  mar 


22b  TcL^.PhO'TE  (Include  Area  Code)  .2.  OFF.C:  SYMBOL 

408-927-2426 


3 3  APR  edition  may  oe  used  un*n  exnausteo 
All  other  editions  re  oosoiete 


SECURITY  CLASSIFICATION  OF  T-"S  RAGE 

unclassified 


OFFICE  OF  NAVAL  RESEARCH 


a 


a 


9 


Contract  N00014-84-C-0708 


R&T  Code  413a001— 01 


Technical  Report  No.  16 


Pseudo-Stark  Effect  and  FM/Stark  Double  Modulation  Spectroscopy 
for  the  Detection  of  Statistical  Fine  Structure  in  Alexandrite 


CO*** 

INSPECTED  , 


T.  P.  Carter,  D.  E.  Home,  and  W.  E.  Moemcr 


Acccsio.",  For 


Prepared  for  Publication 


NTIS  CRA&i 
DTIC  TAB 

Unannounced 

Justification 


Chemical  Physics  Letters 


IBM  Research  Division 
Almadcn  Research  Center 
650  Harry  Road 
San  Jose,  California  95120-6099 


AVa.ifblliry  CudiiS 


Avr.'l  ant!  for 


A -I 


LI _ 1. 


June  20,  1988 


Reproduction  in  whole,  or  in  part,  is  permitted  for  any  purpose  of  the  United  States 
Government. 


*  This  document  has  been  approved  for  public  release  and  sale;  its  distribution  is  un¬ 
limited. 


88  6  28  108 


PSEUDO-STARK  EFFECT  AND  FM/STARK  DOUBLE 
MODULATION  SPECTROSCOPY  FOR  THE  DETECTION 
OF  STATISTICAL  FINE  STRUCTURE  IN  ALEXANDRITE 

* 

T.  P.  Carter  ,  I).  E.  Horne,  and  W.  E.  Moerner 
IBM  Research  Division 
Almadcn  Research  Center 
San  Jose,  California  95120 

ABSTRACT:  Using  laser-FM  spectroscopy  and  transient  spectral  hole-burning,  the  R,m 
transition  of  Cru  ions  in  alexandrite  is  shown  to  exhibit  a  linear  pseudo-Stark  effect  with 
coefficients  of  0.141  MUz-cm-V  1  and  0.0538  MHz-cm-V  '  at  1.6  K  for  electric  fields  along 
the  crystal  a  and  c  axes,  respectively.  This  result  was  used  to  develop  a  sensitive  FM/Stark 
double  modulation  method  for  the  detection  of  statistical  fine  structure  (SFS)  on  the 
inhomogeneously  broadened  Rlm  line.  While  the  final  signal-to-noise  ratio  is  not  much 
greater  than  unity,  autocorrelation  of  the  measured  spectra  provides  the  first  strong  evidence 
for  the  presence  of  SFS  in  an  inorganic  material. 

• 

IBM  Visiting  Scientist,  present  address:  Chemistry  Department,  Florida  State  University, 
Tallahassee,  FL  32306 


I.  Introduction 


Recently,  we  reported  the  first  observation  of  statistical  fine  structure  (SFS)  on  an 

inhomogeneously  broadened  absorption  profile  ’  using  the  S,  <—  Sn  zero-phonon  transition 

of  pentacenc  in  /Mcrphenyl  mixed  crystals  and  the  technique  of  laser  frequency-modulation 

(FM)  spectroscopy.  ?  SFS  is  time-independent  structure  on  the  inhomogeneous  line  caused 

by  fundamental  statistical  variations  (number  fluctuations)  in  the  spectral  density  of  absorbers 

in  each  frequency  interval.  Observation  of  SFS  in  the  spectra  of  pentaccne  in  p-tcrphenyl 

“10  2 

was  facilitated  by  the  large  low  temperature  absorption  cross-section  (~  10  cm  )  and  the 

2 

very  small  persistent  hole-burning  quantum  yield.  In  many  ways  this  system  is  ideal  for  the 
task  of  observing  SFS,  and  many  other  organic  mixed  crystal  systems  where  the  guest  has 
high  oscillator  strength  and  hole-burning  is  weak  will  have  similarly  favorable  properties. 
However,  in  order  to  illustrate  the  generality  of  the  SFS  phenomenon,  we  report  in  this  paper 
the  results  of  our  measurements  on  a  transition  metal  ion  (with  low  oscillator  strength)  in 
an  inorganic  host  crystal. 

The  system  chosen  for  this  study  is  the  lowest  energy  electronic  absorption  of  Cru  ions 
in  alexandrite  (chrysoberyl)  crystals.  As  we  will  show,  this  system  has  very  different 
properties  than  pentacenc  in  /?-terphenyl,  and  presents  challenging  experimental  problems  in 
the  observation  of  SFS.  The  principal  problems  are  that  the  small  oscillator  strength  in 
alexandrite  (^  IQ5  compared  to  «  0.1  for  pentacenc)  leads  to  small  absolute  absorption 
changes  due  to  SFS,  and  that  the  alexandrite  excited  state  is  easily  saturated  due  to  the 
relatively  long  excited  state  lifetime.  We  chose  to  detect  SFS  using  a  combination  of  FM 
spectroscopy  and  Stark  modulation,  and  since  the  pseudo-Stark  effect  for  the  electronic 
transitions  in  alexandrite  has  not  been  previously  reported,  it  was  necessary  to  perform  some 
preliminary  studies  on  the  electric  field  behavior  of  the  spectra.  We  report  here  our 
measurements  of  the  pseudo-Stark  effect  on  transient  holes  in  the  spectra  of  alexandrite  and 
its  use  as  a  method  of  performing  double-modulation  spectroscopy  for  the  detection  of  SFS. 


H.  Experimental 


4 

The  alexandrite  (BeAl204:Cr3+)  crystal  structure  is  orthorhombic  Puma  with  four 
molecules  per  unit  cell.  There  are  two  crystallographically  inequivalent  sites  occupied  by 
Al3+  ions  in  the  lattice:  one  site  having  inversion  (Q)  symmetry  and  the  other  having  mirror 
(Cs)  symmetry,  with  the  mirror  along  the  crystal  ac  plane.  The  Cr34  ions  substitute  for  Al34 
ions  with  about  78%  going  into  the  mirror  site.  5' 6  The  alexandrite  crystal  used  in  all 
experiments  was  a  highly  polished  nearly  cubic  single  crystal  about  5  mm  on  a  side,  with 
faces  cut  parallel  to  the  crystallographic  axes.  The  Cr34  doping  level  was  %  0.03  atom  % 
Cr/Al. 

The  Cr34  transition  studied  in  this  work  was  the  lower  energy  component  of  the 
2Fg  *-4A2g  transition  for  the  mirror  site  ions,  which  is  split  by  the  distortion  to  C  symmetry 

7  8 

and  is  labeled  the  RIm  line.  ’  At  1.6  K  this  zero-phonon  line  is  centered  at  ss  679.62  nm 
(vacuum  wavelength  measured  by  a  Burleigh  WA-10  wavemeter)  with  a  full-width  at 
half-maximum  (FWIIM)  of  «  2.2  cm'  and  is  strongly  6-axis  polarized  with  a  peak  optical 
density  greater  than  4  at  this  Cr3t  concentration.  In  all  experiments,  the  laser  polarization 
was  maintained  along  the  6-axis  and  the  laser  wavelength  was  located  on  the  long-wavelength 
edge  of  the  line. 

Laser  FM  spectroscopy  was  implemented  in  the  manner  described  fully  in  Reference  2. 
Figure  1  shows  a  simplified  apparatus  diagram:  the  output  of  a  single-frequency  cw  dye  laser 
DL  (Coherent  599-21,  DCM  dye)  was  sent  through  an  electro-optic  phase  modulator  HO 
(Lascrmetrics  1097,  LiTaO,  )  driven  by  an  rf  oscillator  and  amplifier  at  148  MHz,  using  up 
to  16  V  rms  into  50  £2.  This  resulted  in  modulation  indices  between  0.3  and  1.  The  laser 
beam  was  first  expanded  and  then  focused  by  a  350  mm  focal  length  lens  to  give  a  « 

65^m  diameter  spot  throughout  the  entire  5  mm  length  of  the  sample.  The  sample  was  held 
between  polished  stainless  steel  electrodes  in  a  liquid  helium  immersion  cryostat,  (7  The 
transmitted  beam  intensity  was  measured  by  a  Si  avalanche  photodiode-preamplifier 


combination  whose  output  was  sent  to  a  double-balanced  mixer  M  referenced  to  the  rf 
oscillator.  The  mixer  output  near  dc  was  then  either  applied  to  the  input  of  a  lock-in 
amplifier  (Princeton  Applied  Research  HR-8)  whose  output  was  digitized  and  averaged  by  a 
digital  scope  DS  (Data  Precision  6100  with  a  610  plug-in)  for  double-modulation  detection, 
or  sent  to  DS  directly  for  ordinary  laser- FM  detection.  In  the  double-modulation 
experiments,  a  modulated  Stark  field  perpendicular  to  the  propagation  direction  of  the  laser 
beam  was  applied  to  the  sample  using  a  unipolar  square  wave  of  up  to  400  V  at  10  kHz. 
This  was  accomplished  using  a  high  voltage  power  supply  and  an  optically  coupled  high 
voltage  switcher,  II V.  A  signal  synchronous  to  the  high  voltage  modulation  was  used  as  a 
reference  for  lock-in  detection.  In  this  way,  only  those  components  of  the  Iascr-FM  signal 
which  were  modulated  at  10  kHz  were  detected.  Not  shown  in  the  diagram  are  provisions 
for  switching  the  FM  sidebands  and  Stark  modulation  on  and  off  as  necessary  in  synchrony 
with  the  0.25  s  laser  scan. 

Transient  holes  were  burned  and  probed  by  controlling  the  dye  laser  scan  externally  in 
the  following  manner.  The  laser  was  held  at  a  fixed  frequency  a  few  hundred  MIIz  from  one 
end  of  the  scan  range  for  about  200  ms  with  the  FM  sidebands  and  Stark  modulation 
switched  off.  This  process  burned  a  single  hole  in  the  spectrum,  uncomplicated  by 
pseudo-Stark  splittings  or  holes  burned  by  the  FM  sidebands.  The  laser  was  then  rapidly 
scanned  to  the  near  end  of  its  range,  the  FM  and  Stark  modulations  were  turned  on,  and 
then  the  laser  was  scanned  continuously  to  the  opposite  end  of  its  range  in  250  ms  while 
digitizing  the  signal.  After  this,  the  laser  was  scanned  rapidly  back  to  the  hole  frequency, 
and  a  new  hole  was  burned.  This  sequence  was  carried  out  repetitively  and  successive  scans 
were  averaged  by  DS.  The  total  scan  range  was  typically  2.5  GHz.  No  evidence  for 
persistent  hole-burning  was  observed  under  any  conditions. 

For  the  measurement  of  SFS  spectra  (as  opposed  to  the  measurement  of  transient  holes), 
the  laser  was  scanned  forward  and  backward  smoothly  and  continuously  across  its  range  with 


the  Stark  and  FM  modulations  on  all  the  time.  This  was  done  to  reduce  interfering  signals 
from  the  turn-on  of  the  FM  sidebands  and  to  prevent  the  formation  of  transient  holes  within 
the  scan  range.  Even  so,  weak  transient  holes  arc  produced  when  the  laser  frequency  is 
stationary  for  a  brief  time  at  the  limits  of  the  scan.  To  prevent  these  holes  from  distorting 
the  SFS  spectra,  only  the  central  half  of  the  scan  range  was  digitized  and  analyzed.  The  total 
scan  range  in  these  experiments  was  typically  1.0  GHz. 

III.  Results  and  Discussion 

A.  Pseudo-Stark  effect 

9 

The  ground  state  pseudo-Stark  effect  for  Cr3+  ions  in  the  mirror  site  of  alexandrite  has 
been  reported  previously  in  ESR  spectra  at  room  temperature.  '°  We  demonstrate  here  that 
the  Rlm  line  exhibits  a  linear  pseudo-Stark  effect  when  an  electric  field  is  applied  along  the 
crystal  a  or  c  axes.  This  is  a  pseudo-Stark  effect,  because  there  arc  two  inequivalcnt  mirror 
sites  in  an  electric  field,  one  of  which  shifts  in  one  direction  with  applied  field  and  the  other 
in  the  opposite  direction.  As  has  been  shown  before  for  other  zero-phonon  transitions,  the 
Stark  effect  can  be  precisely  measured  at  high  resolution  using  hole-burning  techniques.  ’ 
Figure  2  shows  the  results  of  a  Stark  experiment  with  several  values  for  the  applied  (dc) 
Stark  field  E„  with  spectra  measured  by  standard  FM  spectroscopy  (the  signal  from  only 
one  FM  sideband  is  shown,  and  the  dc  Stark  field  was  turned  on  only  during  hole  scanning). 
These  spectra  were  obtained  with  E,  ||  c ;  similar  results  were  obtained  for  Es  ||  a.  Each 
spectrum  is  the  average  of  64  scans.  Trace  2a  shows  the  transient  spectral  hole  with  E,  =  0 
burned  at  679.679  nm  (=  0  GHz)  at  1.6  K  with  about  6.5  n W  on  the  sample.  Traces  2b-2f 
were  similarly  obtained  using  applied  voltages  of  50  V,  100  V,  200  V,  300  V,  and  400  V, 
respectively.  As  the  figure  shows,  the  hole  splits  symmetrically  into  two  components,  each 
with  approximately  the  same  width  as  the  unshifted  line.  We  have  plotted  these  shifts  as  a 
function  of  the  applied  field  to  determine  the  pseudo-Stark  coefficients  for  I7.,  applied  along 


each  of  the  three  axes,  as  shown  in  Figure  3.  Circles,  squares  and  triangles  correspond  to 
applying  E,  along  the  a,  b  and  c  axes,  respectively.  No  efTcct  was  seen  for  the  case  of  E<  || 
b,  as  expected  for  the  A”  — ►  A'  transition  of  a  3d  ion  at  a  site  of  C,  symmetry.  The  straight 
lines  were  obtained  from  least-squares  fits  to  the  points,  clearly  demonstrating  a  linear 
dependence  on  field.  The  resulting  pseudo-Stark,  coefficients  for  the  shift  of  a  single 
component  are  0.141  Mllz-cm-V  1  and  0.0538  Mllz-cm-V  1  for  Es  along  the  a  and  c  axes, 
respectively.  These  coefficients  are  similar  to  those  for  ruby",  as  expected. 

B.  EM  detection  and  transient  hole  behavior 

Figure  4a  shows  a  typical  trace  for  lascr-FM  only  detection  of  a  transient  hole  burned 

at  1.6  K  into  the  Rlm  line  at  679.684  nm  (s  0  MHz)  using  «  7.5  /AV  laser  power.  For  this 

experiment,  the  Stark  field  was  left  off.  The  features  at  ±  148  MHz  are  the  FM  signals  of 

2 

the  transient  hole  using  cosine  (F,  )  detection,  and  the  spacing  between  these  replicas  of 
the  hole  allows  calibration  of  of  the  frequency  scale  to  1  part  in  106.  The  large  transient 
at  the  left  edge  of  the  trace  is  due  to  a  shift  of  the  mixer  dc  output  level  caused  by  dc  residual 
amplitude  modulation'3  (dc-RAM)  due  to  the  FM  sidebands  being  turned  on  at  this  point 
in  time.  The  other,  smaller  features  (which  are  most  obvious  between  the  two  signals  from 
the  transient  hole)  are  due  to  frequency-dependent  RAM.  This  hole  was  power  broadened 
considerably  due  to  the  conditions  chosen  for  this  experiment,  and  exhibits  a  width  of  14.8 
MHz  (vide  infra).  Since  the  hole  is  transient  and  therefore  constantly  decaying,  the  two 
replicas  have  different  intensities;  only  a;  60%  of  the  hole  intensity  seen  by  the  leading 
sideband  is  left  at  the  time  when  the  trailing  sideband  scans  the  hole.  The  time  difference 
between  when  the  passage  of  the  two  sidebands  through  the  hole  is  32.7  ms  which,  together 


with  the  observed  hole  decay  and  assuming  a  single-exponential  decay  process,  implies  a 

lifetime  of  «  35  ms.  This  value  is  much  greater  than  the  2.3  ms  low-temperature  lifetime 

1  8 

of  the  2E?  state  as  measured  by  fluorescence  decay.  '  However,  since  the  ground  state  is 
degenerate  with  a  splitting  14  of  0.54  cm  ',  it  is  reasonable  to  expect  that  there  might  be 


ground  state  cross-relaxation  processes  occurring  which  increase  the  hole  recovery  time,  as 

has  been  observed  for  transient  holes  burned  in  the  R  line  of  rubv.  15  No  significant 

1 

difference  in  width  was  observed  for  the  two  FM  components  cf  a  transient  hole,  implying 
there  are  no  important  spectral  diffusion  processes  occurring  on  a  time  scale  faster  than 
several  tens  of  milliseconds. 

Although  it  is  not  the  principal  goal  of  this  paper  to  provide  a  careful  measurement  of 
the  homogeneous  width  for  the  Rlm  transition,  at  lower  laser  intensities  we  did  observe 
shallow  transient  holes  whose  widths  were  v  4.6  MHz,  implying  an  upper  limit  to  the 
zero-field  homogeneous  linewidth  of  onc-half  of  this  value,  or  2.3  MHz.  Since  this  width  is 
comparable  to  the  laser  jitter  linewidth  for  a  250  ms  period,  we  assume  that  the  actual 
homogeneous  width  is  smaller.  The  only  other  report  of  the  homogeneous  width  at  zero  field 
for  Cr3+  in  alexandrite'6  is  44  MHz,  as  determined  by  fluorescence  line-narrowing  techniques. 
The  difference  in  the  two  measurements  may  be  due  to  the  fact  that  the  crystal  used  in  our 
measurements  is  more  dilute  (0.03  atom  %  versus  0.05  atom  %). 

C.  FM/Stark  double  modulation  spectroscopy 

The  purpose  of  laser- FM/Stark  double  modulation  spectroscopy  is  to  achieve  true 
zero-background  conditions  in  order  to  detect  the  very  weak  spectral  features  due  to  SFS. 
Without  double  modulation,  the  sensitivity  threshold  of  FM  spectroscopy  can  be  limited  to 

_4  j  7 

features  greater  than  10  in  absorbance  change  due  to  RAM  effects  in  the  modulator. 

The  approach  here  is  to  modulate  the  spectral  features  at  some  low  frequency  vs  far  from 
the  FM  modulating  frequency,  and  detect  the  v,  component  of  the  mixer  output  using  lock-in 
detection.  In  this  way,  none  of  the  RAM  background  is  detected,  allowing  an  increase  in  the 
gain  of  DS  without  overload  from  spurious  background  signals. 

Trace  4b  shows  a  simulation  of  double  modulation  detection  of  a  hole  similar  to  the  one 
shown  in  Trace  4a,  using  a  Stark  shift  appropriate  for  an  applied  field  of  50  V  along  the 


(/-axis.  In  double  modulation  detection,  the  LIA  sees  an  unshiftcd  hole  in  one  detection 
half-cycle,  and  two  pseudo-Stark-split  holes  in  the  other  detection  half-cycle.  When  the  1. 1 A 
rectifies  the  signal  at  the  reference  input  frequency,  these  two  half-cyclcs  arc,  in  effect, 
subtracted  from  each  other,  resulting  in  lineshapes  similar  to  those  in  Trace  4b.  This  is  seen 
in  the  measured  data  in  Figure  4c,  which  was  obtained  using  Es  =  50  V  along  the  a-axis. 
Note  that  the  trailing  sideband  signal  at  -f  148  MHz  is  shallower  as  expected.  For  this  value 
of  E,.  the  Stark  lines  strongly  overlap  the  unshifted  line  (Stark  shift  =  14.4  MHz,  Iinewidth 
=  14.8  MHz’).  At  higher  values  of  applied  field,  the  Stark  components  shift  farther  from 
the  unshifted  hole  position,  and  the  signal  increases  because  the  oppositely  signed 
components  no  longer  overlap.  As  can  be  seen  by  comparing  traces  4a  and  4c,  the  double 
modulation  technique  very  efficiently  suppresses  the  RAM  background  seen  using  simple  FM 
detection,  effectively  increasing  the  ultimate  detection  limit.  17 

D.  SFS  measurements 

Acquisition  of  SFS  spectra  in  alexandrite  is  complicated  by  several  factors.  First  of  all, 

-4 

the  signal  is  expected  to  be  small  (10  in  absorbance  change).  Usually  the  laser  power  is 
increased  in  order  to  lower  the  relative  contribution  of  shot  noise  and  improve  the  sensitivity 
to  weak  signals,  but  in  the  present  case  the  laser  power  had  to  be  kept  below  a  few  n W  in 
order  to  prevent  power  broadening.  This  is  the  majoi  difference  between  the  present  FM 
measurements  and  previous  studies'7  in  gases  in  which  several  mW  of  power  could  be 
presented  to  the  detector.  The  long  lifetime  of  the  transient  holes  which  may  be  burned  in 
the  spectrum  also  interferes  in  the  following  way.  As  the  laser  carrier  frequency  and  the  two 
sidebands  sweep  the  spectrum,  each  component  excites  some  of  the  ions  into  the  excited 
state.  Because  of  the  lifetime  of  the  transient  hole  and  the  scan  rates  and  modulation 
frequencies  used,  the  trailing  sideband  secs  a  spectrum  which  has  been  changed  by  the 
passage  of  the  leading  sideband  and  the  carrier  center  frequency.  More  importantly,  this 
"perturbed"  spectrum  can  in  principle  be  different  for  each  laser  scan,  unless  the  exact  same 


ions  undergo  excitation  by  the  leading  sideband  and  center  frequency  every  time  the  spectrum 
is  scanned,  which  is  unlikely.  If' such  a  process  is  occurring,  then  successive  averages  of  many 
scans  can  be  different,  even  in  the  absence  of  spectral  diffusion  or  persistent  hole-burning 
and  trenching  which  might  cause  permanent  changes  in  the  "true"  shape  of  the  SFS.  Also, 
if  ground  or  excited  state  cross-relaxation  is  occurring,  then  it  is  possible  that  the  "true"  SFS 
spectrum  changes  after  each  scan.  These  effects  can  be  minimized  somewhat  by  lowering  the 
laser  power,  but  then  the  overall  detection  signal-to-noisc  would  degrade. 

Figure  5  shows  typical  results  for  attempts  to  detect  SFS  in  the  R,m  line  of  alexandrite 
at  1.6  K  using  laser- FM/Stark  double  modulation.  Trace  5a  shows  two  superimposed 
acquisitions  of  the  SFS  spectrum,  each  of  v  nich  is  the  average  of  4096  laser  scans  (0  GHz 
=  679.676  nm).  The  spectra  arc  plotted  in  units  of  (Aa)F,  where  a  is  the  absorption 
coefficient,  I-  is  the  sample  length  and  (Aa)F  is  a  measure  of  the  difference  in  absorption 
between  the  two  FM  sidebands.  One  can  see  that  there  arc  obvious  correlations  of  both  the 
narrow  and  broad  features  of  these  two  spectra,  but  even  with  the  large  number  of  averages 
taken  the  signal-to-noise  ratio  is  near  unity.  Shown  in  trace  5b  is  a  typical  spectrum  obtained 
when  the  laser  is  tuned  off  the  R,m  line,  again  with  4096  averages  taken  with  constant  laser 
power  at  the  detector.  Due  to  the  large  noise  background  from  laser  shot  noise  and  detector 
avalanche  noise  and  to  the  effects  mentioned  above,  the  presence  of  SFS  is  not  convincing 
when  observing  the  raw  spectra. 

More  convincing  evidence  for  the  presence  of  SFS  structure  is  obtained  by  computing 

1  2 

the  normalized  autocorrelations  ’  of  the  spectra.  These  are  shown  in  Figure  6,  where  traces 
6a  and  6b  correspond  to  the  spectra  in  5a,  and  trace  6c  corresponds  to  5b.  I  he 
autocorrelations  of  spectra  taken  with  the  laser  in  resonance  with  the  R,m  line  show  a 
significantly  broader  origin  feature  than  those  taken  with  the  laser  off-resonance,  suggesting 
the  presence  of  underlying  homogeneous  lincshapcs  recurring  in  the  on-line  spectra.  This 
broadening  is  most  likely  due  to  SFS,  with  the  width  of  the  origin  feature  reflecting  the  width 


of  the  absorption  features  producing  the  SI'S.  *'  2  The  width  at  the  origin  of  the 
autocorrelation  of  ofT-resonancc  spectra  is  determined  by  the  bandwidth  of  the  detection 
electronics  together  with  the  digitization  rate  of  DS.  In  the  previous  reports  of  SFS,  the 
widths  at  the  origin  of  (on-resonance)  autocorrclatcd  spectra  were  determined  principally  by 
the  underlying  homogeneous  linewidth  of  the  features  making  up  the  inhomogeneous  band. 

Here  however,  the  width  at  the  origin  for  traces  6a  and  6b  is  comparable  to  the  short-time 
laser  jitter  linewidth,  ss  1-2  MIIz.  We  therefore  interpret  our  data  as  showing  that  the  zero 
(magnetic)  field  homogeneous  width  of  the  Cru  ions  in  the  R,m  line  of  alexandrite  is  less  than 
2  MHz,  limited  in  our  experiment  by  the  laser  linewidth.  This  result  is  consistent  with  the 
transient  hole  widths  reported  in  Sect.  III.B.  above. 

IV.  Conclusion 

Using  lascr-FM  spectroscopy  and  transient  spectral  hole-burning,  we  have  measured  the 
Stark  coefficients  for  Cr3f  ions  in  alexandrite  for  laser  light  polarized  along  the  b-axis  and 
find  a  linear  pseudo-Stark  splitting  for  the  mirror  site  ions.  Using  this  information,  we  have 
developed  a  FM/Stark  double  modulation  detection  technique  that  removes  the  interfering 
background  due  to  residual  amplitude  modulation.  We  have  used  this  technique  and 
autocorrelation  analysis  to  provide  strong  evidence  for  the  existence  of  statistical  fine 
structure  in  the  R,m  transition  of  alexandrite  at  1.6  K.  This  is  the  first  evidence  for  SFS  in 
an  inorganic  material  and  suggests  that  SFS  will  be  a  general  feature  of  all  inhomogcncously 
broadened  transitions.  In  order  to  improve  the  signal-to-noisc  ratio  of  SFS  measurements 
on  weakly  allowed  transitions  where  excitc-statc  saturation  may  be  a  problem,  a  detection 
technique  is  required  that  will  provide  higher  laser  power  on  the  detector  while  maintaining 
the  laser  power  on  the  sample  below  levels  that  produce  power  broadening.  One  technique 
that  may  achieve  this  is  heterodyne  FM  spectroscopy,  where  the  optical  carrier  is  produced 
by  a  separate,  higher  power  beam  that  does  not  pass  through  the  sample. 


ACKNOWLEDGEMENT 


The  authors  thank  Dr.  R.  C.  Morris  of  Allied  Corporation,  Corporate  Technology 
Division,  for  the  generous  loan  of  the  alexandrite  crystal,  Dr.  R.  M.  Macfarlanc  for  the  use 
of  the  Stark  effect  sample  holder,  and  Dr.  R.  S.  Meltzcr  for  stimulating  discussions.  This 
work  was  supported  in  part  by  the  U.S.  Office  of  Naval  Research. 


REFERENCES 


1.  W.  E.  Mocrner  and  T.  P.  Carter,  Phys.  Rev.  Lett.  59  ,  2705  (1987);  W.  11.  Moerner 
and  T.  P.  Carter,  Bull.  Am.  Phys.  Soc.  32,  1630  (1987). 

2.  T.  P.  Carter,  M.  Manavi  and  W.  E.  Moerner,  to  appear  in  .1.  Chem.  Phys.,  August 
1988. 

3.  G.  C.  Bjorklund,  Opt.  Lett.  5,  1.5  (1980);  G.  C.  Bjorklund,  M.  D.  Levenson,  W. 
Lenth,  and  C.  Ortiz,  Appl.  Phys.  B  32,  145  (1983). 

4.  E.  F.  Farrell,  J.  11.  Fang  and  R.  E..Newnham,  Am.  Mineral.  48,  804  (1963);  E.  F. 
Farrell  and  R.  E.  Newnham,  Am.  Mineral.  50,  1972  (1965). 

5.  R.  E.  Newnham,  R.  Santoro,  .1.  Pearson  and  C.  Jansen,  Am.  Mineral.  449,  427 
(1964). 

6.  M.  L.  Shand,  J.  C.  Walling  and  II.  P.  Jenssen,  IEEE  .1.  Quantum  Electron.  QE-18, 
167  (1982). 

7.  R.  C.  Powell,  L.  Xi,  X.  Gang,  G.  .1.  Quarles  and  J.  C.  Walling,  Phys.  Rev.  B  32,  2788 
(1985). 

8.  A.  B.  Suchocki,  G.  D.  Gilliland,  R.  C.  Powell,  J.  M.  Bowen  and  J.  C.  Walling,  J. 
Lumin.  37,  29  (1987). 

9.  M.  G.  Cohen  and  N.  Bloembcrgen,  Phys.  Rev.  A  135,  950  (1964). 

10.  V.  A.  Vazhenin,  A.  E.  Nikiforov,  B.  K.  Sevast'yanov,  K.  M.  Starichcnko,  A.  K. 
Shevchenko  and  Yu.  A.  Sherstkov,  Sov.  Phys.  Solid  State  29  362  (1987). 

11.  P.  E.  Jessop,  T.  Murarnoto  and  A.  Szabo,  Phys.  Rev.  B  21,  926  (1980). 

12.  R.  T.  Harley  and  R.  M.  Macfarlane,  J.  Phys.  C  |6,  1507  (1983). 

13.  E.  A.  Whittaker,  M.  Gehrtz  and  G.  J.  Bjorklund.  .1.  Opt.  Soc.  Am.  B2,  1320  (1985). 

14.  .1.  C.  Walling,  O.  G.  Peterson,  II.  P.  Jenssen,  R.  C.  Morris  and  E.  W.  O'Dell,  IEEE 
J.  Quantum  Electron.  QE-16,  1302  (1980). 

15.  A.  Szabo,  Phys.  Rev.  B  ]^,  4512  (1975). 

16.  D.  M.  Boyc,  R.  S.  Mcltz.cr  and  J.  E.  Rives,  Bull.  Am.  Phys.  Soc.  33  693  (1988). 


-12- 

17.  M.  Gehrtz,  G.  C.  Bjorklund  and  fi.  A.  Whittaker,  J.  Opt.  Soc.  Am.  B2.  1510  (1985). 


FIGURE  CAPTIONS 


Figure  I.  Simplified  apparatus  diagram.  See  text  for  description  of  symbols.  A  detailed 
description  of  the  FM  apparatus  is  presented  in  Ref.  2. 

Figure  2.  Pseudo-Stark  efTect  on  transient  holes  burned  into  the  R)m  line  of  alexandrite  at 
1.6K,  detected  using  laser-FM  spectroscopy.  The  signal  from  only  one  FM  sideband  is  shown 
here.  Traces  a)  through  Q  correspond  to  applied  voltages  of  0,  50,  100,  200,  300  and  400 
V,  respectively. 


Figure  3.  Pseudo-Stark  splittings  as  a  function  of  applied  electric  field.  The  laser  polarization 
is  always  along  the  crystal  b  axis.  Circles,  squares,  and  triangles  correspond  to  applied 
electric  fields  along  the  ax  and  b  axes,  respectively.  The  straight  lines  are  least-square  fits 
to  the  data.  No  splitting  was  observed  for  electric  fields  applied  along  the  b  axis. 

Figure  4.  Transient  hole  burning  and  laser- FM/Stark  double  modulation  detection  in  the 
Rlm  line  of  alexandrite  at  1.6K.  Trace  a)  is  the  signal  from  a  transient  hole  detected  with 
laser-FM  spectroscopy;  trace  b)  is  a  simulation  of  the  lineshape  expected  for  double 
modulation  detection;  trace  c)  is  the  actual  double  modulation  signal  of  a  transient  hole 
obtained  using  identical  conditions  to  those  used  to  produce  trace  a). 


Figure  5.  SFS  in  alexandrite  at  1.6  K  using  double  modulation  detection  with  a  10  kHz  signal 
at  200  V  applied  along  the  a  axis.  Trace  a)  shows  the  superimposed  results  of  successive 
averages  of  4096  laser  scans  across  a  small  portion  of  the  R,m  line.  Trace  b)  is  an  average 
of  4096  scans,  with  the  laser  tuned  off  of  the  line. 


Figure  6.  Normalized  autocorrelations  of  the  SFS  data  presented  in  Figure  .5.  Traces  a)  and 
b)  were  obtained  from  the  on-line  SFS  spectra,  trace  c)  was  obtained  From  the  ofT-linc 
spectrum. 


DL/1113/87/2 


TECHNICAL  REPORT  DISTRIBUTION  LIST,  GEN 


No. 

Copies 

Office  of  Naval  Research  2 

Attn:  Code  1113 
800  N.  Quincy  Street 
Arlington,  Virginia  22217-5000 

Dr.  Bernard  Douda  1 

Naval  Weapons  Support  Center 
Code  50C 

Crane,  Indiana  47522-5050 


Naval  Civil  Engineering  Laboratory  1 

Attn:  Dr.  R.  W.  Drisko,  Code  L52 
Port  Hueneme,  California  93401 

Defense  Technical  Information  Center  12 
Building  5,  Cameron  Station  high 

Alexandria,  Virginia  22314  quality 

DTNSRDC  #  1 

Attn:  Dr.  H.  Singerman 
Applied  Chemistry  Division 
Annapolis,  Maryland  21401 

Dr.  William  Tolies  1 

Superintendent 

Chemistry  Division,  Code  6100 
Naval  Research  Laboratory 
Washington,  D.C.  20375-5000 


No. 

Copies 

Dr.  David  Young  1 

Code  334 

N0RDA 

NSTL,  Mississippi  39529 

Naval  Weapons  Canter  1 

Attn:  Dr.  Ron  Atkins 
Chemistry  Division 
China  Lake,  California  93555 

Scientific  Advisor  1 

Commandant  of  the  Marine  Corps 
Code  RD-1 

Washington,  D.C.  20380 

U.S.  Army  Research  Office  '  1 

Attn:  CRD-AA-IP 
P.0.  Box  12211 

Research  Triangle  Park,  NC  27709 

Mr.  John  Boyle  1 

Materials  Branch 
Naval  Ship  Engineering  Center 
Philadelphia,  Pennsylvania  19112 

Naval  Ocean  Systems  Cmter  1 

Attn:  Dr.  S.  Yamamoto 
Marine  Sciences  Division 
San  Diego,  California  91232